Toxicological consequences of toxaphene ingestion by cynomolgus (Macaca fascicularis) monkeys. Part 1: pre-mating phase

Food and Chemical Toxicology 39 (2001) 467±476

www.elsevier.com/locate/foodchemtox

Research Section

Toxicological consequences of toxaphene ingestion by cynomolgus (Macaca fascicularis) monkeys. Part 1: pre-mating phase D.L. Arnold a,*, F. Bryce a, C. Baccanale b, S. Hayward c, J.R. Tanner a, E. MacLellan a, T. Dearden a, S. Fernie a a

Toxicology Research Division, Bureau of Chemical Safety, Food Directorate, Health Protection Branch, Address locator 2202D1, Health Canada, Ottawa, Ontario, Canada K1A 0L2 b Animal Resources Division, Food Directorate, Health Protection Branch, Address locator 2202D1, Health Canada, Ottawa, Ontario, Canada K1A 0L2 c Biostatistics and Computer Application Division, Food Directorate, Health Protection Branch, Address locator 2202D1, Health Canada, Ottawa, Ontario, Canada K1A 0L2 Accepted 10 September 2000

Abstract A total of 40 menstruating cynomolgus monkeys (Macaca fascicularis) with an average age of 7.251.06 years (standard deviation), ®ve male cynomolgus monkeys with an average age of 12.60.66 years, and ®ve male cynomolgus males with an average age of 6.20.23 years were obtained from the Health Canada breeding laboratory. The females were initially randomized to the four test groups in accordance with their previous reproductive success and body weight. They were then randomly allocated between two similar environmentally-controlled rooms (20 females/room). The males were randomly assigned to one of the test rooms (six or four males/room). The female test groups self-ingested capsules containing doses of 0, 0.1, 0.4 or 0.8 mg (Groups A, B, C, D) of technical grade toxaphene/kg body weight/day (i.e. ®ve females/dose group/room). The older males (Group E) were proven breeders and were used exclusively for mating and their capsules contained no toxaphene. The younger males (Group F) ingested capsules containing 0.8 mg of technical grade toxaphene/kg body weight/day. After 20 weeks of daily dosing, it was assumed, based on the results of a pilot study [Andrews P., Headrick K., Pilon J.-C., Bryce F., Iverson F. (1996) Capillary GC±ECD and ECNI GCMS charcterization of toxaphene residues in primate tissues during a feed study. Chemosphere 32, 1043±1053], that the treated monkeys had attained a qualitative pharmacokinetic steady state regarding the concentration of toxaphene in their adipose tissue and blood. On a daily basis, each monkey's feed and water consumption as well as its health were monitored. In addition, the females were swabbed daily to determine menstrual status. On a weekly basis, each monkey's body weight was determined and its dose of toxaphene adjusted. Detailed clinical examinations were conducted at intervals of 4 weeks or less. Periodically, starting prior to the initiation of dosing, blood samples were taken for serum biochemistry, haematology and toxaphene analysis. In addition, specimens from the nuchal fat pad were also obtained for toxaphene analysis. Statistical analysis did not reveal any e€ect of treatment on body weight gain, feed consumption, water consumption or haematological parameters during the 75-week pre-mating phase. The only serum biochemistry parameter which was consistently a€ected by treatment was cholesterol, the level of which decreased in a linear fashion as a consequence of dose, and this e€ect increased with time on test (P=0.037). No other biological e€ects of toxaphene ingestion were found during the premating phase of this toxicological±reproduction study. # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Toxaphene; Cynomolgus monkeys; Toxicity

* Corresponding author. Tel.: +1-613-957-0992; fax: +1-613-941-6959. E-mail address: doug_arnold@hc- sc.gc.ca (D.L. Arnold). 0278-6915/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0278-6915(00)00151-4

468

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

1. Introduction Toxaphene, a mixture of polychlorinated monoterpenes, was initially sold by Hercules Co., as Hercules 3956, in 1945. It was the most heavily-used pesticide in the United States and in many parts of the world prior to its being banned in the US (Voldner and Schroeder, 1990; Saleh, 1991). It was registered for use on more than 168 agriculture commodities and crops, and to control ectoparasites on cattle and sheep in addition to its use for ®sh eradication programs (Korte et al., 1979). It has been reported that 65±90% of all the toxaphene used in the US was applied to cotton in the mid-1970s (von RuÈmker et al., 1974; Eichers et al., 1978; IARC, 1979; Korte et al., 1979; EPA, 1982a), but the use of toxaphene in general and on cotton in particular decreased dramatically in subsequent years (EPA, 1982a). Most registered usages of toxaphene in Canada were cancelled in 1980 (Agriculture Canada, 1980). Similar actions were undertaken in such countries as Algeria, Denmark, Egypt, England, Finland, France, Hungary, Italy, Sweden and Switzerland (Cohen et al., 1982) prior to the US Environmental Protection Agency (EPA) cancelling most toxaphene registrations in 1982 (EPA, 1982b). The EPA had concluded that the use of toxaphene posed risks of: (1) chronic e€ects to wildlife; (2) acute toxicity to aquatic organisms; (3) population reduction in non-target organisms; and (4) increased potential for oncogenicity to humans (EPA, 1982b). At the time toxaphene was banned, it was well known that the prevailing winds had transported toxaphene for long distances (C&EN, 1982; ATSDR, 1996). The ®rst documented evidence for this phenomenon was attributed to Bidleman and Olney (1975; EPA, 1982a). Within a few years, toxaphene residues were found consistently in samples where toxaphene was either never used or not used to any signi®cant extent (Jansson et al., 1979; SundstroÈm, 1981; Schmitt et al., 1983). For example, toxaphene has been found in peat bogs from the Great Lakes area through north-eastern Canada, even though the predominant use of toxaphene was in the southern US (Rapaport and Eisenreich, 1986) and its use in Canada was mainly as a pesticide for domestic animals (Agriculture Canada, 1980). However, whether the toxaphene found in the various Arctic samples is due to volatilization of North American soil residues or as a result of transport from other sources is an open question (Andersson et al., 1988; Bidleman et al., 1990), although some have suggested that non-North American sources may be important contributors to the amount of toxaphene found in the Canadian Arctic (Rahn, 1981; Barrie, 1986; Muir et al., 1990). The implications of these observations were evident in a report by DeWailly et al. (1993), wherein the mean concentration of organochlorines in the milk fat samples obtained from Inuit people living in Arctic QueÂbec

were two to 10 times greater than in similar samples collected in southern QueÂbec. Subsequently, Kuhnlein et al. (1995) reported that 50% of the dietary intake recalls collected from an indigenous population on Ban Island in the eastern Arctic exceeded the ``tolerable daily intake levels'' for toxaphene. Because of these ®ndings, and the general lack of toxicological data regarding toxaphene that conforms to the present standards for toxicological testing, the study described herein was undertaken. 2. Materials and methods 2.1. Animals More than 40 female and 10 male cynomolgus (Macaca fascicularis) monkeys from the Health Canada (HC) monkey breeding laboratory were obtained for evaluation. All of the monkeys tested negative for Herpes B Virus (serum sample) and tuberculosis (eyelid test) prior to being included in the study population and in yearly tests thereafter. Birth dates were known for all of the female monkeys and all of the treated male monkeys. All monkeys were acclimatized to the various test procedures for at least 5 months. They were housed in individual stainless-steel cages, the dimensions of which complied with the Canadian Council on Animal Care (1984) guidelines Environmentally, the room temperature was 222 C, relative humidity 4010%, with 15 air changes/h and a light/dark cycle of 12 h. After being acclimatised for at least 5 months, a total of 40 female and 10 male monkeys were selected based on: (a) good health, (b) no aberrant haematological or serum biochemistry ®ndings (i.e. 2 standard deviations from HC's historical average for this species), (c) temperament and amenability to frequent handling, (d) no detectable toxaphene in blood or nuchal fat specimens, and (e) the females had regular menstruations. The females were initially age-matched and then randomly distributed to one of the four test groups in accordance with their previous reproductive success and body weight. They were then randomly assigned to one of the two test rooms (20 females/room). The older males, all proven sires, were assigned to the control group while the younger monkeys, who had not been previously mated and were similar in age to the treated females, were assigned to the treated group. The males were then randomly assigned to either test room (six or four males/room). The average age and weight of the test groups were as follows: control females/Group A Ð 2723.6448.76 days (standard deviation), 3721.6222.5 g; Group B Ð 2646.6347.57 days, 3757.1384.1 g; Group C Ð 2591.5429.88 days, 3761.5416.6 g; Group D Ð 2617.3325.24 days, 3938.7745.0 g; control males/Group E Ð 4584.6239.48 days, 8932.4616.4 g; treated males/ Group F Ð 2256.284.10 days, 6448.2628.5 g.

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

2.2. Husbandry The female monkeys were allotted approximately 165 g of Laboratory Monkey Chow (15% protein, product #5034, Purina Canada, Strathroy, ON, Canada) each day while the male monkeys were allotted approximately 200 g each day. All monkeys were allowed ad lib. access to Municipality of Ottawa water. The water bottles were changed daily, while each cage was washed weekly. Small amounts of fruits and/or vegetables (i.e. oranges, grapes, apples, bananas, carrots and peanuts) were used daily as rewards and dietary variation. Each monkey spent at least 1 day per week with a second monkey from its dose group in an exercise cage which was more than 5 times larger than their ``home'' cage. All males were exercised individually at least once a week. 2.3. Animal procedures During the pre-dosing period, each monkey was trained to ingest a capsule containing corn oil and glycerol. The following parameters were monitored at the frequencies indicated during the training period and throughout the pre-mating phase which lasted for a total of 75 weeks; however, not all specimens were analysed. 1. General health status (daily). Visual examination of the monkey and its environment. 2. Feed and water consumption (daily, except for the 1 day per week that a monkey was in the exercise cage). 3. Determination of menstrual status by swabbing (daily) Ð detection of menses consisted of a daily vaginal swab wherein each monkey was trained to back up to the front of the cage and present her external genitalia (``present position'') to facilitate the insertion of a cotton-tipped swab approximately 5 cm into the vagina. The following macroscopic ®ndings were recorded in the monkey's clinic ®le: no discharge, mucus discharge, dried blood, positive for blood Ð slight, positive Ð medium and positive Ð heavy. Only menstrual cycles occurring prior to the pre-mating phase of this study were included here (i.e. the ®rst 74 weeks of the study). Other details and de®nitions were previously published (Bryce et al., 2000). 4. Body weight (weekly) Ð all monkeys were trained to enter a ``transfer box'' to be weighed. 5. Detailed clinical evaluation (intervals of 4 weeks or less). Each monkey was removed from its cage and examined as previously described (Tryphonas et al. 1986a,b). The males, but not the females, were administered ketamine (10 mg/kg body weight, im) to enhance handling and personal safety. The monkeys were speci®cally examined for the following halogenated hydrocarbon signs of

469

toxicity: skin/coat Ð alopecia, dry skin, erythema; nails Ð prominent beds, nail changes (see Arnold et al., 1993, for example); lymph gland enlargements; eyelids Ð changes to tarsal glands. In addition, the following was also assessed: hydration; ears; mouth; teeth; nares; heart rate; respiration rate; body temperature; cardiac, pulmonary and abdominal auscultations; abdominal status by palpation; and uterine or prostate status by transrectal palpation. 6. Haematology Ð Coulter Counter model S-Plus IV (Coulter Electronics Inc., Hialeah, FL, USA) and the method of Fernie et al. (1994) were used to determine the following parameters: total leukocyte count (WBC109/l), erythrocyte count (RBC1012/ l), mean corpuscular (erythrocyte) volume (MCV ¯), mean corpuscular (erythrocyte) haemoglobin (MCH pg), mean corpuscular (erythrocyte) haemoglobin concentration (MCHC g/l), haemoglobin concentration (Hb g/l), hematocrit (Hct L/l), red cell (erythrocyte volume) distribution width (RDW%), mean platelet volume (MPV ¯), platelet count (Pl109/l) and platelet volume distribution width (PDW%). Leukocyte di€erential counts were performed and the results reported in absolute numbers (i.e. basophils, eosinophils, neutrophils, lymphocytes and monocytes). In addition, percent and absolute numbers of reticulocytes were determined. All blood samples were obtained from the femoral vein or artery. The haematology samples which were statistically analysed for this portion of the study were obtained 4 months prior to dosing and during weeks 5, 11, 19, 23, 41 and 64 after the start of dosing. 7. Serum biochemistry Ð the following parameters were determined using a Beckman SYNCHRONCX1 SYSTEMS (Beckman Instruments Inc., Brea, CA, USA) and the methods of Fernie et al. (1994): albumin (g/l), blood urea nitrogen (BUN, mg/dl), potassium (nmol/l), calcium (nmol/l), sodium (nmol/l), chloride (nmol/l), total bilirubin (nmol/l), creatinine (nmol/l), total protein (g/l), glucose (nmol/l), phosphorus (mg/dl), magnesium (mg/dl), cholesterol (nmol/l), creatinine phosphokinase (CPK, IU/l), lactate dehydrogenase (LDH, IU/l), triglycerides (nmol/l), gamma-glutamyl transferase (gGT, IU/l), alanine aminotransferase (ALT, IU/l), alkaline phosphatase (AP, IU/l), amylase (U/l), aspartate aminotransferase (AST, IU/l), osmolarity, uric acid (nmol/l), albumin/globulin ration, thyroxine (T4; nmol/l) and thyroxine uptake (TU-thyroxine binding capacity;%). The serum samples which were statistically analysed for this portion of the study were obtained 4 months prior to dosing and during weeks 9, 18, 36 and 63 after the start of dosing.

470

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

8. Oestrogen and progesterone: during one menstrual cycle after the assumed qualitative pharmacokinetic steady state had been attained, daily 1-ml samples of blood were obtained during one menstrual cycle for analysis by radioimmunoassay (method of Truelove et al., 1990). Blood samples obtained for these determinations were collected during test weeks 22 through 31 inclusive. Results will be reported separately. 9. Serum hydrocortisone (monthly) using a diagnostic radioimmunoassay kit (QuanticoatTM #825 KallestadTM Diagnostic Inc. Montreal, Quebec, Canada; Tryphonas et al. in preparation). 10. Immunological testing, using method of Tryphonas et al. (1989), were initiated on study week 34 and completed on study week 60. Results will be reported separately by Tryphonas et al. 11. Specimens for toxaphene analyses (at least monthly). A specimen of blood was collected from the femoral vein or artery and an adipose sample from the nuchal fat pad was obtained using a 2% solution of lidocaine (lignocaine) HCl as a local anaesthetic. The GC±ECD analysis was undertaken using the method of Andrews et al. (1996). The blood or adipose specimen was ®rst extracted with acetone:hexane (2:1; all solvents were analytical grade or better) and then ®ltered through glass wool using three acetone:hexane washes, evaporated to approximately 1 ml. Hexane added and evaporated to approximately 1 ml. Hexane was again added to the residue and then ®ltered through sodium sulfate (Na2SO4). The eluant was evaporated to approximately 1 ml and then poured onto a column of 0.5% deactivated Florisil and eluted with 75 ml of 20% dichloromethane (CH2Cl2). The eluant was evaporated to approximately 1ml and then hexane was added continuously and evaporated until no traces of CH2Cl2 remain, and this mixture was subsequently evaporated to dryness. After the addition of 1 ml hexane, a GC analysis was undertaken (Andrews et al., 1996). Results will be reported separately. 2.4. Test chemical The toxaphene was manufactured by the Hercules Powder Company Inc. (Wilmington, DE, USA). It was technical grade, Lot #71-132-B8-17-70. The toxaphene was ®rst dissolved in corn oil and then added to a gelatin capsule containing a constant volume of glycerol as a ¯avouring agent. The test groups received capsules containing daily dosages of 0 (Group A), 0.1 (B), 0.4 (C), 0.8 (D), 0 (E) or 0.8 (F) mg of technical grade toxaphene/kg body weight/day, which they self-ingested. The doses were recalculated on a weekly basis to re¯ect body weight changes.

2.5. Statistical procedures 2.5.1. Feed and water consumption As the individual feed and water data were variable from day to day, linear polynomials were used to summarize these data. Therefore, straight lines were ®t to these data. For the females, the mean consumption and the slope were analysed using an analysis of variance (ANOVA) model with a main e€ect for room and linear and non-linear e€ects for dose. For each of the male groups (E and F), an ANOVA model was used in which the main e€ect was for room. 2.5.2. Menstrual status These data were analysed as previously reported (Bryce et al., 2000). Brie¯y, the calendar was divided into quarters starting with 1 April so that comparisons could be made with the results of Bryce et al. (2000); however, the ®rst day of dosing for this study was 21 May 1996 and the ®rst day of mating was 3 November 1997. Therefore, there were no menstrual data during the ®rst quarter from 1 April±20 May 1996 and there were no data from 3 November±31 Decembeer 1997 for the seventh and ®nal premating quarter. Menstrual cycle length (i.e. the number of days from the ®rst day of menstruation until the day prior to the start of the next menses), menses duration (i.e. the number of days from the ®rst through the last day of menses) and the frequency or number of cycles in each quarter of the year was determined for each female. The month in which the menses started determined which month the data were recorded for, regardless of which month the menses cycle ended. A repeated measures ANOVA variance model with quarter as the withinsubject e€ect and treatment as the between-subject e€ect was used to analyse the following variables: square root of the frequency; log10 mean menses duration; log10 mean menstrual cycle length; and 1/(mean menstrual cycle length)2. These transformations were undertaken to stabilize the variance and normalize the data. It should be noted that year was not included as one of the within-subject e€ects since there were data for only ®ve complete quarters. In addition, since the ®rst and seventh quarters did not have three full months of data, the analyses were undertaken both using these data and excluding these data. The frequency estimates for the ®rst and seventh quarters underestimated the true frequency since complete data were not available for these quarters. As a result, the test for quarter e€ect would be more signi®cant when these two quarters were included in the analyses; therefore, the results for this e€ect were ignored when these quarters were included. 2.5.3. Body weight Orthogonal cubic polynomials were ®tted to the growth data from the beginning of the study until the

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

start of breeding Ð a period of 74 weeks. The mean weight and the weight gain for this period and the coef®cients for the ®tted orthogonal cubic polynomials were the response variables analysed. For the females, the variables were analysed using an ANOVA model with a main e€ect for room and linear and non-linear e€ects of dose. The test for the linear e€ect of dose with a non-linear term included in the model, provides a test for linear trend as described by Armitage (1973). As the two groups of males were of di€erent ages, their body weight changes were not compared for treatment e€ects. Therefore, the mean weights, weight gain and the polynomial coecients for the two male groups were analysed separately using an ANOVA model with a main e€ect for room. 2.5.4. Haematology Pro®le analyses (Morrison, 1976), with sample period as the within-subjects factor, were carried out across all samples using an ANOVA model, with a main e€ect for room, and linear and non-linear terms for dose. A log transformation was applied to the basophil counts, eosinophil counts,% reticulocytes, reticulocyte counts and segmented neutrophil counts, and a square root transformation was applied to the monocyte counts to stabilize variance and normalize the data. Due to the number of variables analysed (18 variables), results were not considered signi®cant unless the P values were less than 0.05/180.0028 (Bonferoni's adjustment; Miller, 1966). 2.5.5. Serum biochemistry Pro®le analyses (Morrison, 1976), with week or sample period as the within-subject factor, were conducted separately for the pre- and post-dosing samples. These analyses were only performed on the female samples, since it was not valid to compare the male samples due to the age di€erence between the two groups. An ANOVA model with the main e€ects for room and dose group was used for the pre-dosing samples to con®rm that there were no signi®cant di€erences between the groups following randomization. The ANOVA model for the post-dosing samples included lines and non-linear terms for dose thereby allowing for the use of Armitage's (1973) test for trend. A square root transformation was applied to the total bilirubin levels to stabilize variance and normalize the data. A log transformation was also applied to the following variables for the same reason: potassium, amylase, AST, ALT, LDH, AP, CPK, gGT, triglycerides and phosphates. As there were 22 serum biochemistry variables, results were not considered signi®cant unless the P value was less that 0.05/22 40.0023 (Bonferoni's adjustment; Miller, 1966). For T4, TU and uric acid, an ANOVA with a main e€ect for room, and linear and non-linear terms for dose

471

was used to analyse the data for each sampling period. The inclusion of both linear and non-linear terms for dose allows for the use of Armitage's test for trend (Armitage, 1973). Pro®le analyses (Morrison, 1976), with week or sample period as the within-subjects factor, were carried out separately for the post-dosing samples. An inverse transformation for the uric acid data was found to stabilize the variances and normalize the data. The best transformation for the T4 data was found to be the inverse of the 4th root; i.e. 1/0.25. No transformation was necessary for the TU data. 3. Results 3.1. Feed and water consumption There was no evidence (P >0.05) of a room or treatment e€ect for mean feed or water consumption, nor for the slopes of the lines ®tted to feed and water consumption, for either sex. When a test for ®tted slopes ®nds that they are signi®cantly di€erent from zero, such a ®nding is interpreted to mean that there was evidence for either an increase or decrease in consumption across time. This test provided signi®cant evidence (P40.020) of an increase in feed and water consumption for all the females groups across time and, save for water consumption by the control males (P=0.55), similar evidence was found for the males. When group means of the feed and water consumption data were plotted against time, it was noted that feed consumption did not stabilize until the monkeys had been on test for 10±20 weeks. Also, a dip in feed consumption was noted around week 54, when the monkeys were administered ketamine (10 mg/kg body weight, im) as part of the immunological testing procedures. In view of these two situations, the feed consumption data were re-analysed by evaluating the period from the start of the study until week 54, termed pre-immunization study, and week 55 through the remainder of pre-mating portion of the study (i.e. week 75). This re-analysis of the data did not reveal any e€ect of treatment (P>0.05). However, there were some room e€ects on the di€erences in mean feed consumption between these two periods for the females (P=0.016), and on mean feed consumption (P=0.013) and the slopes (P=0.0072) after the start of the immunization testing of the treated males. Although there was no evidence for any di€erence in average consumption between the two periods for either sex (P>0.40), there was strong evidence of a di€erence in slopes between the two periods for both the females (P40.001) and treated males (P=0.001). There also were signi®cant increases in feed consumption across both periods for the females and treated males (P <0.01). While the untreated males exhibited similar pro®les to the treated males, due to a

472

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

greater variability in consumption, the increases in feed consumption were less signi®cant (pre-immunization: P=0.89; post start of immunization: P=0.017) and the results for the di€erence of the slopes between the two periods was also less signi®cant (P=0.060). In summary, there were no e€ects of treatment on feed or water consumption which generally increased during the premating phase. Upon initiation of the immunological testing, there was a noticeable decrease in feed consumption. Subsequently, there was an increase in feed consumption which was somewhat greater than the increasing feed consumption observed prior to the start of immunological testing, suggesting that the animals may have attempted to compensate for their previous decrease in feed consumption. 3.2. Menstrual status There were no signi®cant e€ect (P<0.05) of dose for any of the parameters studied, but there was a marginally signi®cant non-linear e€ect of dose (P=0.047) on mean menses duration for the ®rst quarter. Additionally, no signi®cant interactions between dose and quarter were found (P>0.05) for any of the parameters. There were signi®cant e€ects of quarter for log10 mean menses duration with and without the data for quarters one and seven (P=0.066; P=0.016, respectively). There was a slight increase in menses duration during the fourth (January±March) and ®fth quarters. For the log transformation of mean menstrual cycle length, there were no signi®cant e€ects of quarter, regardless of whether the data for quarters one and seven were included (with P=0.25; without P=0.13). Menstrual cycle length was longest in quarters one through four and shortest in the sixth quarter (July± September). For the transformation 1/(mean menstrual cycle length)2, there was a signi®cant e€ect of quarter with and without the data from quarter 7 (P=0.030). Therefore, while the latter transformation is uncommon, the results were similar with both transformations in that there was a signi®cant quarter e€ect, thereby, precluding choice of transformation as an issue. 3.3. Detailed clinical evaluation The only treatment-related conditions were a higher incidence of a slight nail bed prominence, a slight edema of the eyelids and a low incidence of dry skin which was found only in the Group D females. 3.4. Body weight There was no statistical evidence (P>0.10) of a linear or non-linear e€ect of treatment on any of the variables analysed. However, due to the grouping and randomization Scheme employed for the females (i.e. initially by

previous breeding performance and secondarily by body weight), the initial body weight of the Group D females was 150 g, on average, heavier than the average body weight in the other female groups. While the average body weight decreased slightly for the females in Groups A, B and C during the initial 12 weeks of the study, the Group D monkeys were able to maintain their initial heavier body weight during this period. For the duration of the premating phase, the Group D females gained weight at a slightly faster rate than the other groups of females. Strong evidence for a body weight gain across the premating phase was indicated by the highly signi®cant (P40.0001) di€erence of weight gain and the linear coecients from zero (P40.001). There was also evidence that the weight gain was not linear but somewhat sigmoidal in shape since the cubic coecients were also signi®cantly di€erent from zero (P=0.0074). The observed weight gain was greatest from week 24±42. For the control males, there was a signi®cant room e€ect on mean weight (P=0.009). However, there was no evidence of a weight gain with the control males (P=0.91) nor were the linear coecients di€erent from zero (P=0.76). Furthermore, there was no evidence that the quadratic (P=0.38) or cubic coecients (P=0.33) were di€erent from zero, indicating a ¯at growth curve. The signi®cant room e€ect was attributed to the randomization of ®ve monkeys to two rooms. Analyses were adjusted for the e€ect of room by including it as a blocking factor. For the treated males, there was strong evidence of a weight gain during the premating phase, which was indicated by the signi®cant di€erences in weight gain (P=0.011) and linear coecients (P=0.087) from zero. However, there was no evidence of any deviation from a straight growth curve since the quadratic coecients (P=0.38) and the cubic coecients (P=0.33) were not signi®cantly di€erent from zero. Finally, correlation coecients were determined for each monkey regarding weight gain versus mean feed consumption and average water consumption. The only signi®cant correlation found was for the control males (r=0.91, P=0.030). The correlation for water consumption with body weight was also high for this group (r=0.76), but not signi®cant (P=0.14). The correlation with feed consumption for the females (r=0.13) and treated males (r=0.44) was not signi®cantly di€erent from zero (P>0.40), while the correlation with water consumption was considerably smaller (females r= 0.081; treated males r=0.039) and not signi®cantly di€erent from zero (P>0.60). Therefore, the only signi®cant correlation was for weight gain with feed consumption for the control males. 3.5. Haematology At the 0.0028 level of signi®cance, there were no signi®cant room or dose group e€ects, with the exception

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

of the last sample regarding platelet distribution width wherein a signi®cant non-linear e€ect of dose was evident (P=0.0003). There were no signi®cant linear e€ects for this sampling period, no signi®cant treatment e€ects for any other sampling period, nor were there any signi®cant di€erences evident with the multivariate test for treatment group e€ects. These ®ndings suggest that the platelet ®nding was spurious. However, there was a suggestion of a linear decrease in platelet distribution width with an increase in dose for three of the post-dosing periods (P<0.020) and for the average across the post-dosing period (P=0.038). There were signi®cant e€ects for the sampling period by room interaction for mean cell volume (P40.0001) and mean platelet volume (P=0.006). This was evident because the mean for each room changed from sampling period to sampling period or with time on test, but there was no pattern associated with these changes. However, there were no signi®cant room e€ects for any sampling period when they were analysed separately. In conclusion, there was no appreciable evidence regarding any treatment e€ect on any of the haematology variables. 3.6. Serum biochemistry For the pre-dosing samples, there were no signi®cant di€erences between the control and treated groups nor any signi®cant di€erences regarding room e€ects after performing the Bonferoni adjustment. For the postdosing samples, only a few statistical di€erences were found after performing the Bonferoni adjustment. For example, there was a signi®cant dose (linear) e€ect regarding cholesterol levels (P=0.0021) for post-dosing week 63, while the di€erences were not signi®cant at the other sampling periods (i.e. week 9, P=0.21; week 18, P=0.059; week 36, P=0.058). It appears that there was an increasing linear e€ect of dose resulting in a decrease in cholesterol levels as the study progressed, which is re¯ected in the multivariate e€ect of dose (P=0.011) and in the P value test for parallelism for the week by dose interaction (P=0.037). The latter value indicates a signi®cant change over time regarding the e€ect of treatment. Generally speaking, the mean cholesterol level for the control group increased almost linearly over time while there was only a minimal increase in the mean values for the treated groups (Fig. 1). For the log lactate dehydrogenase data, there was a signi®cant room e€ect for post-dosing week 63 (P=0.0019), but not for the preceding sampling periods (week 9, P=0.013; week 18, P=0.031; week 36, P=0.12). The multivariate P value for overall room e€ect was 0.0050. Although there is some evidence of a room e€ect at the 0.05 level of signi®cance for weeks 9, 18, 63, and for the multivariate test, only the univariate result for week 63 was signi®cant after adjusting for the number of variables. While there may be a room e€ect

473

Fig. 1. The mean cholesterol levels ( standard error) for female cynomolgus monkeys receiving 0 (Group A), 0.1 (Group B), 0.4 (Group C) or 0.8 (Group D) mg of technical grade toxaphene/kg body weight/day during the 70 weeks prior to mating.

on the log lactate dehydrogenase levels, the evidence was not overwhelming. In addition, there was a signi®cant (non-linear) e€ect (P=0.0006) for osmolarity at week 18, but only nonsigni®cant results (P>0.50) for weeks 9 and 36 and a marginally signi®cant result for week 63 (P=0.032). The week 18 results were attributable to the Group B females having a higher value than the controls or Groups C and D and, while the Group B females consistently had a higher value than the other groups, there was no consistent ordered treatment e€ect at all sampling times. In summary, the mean values for the treated groups tended to increase with time on test while the control values remained constant. 3.7. T4, TU and uric acid There were no signi®cant di€erences regarding T4 or uric acid results except for the signi®cant e€ects of sampling week (P40.0001); that is, there were no statistical di€erences between the control and treated groups for any of the four sampling periods but all of the mean values were markedly higher for the ®nal two sampling periods vs the initial two sampling times. For TU values, there were no signi®cant room e€ects (P50.15) or non-linear e€ects of treatment. However, signi®cant linear e€ects of treatment for all post-dosing samples were found (week 9, 18, 36, 63) and the overall dose e€ect was P=0.0048. There was no signi®cant week by room or week by dose interaction (P>0.10). The only within-subjects e€ect that was statistically signi®cant was the e€ect of sampling week (P40.0001). Generally, TU levels decreased with increasing toxaphene, but this generality was not consistent from one sampling period to the next.

474

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

4. Discussion A literature review did not reveal many studies in which toxaphene had been administered to monkeys, and most of the studies with monkeys were performed by the US manufacturer of toxaphene or its contractors. These studies used a minimal number of monkeys, and the reports contained few experimental details. It also appears that the same study may have been attributed to more than one research group. For example, Hercules Inc. fed monkeys of an unspeci®ed strain diets containing 10± 15 ppm (0.64±0.78 mg/kg/day) toxaphene for 2 years with no clinical or histological e€ects being reported (EPA, 1980, 1987). In a similar study, Treon et al. (1952; FAO/WHO, 1969; EPA, 1982b; IPCS, 1984) fed two adult female monkeys diets containing 0.64 to 0.78 mg of toxaphene/kg body weight, 6 days/week for 2 years. A third female served as the control. There were no signs of toxicity as judged by growth rate, ratios of liver or spleen to body weight, and histological examination of tissues. Subsequently, Maier-Bode (1965) fed two female monkeys diets containing 10 ppm toxaphene for 2 years and no di€erences regarding growth or organ weights were evident when compared to controls. Citing a Kettering Laboratory Report, Lehman (1965) stated that two female monkeys were fed diets containing 10 ppm toxaphene while a third female served as the control. The study lasted 2 years, with all of the monkeys surviving and gaining weight satisfactorily. The kidney and liver/ body weight ratios for the treated monkeys were comparable to the control and no relevant histological changes were observed in the treated monkeys. Korte et al. (1979) reported that dietary levels of 15 ppm for the monkey did not result in any observed toxic e€ects. Santolucito (1975) reported that there were EEG changes after squirrel monkeys (Saimiri sciureus) had received 1.0 mg of toxaphene/kg body weight, 6 days/ week for 3 years; however, no de®ciency in neurological function was found. Without indicating what data they used in their evaluation, von RuÈmker et al. (1974) concluded that the chronic ``no e€ect'' level for monkeys regarding toxaphene ingestion was 0.7 mg/kg body weight/day (dietary concentration of approx. 14 ppm). During the pre-mating phase of our study, no statistical di€erent e€ects attributable to toxaphene treatment were evident regarding feed or water consumption and body weight gain. The statistical analysis of the menstrual data did not indicate any e€ects due to treatment. However, the slight, non-statistical signi®cant increases observed in the mean menses duration and menstrual cycle lengths were not attributable to seasonality as we had previously found in a chronic reproduction study in which rhesus monkeys (Macaca mulatta) ingested Aroclor1 1254 (Bryce et al., 2000). Changes in several serum biochemistry parameters have been reported when various mammals ingested

toxaphene (Kennedy et al., 1973; Baeumler, 1975; Gertig and Nowczyk, 1975; GarcõÂa and Mourelle, 1984; Chu et al., 1986, 1988; Kuntz et al., 1990), but some of these changes were transient (Gertig and Nowczyk, 1975; Chu et al., 1988). Our monkeys were the only test animals that had a consistent decrease in cholesterol levels with increasing dose of toxaphene and time on test. It should be noted that very few of the cholesterol values were outside of the historical range found with the cynomolgus monkeys in our colony. When these data were plotted, it was found that the cholesterol values for the control group tended to increase with time on test while the values for the treated groups changed minimally over time. However, since all of the mean cholesterol values, one standard error, were well within the historical range for this species in our monkey colony, the decrease in cholesterol values during the pre-mating phase of this study is not believed to have biological signi®cance. Data from rodent studies have demonstrated inconsistencies between the sexes and among generations during reproduction studies regarding the e€ects of toxaphene on cholesterol values ( Kennedy et al., 1973; Chu et al., 1986, 1988). Therefore, it would appear that toxaphene may have some minimal e€ect on cholesterol metabolism in several laboratory species but the biological signi®cance of these observations is not apparent. Toxaphene was never used as a pesticide on Canadian-grown food crops but toxaphene was used to control pests a€ecting domesticated mammals. In 1980, Agriculture Canada changed the regulatory status of toxaphene in an attempt to minimize human exposure to it by suspending its registration (Agriculture Canada, 1980). Because toxaphene was never used on Canadiangrown food crops, a maximum residue level (MRL) for toxaphene was never established. However, Canada's Food and Drug Regulations specify that if a MRL has not been established for a pesticide, and it is found to be present in food at a level in excess of 0.1 ppm, it is in violation of the Regulations. For purposes of risk assessment, Health Canada has established a TDI (total daily intake) of 0.2 mg/kg body weight/day for toxaphene (J. Salminen, pers. commun.). It is due to this policy that Kuhnlein et al. (1995) concluded that 50% of the quanti®ed dietary intake recalls collected from Arctic indigenous women living on Ban Island exceeded what they termed the ``acceptable or tolerable daily intake levels'' established for toxaphene by Health Canada. However, due to what these authors term the ``multifactorial and multidisciplinary'' nature of a risk± bene®t analysis, they concluded that ``regulations or advice to alter traditional food consumption patterns to avoid exposure to OC (organochlorines) does not appear to be warranted for the communities studied.'' In conclusion, the only statistically signi®cant e€ect associated with ingestion of toxaphene at levels up to

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

0.8 mg/kg body weight during the ®rst 75 weeks of this study was a decreasing level of serum cholesterol with time on test. However, this observation was not believed to be biologically signi®cant. Acknowledgements The authors wish to express their sincere appreciation to the following individuals for their assistance regarding veterinary support: Dr. J. Fournier and technical assistance; Ms. D. Beauchamp, L. Coady, C. Leblanc and M. Norman; Messrs. R. Henry, P. Huard, D. Johnson, R. Miller and D. Patry. The comments of Dr. G. Bond and Mr. M. Feeley during the preparation of this manuscript were greatly appreciated. References Agriculture Canada (1980) Trade Memorandum: Changes in the Regulatory Status of Toxaphene, 31 October, 1980. Andersson, O., Linder, C.-E., Olsson, M., ReutergaÊrdh, L., Uvemo, U.-B., Wideqvist, U., 1988. Spatial di€erences and temporal trends of organochlorine compounds in biota from the Northwestern hemisphere. Archives of Environmental Contamination and Toxicology 17, 755±765. Andrews, P., Headrick, K., Pilon, J.-C., Bryce, F., Iverson, F., 1996. Capillary GC±ECD and ECNI GCMS characterization of toxaphene residues in primate tissues during a feeding study. Chemosphere 32, 1043±1053. Armitage, P. (1973) Statistical Methods in Medical Research. pp. 271± 275. Blackwell Scienti®c, London. Arnold, D.L., Bryce, F., Stapley, R., McGuire, P.F., Burns, D., Tanner, J.R. and Karpinski, K. (1993) Toxicological consequences of Aroclor 1254 ingestion by female rhesus (Macaca mulatta) monkeys. Part 1A. Prebreeding phase: clinical health ®ndings. Food and Chemical Toxicology 31, 799-810. ATSDR (Agency for Toxic Substances and Disease Registry), 1996. Toxicological Pro®le for Toxaphene. Baeumler, W. (1975) Side e€ects of toxaphene on mice. Anzeiger fuer Schaedlingskunde, P¯anzenschutz, Umweltschutz 48, 475±477. Barrie, L.A., 1986. Arctic air pollution: an overview of current knowledge. Atmospheric Environment 20, 643±663. Bidleman, T.F., Olney, C.E., 1975. Long transport of toxaphene insecticide in the atmosphere of the western North Atlantic. Nature 257, 475±477. Bidleman, T.F., Patton, G.W., Hinckley, D.A., Walla, M.D., Cotham, W.E., Hargrave, B.T., 1990. Chlorinated pesticides and polychlorinated biphenyls in the atmosphere of the Canadian Arctic. In: Kurtz, D.A. (Ed.), Long Range Transport of Pesticides. Lewis Publishers, Chelsea, MI, pp. 347±372. Bryce, F., Hayward, S., Stapley, R., Arnold, D.L., 2000. Consequences of Aroclor1 1254 ingestion on the menstrual cycle of rhesus (Macaca mulatta) monkeys. Food and Chemical Toxicology 38, 1053±1064. Canadian Council on Animal Care, 1984. Guide to the Care and Use of Experimental Animals, Vol. 2, pp. 163±173. C&EN, 1982. EPA bans most uses of pesticide toxaphene. Chemical and Engineering News, 25 October, p. 6. Chu, I., Villeneuve, D.C., Sun, C.-W., Secours, V., Procter, B., Arnold, E., Clegg, D., Reynolds, L., Valli, V.E., 1986. Toxicity of toxaphene in the rat and beagle dog. Fundamental and Applied Toxicology 7, 406±418.

475

Chu, I., Secours, V., Villeneuve, D.C., Valli, V.E., Nakamura, A., Colin, D., Clegg, D.J., Arnold, E.P., 1988. Reproduction study of toxaphene in the rat. Journal of Environmental Science and Health B23, 101±126. Cohen, D., Bowes, G., Ali, S., 1982. Toxaphene California State Water Research Control Board Toxic Substance Control Program Special Projects Report 82-4Sp (cited in Saleh, 1991). DeWailly, E., Ayotte, P., Bruneau, S., LaliberteÂ, C., Muir, D.C.G. and Norstrom, R.J. (1993). Inuit exposure to organochlorines through the aquatic food chain in Arctic QueÂbec. Environmental Health Perspectives 101, 618-620. Eichers, T.R., Andrilenas, P.A., Anderson, T.W., 1978. Farmers' use of pesticides in 1976. Agriculture Economics Report No. 418. US Department of Agriculture, Economics, Statistics and Cooperative Service, Washington, DC. EPA (US Environmental Protection Agency), 1980. Ambient Water Quality Criteria for Toxaphene. EPA 440/5-80-076. NTIS PB81117863. Prepared by the Oce of Health and Environment Assessment, Environmental Criteria and Assessment Oce, Cincinnati, OH, for the Oce of Water Regulations and Standards, Washington, DC. EPA (US Environmental Protection Agency), 1982a. Toxaphene; intent to cancel or restrict registrations of pesticide products containing toxaphene; denial of applications for registration of pesticide products containing toxaphene; determination concluding the reputable presumption against registration; availability of decision document. Federal Register 47, 53784±53793. EPA (US Environmental Protection Agency), 1982b. Toxaphene: Decision Document and Federal Register Notice. NTIS Accession Number: PB83-144204, Report no. EPA-540/9-82-027, Oce of Pesticide Programs, Washington, DC. EPA (US Environmental Protection Agency), 1987. Health E€ects Assessment for Toxaphene. NTIS Accession Number: PB88182886/XAB, Report no. EPA/600/8-88/056. Environmental Criteria and Assessment Oce, Cincinnati, OH. FAO/WHO (Food and Agriculture Organisation of the United Nations/World Health Organization), 1969. Evaluation of Some Pesticide Residues in Food. FAO/WHO, Geneva. Fernie, S., Wrenshall, E., Malcolm, S., Bryce, F., Arnold, D.L., 1994. Normative hematologic and serum biochemical values for adult and infant rhesus monkeys (Macaca mulatta) in a controlled laboratory environment. Journal of Toxicology and Environmental Health 42, 53±72. GarcõÂa, M.B., Mourelle, M., 1984. Gamma-glutamyl transpeptidase: a sensitive marker in DDT and toxaphene exposure. Journal of Applied Toxicology 4, 246±248. Gertig, H., Nowczyk, W., 1975. The in¯uence of carathane and toxaphene on the activity of some enzymes in rat's tissues in the studies in vivo. Polish Journal of Pharmacology and Pharmacy 27, 357±364. IARC (1979) Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Vol. 20. Some Halogenated Hydrocarbons. pp. 327±348. International Agency for Research on Cancer, Lyon. IPCS (International Program on Chemical Safety), 1984. Environmental Health Criteria 45, Camphechlor. World Health Organization, Geneva. p. 26. Jansson, B., Vaz, R., Blomkvist, G., Jensen, S., Olsson, M., 1979. Chlorinated terpenes and chlordane components found in ®sh, guillemot and seal from Swedish waters. Chemosphere 4, 181±190. Kennedy Jr, G.L., Frawley, J.P., Calandra, J.C., 1973. Multigeneration reproduction e€ects of three pesticides in rats. Toxicology and Applied Pharmacology 25, 589±596. Korte, F.I., Scheunert, I., Parlar, H., 1979. Toxaphene (camphechlor): a special report. Commission on Terminal Pesticide Residues. International Union of Pure and Applied Chemistry 51, 1585±1601. Kuhnlein, H.V., Receveur, O., Muir, D.C.G., Chan, H.M., Soueida, R., 1995. Arctic indigenous women consume greater than acceptable levels of organochlorines. Journal of Nutrition 125, 2501±2510.

476

D.L. Arnold et al. / Food and Chemical Toxicology 39 (2001) 467±476

Kuntz, D.J., Rao, N.G.S., Berg, I.E., Khattree, R., Chaturvedi, A.K., 1990. Toxicity of mixtures of parathion, toxaphene and/or 2,4-D in mice. Journal of Applied Toxicology 10, 257±266. Lehman, A.J. (1965) Summaries of Pesticide Toxicity. The Association of Food and Drug Ocials of the United States, Topeka, KS, p. 38. Maier-Bode, 1965. Ofkabzebscgytzuttek-RuÈckstaÈnde, Insektizide. Verlag Eugene Ulmer, Stuttgart (cited in Korte et al., 1979, p. 1593). Miller, R.G., 1966. Simultaneous Statistical Inference. McGraw-Hill, New York. Morrison, D.F., 1976. Multivariate Statistical Methods. McGrawHill, New York. Muir, D.C.G., Ford, C.A., Grift, N.P., Metner, D.A., Lockhart, W.L., 1990. Geographic variation of chlorinated hydrocarbons in burbot (Lota lota) from remote lakes and rivers in Canada. Archives of Environmental Contamination and Toxicology 19, 530±542. Rahn, K.A., 1981. Relative importances of North America and Eurasia as sources of Arctic aerosol. Atmospheric Environment 15, 1447±1455. Rapaport, R.A., Eisenreich, S.J., 1986. Atmospheric deposition of toxaphene to eastern North America derived from peat accumulation. Atmospheric Environment 20, 2367±2379. Saleh, M.A., 1991. Toxaphene: chemistry, biochemistry, toxicity and environmental fate. Reviews of Environmental Contamination and Toxicology 118, 2±85. Santolucito, J.A., 1975. The use of quantitative EEG for detecting low-level prolonged exposure to pesticides. Commission of the European Communities, Europe; Issue Europe 5360. Proceedings Ð International Symposium: Recent Advances in the Assessment of Health E€ects of Environmental Pollution 4, 2387±2394. Schmitt, C.J., Ribick, M.A., Ludke, J.L., May, T.W., 1983. National Pesticide Program: Organochlorine Residues in Freshwater Fish, 1976±79. U.S. Fish and Wildlife Service, Washington, DC.

SundstroÈm, G., 1981. Toxaphene in the Swedish environment: transport via aerial fallout. Proceedings of the 17th Nordic Water Research 1, 331±336. Treon, J. F., Cleveland, F., Poynter, B., Wagner, B., Gahegan, T., 1952. The physiologic e€ects of feeding experimental animals on diets containing toxaphene in various concentrations over prolonged periods. Unpublished report, Kettering Laboratory (cited in FAO/ WHO, 1969, p. 270; EPA, 1982b, pp. IV-12±13; IPCS, 1984, p. 26). Truelove, J.F., Tanner, J.R., Langlois, I.A., Stapley, R.A., Mes, J.C., Arnold, D.L., 1990. E€ect of polychlorinated biphenyls on several endocrine reproductive parameters in the female rhesus monkey. Archives of Environmental Contamination and Toxicology 19, 939± 943. Tryphonas, L., Arnold, D.L., Zawidzka, Z., Mes, J., Charbonneau, S., Wong, J., 1986a. A pilot study in adult rhesus monkeys (M. Omulatta) treated with Aroclor 1254 for two years. Toxicologic Pathology 14, 1±10. Tryphonas, L., Charbonneau, S., Tryphonas, H., Zawidzka, Z., Mes, J., Wong, J., Arnold, D.L., 1986b. Comparative aspects of Aroclor 1254 toxicity in adult cynomolgus and rhesus monkeys: a pilot study. Archives of Environmental Contamination and Toxicology 15, 159±169. Tryphonas, H., Hayward, S., O'Grady, L., Loo, J.C.K., Arnold, D.L., Bryce, F., Zawidzka, Z.Z., 1989. Immunotoxicity studies of PCB (Aroclor 1254) in the adult rhesus (Macaca mulatta) monkeyÐpreliminary report. International Journal of Immunopharmacology 11, 199±206. Voldner, E.C., Schroeder, W.H., 1990. Long range atmospheric transport and deposition of toxaphene. In: Kurtz, D.A. (Ed.), Long Range Transport of Pesticides. Lewis Publishers, Chelsea, MI, pp. 223±231. von RuÈmker, R., Lawless, E.W., Meiners, A.F., Lawrence, K.A., Kelsco, G.L., Horay, F., 1974. Production, Distribution, Use and Environmental Impact of Selected Pesticides. Contract No. EQC311, for the US Environmental Protection Agency, EPA 540/1-74001, pp. 196±204.