Registration Dossier

Administrative data

neurotoxicity: sub-chronic dermal
Type of information:
experimental study
Adequacy of study:
key study
1 (reliable without restriction)
Rationale for reliability incl. deficiencies:
other: GLP Guideline study

Data source

Materials and methods

Test guideline
equivalent or similar to
other: OECD Guideline 411 (1981)
GLP compliance:

Test material

Details on test material:
Alkyl (C12-C14) glycidyl ether
Label: Epoxide 8
51% C12 Glycidyl ether
21% C14 Glycidyl ether
5% C16 Glycidyl ether
13% N +1's
4% alpha addition Products
94% total active (label claim)
- Physical state: Clear colorless liquid
- Storage condition of test material: ambient

Test animals

Fischer 344
Details on test animals and environmental conditions:
Supplier : Charles River Laboratories Inc. (Raleigh, NC, USA)
This strain of rat was chosen because of its general acceptance in neurotoxicity testing and availability of historical data.

Animal Welfare
In response to the Final Rules amending the U.S. Animal Welfare Act that were promulgated by the U.S. Department of Agriculture effective October 30, 1989, the
Animal Care and Use Activity (ACUA) that was required for the conduct of this study was submitted and reviewed prior to the study start and given full approval by the Institutional Animal Care and Use Committee (IACUC). The IACUC determined that the proposed Activity was in full accordance with these Final Rules.

Physical Acclimation
Upon arrival at the laboratory, the general health of each animal was evaluated by a laboratory veterinarian. The animals were housed two per cage in stainless steel wire cages and acclimated to the laboratory for at least one week prior to the study start day.

The animals were housed one per cage after randomization in suspended stainless steel cages which had wire-mesh floors. The holding room was maintained at a temperature range of 21-23°C (minimum of 21.7°C and maximum of 22.4°C); at a humidity range of 40-60% (minimum of 48% and maximum of 54%), and on a 12-hour light-dark cycle (7 AM - 7 PM). Animal room temperature and humidity data can be found in the study file.

Approximately 7 weeks at pre-exposure

Rats were assigned one per cage to study groups using a computer-generated randomization program based on body weights. The dispositions of rats not placed on study were documented in the study file. Rats were uniquely identified via subcutaneously implanted transponders which were correlated to a unique identification number.

Feed and Water
Animals were provided Purina Certified Rodent Chow #5002 (Purina Mills, Inc., St. Louis, MO, USA) in pelleted form. Feed and tap water were provided ad libitum. Analysis of the chow was performed by Purina Mills Inc. to confirm the diet provided adequate nutrition and to quantify the levels of selected contaminants associated with the formulation process. Drinking water obtained from the City of Midland, Michigan was analyzed for chemical parameters and biological contaminants by
the City of Midland Water Department. In addition, specific analyses for chemical contaminants were conducted at periodic intervals as stated in the Standard Operating Procedures of The Toxicology Research Laboratory, The Dow Chemical Company. Copies of these analyses are maintained in this laboratory. The results of the feed and water analyses were judged to be within acceptable limits.

Administration / exposure

Route of administration:
Details on exposure:
The dermal route of exposure represents a likely route of occupational exposure to the test material (i.e. via contact during manufacture or use of Alkyl Glycidyl Ether). Dose solution concentrations were based on the expected average growth patterns (body weight) of male and female Fischer 344 rats over the duration of a 13-week study. This dosing pattern resulted in delivery of slightly higher than target concentrations at the be inning of the study and slightly lower than target concentrations at the end of the study in order to achieve a time-weighted average close to the original targeted dose. Dose solutions were prepared by mixing test material in acetone. The dosing solution concentration for each dose level remained constant throughout the duration of the 13 week study for females (5%, 0.5%, and 0.05% AGE in acetone). The dosing solution concentration had to be increased at about 8 weeks into the dosing period for males (from 7.5%, 0.75%, and 0.075% AGE in acetone to 8.5%, 1.0%, and 0.1 %). The following percentage-of-target dosage were achieved for the 1, 10 and 100 mg/kg groups: 99.2%, 97.7%, 94.2% for males and 96%, 94.4% and 96.8% respectively for females.

Dose solutions were administered as a fixed volume (300 µl/rat/application). Doses were delivered by pipette to the clipped area of the back (approximately 10% surface area) of each rat. The treated area was not covered nor was the skin cleansed following treatment.
Analytical verification of doses or concentrations:
Details on analytical verification of doses or concentrations:
Stability and Homogeneitv
Stability of AGE in acetone was determined over a course of 63 days for the low dose of 1 mg/kg and the high dose of 100 mg/kg. The mixtures were found to maintain stability for over 2 months. At sixty three days, the mixtures were at 100% and 96% of the concentration of day 1 for the low and high doses respectively. Homogeneity of the test material in the low and high dose-level solutions was determined concurrently with the conduct of the study and both solutions were found to be very homogenous.

Analysis of dosing solutions by high-performance liquid chromatography (HPLC) to verify the concentration of the test material was conducted on all dose levels at the beginning of the dosing period, followed by two other additional concentration analyses at about monthly intervals. All concentrations were found to be acceptable and were within 6% of the target concentrations.
Duration of treatment / exposure:
13 weeks
Frequency of treatment:
5 applications/week (Monday through Friday)
No. of animals per sex per dose:
Groups of 12 Fischer 344 rats/sex/dose level were dosed dermally with AGE in acetone.
Details on study design:
Dose Levels
Dose levels for this study were approximately 0, 1, 10, and 100 mg/kg/application. The high-dose level of 100 mg/kg/day in the 13-week study was expected to be at or slightly exceed a maximum tolerated dose based on the integrity of the skin as determined by a 2 week range finding study (McGuirk and Johnson, 1997) and defined by the EPA Workshops on Carcinogenesis Bioassay via Dermal Route (EPA 1987a, 1988). The remaining dose levels were expected to provide dose-response data for any treatment-related effect(s) observed in the high-dose group and to ensure definition of a no observed effect level (NOEL).
The volume of material to be applied represented the maximum levels recommended by the Environmental Protection Agency (EPA, 1987a and 1988). The dosing regime of 5 days/week was based upon the results of an evaluation of the skin and the feasibility of continuing at these dose levels following 4 weeks of treatment in a 13-week dermal toxicity study (McGuirk and Johnson, 1997).

Test Material Administration
Study Design
Groups of 12 Fischer 344 rats/sex/dose level were dosed dermally with AGE in acetone.

The dosing volume was 300 µl per application and dosing solution concentrations were set to achieve approximate doses of 0, 1, 10, or 100 mg/kg body weight/application. The dermal dosing regimen involved 5 applications/week (Monday through Friday, dose site not washed between applications) for a scheduled period of 13 weeks which was extended to include an additional week. This treatment extension resulted due to technical problems with the electrodiagnostic equipment which caused a delay in electrodiagnostic testing originally scheduled to be conducted at the end of 13 weeks of dosing.

Daily cageside and weekly body weights, clinical and dermal exams were conducted during the pre-exposure and dosing periods. The first week of exposure entailed daily dermal grading. Motor activity and a functional observational battery (FOB), including grip performance and hindlimb landing foot splay testing, were conducted before exposure and monthly thereafter. A battery of electrodiagnostic tests was conducted after 14 weeks of exposure on 10 rats/sex/exposure group. Electrodiagnostic tests were used to screen for dysfunction of peripheral nerves, spinal cord, brainstem, cerebellum and
cerebrum by evaluating evoked potentials from several modalities. Mid- and high frequency tone-pip auditory brainstem responses were conducted to evaluate rats for hearing loss. FOB testing was conducted on all rats in a random order, whereas motor activity and electrodiagnostic testing were counterbalanced across dose groups. Electrodiagnostic tests were divided into principal and ancillary. Principal tests included flash evoked potential (medium flash intensity) recorded from the visual cortex (FEP-V), somatosensory evoked potential recorded from the somatosensory cortex (SEP-S), auditory brainstem response to clicks (ABRc), and caudal nerve action potentials (CNAP). Ancillary tests included: Low intensity FEPs, FEPs recorded from the cerebellum (FEP-C), SEPs recorded from the cerebellum (SEP-C), and tone-pip ABRs.

Ten rats (5 males and 5 females) per exposure level were necropsied after electrodiagnostic testing for histopathologic evaluation. The remaining rats were maintained for about 5 more weeks with daily cageside examinations and weekly body weights.

Based on evaluation of FEPs conducted following the end of exposure, FEP-C waveforms were analyzed in the same way as PEP-V waveforms. In addition, postexposure testing was considered necessary and all male rats were tested about 5 weeks post-exposure for PEPs. Control- and high-dose male rats were tested for electroretinograms (ERGs) also. Eyes and brains from all male rats were saved after the post-exposure testing but were not used for histopathologic evaluation.


Observations and clinical examinations performed and frequency:
- Time schedule: daily
A cage side visual evaluation for moribundity, mortality, the availability of feed and water, and treatment-related effects was conducted twice a day. Because of cage marking, the observer was aware of the treatment group. Animals were not hand-held for these observations. To the extent possible, cage side observations included evaluation of the fur, skin, mucous membranes, respiration, nervous system function, tremors, convulsions, diarrhea, swelling, masses and animal behavior. Animals found dead were refrigerated until necropsied. When FOBs or clinical exams were conducted, a cage side observation was not necessary for that corresponding time of day (i.e. FOB conducted in a.m., no need to do a.m. cage side observations).

Clinical observations were comparable to the hand-held portion of the FOB. Clinical observations were conducted once a week during the pre-exposure and exposure periods. Clinical observations included hand-held observations of the fur, skin, mucous membranes, respiration, nervous system function, tremors, convulsions, diarrhea, swelling, masses and animal behavior.

Evaluation of Dermal Application Site
The dermal test site was subjectively evaluated daily prior to each application for the first week; then weekly thereafter during the exposure period using this laboratory's modification of the acute dermal irritation scoring system recommended by the Organisation for Economic Co-operation and Development (OECD, 1981)

All rats were weighed during the pre-exposure period and at weekly intervals during the
exposure period.

The eyes of all animals were examined by indirect ophthalmoscopy by a veterinarian pre-exposure and pre-necropsy. One drop of 0.5% tropicamide ophthalmic solution was instilled in each eye to produce mydriasis prior to examination.
Neurobehavioural examinations performed and frequency:

The FOB was conducted under red-light illumination by the same observer on all rats during the week of pre-exposure and during weeks 4, 8, and 13 of treatment about the same time of day (most of the battery completed by noon). The FOB included hand-held and open-field observations and measurements of grip performance, rectal temperature and landing foot splay. The FOB was conducted in such a manner that the observer did not know the treatment status of the animal (blind observations). Expanded FOB methods, scoring definitions and proficiency data for the observer is included in the report.

Hand-held and Open-Field Observations
Hand-held and open-field observations included a careful physical examination and sensory evaluation according to an established format (Table 4).

Grip Performance
Hindlimb grip performance was tested according to the procedure described by Mattsson et al. (1986). Briefly, the rats were selected in a random manner and given to the observer in such a way that the observer did not know the treatment status of the animal. The observer then placed the rat's forelegs on a bench and the hindfeet were set on a horizontal screen attached to a strain gauge. The observer then smoothly but firmly pulled backward on the tail until the rat's grip on the screen was broken. The electronic strain gauge, linked to a personal computer, recorded the rat's resistance to the pull in grams. The average of three trials was used for statistical analysis. Forelimb grip performance was similarly tested. In this application, a bench was not used, and the rats were placed so that the forefeet were on the screen. The test sequence was the same as for hindlimb testing.

Landing Foot Splay
The landing foot splay procedure was similar to the procedure published by Edwards and Parker (1977). The rats were selected in a random manner and given to the observer in such a way that the observer did not see the treatment status of the animal. The outermost toe of each hindfoot was marked with ink. The animal was then dropped from a height of approximately 30 cm onto the recording sheet. This was repeated three times, and the toe was re-inked, as necessary. The distance from center-to-center of the ink marks was measured (cm) and the average of the three splay values was used for statistical analysis.
Motor Activity System
Twenty-four motor activity (MA) cages visually isolated from each other were located in a quiet, light-attenuated room. Each MA cage consisted of a clear plastic circular alley similar to the one developed by Fontaine et al. (1966) and Richelle et al. (1967). An infrared photobeam bisected the cage so that the beam crossed the alley in 2 locations. Each animal was tested individually for motor activity and no entry into the test room was allowed during the testing period. All test sessions consisted of six 8-minute epochs, totaling 48 minutes of testing per animal per test session. This duration was chosen based on the results of a validation study indicating that performance of control animals approached asymptote in 30-40 minutes (Shankar, 1996). Total activity counts for each epoch were recorded. Each beam break that lasted more than 100 msec, following at least a 100-msec interval between beam breaks, constituted an activity count. These minimum durations were set to discount activities such as tail-flicking, head-bobbing, etc. Motor activity was monitored by a computerized system (DEC PDP11 microcomputer with the SKED-11 Software System and the Micro/RSX Operating System) located in an adjoining room.
Positive control data for the motor activity test is included in the study report.

Motor Activity Chamber Calibration
Cages to be used for testing were calibrated prior to testing each day. Calibration was performed with a rod (attached to a motor rotating at a constant speed) that broke the infrared beam several times in each chamber. Equivalence of photocell function across chambers was ensured by ascertaining that the differences between any two of the individual average times that the rod passed by each photocell did not exceed 4 centiseconds.

Motor Activity Cage Allocation
The rats were allocated to the motor activity cages in such a way that counterbalancing of treatment groups and sexes across cages and across test times was maximized.
Surgical implantation of epidural electrodes was scheduled about 12 weeks into the study. Surgery was completed in counterbalanced order and was accomplished in 5 days. Anesthesia was induced by methoxyfluorane inhalation and maintained with isoflurane. A sterile ophthalmic ointment was applied to the cornea and the eyelids were closed with skin clips to prevent dehydration. Once placed in the stereotaxic instrument (#900, David Kopf Instruments, Tujunga, CA), a rectal thermistor was inserted about 7 cm into the rectum to monitor core temperature during surgery. Body temperature was supported with a water blanket Epidural electrodes (7 rnm long, #0-80, stainless steel set screws) were surgically implanted into the skull and supported with dental acrylic. The top 3mm of the set screws remained exposed above the acrylic so recording leads could be directly attached to the screws (an electrical plug was not used). The somatosensory electrode was placed 2.0 mm posterior and 2.0 rnm lateral left of bregma, the visual cortex electrode was placed 1.5 mm anterior and 3.0 mm lateral right of lambda, and the cerebellar electrode was located 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.

Electrodiagnostic System
The electrodiagnostic system was a Nicolet Pathfinder II (Nicolet Biomedical Instruments, Madison, WI, USA). Data sweeps (msec segments of EEG) were digitally sampled 512 times and averaged by an online computer. Rats were physically restrained during electrodiagnostic testing (approximately 40 minutes). Rats were tested two at a time, counterbalanced for sex and exposure level. Rectal temperature was recorded immediately before and after electrodiagnostic testing. In addition, tail temperature was recorded before the caudal nerve test.

Electrodiagnostic Tests
The following electrodiagnostic data were collected, in sequence, within 1 to 4 days after the last exposure:
flash evoked potentials (FEPs),
auditory brainstem responses (ABRs),
somatosensory evoked potentials (SEPs),
caudal nerve action potentials (CNAPs). (see Table 5)

Flash Evoked Potentials (FEP): Low and medium intensity FEPs were recorded from the visual cortex (FEP-V) and from the cerebellum (FEP-C). The source of the rat FEP-V appears limited to the cerebral cortex, with no waves linked directly to input pathways. Therefore, alterations in early components reflect changes either peripheral to the first cortical neurons (i.e., retina, optic nerve, lateral geniculate), or at the first cortical neuron itself. Altered later components indicate functional changes in cortical-subcortical networks. Thus, FEPs provide a simple functional evaluation of the visual system, but limited anatomic localization.

Auditory Brainstem Response to Clicks (ABRc): ABRc was recorded from the electrode over the cerebellum. The recording method allows isolation of far-field electrical activity, and therefore, ABRs provide information on the functional integrity of 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 (Mattsson et al. 1992).

Auditory Brainstem Response to Tone Pips (ABRt): ABRs to tone pips were recorded from the electrode over the cerebellum. The following frequencies were tested: 10kHz (ABR10) and 30 kHz (ABR30). Tone-pip ABRs provide information on auditory function at higher frequencies, and are used to screen for mid- to high-frequency hearing dysfunction.

Somatosensory Evoked Potentials (SEP-S and SEP-C): SEPs were recorded over the sensory cerebral cortex and cerebellar vermis. The SEP is primarily conducted along ascending fibers in the spinal cord dorsal column that lead to the thalamus and then to the cerebral cortex. Parallel processing also occurs in spinocerebellar and spinothalamic pathways. Early components of the SEP relate to input pathways, brainstem and thalamus. Longer latency potentials relate to cortical and cortical-subcortical networks. SEPs are generated by electrical stimulation of ventrolateral nerves at the base of the tail, and recorded at the somatosensory cortex (SEP-S) and from the cerebellum (SEP-C). The stimulating electrodes are a pair of small needles separated by 1 cm and are placed subcutaneously on the ventral aspect of the tail. The cathode is placed at the margin of the fur at the base of the tail and the anode is 1cm distal to the cathode.

Caudal Nerve Action Potentials (CNAP): Ventrolateral caudal nerves were stimulated near the tip of the tail and mixed nerve action potentials from single stimuli (CNAP1) were recorded at the base of the tail (Rebert, 1983). The stimulating and recording electrodes (separated by 9 cm) were mounted in a plastic tray. Subsequently, to provide data on the efficiency of peripheral nerve recovery (ability to generate a second action potential), the caudal nerves were stimulated with a pair of pulses (CNAP2) with an interstimulus interval of 3 msec.

Waveform Analyses
Digital Filtering
The following waveforms (primary data), collected with a broad-band analog filter, were digitally filtered using a computer routine (Nicolet Biomedical Instruments, Madison, WI, USA). The digital filters (high pass-low pass) were respectively set at 100-3000 Hz (ABRs); 1-250 Hz (FEP); 1-750 Hz and 85-750 Hz (SEP recorded with a 35-msec time base); and 1-500 Hz (SEP recorded with a 200-msec time base).

Qualitative Analyses
All individual and group composite waveforms were visually examined for changes in waveform shape, power, and latency.

Quantitative Analyses
An automated computer technique was utilized to analyze primary data (medium intensity FEP-V, ABRc, SEP-S and the CNAP) by quantification of differences from a template in waveform shape (optimal cross-correlation) and latency. In addition, the FEP-C was analyzed in the same way as the FEP-V. Power (area under the curve) was calculated for each waveform. For more analytical details, see Mattsson and Albee (1988). The template for each type of response was created by making a 'grand average' of all the control responses for males and females. A window (starting and ending point in msec) was established for each type of response. The window widths for automated computer scoring were:
FEP-V (med. inten.) = 27.6 - 109.2 msec. Captures components that reflect input to the visual cortex and early complex processing (Peaks N1 and N2).
FEP-V (med. inten.) = 109.2 - 258 msec. Captures mid-latency components, including N3, that reflect complex cortical-cortical and cortical-subcortical-cortical processing.
ABR (click) = 1.0 - 5.0 msec. Includes peaks I to V, acoustic nerve to upper brainstem.
SEP-S = 4.2 - 16.8 msec. Somatosensory input pathway, brainstem, thalamus and first cortical neuron.
SEP-S = 16.8 - 90 msec. Captures complex cortical-cortical and cortical-subcorticalcortical
CNAPI (single stimuli) = 3.2 - 8.4 msec. Captures the entire caudal nerve action potential.
CNAP2 (paired stimuli) = 6.6 - 11.8 msec. Captures only the second of the pair of action potentials.

The computerized analyses included similarity of shape (optimal cross correlation; reported as an r value) of an individual waveform to the template waveform, latency difference (phase shift in msec required to reach optimal cross-correlation), and amplitude measured as the total power of the peaks contained within the window (µvolts). Subsequent to computerized waveform evaluation, the computer's measurements of latency difference for each waveform for each animal were visually confirmed by a trained person 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 appropriate value as determined by the trained person was accepted and a new cross-correlation coefficient was calculated.

Five Week Post-Exposure FEP and ERG Testing (Males only)
FEP testing was conducted according to the methods used in the main study for males of all dose levels. Sample sizes were control = 6; low dose = 7; mid dose = 6; and high dose = 7. All waveforms were evaluated qualitatively but only FEP-C early component waveforms were statistically analyzed. Waveform analyses were the same as for the main study.

ERGs were conducted on control and high-dose male rats only. Animals were lightly anesthetized with isoflurane vapors. Electrodes were fashioned from EEG needle electrodes; the needles were blunted and shaped into a curvature to fit under the eyelid and behind the eye. The reference electrode was a subcutaneous needle on the shoulder, the ground electrode was a subcutaneous needle on the chin. The rat was placed in a box (15 in length x 10 in width x 8 in height) that was lined with white paper. The strobe (Grass photic stimulator Model PS 22D, Grass Medical Instruments) was set into the back of the box and the rat was facing away from the strobe (exposed to reflected light on the opposite wall). The strobe intensity was set on high and the rate was 0.7 flashes/sec and the number of data sweeps was 200. The ERG was band passed at 1-500 Hz with a 60Hz notch filter. Data were collected from individual eyes simultaneously (generating a left and right ERG).

ERG quantitative analyses
All ERGs were qualitatively evaluated. Waveforms from right and left eyes were averaged together for each animal. There was one exception; one high-dose rat had a flat ERG in one eye and the ERG from the 'good' eye was used for analysis. An ERG from one high-dose rat was extremely different from all other control and high-dose ERGs (very large with pronounced oscillations and a prolonged b-wave). Quantitative analyses were conducted both with and without this animal's ERG. ERGs were high-pass filtered above 50 Hz to maximize the resolution of the a-wave and the b-wave oscillations. The amplitude of the a-wave was measured from baseline to the first negative peak and the first b-wave oscillation was measured from valley to peak. The amplitudes of the a-wave and the b-wave were statistically analyzed by Wilcoxon-Mann-Whitney Test (Siegel and Castellan, 1988).

Waveform Illustrations
Waveforms recorded from individual animals were averaged together to make waveform composites. A composite was made for each electrodiagnostic test at each exposure level for males and females.
Sacrifice and (histo)pathology:

At the end of the electrodiagnostic testing that immediately followed the last exposure (Week 14), 5 rats/sex/dose that were randomly preselected for neuropathology segment of the study, were heparinized approximately 10 minutes prior to perfusion (0.2 ml heparin 10,000 USP/ml/100 gm body weight intraperitoneally) and were anesthetized by methoxyflurane vapor inhalation. Rats were perfused intracardially with 0.05M phosphate buffer containing 0.7% sodium nitrite, followed by a phosphate-buffered solution of 1.5% glutaraldehyde - 4% formaldehyde (c. 540 mOsM). A complete gross examination of tissues was conducted on all animals by a veterinary pathologist.

Representative samples of tissues as listed in the protocol were collected and preserved in 1.5% glutaraldehyde - 4% formaldehyde (See Table 6).

The necropsy consisted of an examination of the external tissues and orifices. The head was removed, the cranial cavity opened and the brain, pituitary and adjacent cervical tissues were examined. The nasal cavity was flushed with the same fixative used for perfusion and immersion. The skin was reflected from the carcass, the thoracic and abdominal cavities were exposed and the viscera were examined in situ. All visceral tissues were dissected from the carcass and re-examined. The brain, head, spinal column with spinal cord, fore- and hindlimbs, and tail were trimmed to remove excessive skin and muscle as necessary and immersed in fixative (phosphate-buffered solution of 1.5% glutaraldehyde - 4% formaldehyde). Muscles from the hindlimbs were reflected to further expose the nerves. Thoracic and abdominal viscera and the dermal test site were also saved in the fixative.

Tissues for neuropathologic evaluation (Table 7) were prepared from all perfusion-fixed rats in the control and high-dose groups. Tissues from animals exposed to 1 and 10 mg/kg/day were not examined microscopically. Nine transverse sections of the brain were prepared from the: olfactory lobe, cerebral cortex (frontal, parietal, temporal and occipital lobes), thalamus/hypothalamus, midbrain, pons, cerebellum, medulla oblongata, and nucleus gracilis/cuneatus. The following tissues were also prepared: trigeminal ganglia and nerve, pituitary gland, eyes with retina and optic nerves, spinal cord (cervical and lumbar), olfactory epithelium and skeletal muscles (gastrocnemius and anterior tibial). Additional tissues (bone, lacrimal gland and oral tissues) accompanied the required tissues and were also evaluated. Tissues from the central nervous system and sections of skeletal muscle were embedded in paraffin, sectioned approximately 6 µm thick, and stained with hematoxylin and eosin. Peripheral nerves (sciatic, tibial and sural) and dorsal root ganglia with spinal nerve roots (cervical and lumbar) were osmicated, embedded in plastic, sectioned approximately 2 to 3 µm thick, and stained with toluidine blue. All tissues were examined by a veterinary pathologist using a light microscope.

Histopathologic findings were subjectively graded to reflect the severity of specific lesions to evaluate the contribution of a specific lesion to the health status of an animal. Very slight and slight grades were used for conditions that were present in excess of the normal textbook appearance of an organ/tissue, but were of minimal severity and usually with less than 10% involvement of the parenchyma. This type of change would not be expected to significantly affect the function of the specific organ/tissue involved nor have a significant effect on the overall health of the animal. A moderate grade was used for conditions that were of sufficient severity and/or extent (up to 50% of the parenchyma) that the function of the organ/tissue may have been adversely affected, but not to the point of organ failure. The health status of the animal may or may not have been affected, depending on the organ/tissue involved, but generally lesions graded as moderate would not be life threatening. A severe grade which would have been used for conditions of even greater severity than moderate was not required on this study.

Following EP testing after 5 weeks post-exposure, eyes and brains from all male rats were saved in Bouin's solution and Formaldehyde respectively but were not histopathologically evaluated.
For overall FOB summarization and subjective evaluation, the data were the average ranks for each FOB observation (for males and females at each dose level). For statistical analyses, the incidence of ranked FOB observations, between control and each treated group (for each sex separately), were evaluated by a test of proportions at a = 0.02 (Bruning and Kintz, 1977). Due to the high correlation between different ranks (within an observation), if more than one rank within the same observation had significant z scores, the toxicologically most significant rank was reported.

Means and standard deviations were calculated by sex for all continuous data and homogeneity of variance was evaluated with the Bartlett's test (a = 0.01; Winer, 1971). No instances of heterogeneity occurred that, upon examination of the data, merited inferential statistical analyses of transformed data or removal of outliers.

The study design had two sexes and four major data collection periods: pre-exposure, months 1, 2 and 3. Initial statistical analyses, therefore, were factorial repeated-measure analyses to account for data from both sexes at all time periods in one statistical analysis. By using sex as a factor, the statistical power of the test was increased. The inclusion of pre-exposure data in the analysis makes relevant only the analyses which include factors of both treatment and time. The following interactions were studied:
Treatment x Time -- A significant p value indicates that, taken together, both males and females were affected by treatment at some time interval.
Treatment x Time x Sex -- A significant p value indicates that treatment effects were different between males and females at some time interval.
Treatment x Time x Epoch (motor activity only) - A significant p value indicates that treatment effects were different amongst the different epochs at some time interval.

Results and discussion

Results of examinations

Clinical signs:
effects observed, treatment-related
mortality observed, treatment-related
Body weight and weight changes:
effects observed, treatment-related
Ophthalmological findings:
effects observed, treatment-related
Behaviour (functional findings):
effects observed, treatment-related
Gross pathological findings:
effects observed, treatment-related
Neuropathological findings:
effects observed, treatment-related
Details on results:
Cageside and Clinical Observations
One female high-dose rat had whole-body tremors for several minutes during handling and treatment during one treatment session of the second week of treatment. Neither this rat nor any other had tremors at any other time period, and the cause remained unknown but was considered to be unlikely to be a response to treatment.
There were no other significant observations in any rat at any other time period. Cageside and clinical observation data and individual animal data can be found in the final study file.

Dermal Grading
Male high-dose rats had well-defined erythema, edema, and scaling during the first week of treatment. Male high-dose rats had a small to moderate amount of scabbing and moderate to severe scaling during the second week of treatment. The severity of the high-dose skin lesions decreased over subsequent weeks to very slight erythema, edema and scaling; the incidence of scabbing dropped to a low level. With only occasional exceptions, female high-dose rats had only very slight erythema, edema, a small amount of scabbing, and slight to severe scaling. In general, mid-dose male and female rats had low incidence of only very slight erythema and slight scaling. Male mid-dose rats also had very slight edema during the eighth and ninth weeks of the study. Skin condition was comparable in both control and low-dose rats; both groups had a low incidence of very slight erythema, slight scaling and a small amount of scabbing. Although control (and low-dose) rats had a small amount of scabbing, as did high-dose rats, only high-dose rats approached the upper end of the scale (i.e., 25% of treated area). Dermal scoring summary data are presented in “Other Findings”. Individual dermal scoring data are reported in the final study file.

From week 3 of the study, a tan color on the skin at the site of dosing was observed in most of the rats. While these coloration effects may have been related to AGE (although male controls had some coloration noted), this finding was believed to not be toxicologically significant.

No treatment-related effects were seen in body weights at any time during the study (Treatment x Week p = 0.0524). The sexes did not respond differently as a function of treatment (Treatment x Week x Sex p = 0.1680). Results of the data analysis and individual animal data can be found in the final study report.
Although the primary interaction term of Treatment x Time was not statistically significant, visual examination of the descriptive statistics revealed an apparent reduction in body weights of all treated males compared to controls (about 6%, 5% and 9% below controls for low-, middle- and high-dose groups respectively). These body weight differences in treated males were not attributed to treatment because of the lack of dose response and the lack of body weight effects in a nearly concurrent 13week dermal toxicity study of the same test material and dose levels in Fischer-344 rats
(McGuirk and Johnson. 1997).

Prior to the start of the study and prior to necropsy, eye examinations were conducted via indirect ophthalmoscopy. Animals with eye anomalies were excluded from the population prior to randomization and assignment to study based on the pre-study eye exams. Pre-necropsy eye exams did not reveal any observation suggestive of a treatment-related effect.

Functional Observational Battery
Hand-held and Open-field Observations
FOB observations are summarized as average ranks; incidence distribution of observations across all ranks; and individual animal observations can be found in the final study file.
There was a low incidence of categorical (present or absent, non-ranked) observations. A few rats in control and different dose groups had minor 'abnormalities' of the eyes. These eye abnormalities were white spots and streaks that generally relate to histopathologic observations of corneal mineralization (discussed in neurohistopathology results section). The lack of dose response relationship indicates that these findings in the eyes were due to normal variation and not due to treatment.

Abstracted Text Table 1, Abstracted Overview of FOB Observations, shows (a) those categorical (+ or -) observations that may be treatment related, (b) average ranks that differed from controls by more than 0.5 rank, (c) those ranked observations that had incidence patterns significantly different from control, and (d) data on response to sharp noise from Week 8 presented because they indicated a lower responsiveness to noise for control rats.
Overall, there were no FOB observations that appeared to be due to treatment. There were no categorical observations of note, and for average ranks of observations, there was only one value that differed from control by greater of 0.5 rank. This difference occurred in 'responsiveness to tail pinch' at Week 8, and occurred between male control and the low-dose group. This difference was attributed to chance. Other differences between average ranks were less than 0.5 and showed no patterns suggestive of a treatment-related effect.
There was only one statistically significant (z-test for proportions) difference for ranked observations, and this occurred in the pre-exposure period. In the pre-exposure period, male control rats were less 'reactive to handling' than were mid-dose male rats. The only other difference in ranked observations that came close to statistical significance was in Week 8, when control male rats appeared to be slightly less responsive to sharp noise than rats in the treated groups. This difference in responsiveness was considered random since it occurred only at Week 8, only in males, did not have dose-response. and did not attain statistical significance.
Considering that there were at least 384 pairwise comparisons, the 'positive' FOB findings in this study were well within the expected false-positive rate.

Text Table 1. Abstracted Overview of FOB Observations

Male Female
Dose(mg/kg/appl) 0 1 10 100 0 1 10 100
Number of rats 12 12 12 12 12 12 12 12
reactivity to handling (ave. rank)
2.6 2.7 3.0 2.8 2.8 3.0 2.8 2.6
none (rank 1) 0 0 0 0 0 0 0 0
minimal (rank 2) 5 4 0c 3 2 1 3 5
Moderate (rank 3) 7 8 12 9 10 10 9 7
pronounced (rank 4) 0 0 0 0 0 1 0 0
severe (rank 5) 0 0 0 0 0 0 0 0

Response to sharp noise (ave. rank) 2.2 2.6 2.5 2.6 2.8 2.8 2.8 2.6
incidence :
none (rank 1) 0 0 0 0 0 0 0 0
minimal (rank 2) 10d 6 6 5 3 4 4 5
moderate (rank 3) 2 5 6 7 8 7 6 7
pronounced (rank 4) 0 1 0 0 1 1 2 0
severe (rank 5) 0 0 0 0 0 0 0 0
Response to tail pinch (ave; rank)
3.6 3.0b 3.3 3.5 3.3 3.3 3.3. 3.3
none (rank 1) 0 0 0 0 0 0 0 0
minimal (rank 2) 0 4 0 0 1 1 0 1
moderate (rank 3) 5 4 9 6 7 6 8 6
pronounced (rank 4) 7 4 3 6 4 5 4 5
severe (rank 5) 0 0 0 0 0 0 0 0

Data selected for overview if (a) relevant clinical observations (none found). (b) average rank difference from control was 0.5 rank or greater, or (c) a ranked observation was statistically different from control (z-test of proportions, a =0.02), or (d) visual examination of the incidence tables indicated a somewhat lower responsiveness to noise for control rats.

Grip Performance
No treatment-related effects were seen in hindlimb (Treatment x Week p = 0.5877) or forelimb (Treatment x Week p =0.5140) grip performance at any time during the study. The sexes did not respond differently as a function of treatment (Treatment x Week x Sex, hindlimb p = 0.4334; forelimb p = 0.7347).

Landing Foot Splay
No treatment-related effects were seen in landing foot splay at any time during the study (Treatment x Week p = 0.6932). The sexes did not respond differently as a function of treatment (Treatment x Week x Sex p =0.5296).

Rectal Temperature
No treatment-related effects were seen in rectal temperature at any time during the study (Treatment x Week p = 0.6903). The sexes did not respond differently as a function of treatment (Treatment x Week x Sex p = 0.1219).

Motor Activity
No treatment-related effects were seen in motor activity at any time during the study (Treatment x Week p = 0.8962; Treatment x Week x Epoch p = 0.9452). The sexes did not respond differently as a function of treatment (Treatment x Week x Sex p = 0.7824).
Treatment did not affect habituation across time (Treatment x Week x Epoch = 0.9452).

Body temperatures were recorded just before and after collection of electrodiagnostic tests. Because physiological responses are sensitive to differences in body temperature, these individual animal temperatures were used as a covariate for analysis-of-variance testing. Tail temperatures were recorded during the CNAP testing. There were no treatment-related differences in rectal temperature (Treatment p = 0.9424; Treatment x Time p = 0.2715) or tail temperature (Treatment p = 0.5671). The sexes did not respond differently as a function of treatment (rectal, Treatment x Time x Sex p = 0.8041; tail, Treatment x Sex p = 0.4114). A significant value for the covariate indicated that temperature had a significant impact on the evoked potential data.

Flash Evoked Potentials (FEPs)
FEP-V (visual cortex) and FEP-C (cerebellum) data can be found in the study report.
Overall, FEP waveforms from both sexes and all treatment levels were very well formed and 'normal' looking. Although statistically significant differences did occur, none of the waveforms appeared distorted or 'pathological' per se. The results for the FEPs are complicated and are presented in Text Table 2. In general, mid- and high-dose male FEPs were smaller and female high-dose PEPs were larger than controls. The divergence in amplitudes, males small and females large, triggered several statistically significant treatrnent-by-sex interactions that caused subsequent analyses to be done for each sex separately.

Early-component FEPs: Although well formed and 'normal' in structure, visual examination of FEPs indicated dose-related and sex-related qualitative differences in waveforms. The early-latency components of the cerebellar flash evoked potential (FEPC) were significantly smaller in mid- and high-dose male rats. There were similar but less distinct differences (statistically non-significant) in the visual cortex (FEP-V). The overall (male and female) statistically significant difference in FEP-C early-component shape (p = 0.0319) was attributed to slightly poorer shaped FEP-Cs of male rats, since examination of shape correlation values indicates dose-response decrements in shape for males and only minor differences in female correlation values. In addition, female high-dose correlation values were slightly better than for controls. Although some statistical changes occurred in shape, these differences in shape likely were related to the amplitude differences and, therefore, amplitude was considered the main feature for diagnostic attention.

Text Table 2: Descriptive Summary of FEP ResuIts (hligh-dose compared to control)
Variable Early components Mid-latency components
Male FEP-C (cerebellar) 29% smaller (sig.) 24% smaller (not sig.)
Female FEP-C 37% larger (not sig.) 55% larger (sig.)
Male post-expo FEP-C 24% smaller (not sig.)
Male post-expo ERG 38% smaller (sig.)
Male FEP-V (vis. cortex) 25% smaller (not sig.) 31%smaller (not sig.)
Female FEP-V about the same (not sig.) 67% larger (sig.)

Retinal hypothesis and electroretinograms (ERGs): Because the first negative waves of FEPs at about 30 msec in male high-dose rats appeared to be smaller in both the FEP-V (visual cortex) and FEP-C (cerebellum), this raised a possibility that the source of the difference was the eye or optic nerve as a common source to both waveforms. The effect was most apparent in the PEP-C. Consequently, at about 5 weeks post-exposure; the PEPs of remaining (n = 6 or 7 per group) male rats were re-tested, and early FEP-C components analyzed quantitatively. Although no statistically significant differences persisted, the overall dose-response patterns were still evident. Because the dose-response patterrn was still present, ERGs were collected from control and high-dose male rats. It was hypothesized that if the retina was involved, the ERG would be smaller in high-dose male rats (a one-sided hypothesis). The initial small negative wave of the ERG (the a-wave) is conceptually attributed to photoreception, and the next, large positive wave (b wave) to signal processing in inner segments of the retina.
ERGs were quite variable from rat to rat, and sometimes between eyes of individual rats. The a-waves were quite small, possibly due to albinism, but the b-waves were very distinct. One high-dose rat (6592) had a clear ERG in the left eye, and a flat ERG in the right eye (presumed blind in the right eye). Ophthalmic examinations revealed that the right eye of this rat had a pale fundus and an absence of fundic blood vessels. One high-dose rat had a very large and oscillating ERG with a prolonged b-wave, which was quantitatively different from any of the other ERGs of treated or control rats. This exceptional ERG was not included in the composite ERG, but was shown separately. Statistical analyses were done both with and without this exceptional ERG. It should be noted, however, that individual ERGs of two control rats had a hint of the same shape; i.e., a slightly prolonged b-wave (waveforms in study file). Consequently, it is not known if the exceptional ERG of the high-dose rat is related to treatment or is simply an exaggerated stage of a phenomenon that occurs naturally in albino rats. Individual ERG waveforms can be found in the study report.
Mathematical filtering was done on the ERGs (band-pass 50 to 100 Hz) to isolate and enhance the a-wave and the oscillations on the b-wave. The a-wave of high-dose rats was about 20% smaller than controls (Wilcoxon, p = 0.0364 one-sided), in spite of the fact that the one large oscillating ERG in the treated group was much larger (146%) than the largest control. It seemed more appropriate to consider this very large oscillating ERG as an outlier, and conduct the statistical analysis without this value. Removing this outlier, with sample sizes of 7 control and 6 treated, the high-dose rat ERGs were 38% smaller than control (Wilcoxon p = 0.0041 one sided).
It was concluded that a smaller ERG suggested a retinal source for the smaller FEP-C
early components in high-dose (and by association, mid-dose) male rats.

Female mid-latency FEPs: Although the FEPs of all treatment groups were well formed and 'normal' in appearance, the mid-latency components of the female FEP-V and FEP-C were larger in the high-dose rats Even though not statistically significant, it is likely that the larger FEP-C early components were functionally linked to the significantly larger FEP-C mid-latency components. This early-late component amplitude relationship also existed for the female FEP-V, but not to the same degree as for the FEP-C.

Male and female early FEP processing: Although the data are much weaker in female
high-dose rats, it appeared that both male high- and mid-dose rats and female high-dose rats had some form of altered early waveform processing. There was strong evidence for divergent responses in males and females, with smaller male waveforms and larger female waveforms. Smaller ERGs in high-dose males pointed to the retina as a plausible source of the smaller male waveforms. It is important to note that histopathologic examination of retinas from high-dose male and female rats did not show any treatment-related pathologic. Consequently, although the PEP and ERG data are suggestive of altered retinal function (at least in males), the data supporting this mechanism are not strong and both male and female FEPs need to be interpreted with caution.

Ancillary data (low intensity FEPs): Qualitative evaluation of ancillary PEPs did not reveal any new dimension to the exposure-related differences in medium intensity FEPs.

Auditory Brainstem Response (ABR)
Visual inspection of ABR clicks and tone-pips did not reveal qualitative differences between treatment and control groups in either males or females.

Click Auditory Brainstem Response (ABRc)
Quantitative evaluation did not reveal any treatment-related differences for either males or females. Waveforms were robust, well shaped and had similar latencies among all groups.

Tone-pip Auditory Brainstem Response (ABR10 kHz & ABR30 kHz)
There was no evidence of ototoxicity based upon visual inspection of Peak I (acoustic nerve) of tone-pip ABRs. Waveforms were robust, well shaped and had similar latencies among all groups.

Somatosensory Evoked Potentials (SEP)
Visual inspection of SEPs recorded from the cerebellum (SEP-C) and from the somatosensory cortex (SEP-S) showed robust and well shaped waveforms that had similar peak latencies for males and females in all groups. No quantitative differences were statistically identified between treatment and control groups.

Caudal Nerve Action Potentials (CNAP)
Visual inspection of CNAPs (CNAP Single-pulse; CNAPs paired-pulse) showed robust and well shaped waveforms that had similar peak latencies for males and females in all groups. No quantitative differences were statistically identified between treatment and control groups.

SEP early-latency waveforms from two control rats (control male 97A6550; control female 97A6595) were too poorly formed to have recognizable peaks (shape-correlation values of 0.00) and therefore neither latency nor shape correlation values could be used in statistical analyses.

There were no treatment-related gross pathologic observations in rats exposed to AGE.

In addition, there were no treatment-related histopathologic observations in the central or peripheral nervous systems of rats exposed to AGE. Individual animal pathology data are presented in the study report. All histopathologic observations (not within norrnal limits) were interpreted to be spontaneous or iatrogenic findings and were not attributed to AGE. Many of the neural lesions observed were consistent with common spontaneous lesions of Fischer 344 rats (Eisenbrandt et al., 1990). These lesions consisted of very slight degeneration of individual nerve fibers in the trapezoid body of the medulla oblongata, trigeminal nerve, white matter of the cervical and lumbar spinal cord, and swollen axons in the nucleus gracilis/cuneatus, pituitary, or spinal cord. Mineralization of the cornea of the eye and nasal mucosa were also noted in the majority of control and AGE-treated rats. Other less common spontaneous neural and non-neural lesions were also observed in the tissues of these rats.
Numerous rats had slight to moderate inflammation (subacute to chronic) of the calvarium where the epidural electrodes had been surgically implanted. In some rats, very slight electrode-associated inflammation (subacute to chronic) also involved the dorsal meninges of the brain (cerebrum). These lesions were interpreted to be iatrogenic.

In summary, there were no gross or histopathologic alterations observed in the central or peripheral nervous system tissues which were attributed to AGE.

Study Overview
Overall, the only treatment-related findings were skin irritation in mid- and high-dose rats, more evident early in the study, and some alterations in the flash evoked potentials (FEPs) of mid- and high-dose male and high-dose female rats. Male rats were tested for FEPs five weeks post-exposure, and electroretinograms (ERGs) also were recorded from control and high-dose male rats at this time. The same FEP dose-response pattern still was present post-exposure, and the ERGs suggested a mild decrease in activity of the retina. Other findings were sporadic and did not demonstrate exposure-response nor time-response patterns suggestive of treatment.

Feed and Water Analyses
Feed was determined to be nutritionally adequate for rats, and the feed or water did not contain unacceptable levels of contaminants.

Summary of Treatment-related Results
Parameter Dose (mg/kg/application)
0 1 10 100
Skin (treated area) - - + ++(m) + (f)
Body weight - - - -
Clinical Observations - - - -
FOB (observations) - - - -
Grip Performance - - - -
Landing Foot Splay - - - -
Rectal Temperature - - - -
Motor Activity - - - -

Evoked Potentials:
Flash evoked (FEP) - - + (m) +
Somatosensory (SEP) - - - -
Auditory brainstem (ABR) - - - -
Click ABRs - - - -
Tone-pip ABRs - - - -
Caudal nerve (CNAP) - - - -
Flash evoked (FEP)*
5wk post exposure - - + (m) ** + (m) **
(5 week post exposure)* - nd nd + (m)

Neuropathology (lesions) - nd nd -

Effect levels

Dose descriptor:
Effect level:
Remarks on result:

Applicant's summary and conclusion

Skin effects, particularly of male high-dose rats were significant and indicated that a maximum tolerated dose was achieved. Other than skin, no treatment-related effects were noted on FOB, MA, clinical examinations or neurohistopathology. Mild treatment-related changes occurred in the flash-evoked potentials (FEP) in mid- and high-dose male rats, and high-dose female rats (at doses irritating the skin). Electroretinograms of high-dose male rats were smaller than control male rats. However, both male and female FEPs need to be interpreted with caution and the significance of these FEP findings for human risk assessment is uncertain. The NOEL for the study was 1 mg/kg/application, based on mild skin effects in mid-dose rats and mild FEP alterations in mid-dose male rats.
Executive summary:

Skin effects, particularly of male high-dose rats were significant and indicated that a maximum tolerated dose was achieved. Other than skin, no treatment-related effects were noted onFOB, MA, clinical examinations or neurohistopathology. Mild treatment-related changes occurred in the flash-evoked potentials (FEP) in mid- and high-dose male rats, and high-dose female rats (at doses irritating to the skin). Electroretinograms of high-dose male rats were smaller than control male rats. However, both male and female FEPs need to be interpreted with caution and the significance of these FEPfindings for human risk assessment is uncertain.

The NOEL for the study was 1 mg/kg/application, based on mild skin effects in mid-dose rats and mildFEPalterations in mid-dose male rats.