Lack of Selective Developmental Neurotoxicity in Rat Pups from Dams ...

57, 250 –263 (2000) Copyright © 2000 by The Dow Chemical Company

TOXICOLOGICAL SCIENCES

Lack of Selective Developmental Neurotoxicity in Rat Pups from Dams Treated by Gavage with Chlorpyrifos Jacques P. J. Maurissen,* ,1 Alan M. Hoberman,† Robert H. Garman,‡ and Thomas R. Hanley, Jr.§ *The Dow Chemical Company, Midland, Michigan 48674; †Primedica, Argus Research Laboratories, Inc., Horsham, Pennsylvania 19044; ‡CVP, Murrysville, Pennsylvania 15668 – 0068; and §Dow AgroSciences, LLC, Indianapolis, Indiana 46268 Received October 30, 1998; accepted June 28, 2000

Pregnant Sprague-Dawley rats were given chlorpyrifos (O,Odiethyl-O-[3,5,6-trichloro-2-pyridinyl] phosphorothioate; CPF) in corn oil by gavage from gestation day 6 (GD 6) through lactation day 10 (LD 10) at dosages of 0, 0.3, 1, or 5 mg/kg/day in a developmental neurotoxicity study that conformed to U.S. Environmental Protection Agency 1991 guidelines. GD 0 was the day when evidence of mating was observed and postnatal day 0 (PND 0) was the day of birth. Toxicity was limited to the highest dosage level (5 mg/kg/day) and, in the dams, consisted of muscle fasciculation, hyperpnea, and hyperreactivity. A nonsignificant overall trend toward weight gain and feed consumption was also observed in the high-dosage dams, with a statistically significant Group ⴛ Time interaction for reduced weight gain in the 5-mg/kg/day group near the end of gestation. Although many developmental indices were normal, pups from high-dosage dams had increased mortality soon after birth, gained weight more slowly than controls, and had several indications of slightly delayed maturation. The early deaths and delayed maturation were attributed to maternal toxicity, though a possible contributing role of direct pup toxicity in delayed development cannot be eliminated. In spite of the apparent delay in physical development, high-dosage pups tested just after weaning had normal learning and memory as tested on a T-maze spatial delayed-alternation task. Habituation, a primitive form of learning, was tested in 2 tasks (motor activity and auditory startle) and was not affected. No overt effects were noted in either dams or pups at 1 or 0.3 mg/kg/day. Based on these data, chlorpyrifos produced maternal and developmental toxicity in the 5-mg/kg/day-dosage group. There was no evidence of selective developmental neurotoxicity following exposure to chlorpyrifos. Key Words: chlorpyrifos; developmental neurotoxicity; spatial delayed alternation; motor activity; auditory startle; habituation; learning; memory; safety evaluation.

Chlorpyrifos (O,O-diethyl-O-[3,5,6-trichloro-2-pyridinyl] phosphorothioate; CPF) is a commonly used organophosphorous insecticide that has been studied extensively. The database on CPF includes, among other studies, a developmental toxicity study and 2-generation reproduction study in rats (Breslin et 1 To whom correspondence should be addressed at The Dow Chemical Company, Health and Environmental Research Laboratories, 1803 Building, Midland, MI 48674. Fax: (517) 638-9863. E-mail: [email protected].

al., 1996) and a developmental toxicity study in mice (Deacon et al., 1980). There was no evidence from these studies that CPF causes a teratogenic response or is a selective developmental toxicant. The study described in this paper was conducted to collect supplementary information about the potential selective developmental neurotoxicity of chlorpyrifos to be used in risk assessment. This developmental neurotoxicity study expands upon the existing database for chlorpyrifos in that dosing was conducted during both gestation (GDs 6 –21) and the first 10 days of lactation (LDs 1–10), and that extensive neurobehavioral and neuropathological evaluations were conducted on the offspring. MATERIALS AND METHODS Overall Study Conduct The potential developmental neurotoxicity of CPF was evaluated in 2 major studies, conducted concurrently in different laboratories. The study reported here evaluated pups for developmental neurotoxicity after exposure of their dams to CPF during gestation and lactation. This study was designed following the requirements of the U.S. Environmental Protection Agency Pesticide Assessment Guidelines (US EPA, 1991). It followed the Guide for the Care and Use of Laboratory Animals (NRC, 1996), and complied with Good Laboratory Practice regulations (US EPA, 1989). In response to the U.S. Animal Welfare Act that was promulgated by the U.S. Department of Agriculture (CFR, 1985), the Animal Care and Use Activities that were required for the conduct of this study were reviewed and approved by the Institutional Animal Care and Use Committee. A companion study evaluated CPF and CPF-oxon content of blood and milk, 3,5,6-trichloro-2-pyridinol (TCP) in blood, as well as cholinesterase (ChE) inhibition of plasma, erythrocytes, heart, and brain in dams, fetuses, and pups at several time points. These data are reported in a separate publication (Mattsson et al., 2000). Young adult Sprague-Dawley (Crl:CD威BR VAF/plus威) rats were obtained from Charles River Laboratories, Inc., Portage, MI, and bred in-house to provide approximately 20 pregnant dams per dosage level for the developmental neurotoxicity element of the study. An extra (satellite) group of 5 pregnant dams per dosage level was added for the determination of maternal plasma, erythrocyte, and brain ChE inhibition on GD 20 (to document exposure). GD 0 was defined as the day when spermatozoa were observed in a smear of the vaginal contents and/or a copulatory plug observed in situ; and postnatal day 0 (PND 0) was the day of birth. Bred rats were randomly assigned to the various elements of this study. The test material (CPF) was 99.8% pure and was administered daily to dams by gavage in corn oil, from GD 6 through LD 10, at dosage levels of 0, 0.3 (100

250

CHLORPYRIFOS DEVELOPMENTAL NEUROTOXICITY

FIG. 1.

251

Schematic time line of the main events of the developmental neurotoxicity study.

times the chronic RfD), 1.0, and 5.0 mg/kg/day. Dosage volume was 1 ml/kg. The dose was adjusted daily for body-weight changes and given at approximately the same time each day (around 11 A.M.). Bred females were initially housed singly in stainless steel, wire-bottom cages beginning on GD 0. Beginning no later than GD 20, bred female rats were individually housed in nesting boxes with corn cob bedding (Bed-o’Cobs, The Andersons Industrial Products Group, Maumee, Ohio). Each dam and litter were housed in a common nesting box during the postpartum period. After weaning, the pups were individually housed in stainless steel, wirebottom cages. An automatically controlled fluorescent light cycle was maintained at 12-h light:12-h dark, with each dark period beginning at 7 P.M. Rats were given Certified Rodent Diet #5002 (PMI Nutrition, International, St. Louis, MO) available ad libitum from individual feeders. Local water that had been processed by passage through a reverse osmosis membrane was available to the rats ad libitum from an automatic watering system or individual water bottles. Chlorine was added to the processed water as a bacteriostat. On PND 0, all pups in a litter were weighed (pup body weights were recorded after all pups in a litter were delivered and groomed by the dam). On PND 4, pups to be culled were selected with a table of random units. The remaining pups were individually marked with a tattoo on the paw, litters were reduced to 10 pups each, and one male and/or one female pup from each litter

was assigned (when possible) to subsets of rats that were grouped for specific evaluations. Whenever possible, the same number of male and female pups per litter were continued on study. Due to early pup deaths in the high-dosage group, on PND 4, 6 high-dosage litters had fewer than 10 pups. In an attempt to evaluate 20 litters/group with similar maternal and lactational demands in each dosage group, extra pups randomly culled from other high-dosage litters were added to those litters of less than 10 pups. When no other high-dosage pups were available, culled pups from other litters were used as needed to increase the litter size to 10 pups (7 pups from 0.3 mg/kg/day litters and 2 pups from control litters). When pups were taken from other litters and placed with high-dosage litters, their sole purpose was to keep all of the evaluated litters equal in size. These replacement pups were uniquely identified with a tattoo mark on the tail and were not used for any subsequent evaluation. The time line of the main study activities is presented in Figure 1. The subsets of rats formed for specific evaluations were as follows: Subset 1. Immersion-fixed brain weights, morphometrics, and neuropathology were evaluated in pups on PND 11. One male and one female pup/litter were first selected (i.e., 80 male and 80 female pups). Half of these (one male or one female/litter) were culled and examined for gross lesions at necropsy. The other half, i.e., 10/sex/dosage (one male or one female per litter) were

252

MAURISSEN ET AL.

retained for brain weight; of these brains, 6/sex/dosage were examined neuropathologically and morphometrically. The sample size per litter was therefore reduced to 4 males and 4 females on PND 11. Subset 2. Learning and memory were evaluated by a test of Spatial Delayed Alternation, which was conducted post-weaning (PNDs 22 to 24) and in the same pups as adults (PNDs 61 to 90). Sample size was 8/sex/dosage with only one male or one female pup/litter selected from 16 randomly chosen litters/dosage. Subset 3. An automated test of motor activity was conducted on PNDs 13, 17, 21, and subsequently on PND 60. Body temperature was measured on PNDs 21 and 60. Auditory startle was evaluated on PNDs 22 and 61. One male and one female pup from each litter were tested (n ⫽ 20/sex/dosage). Subset 4. Rats (n ⫽ 20/sex/dosage) were evaluated for body weights (PNDs 0, 4, 11, 17, 21, 39, and 65), clinical signs, day of pinna detachment, eye opening, and preputial separation or vaginal opening. On PNDs 65 to 70, 10 rats/sex/dosage (one male or one female/litter) were perfused and their brains weighed. Of these, the brains of 6/sex/dosage were examined neuropathologically and morphometrically. Evaluation of Dams Main study. Dams were examined before the daily treatment, inside and outside the cage, 3 to 4 h after dosing. An individual unaware of treatment group examined the rats for signs of autonomic dysfunction (lacrimation, salivation, palpebral closure, prominence of the eye, piloerection, respiration, urination, and defecation), abnormal postures, abnormal movements, or abnormal behavior patterns. The dams were evaluated for duration of gestation (GD 0 to the time the first pup was delivered), litter sizes (all pups delivered), and pup viability at birth. The dams’ behavior was observed daily when the pups were examined during the 22-day postpartum observation period. Body weights were recorded on GD 0, daily during the dosage period, and on the day sacrificed. Feed consumption values were recorded daily from GD 0 through LD 14, when it was expected that pups would begin to consume solid feed, confounding this parameter. On LD 21, rats that delivered a litter were sacrificed. A gross necropsy of the thoracic, abdominal, and pelvic viscera was performed and the number and distribution of implantation sites were recorded. Dams with no surviving pups were sacrificed after the last pup was found dead or missing and presumed cannibalized. Rats that did not deliver a litter were sacrificed on GD 25 and examined for gross lesions, pregnancy status and uterine contents. Uteri of apparently non-pregnant rats were stained with 10% ammonium sulfide to confirm the absence of implantation sites (Salewski, 1964). Satellite study. On GD 20, 5 dams per dosage level from the ChEinhibition satellite group were sacrificed via carbon dioxide inhalation 4 to 5 h after dosage, to document exposure. A more extensive study of ChE inhibition and quantitative analysis of chlorpyrifos, its oxon, and trichloropyridinol has been reported elsewhere (Mattsson et al., 2000). Blood was collected from the inferior vena cava, and the brain was removed. The maternal blood (5 ml) was collected into heparinized centrifuge tubes and centrifuged at 2500 g for 5 min, and the plasma was removed and stored undiluted at –70°C. The packed erythrocytes were washed with normal saline; the total volume was brought to 5 ml and re-centrifuged at 2500 g for 5 min. The supernatant was discarded and 500 ␮l of the packed cells was diluted to 5 ml with 0.1 M sodium phosphate buffer (pH 8.0 with 1% Triton X-100). Maternal brains were homogenized on ice [1 g tissue per 9 ml 0.1 M sodium phosphate buffer (pH 8.0 with 1% Triton X-100)]. The brain, plasma, and erythrocyte samples were frozen at –70°C until analyzed. Cholinesterase activity was measured using a radiometric assay (Johnson and Russell, 1975) with a final substrate concentration of 1.2 mM. The [ 3H]acetate produced by the hydrolysis of [ 3H]acetylcholine was measured using a Wallace 1410 scintillation counter with a counting efficiency of approximately 50%. Evaluation of Pups All pups were examined for vital status at birth (liveborn vs. stillborn). Pups that were found dead during the initial examination of the litters were evalu-

ated by removing and immersing the lungs in water. Pups with lungs that sank were considered stillborn; pups with lungs that floated were considered liveborn, and to have died shortly after birth. Each litter was evaluated for viability at least twice each day of the 22-day postpartum period. Dead pups were removed from the nesting box. When not precluded by autolysis or cannibalization by the dam, any pup found dead was necropsied and examined for the cause of death. The pups present in each litter were counted and physical signs (including gross external physical anomalies) were recorded once each day during the 22-day postpartum period. Pup body weight and sex were recorded on PNDs 0, 4, 11, 17, and 21 (all pups on PNDs 0 and 4 were used for body weight measurement). Pups were observed for viability at least twice daily during the postweaning periods. Body weights and clinical observations were also recorded on PNDs 39 and 65. Feed consumption was measured on PNDs 22 to 29, 39 to 46, and 58 to 65. Learning and memory. Spatial-delayed alternation was tested in rats in a manner similar to the method used by Stanton et al. (1994). The spatialdelayed alternation apparatus (Coulbourn Instruments, Inc., Allentown, PA) consisted of a Plexiglas威 T-maze that included a start box, a runway leading from the start box to the choice-point, and left and right runways extending from the choice-point to the left and right goals (supplied with liquid dippers). Light cream (with sweet condensed milk) or water was used as a reinforcer for pups or adults, respectively. All runways were fitted with hinged tops made of clear Plexiglas威. Automated guillotine-type doors were located at the start box and the left and right sides of the choice-point. The test was divided into discrete trials. The latency and the performance of the rat for each trial were computer-recorded. Pups were tested on PNDs 22–24, and the same pups were tested as adults on PNDs 61–90. The spatial-delayed alternation testing was conducted blind to treatment on feed and/or water-deprived rats. Pups were deprived of feed and water, but had access to both in their home cages for 3 h after the end of the last test session of the day, in an attempt to achieve a target of 85% free-feeding weight. At the end of each session, the pups were also eventually fed supplementary light cream to equate the amount of cream received as a reinforcement by all the groups during the test session. They were also supplemented with light cream when they fell below the target weight. Adult rats were deprived of water only, and were given access to it for half an hour, approximately one h after the last test session of the day. The 3 phases of testing included maze acclimation, acquisition training, and delay testing. Maze acclimation took place on PND 22 and PNDs 61 to 72, and was divided into 2 parts: goal box training and forced runs (12 trials). Goal box training allowed the rat access to the reward in both the right and left arm of the maze. Forced runs allowed the rat to access only one arm of the maze, where the reward was available (either the right or the left). Acquisition training was comprised of a forced run immediately followed by a choice run (7 blocks of 12 trials on PNDs 23 and 24, and 10 blocks of 6 trials on PNDs 73 to 82). In the forced run, the rat was given access to one arm of the T-maze that was baited. In the choice run, the doors to both arms of the maze were raised simultaneously. To have access to the reward, the rat had to choose the opposite arm from the one previously rewarded in the forced run (spatial alternation). The percent correct was calculated in blocks of 12 consecutive trials, and the learning curve was represented by the average percent correct (groups of 8 rats each) over successive blocks. Delay testing was conducted under the same conditions as acquisition training, with time delays of various durations added between the forced run and the choice run (4 blocks of 12 trials on PND 24, and 6 blocks of 6 trials on PNDs 85 to 90). Delays alternated every 6 trials. Each animal was evaluated every test day. The percent correct retention could be plotted at each delay (“forgetting” function). The memory component was evaluated by the retention curve and was represented by the slope corresponding to the percent correct retention at the tested delays. The intercept of the slope was an extrapolated value (time zero) and gave information about the non-mnemonic aspects of the task (e.g., changes in motivation, sensory-motor functions, attention, and encoding of information). On PND 24, the delays were set at 5, 15, and 25 s, while they were set at 5, 65, and 125 s for PNDs 85 to 90.

CHLORPYRIFOS DEVELOPMENTAL NEUROTOXICITY Motor activity. Motor activity was evaluated on PNDs 13, 17, 21, and 60 by counting beam breaks of a passive infrared sensor mounted outside a stainless-steel wire-bottom cage with Plexiglas威 flooring (40.6 ⫻ 25.4 ⫻ 17.8 cm). Each test session was one h in duration with the number of beam breaks tabulated for each 5-min interval. The apparatus monitored a rack of up to 32 cages and sensors during each session, with each rat tested in the same location on the rack across test sessions. Groups were counterbalanced (sex, dosage group) across testing sessions and cages. Data were collected to demonstrate that the test system was capable of detecting increases or decreases in activity produced by positive control substances. The calibration procedure for the motor activity equipment was performed at least semiannually. Auditory startle. The auditory startle test was conducted on PNDs 22 and 61. Animals were evaluated in sets of 4 within a sound-attenuated chamber, and the amplitude and latency of the startle response were measured (Coulbourn Instruments, Inc., Allentown, PA). Each rat was placed inside a cage situated above a platform containing a pressure transducer in its base. The cage had the overall shape of a hexahedron with an 8-cm base, 8-cm height and 16-cm length for younger rats, and with a 9-cm base, 9.5-cm height and 18.5-cm length for older rats. A microcomputer sampled the output of the pressure transducer and controlled the test session. The rats were initially given an adaptation period of five min. During the last minute of this period, ten “blank” trials (no startle auditory stimulus) were given to sample the baseline pressure exerted by the rat. The rats were then presented with 20-ms, 120-dBA bursts of white noise at 10-s intervals for 50 trials. An additional ten blank trials followed. The peak amplitude of each response was recorded, and the average response on baseline trials was subtracted to calculate the peak response. The average peak responses and latencies over 10-trial blocks were compared among the dosage groups. The calibration of auditory signals was performed prior to the start of the study. The calibration of the auditory startle pressure transducers was performed daily prior to usage. Groups were counterbalanced (sex, dosage group) across testing sessions and cages. Body temperature. Body temperature was measured on PNDs 21 and 60, immediately following the completion of motor activity testing. The measurement was made with a rectal probe (Physitemp Instruments Model RET-2) and digital thermometer (Physitemp Instruments Model Bat-10 R LOP). The digital thermometer was calibrated daily prior to usage. Physical signs of maturation. Pinna unfolding was monitored daily beginning on PND 1 (all rats in all litters) and continued until all pups (100%) had the pinna unfolded. In rats of subset 4, eye opening was monitored beginning on PND 11 and continued until each pup reached the criterion. Male rats were evaluated for the age at preputial separation beginning on PND 38. Female rats were evaluated for the age at vaginal patency beginning on PND 27. A body weight was recorded for the rat on the day the criterion was attained. Brain weight, histopathology and morphometrics. At PND 11, 10 male and 10 female pups (one per litter) were euthanized by carbon dioxide overexposure. The head of each pup was severed between the first two cervical vertebrae, the calvaria removed from the top of the skull, and the entire head placed into neutral buffered 10% formalin. After fixation, brains were removed and weighed. Brains from 6/sex/dosage were subsequently examined histopathologically. During PNDs 65 to 70, ten rats/sex/dosage group were randomly selected for brain weight measurements. These grown pups were administered a combination of heparin and an anesthetic (39 mg/kg sodium pentobarbital and 20,000 USP/kg sodium heparin) and perfused in situ with neutral buffered 10% formalin. Of these 10, 6 were randomly selected for neuropathologic evaluation. In addition to neuropathologic examinations on PND-11 and PND-65 rats, morphometric measurements (6 rats/sex/dosage) were made of the anterior to posterior length of the cerebrum and cerebellum, height of the cerebellum, thickness of the corpus callosum, frontal cortex, parietal cortex, caudateputamen, hippocampus, and in pups only, the thickness of the external germinal layer of the cerebellum. Linear measurements of the anterior-posterior dimensions of the cerebrum and cerebellum were made with a Vernier caliper. The brains were then weighed and divided into 6 coronal slices. The cuts were

253

made half way between the ventral base of the olfactory bulbs and the optic chiasm, through the optic chiasm, through the infundibulum, through the midbrain just posterior to the mammillary body, through the cerebellum just anterior to its midpoint, and finally through the anterior portion of the medulla. The anterior faces of these coronal brain slices were sectioned (7 ␮m), paraffin-embedded, and stained with hematoxylin and eosin. After processing, the slides were coded and evaluated, blind to treatment. Statistical Analyses Statistical analyses were performed with SAS/STAT娀 (SAS Institute, 1989) programs or by proprietary programs developed for computer use by the Testing Facility. All tests were 2-tailed. Individual motor activity counts were transformed to square roots. Bartlett’s Tests (Sokal and Rohlf, 1969) for homogeneity of variance were run at ␣ ⫽ 0.001. Parametric data were analyzed by factorial analysis of variance (ANOVA, factors of group and sex) or factorial repeated-measure analyses (RepANOVA, factors of group, sex, repeated for time), using the multivariate approach and the Pillai trace statistic (SAS Institute, 1989). The main effects and interactions of interest were: Group: A significant p value indicated that both male and female pups, taken together, differed in some way among the different groups, independent of time. Group ⫻ Sex: A significant p value indicated that male and female pups differed among the different groups (independent of time) in the way they reacted to the test compound. Group ⫻ Time: A significant p value indicated that both male and female pups, taken together, differed in some way among the different groups at some time interval. Group ⫻ Time ⫻ Block (motor activity and auditory startle): A significant p value indicated that group effects were different among the different blocks at some time interval. The Group ⫻ Sex ⫻ Time factor was also added by request, and examined post hoc. A significant p value indicated that group effects were different in males and females at some time interval. In the event of a statistically significant Group main effect (in the presence or absence of any other statistically significant interaction), one dosage level (beginning with the high-dosage group) was removed, and the same analysis was re-run. If an overall statistically significant effect was present, but disappeared when the high-dosage group data were removed, it was inferred that the statistically significant effect was due to the high-dosage group (Tukey et al., 1985). This approach minimizes the number of comparisons (and, therefore, of Type-I errors) and also of Type-II errors (by maximizing the sample size). The following sequential approach was used in an effort to further reduce Type-I and Type-II errors: If Group ⫻ Sex was significant, the same analysis was rerun on separate sexes. If the Group ⫻ Time interaction was significant, the data were examined to identify the time period responsible for the statistical significance. The Type I-error rate was set at 0.02 for all primary analyses. Sequential analyses were conducted at ␣ ⫽ 0.02, following a statistically significant primary planned analysis (consistent with the recommendations proposed by Tukey et al., 1985). When one male and one female from the same litter were used in a test, the litter was used as the unit for analysis. All p values were reported uncorrected. Brain morphometric data were analyzed by ANOVA for males and females separately, followed by Dunnett’s tests if the ANOVA was statistically significant. This approach was taken because the number of dosage groups examined was not the same, either in males and females, or at the 2 time points analyzed (i.e., male and female pups: morphometric data from control, low-, middle-, and high-dosage groups; male adults: control and high-dosage groups, female adults: control, middle- and high-dosage groups). Such a difference among the number of groups selected for morphometric analysis came from the fact that, at first, the pups from all dosage groups were prepared for morphometric analyses. However, for adults, the stepwise approach suggested in the EPA developmental neurotoxicity guideline (US EPA, 1991) was used, which recommends that control and high-dosage groups be examined first, and then, if evidence of alterations is found, lower dosage groups be examined in sequence. Analysis was performed on all tissues prepared for morphometric analysis. The Type-I error rate was set arbitrarily at 0.05 because this type of analysis has less statistical power than the repeated-measure analyses.

254

MAURISSEN ET AL.

TABLE 1 Summary of Clinical Observations of Dams Dosage (mg/kg/day) During gestation No. animals Fasciculations Ataxia Tremors Excessive salivation During lactation No. animals Fasciculations Hyperreactivity Hyperpnea

0

0.3

1.0

5.0

25 0 0 1 1

25 0 0 0 0

25 0 0 0 0

25 6a 1 0 0

25 0 2 0

24 0 7 0

24 0 2 1

24 16 a 17 a 8a

Note. Total number of rats showing clinical signs. a Statistically significant (␣ ⫽ 0.02). Developmental landmarks were evaluated using the Kruskal-Wallis (Sokal and Rohlf, 1969) (pinna unfolding, vaginal patency, preputial separation, eye opening; ␣ ⫽ 0.02). Clinical observations were evaluated using the Chi-square test for proportions (Snedecor and Cochran, 1967). More than 200 a priori p values were derived in this study, out of which approximately 40 were statistically significant. These figures provide a context for a better interpretation of the results.

RESULTS

Dams Gavage administration of 5 mg/kg/day of CPF to pregnant/ lactating female rats produced clinical signs of toxicity. Table 1 shows the total number of rats exhibiting clinical signs during GD 6 –21 and LD 1–21. Muscle fasciculation was noted during gestation (p ⫽ 0.001; in all cases these signs occurred postdosing on the last day of gestation) and during lactation (p ⫽ 0.001; primarily during the first 4 days of lactation). In addition, signs of hyperreactivity (p ⫽ 0.001) and hyperpnea (p ⫽ 0.001) were observed in a number of rats in the high-dosage (5 mg/kg/day) group during lactation. These clinical signs were observed in high-dosage animals immediately prior to delivery, when body weight was the highest and consequently the absolute dose in mg administered was the highest, and during the first few days of lactation. There were no clinical signs of treatment-related toxicity in rats given 0.3 or 1.0 mg/kg/day. Though sporadic incidences of hyperreactivity were noted in controls and the lower dosage groups, these were not dosagedependent and not considered related to treatment. All other clinical observations were infrequent, lacked a pattern, and were considered spontaneous occurrences unrelated to chlorpyrifos administration (missing/broken incisors, localized alopecia, hind limb, head, or tail lesions, red perivaginal substance [data not shown]). There were no overall statistically significant differences in dams’ gestation (p ⫽ 0.94; F(3, 93) ⫽ 0.13) or lactation (p ⫽ 0.48; F(3, 75) ⫽ 0.84) body weights (data not shown). Body

weight gains are depicted in Figure 2. Though there were no statistically significant differences in body weight gain over the entire gestation period (p ⫽ 0.09; F(3, 93) ⫽ 2.20; GDs 0 –20) in any dosage group, a statistically significant Group ⫻ Time interaction (p ⫽ 0.006; F(9, 279) ⫽ 2.64) was present in body weight gain, and the 5-mg/kg/day dosage group, toward the end of gestation (GDs 17–20), appeared to be responsible for this effect (Fig. 2). During lactation, no overall statistically significant differences in body weight gains were noted (p ⫽ 0.36; F(3, 75) ⫽ 1.08; LDs 0 –21), and the Group ⫻ Time interaction did not reach statistical significance (p ⫽ 0.04; F(12, 222) ⫽ 1.86). However, a transiently lower mean lactation weight gain was observed in the 5-mg/kg/day dosage group on LDs 0 –3 (6.8 ⫾ 10.7 g in the control group vs. 1.1 ⫾ 6.5 g in the high-dosage group). This reduced body weight gain toward the end of gestation and the beginning of lactation (even though not statistically significant) was considered treatment related. No differences in body weights and body weight gains were apparent in the 0.3 or 1.0 mg/kg/day dosage groups when compared to controls. There were no overall significant differences in dams’ absolute and relative feed consumption (relative to individual body weight) during gestation (p ⫽ 0.78; F(3, 91) ⫽ 0.36, and p ⫽ 0.64; F(3, 91) ⫽ 0.57) or during lactation (p ⫽ 0.04; F(3, 54) ⫽ 3.07, and 0.12; F(3, 54) ⫽ 2.02) in any treatment group (data not shown). During lactation, however, feed consumption was 4 to 15% lower in the high-dosage group, with the greatest effect during the period immediately following parturition. Table 2 presents ChE activity as a percent of control values, measured in a subgroup of dams sacrificed on GD 20 (satellite study). At 5 mg/kg/day of chlorpyrifos, brain and plasma ChE activity was markedly depressed (approximately 90%) when compared to controls, and RBC ChE levels were depressed almost completely. At 1.0 mg/kg/day, dams’ brain ChE activity was depressed approximately 18% relative to controls, while RBC ChE levels were inhibited approximately 84% and plasma levels were depressed approximately 69%. At the low

FIG. 2. Maternal body weight gains during gestation and lactation (mean ⫹ SD). Notice the reduced weight gain in the 5-mg/kg/day group during GDs 17–20 and LDs 0 –3. Maternal dosing took place from GD 6 through LD 10.

255


TABLE 2 Cholinesterase Activity in Dams on GD 20 Dosage (mg/kg/day) a Plasma Erythrocyte Brain

0.3

1.0

5.0

56.7 ⫾ 2.7 58.7 ⫾ 16.1 99.7 ⫾ 1.5

31.1 ⫾ 4.1 15.6 ⫾ 6.8 82.1 ⫾ 2.8

8.5 ⫾ 1.2 0.1 ⫾ 0.2 10.2 ⫾ 1.0

Note. Values are given as percentage of control ⫾ SEM. The control values are 861, 653, and 11,296 nmol acetylcholine metabolized/min/ml (or mg) in plasma, erythrocyte and brain, respectively. a Sample size ⫽ 5 rats/dosage group.

dosage of 0.3 mg/kg/day, there was no indication of effect on brain cholinesterase activity, while plasma and RBC ChE levels were approximately 45% lower than control values. Developmental Indices The developmental parameters measured in dams are presented in Table 3. All pregnant rats (at least 24 per group) delivered litters. The duration of gestation, the number of implantations per dam, the number of dams with stillborn pups, and the mean litter size (surviving pups) on PND 0 were comparable across the groups. A total of 3 dams had total litter loss (accounting for 31 pups) between PNDs 0 and 4. During PNDs 1– 4, there were 2, 2, 3 and 38 pups cannibalized in the 0, 0.3, 1, and 5 mg/kg/day groups respectively. This resulted in

a substantial decrease in litter size on PND 4, and a reduced viability index in the 5 mg/kg/day dosage group. Overall, 30% (88/292) of the fetuses born to dams given 5 mg/kg/day died during the PND 0 – 4 interval, compared to approximately 1–2% in all other groups. Following culling, there were no overall significant differences in mortality in any of the treatment groups. Statistically significant lower male and female pups’ body weights were seen at 5 mg/kg/day on PNDs 0 – 4 (p ⬍ 0.001; F(3, 163) ⫽ 11.15); a significant Time ⫻ Group interaction (p ⬍ 0.001; F(6, 326) ⫽ 5.14) may indicate a larger effect on a particular time period (i.e., PND 4). When pups’ body weights were evaluated over a longer period (i.e., PNDs 0 – 65), the analysis showed that they were lower in the high-dosage group in both males and females compared to controls (p ⬍ 0.001; F(3, 75) ⫽ 14.07) (Fig. 3). A significant Sex ⫻ Group effect (p ⫽ 0.02; F(3, 71) ⫽ 3.52) indicated that males and females were differentially affected by treatment. However, when the data were analyzed separately by sex, the analyses showed that the body weights of both males (p ⬍ 0.001; F(3, 73) ⫽ 11.47) and females (p ⬍ 0.001; F(3, 73) ⫽ 8.13) were significantly decreased at the high dosage. A Day ⫻ Group significant effect (p ⬍ 0.001; F(18, 216) ⫽ 2.80) indicated that the difference between groups for both sexes taken together changed as a function of day (e.g., more of a relative effect on PND 65 compared to PND 0). By PND 65, the weights of high-dosage female offspring were

TABLE 3 Summary of Developmental Parameters Dosage (mg/kg/day) No. animals No. pregnant No. litters Mean gestation length (days ⫾ SD) Mean implantations per dam (⫾ SD) No. dams with stillborn pups No. dams with total litter loss Mean litter size (pups ⫾ SD) Liveborn Stillborn PND 4 (live; pre-cull) Pup mortality PND 0 PNDs 1–4 Mean pup weights (g ⫾ SD) PND 0 Males Females PND 4 (precull) Males Females Viability index (%) b a b

0

0.3

1.0

5.0

25 25 25 23.1 ⫾ 0.5 13.6 ⫾ 3.2 1 0

25 24 24 23.2 ⫾ 0.4 14.5 ⫾ 2.0 1 0

25 24 24 23.0 ⫾ 0.5 14.2 ⫾ 2.4 4 0

25 24 24 23.0 ⫾ 0.2 14.2 ⫾ 1.6 1 3a

12.3 ⫾ 3.1 0.0 ⫾ 0.2 12.2 ⫾ 3.1

13.3 ⫾ 1.9 0.0 ⫾ 0.2 13.1 ⫾ 1.7

13.0 ⫾ 2.4 0.2 ⫾ 0.5 12.7 ⫾ 2.5

12.7 ⫾ 2.4 0.1 ⫾ 0.6 8.9 ⫾ 4.9 a

0/308 4/308

1/319 3/318

2/311 4/309

21/292 a 67/271 a

6.6 ⫾ 0.6 6.3 ⫾ 0.6

6.7 ⫾ 0.6 6.2 ⫾ 0.5

6.4 ⫾ 0.5 6.1 ⫾ 0.5

6.1 ⫾ 0.7 a 5.6 ⫾ 0.6 a

9.9 ⫾ 0.9 9.6 ⫾ 0.9 98.7

10.2 ⫾ 1.1 9.7 ⫾ 1.1 98.7

10.3 ⫾ 1.6 9.7 ⫾ 1.5 98.1

8.8 ⫾ 1.6 a 8.2 ⫾ 1.6 a 69.9 a

Statistically significant (␣ ⫽ 0.02). Number of live pups on PND 4 (preculling)/number of liveborn pups on PND 0.

256

MAURISSEN ET AL.

treated with 5 mg/kg/day (Table 4). The only developmental landmark which was identified as statistically significantly different from the controls was delayed vaginal opening (p ⫽ 0.01). However, pinna detachment (p ⫽ 0.03) and preputial separation (p ⫽ 0.05) also appeared to be somewhat delayed, though there were no statistically significant differences. There were no effects at either 0.3 or 1.0 mg/kg/day in any of these measures. Brain Measurements

FIG. 3. Pup body weights (mean). Overall statistically significant difference in male and female pups of the high-dosage group compared to controls. Maternal dosing took place from GD 6 through PND 10.

comparable to the controls, while the weights in male offspring remained lower. Overall, the body weight changes followed the pattern of effects seen in body weights; however, the effect on males was statistically significant (p ⬍ 0.001; F(3, 73) ⫽ 6.77), but not in females (p ⫽ 0.08; F(3, 73) ⫽ 2.36). The Day ⫻ Sex ⫻ Group was also statistically significant (p ⬍ 0.008; F(15, 207) ⫽ 2.17), and may indicate that males and females reacted differently to treatments mainly after day 22. None of the effects mentioned above were seen when the high-dosage group was removed from the analysis. The feed consumption analyses revealed a main Group effect (p ⬍ 0.001; F(3, 75) ⫽ 8.27) in the high-dosage group without a significant Sex ⫻ Group interaction (p ⫽ 0.22; F(3, 62) ⫽ 1.53). The biological significance of the Day ⫻ Sex ⫻ Group interaction (p ⬍ 0.01; F(6, 124) ⫽ 2.86) is unclear, but may reflect a tendency for a small reduction, over time, of the difference in feed consumption between control and highdosage groups that may be more accentuated in males than in females. The litter losses/reductions and lower body weights of offspring in the high-dosage group (PNDs 0 – 4) corresponded temporally with clinical signs, cannibalization of pups, a trend toward reduced feed consumption, and body weights in the dams. There were no effects of treatment with chlorpyrifos in the 0.3- or 1-mg/kg/day dosage groups. Developmental Landmarks In addition to the changes in pups’ body weights, delays in some developmental landmarks were present in pups of dams

The mean absolute and relative brain weights in offspring measured on PND 11 and PND 65 (adults) are presented in Figure 4. On PND 11, absolute neonatal brain weight (Table 5) was statistically significantly decreased at 5 mg/kg/day, and the brain-to-body weight ratio was increased at the same dosage. There were 9 areas of the brain measured in PND-11 pups (8 for adult rats), either grossly or microscopically, for length and layer thickness (Table 5). Five of these linear measures (cerebrum length, cerebellum length and height, parietal cortex, hippocampal gyrus) were statistically significant overall at the ANOVA level in males, and 2 (caudate-putamen and cerebellum height) in females. The Dunnett’s tests confirmed a statistically significant difference between control and high-dosage groups in cerebellum length and height in males, and in caudate-putamen in females. Another difference found between control and the medium-dosage group in cerebellum was height in females. Examination of the data revealed that most of the absolute measurements (except for the external germinal layer) were smaller in the high-dosage males and females compared to controls (Table 5). It is interesting to note that the majority of the smallest layers were found in the smallest brains, i.e., in the high-dosage pups. The lack of statistical significance in some layer thicknesses is consistent with the small sample size (n ⫽ 6/sex/dosage), the relatively small size of the effect, and the slightly larger coefficient of variation in TABLE 4 Summary of Developmental Landmarks in Pups Dosage (mg/kg/day) Pinna detachment a Eye opening b Males Females Vaginal opening c Preputial separation e

0

0.3

1.0

5.0

2.5 ⫾ 0.7

2.5 ⫾ 0.5

2.6 ⫾ 0.6

3.0 ⫾ 0.6

13.6 ⫾ 0.9 13.4 ⫾ 1.0 31.4 ⫾ 1.0 43.2 ⫾ 1.9

13.6 ⫾ 0.8 13.0 ⫾ 0.8 30.5 ⫾ 1.5 42.3 ⫾ 1.9

13.7 ⫾ 0.7 13.4 ⫾ 0.7 31.1 ⫾ 2.3 44.2 ⫾ 3.2

13.9 ⫾ 1.0 13.6 ⫾ 1.1 32.4 ⫾ 2.2 d 46.0 ⫾ 5.9

Note. Values are given as mean ⫾ SD. a Average PND on which at least 50% of the pups in a litter exhibited unfolding of the pinna. b Average PND on which all pups in every litter were observed with open eyes. c Average PND that the vagina was observed to be patent. d Statistically significant (␣ ⫽ 0.02). e Average day postpartum that the prepuce was observed to be separated.

257


except in the parietal cortex of adult females in medium- and high-dosage groups compared to control. These decreases were about the same (4.2 and 5.1%, respectively), and were seen in adult females but not in adult males. Given that brain and brain-layer thickness normally would be correlated, the data for pup and adult female rats were plotted a posteriori, before and after adjustment for brain weight in Figure 5. The upper panel shows that the 5-mg/kg/day female pup group is significantly different from the corresponding control group. When the same data are plotted after correction for brain weights, the statistically significant difference disappears and a 4.2% increase in relative layer thickness is recorded in the high-dosage group. The lower panel shows that, after correction for brain weight, the 5-mg/kg/day female adult group is not statistically different from the appropriate control group. Neurobehavioral Assessment

FIG. 4. Brain weight in pups (PND 11) and in pups as adults (PND 65) (mean ⫹ SD). Absolute brain weights (upper left panel) were significantly decreased in the high-dosage group on PND 11. Brain to body weight ratios were significantly increased in the same group at the same time (upper right panel). Data from males and females were combined for purposes of illustration only.

layer thickness (in comparison to brain weights’ coefficient of variation). In PND-65 rats, no statistically significant differences were seen in brain weight or in any linear dimensions of the brain,

The results of the neurobehavioral assessment of pups are presented in Figures 6 –10. Learning and memory. There were no effects of exposure on cognitive functions in any dosage group on PNDs 22–24 or in the same rats on PNDs 61–91 (adults). Figures 6 and 7 demonstrate the lack of any effect of CPF on learning acquisition (p ⫽ 0.92; F(3, 55) ⫽ 0.17 for pups, and p ⫽ 0.80; F(3, 54) ⫽ 0.33 for adults), short-term retention (overall p ⫽ 0.45; F(3, 54) ⫽ 0.89 for the slope of the forgetting curve), or non-mnemonic factors (overall p ⫽ 0.67; F(3, 54) ⫽ 0.52 for intercept), e.g., motivation, attention, sensori-motor performance, in any of the dosage groups at either time point. Further a posteriori examination of Figure 6 reveals that the percent correct in Block 1 (consisting of the first 12 trials) in adults was above random performance and that there did not appear to be any differences in long-term memory among dosage groups. It

TABLE 5 PND 11 Brain Measurements

Dosage (mg/kg/day) Males 0 0.3 1 5 Females 0 0.3 1 5

Brain wt (g)

Ant./Post. cerebrum (mm)

Ant./Post. cerebellum (mm)

Frontal cortex (␮m)

Parietal cortex (␮m)

Caudateputamen (␮m)

Corpus callosum (␮m)

Hippocampus (␮m)

Cerebellum ht (␮m)

Ext⬘l germ layer (␮m)

1.28 ⫾ 0.04 1.41 ⫾ 0.07 1.36 ⫾ 0.08 1.17 ⫾ 0.16 a,*

12.5 ⫾ 0.3 13.4 ⫾ 0.5 13.1 ⫾ 0.5 11.8 ⫾ 1.0*

3.3 ⫾ 0.3 3.4 ⫾ 0.4 3.3 ⫾ 0.2 2.5 ⫾ 0.6 a,*

1348 ⫾ 53 1360 ⫾ 100 1352 ⫾ 47 1272 ⫾ 153*

1336 ⫾ 54 1448 ⫾ 58 1448 ⫾ 33 1256 ⫾ 138*

2240 ⫾ 84 2240 ⫾ 108 2312 ⫾ 93 2224 ⫾ 147*

293 ⫾ 25 303 ⫾ 24 290 ⫾ 36* 293 ⫾ 56

904 ⫾ 93 1004 ⫾ 114 972 ⫾ 54 824 ⫾ 66*

3504 ⫾ 129 3456 ⫾ 172 3416 ⫾ 200 3008 ⫾ 504 a,*

37 ⫾ 2* 38 ⫾ 4 40 ⫾ 7 38 ⫾ 3

1.28 ⫾ 0.08 1.28 ⫾ 0.04 1.27 ⫾ 0.11 1.17 ⫾ 0.13 a,*

12.4 ⫾ 0.3 12.7 ⫾ 0.3 12.8 ⫾ 0.7 12.2 ⫾ 0.6*

3.2 ⫾ 0.2 3.0 ⫾ 0.3* 3.3 ⫾ 0.2 3.0 ⫾ 0.3*

1376 ⫾ 92 1388 ⫾ 79 1356 ⫾ 54* 1368 ⫾ 86

1380 ⫾ 54 1376 ⫾ 20 1368 ⫾ 80 1304 ⫾ 72*

2384 ⫾ 131 2224 ⫾ 116 2288 ⫾ 108 2152 ⫾ 134 a,*

307 ⫾ 38 286 ⫾ 27 304 ⫾ 36 274 ⫾ 40*

936 ⫾ 82 912 ⫾ 50 932 ⫾ 96 828 ⫾ 79*

3512 ⫾ 200 3176 ⫾ 130 3120 ⫾ 328 a,* 3208 ⫾ 226

39 ⫾ 3 36 ⫾ 6* 41 ⫾ 6 41 ⫾ 6

Notes. Values given are mean ⫾ SD. Overall statistically significant ANOVA for males (Brain wt, Ant./Post. cerebrum, Ant./Post. cerebellum, Parietal cortex, Hippocampus, and Cerebellum ht) and females (Brain wt, Caudate-putamen, and Cerebellum ht) separately (␣ ⫽ 0.05). a Statistically significant Dunnett’s test (␣ ⫽ 0.05). *Smallest values.

258

MAURISSEN ET AL.

concerned (p ⫽ 0.08; F(3, 73) ⫽ 2.33). The Group ⫻ Time interaction (p ⫽ 0.10; F(9, 225) ⫽ 1.65) was not significant, nor were the Group ⫻ Time ⫻ Block interactions (p ⫽ 0.18; F(99, 135) ⫽ 1.18), i.e., no effect on habituation. Examination of the data showed that high-dosage pups may have had a decrease in activity at PND 13 in the high-dosage group, relative to controls, and a slight increase in activity on PNDs 17, 21, and 60 (Fig. 8). Auditory startle. Litter was used as a factor in the analysis for auditory startle. Evaluation of the auditory startle response on PND 22 and 61 revealed an overall statistically non-significant increase in latency-to-peak response (overall p ⫽ 0.03; F(3, 75) ⫽ 2.95; Fig. 9), and an overall non-significant decrease in response amplitude (overall p ⫽ 0.14; F(3, 75) ⫽ 1.87; Fig. 10). The decreased amplitude was consistent with an increased latency, both were mainly observed in the 5 mg/kg/ day dosage group, and the effects may have been real. No effects appeared to be present in pups at 0.3 or 1.0 mg/kg/day. Examination of these same data on PND 61 failed to produce any evidence of an effect on either of these measures among

FIG. 5. Parietal cortex layer thickness in female pups and pups as adults before and after correction for brain weight. Asterisks indicate a statistically significant difference (␣ ⫽ 0.05) between control and treatment groups (Dunnett’s test). Parietal cortex thickness in the female adult low-dosage group was not measured (see Materials and Methods).

should be added here that, given that the same animals were used for learning and memory as weanling and adult rats, no statement could be made about “original” learning in adult rats. However, the lack of a statement on original learning in adults is offset by the information provided about the potential absence of an effect on long-term memory. Motor activity. Litter was used as a factor in the analysis of motor activity. There were no overall differences in motor activity, measured on PNDs 13, 17, 21, and 60, at any dosage level (p ⫽ 0.80; F(3, 75) ⫽ 0.33). Males did not behave differently from females, as far as the test compound was

FIG. 6. Spatial delayed alternation: Acquisition (mean ⫹ SD). Percent correct choices as a function of block in pups (PNDs 23–24) and pups as adults (PNDs 73– 82). Each block is a set of 12 consecutive trials. There are no overall statistically significant differences across the different dosage groups in the rate of acquisition of the correct response over time. The 50% correct represents the chance level. Data from males and females were combined for purposes of illustration only.

259


any gross pathologic changes attributed to treatment (data not shown). Overall terminal body weights on PND 21 or PND 65 were depressed (p ⫽ 0.004; F(3, 143) ⫽ 4.57). The Sex ⫻ Group interaction was also significant (p ⫽ 0.01; F(3, 143) ⫽ 3.61), as well as the Sex ⫻ Group ⫻ Time (p ⫽ 0.01; F(3, 143) ⫽ 4.04) whereas the Group ⫻ Time interaction was not significant (p ⫽ 0.14; F(3, 143) ⫽ 1.88). When male and female data were analyzed separately with all dosage groups, no effects were seen in females at any time. In males, the Group main effect was statistically significant (p ⫽ 0.004; F(3, 71) ⫽ 4.90). When the male body weights were analyzed separately without the high-dosage group, no statistically significant effects were seen (p ⫽ 0.19; F(2, 53) ⫽ 1.71). As far as the Sex ⫻ Group ⫻ Time interaction is concerned, it may have suggested the presence of a Group effect that was more marked in one sex at a time point (i.e., males on PND 65). Overall, the analysis can be interpreted as reflecting a decrease in the terminal body weights in males in the high-dosage group on PNDs 11 and 65. Examination of the data shows that the group effect on body weight may extend to high-dosage females on PND 11. No gross pathologic changes were associated with lower body weights (data not shown). All observations were considered within background variability and were consistent with the ages and strain of rat examined. Likewise, there were no microscopic neuropathologic alterations in any of the brains (data not shown) from any treatment group. FIG. 7. Spatial delayed alternation: Delay (mean ⫹ SD). Percent correct choices as a function of delay in pups (PND 24) and pups as adults (PNDs 85–90). There are no overall statistically significant differences across the different dosage groups in the slope and intercept of the percent correct response. The 50% correct represents the chance level. Data from males and females were combined for purposes of illustration only.

DISCUSSION

Based on the pharmacokinetic companion study (Mattsson et al., 2000), fetal exposure from high-dosage dams (5 mg/kg/ day) was sufficient to cause about 57% inhibition of fetal brain ChE on GD 20. Exposure to high-dosage pups via milk was estimated at 0.1 mg/kg/day. No inhibition of brain, RBC, or

treatment groups. The decrease in startle amplitude and increase in latency in PND-22 pups of the 5 mg/kg/day group were consistent with a developmental delay (Sheets et al., 1988) also present in other parameters (e.g., body weights) at 5 mg/kg/day. The non-statistically significant interaction Group ⫻ Time ⫻ Block for amplitude (p ⫽ 0.15; F(12, 222) ⫽ 1.43) indicated that the distribution of peak response in blocks within the test session over times did not change as a function of treatment, i.e., there were no effects of chlorpyrifos on habituation of the peak response at any time, another primitive form of learning. Body temperature. There were no statistically significant differences in body temperature among any of the dosage groups (p ⫽ 0.86; F(3, 75) ⫽ 0.25). Pathologic Evaluation Examination at necropsy of dams treated with CPF, and of the offspring examined on PND 21 or PND 65 did not reveal

FIG. 8. Motor activity: Number of beam breaks by rats on PNDs 13, 17, 21, and 60 (mean ⫹ SD). No overall statistically significant differences in total counts as a function of day. Data from males and females were combined for purposes of illustration only.

260

MAURISSEN ET AL.

killed and are not replaced. Furthermore, such a loss of neurons is often followed by abnormal cell migration and differentiation. As Bayer et al. (1993) wrote, “Once ectopia has been produced by abnormal migration, it becomes a permanent anatomical feature of the brain.” The paper continued, stating that hypoplasia and dysplasia are also permanent phenomena in the developing brain. If such events had taken place in the parietal cortex, they would have been present on PND 11. No gross or microscopic neuropathological alterations were observed, however, whether on PNDs 11 or 65, at any dose. Finally, from a statistical point of view, the multiplicity problem should be recalled in light of the numerous statistical tests performed in this study. As Ware et al. (1992) stated, “In that view, one of every 20 tests will produce a p value smaller than 0.05 merely by chance, and this likelihood must simply be kept in mind in the interpretation of any collection of hypothesis tests and their p values.” Similar positions can be found in Tukey (1980) and Wilkinson et al. (1999). For the reasons stated above, the statistically significant middle-dose effect in the parietal cortex of females at PND 65 was isolated, did not support a biologically plausible interpretation, and was, therefore, considered spurious.

FIG. 9. Auditory startle latency (mean ⫹ SD) in ms on PNDs 22 and 61 (pups as adults). No overall statistically significant difference across groups. Data from males and females were combined for purposes of illustration only.

plasma ChE occurred in fetuses or pups from dams gavaged at 0.3 or 1 mg/kg/day. The only statistically significant effect seen in the middle dosage group (1 mg/kg/day) was recorded as a 4.2% thinner parietal cortex (one of the 8 morphometric measurements) in adult females (PND65). No parietal cortex effects were observed either in adult males, or in male and female pups (PND11) at this dosage. The middle- and high-dosage groups differed by 0.9% and such a difference was considered small enough to question the existence of a dose-response relationship. The data a posteriori, corrected for brain weight, showed a clear loss of dose-response relationship and a loss of statistical significance in the high-dosage group. In the middledosage group, there were no behavioral or functional findings, and no inhibition of blood or brain cholinesterase (Mattsson et al., 2000). From the neurogenesis point of view, the rat neocortex, parietal and frontal, develops mainly between GDs 14 and 20 (Bayer et al., 1993). If treatment affected cortical neurogenesis, it would be expected that parietal and frontal cortices would be similarly affected, and this was not the case in the present study. Also, as Bayer et al. (1993) concluded from their studies, using X-irradiation when an environmental insult occurs during neurogenesis, the neuron precursors are

FIG. 10. Auditory startle amplitude (mean ⫹ SD) on PNDs 22 and 61 (pups as adults). No overall statistically significant difference among groups. However, the high-dosage group amplitude is smaller. Data from males and females were combined for purposes of illustration only.


Toxicity was limited to the highest dosage level tested, 5 mg CPF/kg/day, and no effects were seen in the pups in the absence of maternal toxicity. At this dosage level, ChE activity of brain erythrocyte and plasma were depressed by 90% when measured in the dams on GD20. Clinical signs of ChE inhibition, reduced body weight, and feed consumption were seen in high-dosage dams primarily at the end of gestation, and during the first few days of lactation. No treatment-related effects (other than ChE inhibition) were seen in the dams at dosages of 0.3 or 1 mg CPF/kg/day. At 5 mg/kg/day, pup survival was reduced during the first 4 postnatal days, when maternal effects were most evident, while pup survival was unaffected from postnatal day 5 throughout the remainder of the study. Pups from this group also exhibited lower body weight, delayed appearance of some developmental landmarks, and on PND 11, decreased weight and linear dimensions of the brain. Such an effect on brain does not, ipso facto, qualify the chemical as a developmental neurotoxicant when it is related to general growth retardation, as Makris et al. (1998) recognize. Slight changes in the auditory startle response were present on PND 22 (non-statistically significant), but were not detected on PND 61. There was some indication that motor activity measured on PND 13 may have been reduced (although no statistically significant differences were found). It would be consistent with the documented decrease in motor activity caused by chlorpyrifos in young rats (Moser et al., 1998). Cognitive functions measured in pups from exposed dams were not affected by CPF. The effects that were noted in the pups at 5 mg/kg/day were considered a result of the lower body weight gain, were viewed as indicative of a transient maturational delay, and were considered secondary to the maternal effects, although some contribution from pup toxicity could not be completely ruled out. Several lines of evidence support that the effects seen in the pups were secondary to effects on the dam. The fetal effects were observed only at a dosage level that caused clear maternal toxicity and occurred in the same time interval. Absolute brain weight was decreased by 8 –9% when measured on PND 11, although the brain weight relative to body weight was increased, and no effect was seen on brain weight when the pups were 2 months of age. Thus the effects on pup brain weight were transient, and are consistent with what is expected due to decreased nutrition or maternal care. The relative sparing of brain growth has been described by Dobbing and Sands (1971) in undernourished rats produced by creating artificially large (15 or more pups) and small (3 pups) litters. Undernutrition during early development causes slower body growth, slower rate of brain growth to a different degree, and contrary to popular belief, depresses the growth rate of various processes within the brain to the same extent (Peeling and Smart, 1994). At weaning, offspring from these large litters showed approximately a 16% difference in absolute brain weight. This brain weight difference was also transient, but resolved more slowly in undernourished litters than in the present study.

261

Some more arguments about maternal toxicity have been advanced by Schardein and Scialli (1999) who stated in a review of the present data, “The developmental findings in the high-dose pups in this study are consistent with the decreased pup weight that occurred in parallel with decreased maternal feed consumption and lactation weight gain. When maternal toxicity and developmental toxicity occur at the same dose, it is not axiomatic that the maternal toxicity is the cause of the developmental toxicity; however, in this case, the nature of the developmental effects and the lack of persistent neurologic abnormalities strongly suggest that the pup effects were secondary to maternal toxicity. More directly, no evidence of specific toxicity of chlorpyrifos for the developing nervous system was identified. The maternal and developmental NOAELs in this study were 1 mg/kg/day.” Finally, data from Mattsson et al. (2000) showed that brain ChE was inhibited by approximately 85% and 78% in dams on LDs 1 and 5. In contrast, brain ChE was inhibited by 35% in pups on PND 1, but was comparable to control levels on PND 5. Neonatal mortality was highest between PNDs 1 and 5, and thus, mortality did not correlate with inhibition of neonatal brain ChE, which was normal by PND 5. However, neonatal mortality did occur concurrently with clinical signs of maternal toxicity (muscle fasciculation, hyperpnea, hyperreactivity), and the time when dams’ body weight and body weight gains were depressed. All these effects pointed at the dams as responsible for neonatal mortality and slower growth rather than at the pups. The results of the present study are consistent with the majority of scientific data published on chlorpyrifos. Nostrandt et al. (1997) concluded that in adult rats given a single oral dose of CPF, a depression in brain ChE in excess of 60 –70% was necessary before any association with behavioral changes could be measured. They observed classical cholinergic signs (miosis and salivation) only in adult rats with ⬎90% inhibition of brain ChE. In the present study, this degree of brain ChE inhibition was seen only in the dams given 5 mg/kg/day. No cholinergic signs were observed in dams given the 0.3- or 1-mg/kg/day dosage, and maternal brain ChE was inhibited by less than 20% at either dose on GD20. Decreased body weight and reduced survival were also observed previously at a 5-mg/ kg/day dosage level in pups from the first generation of a 2-generation reproduction study (Breslin et al., 1996), but not in the second generation. Reproductive indices were unaffected. In the present study, no effects were observed in the pups at the lower dosage levels (i.e., 0.3 and 1 mg/kg/day). In the present study, cognitive functions, such as learning (i.e., rate of acquisition of a discrimination), short-term memory (i.e., retention), habituation (a primitive form of learning; Cabe and Eckerman, 1982) of motor activity and auditory startle amplitude, and, potentially, long-term memory (a posteriori observation) were not affected by chlorpyrifos. The absence of any cognitive effects in the present study was not unexpected in light of data from a previous cognitive study in

262

MAURISSEN ET AL.

adult rats administered CPF (Maurissen et al., 2000). These authors demonstrated that in adults there were no differences in retention of previously acquired information at dosages of up to 10 mg/kg/day for 28 days, though brain ChE was inhibited by up to 85% and cholinergic effects, as well as motor slowing, were prominent in the high-dosage group. Stanton et al. (1994) injected PND-21 rats, sc, with CPF, and no effects were observed in rats given 90 mg/kg. Rat pups in the 240-mg/kg group could not be evaluated because of overt toxicity. At 120 mg/kg, brain ChE activity was inhibited by approximately 80%, and was accompanied by a small (⬍10%) transient effect on acquisition of information on PND 23, but by PND 26 there was no discernible effect on acquisition. The developmental and cognitive data suggest that the transient changes in brain size have no impact on functional capabilities. A similar conclusion was expressed by Dobbing (1971), who noted there was no evidence that a transient decrease in brain size had any functional significance, even though it has been argued that a temporary retardation of developmental processes could lead to long-term consequences. The results of the present study contrast with the conclusions from Campbell et al. (1997) who administered CPF sc to rat pups during the early neonatal period (PNDs 1– 4 and 11–14) and predicted that the neurochemical changes they reported “can be expected to have a significant impact on nervous system function and may be responsible for neurobehavioral disturbances seen with developmental chlorpyrifos exposure”. The present study did not report such findings in the nervous system of rat pups and adults and did not, therefore, confirm such a prediction. The study design and timing of CPF administration made the Campbell study unsuitable for human risk assessment for several reasons: the study was designed to maximize the rate of absorption rather than to simulate environmental exposure (chlorpyrifos sc in DMSO), the route of administration was subcutaneous, and pups were injected (PND 1– 4) when the nervous system development of the rat brain was at a stage corresponding to the 3rd trimester in humans. Pope and Liu (1997) showed that the increased sensitivity (approximately 6 times) of neonatal rats (relative to adults) to chlorpyrifos is associated with the highest dosage compatible with no mortality (i.e., maximum tolerated dose). At dosage levels that cause no cholinergic effects (e.g., 50% inhibition of brain ChE), the neonate was no more than twice as sensitive as the adult. The authors also concluded that adults may be more sensitive than the pups to sub-lethal alterations following repeated exposures. Gavage (i.e., bolus dosage leading to high pulses of exposure) is often recommended by governmental agencies for developmental neurotoxicology studies, and was the route of administration used in the present study. However, the same dosage given as a fractionated dosage will most likely provide lower peak concentrations and a less pronounced toxicity pro-

file compared to a bolus administration (Wagner, 1971). Conolly et al. (1999) cautions against use of gavage administration of chemicals in these terms: “Experimental convenience has also been used to justify unrealistic methods of exposure such as corn oil gavage and nasal instillation. Both of these methods of exposure have the potential to deliver chemicals to a target site at a rate that far exceeds anything that would occur in the real world. [. . .] The relevance of experiments using doses that are multiples of conceivable human exposures and unrealistic routes of exposure is, at most, quite dubious. Mechanisms of action may be elicited under such conditions that would not occur with relevant routes and exposure levels.” This admonition appears relevant to our study, because both dose and mode of administration may exacerbate toxic effects. Such a caution is applicable to all studies using a bolus mode of administration of the test compound, such as those using gavage or subcutaneous injection, for example. In summary, chlorpyrifos did not cause any effects in the offspring when administered to dams during gestation and lactation at dosage levels of 0.3 or 1 mg/kg/day. Administration of chlorpyrifos at a dosage level of 5 mg/kg/day produced clinical evidence of cholinergic toxicity in the dams. Increased pup mortality, decrease in pup body weight, brain size, brain layer thickness, potential transient changes in the startle response, motor activity, pinna detachment, vaginal opening, and preputial separation were seen only at 5 mg/kg/day in the presence of maternal effects, and were consistent with delayed maturation. Cognitive functions (learning, short-term memory, and habituation in 2 different tasks) were not impaired in the pups at any time at any dosage. The no-observed-effect-level (NOEL) for any effects in pups in this study was 1 mg/kg/day. A NOEL was not determined in dams, due to inhibition of plasma and RBC ChE at 0.3 mg/kg/day. This study, using the testing guidelines established by US EPA, indicates that chlorpyrifos is not a selective developmental neurotoxicant in the rat. ACKNOWLEDGMENTS The authors wish to thank Joel L. Mattsson for the time spent in discussion, consultation and evaluation of the data; Mark E. Stanton (US EPA) for his help in setting the spatial delayed alternation method; Mildred S. Christian for commenting on this manuscript; John F. Barnett, Jr. and Eileen M. Tlush for assistance in managing the conduct of the study; Rosalyn Garman and Alice Campos for the histotechnical preparations; Nancy L. Freshour for the analytical chemistry of the test compound; Stephanie Padilla (US EPA) and Rene´e S. Marshall for the cholinesterase assays; Mamtha R. Shankar for assistance in the preparation of this manuscript; and Ralph R. Albee for preparing the graphics.

REFERENCES Bayer, S. A., Altman, J., Russo, R. J., and Zhang, X. (1993). Timetables of neurogenesis in the human brain based on experimentally determined patterns in the rat. Neurotoxicology 14, 83–144. Breslin, W. J., Liberacki, A. B., Dittenber, D. A., and Quast, J. F. (1996).

CHLORPYRIFOS DEVELOPMENTAL NEUROTOXICITY Evaluation of the developmental and reproductive toxicity of chlorpyrifos in the rat. Fundam. Appl. Toxicol. 29, 119 –130. Cabe, P. A., and Eckerman, D. A. (1982). Assessment of learning and memory dysfunction in agent-exposed animals. In Nervous System Toxicology (C. L. Mitchell, Ed.), pp. 138 –198. Raven Press, New York. Campbell, C. G., Seidler, F. J., and Slotkin, T. A. (1997). Chlorpyrifos interferes with cell development in rat brain regions. Brain Res. Bull. 43, 179 –189. Code of Federal Regulations (CFR) (1985). Title 9: Animals and Animal Products, Subchapter A: Animal Welfare. Office of the Federal Register, Washington, D.C. Conolly, R. B., Beck, B. D., and Goodman, J. I. (1999). Stimulating research to improve the scientific basis of risk assessment. Toxicol. Sci. 49, 1– 4. Deacon, M. M., Murray, J. S., Pilny, M. K., Rao, K. S., Dittenber, D. A., Hanley, T. R., Jr., and John, J. A. (1980). Embryotoxicity and fetotoxicity of orally administered chlorpyrifos in mice. Toxicol. Appl. Pharmacol. 54, 31– 40. Dobbing, J. (1971). Undernutrition and the developing brain: The use of animal models to elucidate the human problem. In Advances in Experimental Medicine and Biology, Chemistry, and Brain Development (R. Paoletti and A. Davison, Eds.), Vol. 13, pp. 399 – 412. Plenum, New York. Dobbing, J., and Sands, J. (1971). Vulnerability of developing brain: IX. The effect of nutritional growth retardation on the timing of the brain growthspurt. Biol. Neonate 19, 363–378. Johnson, C. D., and Russell, R. L. (1975). A rapid, simple radiometric assay for cholinesterase, suitable for multiple determinations. Anal. Biochem. 64, 229 –238. Makris, S., Raffaele, K., Sette, W., and Seed, J. (1998). A retrospective analysis of twelve developmental neurotoxicity studies submitted to the US EPA Office of Prevention, Pesticides, and Toxic Substances (p. 46) (OPPTS). US EPA Internal Document. Mattsson, J. L., Maurissen, J. P., Nolan, R. J., and Brzak, K. A. (2000). Lack of differential sensitivity to cholinesterase inhibition in fetuses and neonates compared to dams treated perinatally with chlorpyrifos. Toxicol. Sci. 53, 438 – 446. Maurissen, J. P., Shankar, M. R., and Mattsson, J. L. (2000). Chlorpyrifos: Lack of cognitive effects in adult Long-Evans rats. Neurotoxicol. Teratol. 22, 237–246. Moser, V. C., and Padilla, S. (1998). Age- and gender-related differences in the time course of behavioral and biochemical effects produced by oral chlorpyrifos in rats. Toxicol. Appl. Pharmacol. 149, 107–119. Nicholas, K. R., and Hartman, P. E. (1991). Milk secretion in the rat: Progressive changes in milk composition during lactation and weaning and the effect of diet. Comp. Biochem. Physiol. 98A, 535–542. Nostrandt, A. C., Padilla, S., and Moser, V. C. (1997). The relationship of oral chlorpyrifos effects on behavior, cholinesterase inhibition, and muscarinic receptor density in the rat. Pharmacol. Biochem. Behav. 58, 15–23.

263

National Research Council (NRC) (1996). Guide for the Care and Use of Laboratory Animals. National Academy Press, Washington, DC. Peeling, A. N., and Smart, J. L. (1994). Review of literature showing that undernutrition affects the growth rate of all processes in the brain to the same extent. Metabolic Brain Disease 9, 33– 42. Pope, C. N., and Liu, J. (1997). Age-related differences in sensitivity to organophosphorus pesticides. Env. Toxicol. Pharmacol. 4, 309 –314. Salewski, E. (1964). Fa¨rbemethode zum makroskopischen Nachweis von Implantationsstellen am Uterus der Ratte. Arch. Pathol. Exp. Pharmakol. 247, 367. SAS Institute. (1989). SAS/STAT User’s Guide, Version 6, 4th ed., Volumes 1 and 2. SAS Institute, Cary, NC. Schardein, J. L., and Scialli, A. R. (1999). The legislation of toxicologic safety factors: The Food Quality Protection Act with chlorpyrifos as a test case. Reprod. Toxicol. 13, 1–14. Sheets, L. P., Dean, K. F., and Reiter, L. W. (1988). Ontogeny of the acoustic startle response and sensitization to background noise in the rat. Behav. Neurosci. 102, 706 –713. Snedecor, G. W., and Cochran, W. G. (1967). Statistical Methods, 6th Ed, pp. 240 –241. Iowa State University Press, Ames, IA. Sokal, R. R., and Rohlf, F. J. (1969). Biometry, pp. 388-389. W. H. Freeman, San Francisco, CA. Stanton, M. E., Mundy, W. R., Ward, T., Dulchinos, V., and Barry, C. C. (1994). Time-dependent effects of acute chlorpyrifos administration on spatial delayed alternation and cholinergic neurochemistry in weanling rats. Neurotoxicology 15, 201–208. Tukey, J. W. (1980). We need both exploratory and confirmatory. Am. Statistician 34, 23–25. Tukey, J. W., Ciminera, J. L., and Heyse, J. F. (1985). Testing the statistical certainty of a response to increasing doses of a drug. Biometrics 41, 295–301. U.S. Environmental Protection Agency (US EPA) (1989). Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA); Good Laboratory Practice Standards; Final Rule. 40 CFR, Part 160. U.S. Environmental Protection Agency (US EPA) (1991). Pesticide Assessment Guidelines. Subdivision F, Hazard Evaluation: Human and Domestic Animals, Addendum 10, Neurotoxicity. Health Effects Division, Office of Pesticide Program. Wagner, J. G. (1971). Biopharmaceutics and Relevant Pharmacokinetics, pp. 149 –150. Drug Intelligence Publications, Hamilton, IL. Ware, J. H., Mosteller, F., Delgado, F., Donnelly, C., and Ingelfinger, J. A. (1992). P values. In Medical Uses of Statistics (J. C. Bailar and F. Mosteller, Eds.), pp. 181–200. NEJM Books, Boston. Wilkinson, L., and the Task Force on Statistical Inference. (1999) Statistical methods in psychology journals. Am. Psychol. 54, 594 – 604.