Lack of Trigeminal Nerve Toxicity in Rats Exposed to Trichloroethylene Vapor for 13 Weeks

Abstract

Male and female Fischer-344 rats were exposed to 1,1,2-trichloroethylene (TCE) at 250, 800, or 2500 ppm for 6 h/day, 5 days/week, for 13 weeks. Weekly body weights and daily clinical observations were recorded and a functional observational battery (FOB) was performed monthly. Postexposure neurotoxicological evaluations included an electrodiagnostic evaluation of auditory function, the trigeminal nerve, and a comprehensive neuropathological examination. After 8 weeks of exposure, female, but not male, rats exposed to 2500 ppm were slightly more reactive to handling than the controls but not after 13 weeks of exposure. After 13 weeks, female rats exposed to 2500 ppm TCE were slightly more active during the 1-min observation period than the controls. There were no treatment-related differences in grip performance, landing foot splay, or on the trigeminal nerve–evoked potential at any dose. At 2500 ppm TCE, mild frequency-specific hearing deficits were observed, including elevated tone-pip auditory brainstem response thresholds. Focal loss of hair cells in the upper basal turn of the cochlea was observed in 2500 ppm–exposed rats. Except for the cochleas of 2500 ppm–exposed rats, no treatment-related lesions were noted during the neuro-histopathologic examination. The no-observable-adverse-effect level for this study was 800 ppm based on ototoxicity at 2500 ppm.

Trichloroethylene (TCE) is a colorless, volatile, unsaturated aliphatic halogenated hydrocarbon with a sweet odor-threshold at about 100 ppm. TCE is primarily used as a metal degreaser, to a much lesser extent in adhesive and aerosol formulations, and as a chemical-process intermediate in polyvinyl chloride and fluorochemical production (Halogenated Solvents Industry Alliance 1996). Inhalation is the expected route of human exposure. Once absorbed, TCE can undergo extensive metabolism to a variety of products including trichloroethanol, trichloroacetate, and the psychoactive drug chloral (reviewed by Lash et al. 2000).

The toxicity of TCE has been recently reviewed by The New York State Department of Health (NYDOH 2005) and The National Academy of Sciences (NRC 2006). A variety of effects, both non-neoplastic and neoplasic, have been reported in a variety of animal models, primarily rodent. The most sensitive effects of inhaled TCE in experimental animals are acute central nervous system sedation and increased liver and kidney weights following subchronic exposure (Kjellstrand et al. 1981; Adams et al. 1951). Acute toxicity studies in animals indicate that TCE is relatively nontoxic by the inhalation route; the LC₅₀ from a 4-h exposure ranged from 12,500 ppm in rats to 8450 ppm in mice (ATSDR 1996). The acute and chronic neurobehavioral effects of TCE exposure have been studied in a number of experimental animal models (NYDOH 2005; NRC 2006). Central nervous system depression was observed in Wistar rats exposed to 3000 ppm TCE for 7 h/day for 27 days (Adams et al. 1951). In another subchronic study, male Wistar rats exposed to TCE vapor concentrations of 500, 1000, or 1500 ppm for 16 h/day, 5 days/week, for 18 weeks showed significant differences in a two-choice visual discrimination task during the exposure period (Kulig 1987). Exposure to 1000 ppm and 1500 ppm TCE caused increase trial response latency beginning in the first week of exposure. The author concluded that the effects were likely to be acute effects of exposure because the effects disappeared once exposures were terminated. Peripheral nerve conduction time was unaffected. Following termination of exposure, there was no carry over of TCE-related functional effects. Pathological examination of brain, liver, or kidney tissues was not performed.

Several TCE studies have found mid-to-high frequency hearing loss in rats (Crofton and Zhao 1993; Crofton, Lassiter, and Rebert 1994; Fechter et al. 1998; Jaspers et al. 1993; Rebert et al. 1991). Although significant hearing loss was reported in Long-Evans rats exposed at 3200 ppm for 12 h/day, 7 days/week, for 12 weeks, no effects on the auditory brainstem response were observed at 1600 ppm, and somatosensory and visual evoked potentials were unaffected (Rebert et al. 1991).

The acute effects of TCE on human performance at low concentrations have been contradictory. TCE at or below 300 ppm produces no adverse behavioral effects (Anger and Johnson 1985). But exposure to high concentrations of TCE have been reported to cause neurotoxicity in humans (Annau 1981; Hartman 1988). The hallmark of the putative TCE neurotoxicity is trigeminal anesthesia (Barret et al. 1982, 1987; Feldman et al. 1992; Feldman, Chirico-Post, and Proctor 1988) and hypesthesia (Barret et al. 1987). Annau (1981) speculated that pure TCE in fact may not be a neurotoxicant. Dichloroacetylene (DCAL) can be formed from the degradation of TCE under strong alkaline conditions (IARC 1986), and is believed to be responsible for the neurotoxicity associated with TCE (Annau 1981; Spencer and Schaumburg 1985; Hartman 1988). Trigeminal nerve/brainstem conduction times were significantly slower in rats exposed to 300 ppm DCAL for about 2 h (Albee et al. 1997).

A series of electrophysiology studies have examined the potential effects of halogenated solvents on nervous system function: dichloromethane (Mattsson, Albee, and Eisenbrandt 1990), 1,1,1-trichloroethane (Mattsson et al. 1993), and 1,1,2,2-tetrachloroethylene (Mattsson et al. 1998). The design and conduct of the current study closely followed these previous studies. The purpose of this study was to evaluate, in Fischer-344 rats, the effects of subchronic exposure to TCE vapor on the function and anatomic pathway of the trigeminal nerve using trigeminal nerve somatosensory-evoked potentials (TSEPs) and histopathologic examination of the trigeminal nerve and its sensory nucleus in the brainstem. To further characterize the effects on the nervous system, as well as verify sufficient exposure, auditory-evoked potentials were recorded and a comprehensive clinical and neuropathological workup was performed.

MATERIALS AND METHODS

Test Material and Chambers

Trichloroethylene (TCE; inhibited with 0.5% butylene oxide, lot no. TB910325AB) was from The Dow Chemical Company (Freeport, Texas). The purity of the test material was determined by gas chromatography with thermal conductivity detection (GC/TCD) prior to study initiation. The sample of TCE was found to be 99.22% pure with the butylene oxide stabilizer present at 0.69%. The identity of the sample was confirmed before and after completion of the study by infrared spectroscopy to be TCE. Stainless steel and glass 4.1-m³ Rochester-type exposure chambers (square with a pyramidal top and bottom) were used with the airflow maintained at 450 L/min. Test atmospheres of TCE were generated according to previously published methods (Miller et al. 1980). The concentration of TCE in each chamber was monitored one to two times per hour with a Miran 1A infrared spectrometer at a wavelength of 11.8 microns.

Experimental Subjects

Approximately 14-week-old male and female Fischer-344 rats were purchased from the Charles River Laboratories (Kingston, New York). This strain was selected based on the suitability of this strain in toxicology testing and the availability of historical control data within our laboratory. The rats were housed one per cage in suspended stainless steel cages which had wiremesh floors. The holding room was maintained at approximately 22°C, 50% humidity, and a 12-h light-dark cycle (lights on at 0700 h). Certified Purina Chow 5002 (Purina Mills, St. Louis, Missouri) and municipal drinking water were available ad libitum, except that feed was not available during exposures. All procedures were reviewed and approved by the Dow Chemical Company Institutional Animal Care and Use Committee (IACUC), which follows the National Institutes of Health Guide for Care and Use of Laboratory Animals (Publication No. 85-23, revised 1985) and were conducted in Dow’s AAALAC-accredited toxicology laboratory. Clinical care was provided and supervised by a clinical veterinarian.

Experimental Design

Ten rats/sex/dose were exposed to TCE vapors at 0, 250, 800, or 2500 ppm for 6 h/day, 5 days/week, for 13 weeks. These concentrations were selected following extensive preliminary testing (using various evoked-potential diagnostic tests), a thorough review of the literature, and were chosen such that the highest dose used (2500 ppm) would cause functional nervous system changes without causing excessive systemic toxicity, which would likely compromise the interpretability of the neurotoxicity data. Two additional rats/sex/group were placed on study to allow for possible losses.

Body Weights and Clinical Observations

Body weights were taken preexposure and weekly during exposure; however, only preexposure and monthly body weights were analyzed statistically and reported. Cage-side observations for general health status were conducted twice daily, for morbidity, mortality, availability of feed and water, and once/week clinical observations were performed on each rat by removing it from its cage and examining it for changes in general health.

Functional Observational Battery (FOB)

An FOB, as described by the United States Environmental Protection Agency (USEPA 1991), was conducted on all rats preexposure (4 days prior to exposure) and during weeks 4, 8, and 13 of exposure. The FOB included hand-held and open-field observations and measurements of grip performance and landing foot splay. The evaluations were done at the same time of the day (morning). All animals were carefully examined, in random order, by the same trained observer throughout the study period. The FOB was conducted in such a manner that the observer did not know the treatment status of the animal. Procedures for grip performance (Mattsson, Johnson, and Albee 1986) and hindlimb splay (Edwards and Parker 1977) have been reported elsewhere. The average of three trials in each test was used for statistical analysis.

Evoked-Potential Testing

Surgical implantation of epidural electrodes was scheduled 11 weeks into the study, and was performed in a counterbalanced order (by dose group and sex) over 5 days. The details of this procedure have been published previously (Mattsson et al. 1998). Briefly, rats were deeply anesthetized with methoxyflurane vapor, and were maintained under isoflurane. Rectal temperature was monitored and body temperature was supported with a warm-water blanket. Rats were placed in a stereotaxic instrument (no. 900; David Kopf Instruments, Tujunga, CA) and epidural electrodes (7 mm long, no. 0–80, stainless steel set screws) were surgically implanted into the skull and supported with dental acrylic. A somatosensory electrode was placed 2.0 mm posterior and 2.0 mm lateral left of bregma, while a cerebellar electrode was placed 3.0 mm posterior and 0.0 mm lateral of lambda. A reference electrode was placed 8.0 mm anterior and 1.0 mm lateral left of bregma.

The electrodiagnostic system was a Nicolet Pathfinder II (Nicolet Biomedical Instruments, Madison, Wisconsin). Data sweeps (millisecond segments of electroencephalogram [EEG]) were digitally sampled 512 times and averaged by an online computer. Ten rats (five/sex) per exposure level were necropsied after electrophysiological testing. Rats were physically restrained during electrodiagnostic testing (approximately 40 min). The restraint device was similar to Rebert (1983) but the nose bar and rubber band head restraints were not used. Rectal temperature was recorded immediately before and after electrodiagnostic testing. Trigeminal somatosensory evoked potentials (TSEPs) and auditory brainstem responses (ABRs) were collected 2 to 3 days after 13 weeks of exposure. Recording parameters are summarized in Table 1.

Auditory Brainstem Response to Clicks (ABR_c)

ABR _c were recorded from the cerebellar electrode. The energy of the click stimulus was predominantly between 2 and 6 kHz. The recording method allows isolation of far-field electrical activity, and therefore, ABRs provide information on the functional integrity of the brainstem and midbrain. The ABR has distinctive peaks that have a general association with anatomic features of the auditory pathway through the brainstem and midbrain. Peak I is associated with the acoustic nerve (VIII cranial nerve), peak II with the acoustic nucleus, peak III with the superior olivary nucleus, peak IV with the lateral lemniscus, and peak V with the inferior colliculus. Responses were collected through a 100 to 3000-Hz high-pass low-pass filter.

Auditory Brainstem Response to Tone Pips (ABR_t )

ABR to tone pips were recorded from the cerebellar electrode. The following frequencies were tested: 4 kHz (ABR₄), 8 kHz (ABR₈), 16 kHz (ABR₁₆) and 30 kHz (ABR₃₀). ABR_t provide information on frequency-specific auditory function, and are used to screen for mid- to high-frequency hearing dysfunction. Auditory thresholds were determined at each frequency using the following method: an initial intensity great enough to see clearly the ABR was presented to the test subject. The intensity was sequentially lowered until the ABR waveform was no longer visible. The lowest intensity that still elicited a visible waveform was recorded as the threshold intensity. Rats were presented to the technician conducting the test in a blind and random manner.

Trigeminal Somatosensory-Evoked Potential (TSEP)

TSEPs provide information on the functional integrity of the trigeminal nerve, trigeminal ganglion, brainstem sensory nuclei, and pathways leading to the thalamus. The trigeminal nerve was stimulated at the vibrissal pad with a pair of subcutaneously placed needle electrodes (Albee et al. 1997). The cathode was placed at the second vibrissae posterior to the tip of the nose. The TSEP was a far-field potential recorded from the somatosensory electrode; a subcutaneous needle electrode was placed in the lower foreleg as a reference. The nasal electrode served as the ground.

Waveform Analyses

An automated computer technique was utilized to quantitatively analyze the TSEP and ABR data by quantification of differences from a template in waveform shape (optimal cross-correlation), latency, and power (Mattsson and Albee 1988). The computerized analyses included optimal cross-correlation (similarity of shape; reported as an r value) of an individual waveform to the template waveform, latency difference (phase shift in milliseconds required to reach optimal cross-correlation), and total power of the peaks contained within the window (μV). Subsequent to computerized waveform evaluation, the computer’s measurements of latency difference for each waveform for each animal were visually confirmed by an experienced auditor having no knowledge of animal identity or treatment. When the computerized analysis was determined to be in error (usually due to mismatching of peaks), the auditor’s value was accepted and a new cross-correlation coefficient was calculated.

Statistical Analyses

Statistical analyses were conducted on body weights, rectal temperature, grip performance, landing foot splay, TSEP, and ABR. FOB data were evaluated post hoc by a test of proportions (Bruning and Kintz 1977). Continuous data were analyzed using repeated-measure analysis of variance (ANOVA) (RM-ANOVA) to account for the data at all time periods in one statistical analysis. The interactions of Treatment × Time and Treatment × Time × Sex were examined. In the event of statistically significant findings in the overall analyses, further visual and/or statistical examination of the data were necessary to locate the source of the statistical significance.

Electrodiagnostic data (optimal correlation, latency difference, and power) were analyzed by a factorial multivariate analysis of variance (MANOVA), using the GLM procedure of SAS (SAS Institute, Cary, NC). Because the evoked potential correlation values are often high (0.8 to 1.0) and, therefore, not well distributed for parametric analysis, r values were arcsine transformed before statistical analysis. Rectal temperature, taken as the average of the before- and after-testing temperature, was used as a covariate. The main effects of treatment and the interaction of Treatment × Sex were assessed by an F-test based on the Pillai’s trace statistic. Where a significant main effect of treatment occurred, univariate contrasts of each treatment level versus control were examined. Temperatures collected during electrodiagnostic testing were analyzed by RM-ANOVA as described above for FOB data. The temperature covariate was tested for homogenicity of slopes and was found to be homogeneous.

Outlier Evaluation

Because extreme departure from homogenicity of variance can affect the F-ratio (Rogan and Keselman 1977), the F _max test for homogeneity of variance was performed (α = 0.01; Bruning and Kintz 1977). When departure from homoscedasticity for a specific variable was judged too extreme by the study director, outlying data points farthest from the mean were removed (one at a time) until homoscedasticity was achieved.

Alpha Adjustment for Multiple Statistical Tests

To reduce the rate of false declarations, the type I error rate (α) per comparison was set at 0.02 for all primary planned analyses. Stepdown analyses (e.g., contrasts after RM-ANOVA) were conducted at α = 0.02 when they followed a statistically significant primary analysis. This overall approach was consistent with the recommendations proposed by Tukey, Ciminera, and Heyse (1985), Mantel (1980), and by the USEPA (1991). The corrections for multiple statistical analyses were applied to α only, and probability values were reported without correction. The final toxicologic interpretation of the data considered other factors, such as dose-response relationships, biological plausibility and consistency, and historical rates.

Neuropathology

Pathological examinations were limited to the nervous system. After 13 weeks on study, 5 rats/sex/group were randomly selected for neuropathologic evaluation. Rats were fasted overnight prior to being presented for necropsy. Rats were injected with 0.2 ml heparin (10,000 USP/ml) per 100 g body weight at least 10 min prior to perfusion. Rats were anesthetized by the inhalation of methoxyflurane vapors. While deeply anesthetized, rats were perfused by placement of a cannula in the left ventricle. The initial perfusate was 0.05 M phosphate buffer containing sodium nitrite followed with a 2.5% glutaraldehyde–2.0% formaldehyde solution in phosphate buffer (300 to 315 mOsM). Tissues were trimmed and immersed in fixative containing 1.5% glutaraldehyde–4.0% formaldehyde. The tympanic bullae were dissected from the skull. Using a dissecting microscope, the bullae were opened, the stapes was removed, the oval and round windows were opened, and the cochleas were retained in the fixative used for perfusion.

Tissues for neuropathologic evaluation were prepared from all perfusion-fixed rats in the control and high-dose groups. Nine transverse sections of the brain were prepared from the: olfactory lobe, cerebrum (frontal, parietal, temporal, and occipital lobes), thalamus/hypothalamus, midbrain, pons, cerebellum, medulla oblongata, and nucleus gracilis. The following tissues were also prepared: trigeminal ganglion, pituitary gland, eyes with optic nerves, spinal cord (cervical and lumbar), nasal tissues with the olfactory epithelium, and skeletal muscles (gastrocnemius and anterior tibial). The above tissues were embedded in paraffin, sectioned approximately 6 μm thick, and stained with hematoxylin and eosin. Peripheral nerves (sciatic, tibial, and sural) and additional dorsal root ganglia (cervical and lumbar) were osmicated, embedded in plastic, sectioned approximately 2 μm thick, and stained with toluidine blue. All tissues were examined by a veterinary pathologist using a light microscope.

The cochleas from two male and two female rats exposed to 0 ppm, and two male and two female rats exposed to 2500 ppm TCE were osmicated, decalcified in 3% EDTA (approximately 2 to 3 weeks) processed using conventional techniques, and were embedded in Epon-Araldite epoxy resin. Following polymerization of the plastic, 17 midmodiolar sections (at least 6 microns apart), 6 microns thick, were cut with a glass knife and an ultramicrotome. Sections were mounted on glass slides and stained with toluidine blue. Cochlear sections were evaluated by a veterinary pathologist using a light microscope.

RESULTS

Exposure Information

Mean analytical concentrations of TCE were 251.4 for the 250-ppm chamber, 802.1 for the 800-ppm chamber, and 2513.6 for the 2500-ppm chamber. The airflow, temperature, and relative humidity values showed only slight variation between chambers and had no impact on study results.

Body Weights and Clinical Observations

There were also no changes in body weight over the course of the study (Table 2). There were no exposure-related effects observed during daily cage-side examinations. Clinical observations revealed an increase in lacrimation immediately after exposure in 2500 ppm–exposed rats, beginning with the females on test day 5. By test day 12, most of the 2500 ppm– and two of the 800 ppm–exposed rats showed increased lacrimation. This observation was apparent immediately following the exposures but was not observed during the FOB, which was conducted prior to the daily exposure, and was suggestive of acute eye irritation.

Functional Observational Battery (FOB)

Minor differences in the reactivity to handling were observed in female rats at exposure weeks 8 and 13. At 8 weeks, female, but not male rats exposed to 2500 ppm were slightly more reactive to handling than the controls. There were no significant exposure-related differences in reactivity after 13 weeks of exposure. After 13 weeks, female rats exposed to 2500 ppm TCE were slightly more active during the 1-min observation period than the controls. There were no treatment-related effects on any other FOB parameter (Table 2).

Electrodiagnostic Results

Average rectal temperatures (collected before and after electrodiagnostic testing) showed no treatment-related effects (RM-ANOVA, p = .1918).

Click Auditory Brainstem Response (ABR_c)

Visual inspection of the waveforms and quantitative evaluation (Table 2) indicated a significant difference in the waveforms of 2500 ppm–exposed rats compared to the controls. The 2500 ppm–exposed rats had click ABRs that were smaller (males 31% and females 23% less power than controls), slightly slower (by 0.03 msec), and less correlated to the template (r) than controls. The most impressive difference was in peak I, which was approximately half the amplitude of the controls. Because peak I was significantly affected by treatment, the amplitudes of the remaining peaks are not interpretable even though they appeared to be different, i.e., altered input affects all subsequent peaks. The degree of slowness of peak V relative to the controls was about the same as peak I indicating a ‘normal’ I–V latency, and a lack of effect on the brainstem. There were no exposure-related differences in rats exposed to ≤800 ppm.

Tone-Pip Auditory Brain Stem Response (ABR_t )

The purpose of the tone-pip ABRs was to evaluate auditory function at multiple frequencies, therefore, the analysis was limited to peak I (acoustic nerve). Significant exposure-related differences were detected only in the 2500-ppm group, and these occurred at all frequencies (Figure 1; Table 2). The effects were greatest at 16 kHz, less at 30 kHz, and least at 4 and 8 kHz (summarized below).

At 4 and 8 kHz, the 2500-ppm group waveforms were slightly slower, whereas, at 16 kHz the 2500-ppm group was slower, had slightly altered shape (smaller r value), and were much smaller (males 59% and females 46% less power). At 30 kHz, the 2500-ppm group had altered ABR₃₀ shapes and significantly smaller peak I (males 34% and females 26% less power) than the controls. The 2500-ppm female 30-kHz data had significant heterogeneity of variance (α = 0.01) that could not be corrected by removing a few outlying data points (see Outliers section, below). The significant heterogeneity of variance was interpreted as a treatment effect.

Tone-Pip Auditory Brain Stem Response Thresholds

To quantify the magnitude of auditory dysfunction, tone-pip ABR thresholds at 4, 8, 16, and 30 kHz were determined for the 0- and 2500-ppm groups. ABR thresholds in the 2500-ppm group were significantly elevated at all frequencies compared to the controls (Figure 2; Table 2). The data clearly indicate that the treatment effect was frequency-specific, i.e., the greatest threshold increase was at 16 kHz (∼15 dB) followed by 30 kHz (∼8 dB), 8 kHz (∼4.5 dB), and 4 kHz (∼4 dB).

Trigeminal Nerve Somatosensory-Evoked Potentials (TSEPs)

There were no significant exposure-related differences in the TSEPs based on qualitative (Figure 3) and quantitative (Table 2) evaluations.

Outliers

Twelve data points (out of 1440 collected) were judged to be outliers based on the F _max test (Bruning and Kintz 1977) and eliminated from further analysis. Eleven outlying data points came from the ABR data. Of the ABR outliers, all were from the 2500 ppm–exposed group and nine were from the female group. The 2500-ppm female 30-kHz ABR data could not be made homogenous with the other groups and, therefore, could not be analyzed with normal parametric statistics. This lack of homogeneity of variance was considered to be a treatment-related effect. A single outlier was detected in the TSEP data (0-ppm male group).

Summary of Electrodiagnostic Tests

A consistent, adverse finding in the electrodiagnostic data was the auditory deficit in the click and tone-pip auditory brain-stem responses (ABRs) recorded from the 2500 ppm–exposed rats. The auditory deficits were somewhat frequency specific because the 16-kHz ABR was the most affected response followed the 30-kHz ABR. The 4- and 8-kHz ABRs were least affected. The 16-kHz ABR threshold of 2500 ppm–exposed rats was elevated about 15 dB, the 30-kHz ABR was elevated 8 dB, and 4- and 8-kHz ABRs were elevated about 4 to 5 dB relative to the control rats.

Neuropathology

In the rats examined for cochlear alterations, an exposure-related effect was observed in three of the four rats exposed to 2500 ppm TCE. The lesion consisted of the loss of one or occasionally two of the outer hair cells from a focal region of the Organ of Corti. All control rats and one male exposed to 2500 ppm TCE had the normal morphology of one inner hair cell and three outer hair cells consistently noted for all six regions of the Organ of Corti (hook region, lower and upper basal turn, lower and upper middle turn and apex). Occasionally, a hair cell was not discerned in 1 or 2 of the 17 sequential sections. This may have been an artifact or a truly missing cell; however, all four hair cells were present in the great majority of the 17 sections for each of these areas from the control rats. In the three affected rats exposed to 2500 ppm TCE, one and sometimes two hair cells were missing in most of the sections from the upper basal level of the Organ of Corti. The other regions of the Organ of Corti of the affected rats contained normal hair cell numbers. There were no effects ascribed to TCE exposure upon the spiral ganglion or cochlear nerve. The nerve fibers and neuronal cell bodies appeared normal. They were well preserved, adequately stained, and had minimal artifact.

All other pathological observations were considered unrelated to treatment and most were typical of spontaneously occurring changes of minimal severity. Many of the diagnoses involved only a single rat from a dose group and often only a single nerve fiber was affected. A normal background incidence, in controls and 2500-ppm rats, was noted for swollen axons in the nucleus gracilis and degeneration of individual nerve fibers of the trapezoid, both in the medulla oblongata, and degeneration of individual nerve fibers in the white matter of the cervical and lumbar areas of the spinal cord (Eisenbrandt et al. 1990).

DISCUSSION

TCE has been reported to cause neurotoxicity in humans. The hallmark of the putative TCE neurotoxicity is trigeminal anesthesia (Barret et al. 1982, 1987; Feldman et al. 1992; Feldman, Chirico-Post, and Procter 1988) and hypesthesia (Barret et al. 1987). There is, however, increasing evidence that a degradation product of TCE under strong alkaline conditions (IARC 1986), dichloroacetylene (DCAL), and not TCE, is the agent responsible for the reported neurotoxicity (Annau 1981; Hartman 1988). Because the trigeminal nerve has been identified as the target tissue, we utilized the trigeminal nerve somatosensory-evoked potential (TSEP) as a diagnostic tool to evaluate trigeminal nerve function.

Previously, we conducted a study to validate the TSEP procedure for DCAL-induced trigeminal dysfunction to demonstrate the suitability of the Fischer-344 rat as an appropriate experimental model (Albee et al. 1997). DCAL caused a slowing of trigeminal nerve/brainstem conduction times in rats exposed to 300 ppm DCAL for approximately 2 h (Albee et al. 1997). This activity plus its apparent causation of degraded TCE-associated trigeminal effects made DCAL a relevant positive control for the present study. DCAL-exposed rats had significantly slower and smaller TSEPs whereas the control and TCE/acetylene-exposed rats were unchanged from baseline (Albee et al. 1997). In the present study, TCE alone did not cause slowing of the TSEP even though exposure lasted 13 weeks, and ototoxicity was demonstrated. Results from the previous DCAL study, and the results of the present study, lead one to conclude that TCE does not cause trigeminal neuropathy under these exaggerated test conditions in a species that has been shown to be at least partially sensitive to DCAL-induced trigeminal neuropathy. These results also indicate the lack of formation of metabolites or levels of metabolites capable of causing trigeminal neuropathy under the conditions of the study.

Exposure to TCE caused only minimal and sporadic changes in activity levels or reactivity to touch throughout the 13-week exposure period; however, the expected mild ototoxicity was noted in the 2500 ppm TCE–exposed group. Our findings were consistent with the TCE ototoxicity literature (Crofton and Zhao 1993; Crofton, Lassiter, and Rebert 1994; Jaspers et al. 1993; Rebert et al. 1991) with one exception (Fechter et al. 1998). Unlike the study by Fechter et al. (1998), we did not find any loss of ganglion cells. Neuropathological examination of midmodiolar sections of the cochlea revealed a loss of one or two outer hair cells in three of four rats examined. These slides were reexamined in light of the findings by Fechter et al. (1998) without confirming ganglion cell pathology. In addition, we found statistically significant increased thresholds at all frequencies evaluated, with the greatest increase at 16 kHz (≈15 dB) followed by 30 kHz (≈8 dB). Fechter et al. (1998) reported significantly increased hearing thresholds only at 8 and 16 kHz, although qualitatively increased thresholds were present at 4 and 32 kHz. Obvious differences between that study and ours was their use of a higher concentration (4000 versus 2500 ppm) and shorter exposure duration (6 h/day for 5 days versus 6 h/day, 5 days/week, for 13 weeks). Perhaps the different outcome can be attributed to the high subacute-type exposure employed by Fechter et al. (1998).

In conclusion, 2500 ppm TCE did not cause trigeminal nerve dysfunction but did cause moderate ototoxicity in rats exposed for 13 weeks. It has been demonstrated that functional testing is slightly more sensitive than histopathology in detecting ototoxicity. Sullivan and Conolly (1988) have shown that elevations in brainstem auditory-evoked response thresholds were detected before observing hair cell loss. Rats exposed to 104 dB white noise had a 50% decrease in functional hearing, whereas 117 dB was required to cause an 50% loss of hair cells. Sullivan, Rarey, and Conolly (1989) also demonstrated a correlation between hair cell loss in specific regions of the cochlea and elevated ABR thresholds. In the present study, histopathology was conducted to confirm the morphologic cause for the elevated ABR thresholds in 1500 ppm–exposed rats. Because there were no functional changes in ABRs from 800 ppm exposed rates histopathology was not conducted. The no-observable-adverse-effect level (NOAEL) for this study was 800 ppm based on ototoxicity.