IPCS INCHEM Home



    PESTICIDE RESIDUES IN FOOD - 1997


    Sponsored jointly by FAO and WHO
    with the support of the International Programme
    on Chemical Safety (IPCS)




    TOXICOLOGICAL AND ENVIRONMENTAL
    EVALUATIONS 1994




    Joint meeting of the
    FAO Panel of Experts on Pesticide Residues
    in Food and the Environment
    and the
    WHO Core Assessment Group 

    Lyon 22 September - 1 October 1997



    The summaries and evaluations contained in this book are, in most
    cases, based on unpublished proprietary data submitted for the purpose
    of the JMPR assessment. A registration authority should not grant a
    registration on the basis of an evaluation unless it has first
    received authorization for such use from the owner who submitted the
    data for JMPR review or has received the data on which the summaries
    are based, either from the owner of the data or from a second party
    that has obtained permission from the owner of the data for this
    purpose.



    ABAMECTIN (addendum)

    First draft prepared by
    D.J. Clegg
    Carp, Ontario, Canada

        Explanation
        Evaluation for acceptable daily intake
           Biochemical aspects
              Tissue distribution of the 8,9-Z isomer of avermectin B1a
              Exploratory study of uptake of the 8,9-Z isomer by fetal 
                CF-1 mice
              Placental P-glycoprotein levels in CF-1 mice of the 
                +/+ genotype
              P-Glycoprotein in immature rats
              Development of P-glycoprotein in rat fetuses
              Immunohistochemistry of P-glycoprotein in rhesus monkeys
              P-Glycoprotein in monkey fetuses
              P-Glycoprotein in human fetuses
           Toxicological studies
              Acute toxicity
              Short-term toxicity
              Reproductive toxicity
                   Multigeneration reproductive toxicity
                   Developmental toxicity
              Special study: Hypersensitivity of dogs to ivermectin
           Observations in humans
        Comments
        Toxicological evaluation
        References

    Explanation

        Abamectin (a mixture containing > 80% avermectin B1a and
    < 20% avermectin B1b) was evaluated toxicologically by the Joint
    Meeting in 1992 and 1994 (Annex 1, references 65 and 71). An ADI of
    0-0.0002 mg/kg bw was allocated in 1994 on the basis of an NOAEL of
    0.12 mg/kg bw per day in a multigeneration study of reproductive
    toxicity in rats, using a safety factor of 500. The safety factor was
    increased because of concern about the teratogenicity of the 8,9- Z 
    isomer (earlier identified as the Delta-8,9 isomer), which is a
    photolytic degradation product that forms a variable part of the
    residue on crops. In 1995 (Annex 1, reference 74), the Joint Meeting
    established a separate ADI of 0-0.001 mg/kg bw for abamectin itself as
    the basis for risk assessment when abamectin is used as a veterinary
    drug and the residue does not contain the 8,9- Z isomer. The present
    Meeting reviewed information that was requested by the 1994 JMPR and
    reconsidered the decision of the 1995 JMPR.

        Data submitted to the 1994 JMPR indicated that the high
    sensitivity of CF-1 mice to the neurotoxicity of avermectins is
    associated with P-glycoprotein deficiency in the small intestine and
    in the capillary endothelial cells of the blood-brain barrier. It was
    speculated that the heterogeneity of the response in CF-1 mice may
    explain the absence of a dose-response relationship for maternal
    toxicity in the studies of teratogenicity. Data submitted to the
    present Meeting resolved the issue of the variability seen in earlier
    studies in CF-1 mice.

        The 1994 JMPR considered additional data, comprising information
    on the photo-oxidative stability of avermectin B1a (the major
    component of abamectin) and a study on the relative sensitivities of
    CF-1 and CD-1 mice. In addition, because of the close structural
    relationship between ivermectin and abamectin and the similarity of
    the toxic effects of these compounds in various animal species, data
    on ivermectin, comprising two studies in primates and one in humans
    were also considered. The JMPR addendum on abamectin does not include
    evaluations of these studies. The Meeting commented that in a study of
    photo-oxidative stability, the half-life of the 8,9- Z isomer was 4.5
    h and that of  avermectin B1a, 6.5 h. In comments on the first study
    in rhesus monkeys, it was indicated that ivermectin had no effects on
    body weight, clinical signs, or ophthalmological, haematological,
    clinical chemical or pathological manifestations at the high dose of
    1.2 mg/kg bw per day administered orally for two weeks to immature
    monkeys. The second study on ivermectin, in neonatal monkeys given a
    maximum of 0.1 mg/kg bw per day for two weeks by nasogastric
    intubation, also showed no effects.

        P-Glycoprotein, a component of the plasma membrane, has been
    associated with multi-drug resistance: cells with high P-glycoprotein
    levels in the membrane show decreased uptake, decreased drug
    steady-state levels, and decreased drug retention. Further, ivermectin
    has been shown in mice genetically engineered for disruption of the
    gene encoding P-glycoprotein to be a substrate for this protein. These
    genetically P-glycoprotein-deficient mice had a dramatic increase in
    sensitivity to a variety of P-glycoprotein substrates, including
    ivermectin. A five-day study by administration in the diet, reviewed
    below and summarized by the 1994 JMPR, indicated the absence of
    adverse neurotoxic effects of abermectin at 0.8 mg/kg bw per day in
    CD-1 mice and in most CF-1 mice; however, 17% of the CF-1 mice were
    highly sensitive, showing severe neurotoxic signs 3-4 h after the
    initial dose. Immunohistochemical investigation showed that the
    sensitive mice had much lower P-glycoprotein levels in the cerebellum,
    cerebral cortex, and jejunum than insensitive CF-1 mice and CD-1 mice.
    Using restriction fragment length polymorphism (RFLP), three
    P-glycoprotein genotypes can be distinguished, +/+, +/- and -/-, the
    incidence being in the 1:2:1 ratio expected on the basis of Mendelian
    inheritance. 

        The 1994 JMPR concluded that CF-1 mice are an inappropriate model
    for studying the toxicity (including teratogenicity) of avermectins.
    Consequently, the ADI was revised and based on a re-evaluation of the
    multigeneration study of reproductive toxicity in rats. The NOAEL in
    this study was assessed in 1992 to be 0.05 mg/kg bw per day, on the
    basis of a reduction in maternal body weight during early lactation.
    The re-evaluation in 1994 was based on the consideration that this
    effect was unlikely to be adverse and the NOAEL was therefore
    increased to 0.12 mg/kg bw per day on the basis of toxicity seen in
    the pups in the same study. Since questions about the potential
    teratogenicity of the 8,9- Z isomer remained unanswered, a safety
    factor of 500 was again used, giving an ADI for abamectin and its
    8,9- Z isomer of 0-0.0002 mg/kg bw.

        Abamectin was considered again by the JMPR in 1995. The Meeting
    concluded that the existing ADI for abamectin and its 8,9- Z isomer
    is appropriate when abamectin is used as a plant protection product
    but is inappropriate for residues after veterinary use, since the
    8,9- Z isomer is not detectable in such residues. The Meeting
    therefore allocated an ADI for abamectin alone, again based on the
    NOAEL in the two-generation study of reproductive toxicity in rats of
    0.12 mg/kg bw per day, but using a 100-fold safety factor, to give an
    ADI of 0-0.001 mg/kg bw.

        The 1994 JMPR indicated a number of studies that would provide
    information useful for continued evaluation of abamectin:

        -- data on P-glycoprotein in other species, including humans;

        -- establishment and validation of a more sensitive method to
           assess the neurotoxic effects of avermectins in rodents;

        -- study of the acute toxicity of the 8,9- Z isomer in CF-1 and
           CD-1 mice, with measurements of P-glycoprotein in blood and
           brain;

        -- study of the teratogenicity of abamectin and the 8,9- Z isomer
           in CD-1 and CF-1 mice with concurrent measurements of
           P-glycoprotein, in order to correlate its presence with
           maternal toxicity and teratogenicity.

        Additional data have been received on the toxicity of single doses
    of abamectin in CF-1 mice and of the 8,9- Z isomer in CF-1 and CD-1
    mice, on the short-term toxicity of abamectin in CF-1 and CD-1 mice,
    on the teratogenicity of the 8,9- Z isomer in CF-1 and CD-1 mice, on
    the effects of ivermectin on reproduction in rats, on the
    immunohistochemistry of P-glycoprotein in rhesus monkeys, on the
    toxicity of a single oral dose of radiolabelled 8,9- Z isomer, on the
    effect of a single dose on the onset of P-glycoprotein expression in
    the developing placenta of CF-1 mice, and on fetal uptake of the
    8,9- Z isomer in CF-1 mice. Studies on P-glycoprotein expression in
    brain and jejunum of human fetuses, in human placentas, in adult and
    fetal rhesus monkeys, and in rat fetuses and pups are also summarized.

    Evaluation for Acceptable Daily Intake

    1.  Biochemical aspects

    (a)  Tissue distribution of the 8,9-Z isomer of avermectin B1a

        Groups of eight 20-28-week-old CF-1 mice of each sex, genotyped
    for P-glycoprotein expression by RFLP, were given a single oral dose
    of 0.2 mg/kg bw of the radiolabelled isomer (dose volume, 10 ml/kg),
    to deliver about 1.2 µCi/mouse. The radiolabelled material was a
    sesame-oil solution of 0.02 mg/ml unlabelled isomer mixed with an
    equal amount of 8,9- Z[53H] avermectin B1a. Four mice of each sex
    of the +/+, +/-, and -/- genotypes were necropsied 8 and 24 h after
    treatment. The brains and blood from the vena cava were removed from
    all mice, and the testes were removed from males. Radiolabel was
    determined in heparinized blood in microfuge tubes and in organs
    (rinsed in saline and blotted dry) in preweighed scintillation tubes.

        The levels of radiolabel in the brains of male and female -/- mice
    were about 60 times higher than those in +/+ mice. By 24 h, the
    difference was even greater, since clearance occurred in +/+ and +/-
    mice but not in -/- mice. At both 8 and 24 h, +/- mice had higher
    levels in the brain than +/+ mice. In the testes, the levels of
    radiolabel were highest in mice of the -/- genotype, accumulation
    resulting in higher levels at 24 h. In mice of the +/- and +/+
    genotypes, radiolabel levels in testes were lower than that in plasma,
    but the levels were in approximate equilibrium at 24 h. Plasma
    radiolabel levels were also highest in mice of each sex of the -/-
    genotype at 8 h, but the difference between the -/- and +/+ genotypes
    was smaller than seen in the organ systems. By 24 h, the plasma
    radiolabel level was higher in males than in females (Lankas et al.,
    1997a).

    (b)  Exploratory study of uptake of the 8,9-Z isomer by fetal CF-1
    mice

        Thirty-five female CF-1 mice, about 20 weeks of age, weighing
    25-52 g, and of the +/- genotype for P-glycoprotein expression, were
    given 1.5 mg/kg bw (about 2.5 µCi/mouse) of radiolabelled 8,9- Z 
    isomer of avermectin B1a on gestation day 17. Ten mice were killed 4,
    8, and 24 h after treatment, and the maternal plasma levels of the
    test material were determined; fetuses were removed, sexed, weighed,
    genotyped for P-glycoprotein expression (three litters at 4 h and two
    at both 8 and 24 h), and analysed for radiolabel. Placentas were
    analysed for P-glycoprotein by western blotting.

        The maternal plasma levels varied significantly among animals,
    even though all of the mice were of the +/- genotype. Maximum plasma
    levels were observed 8 h after treatment, the levels at 4 and 24 h
    being approximately equivalent. The level of radiolabel (dpm/g fetus)
    was clearly highest in -/- fetuses, intermediate in +/- fetuses, and
    lowest in +/+ fetuses at all intervals and in all litters, and these

    differences increased with time of sacrifice after treatment. These
    differences and increases were still clearly apparent when the levels
    were expressed as dpm per g fetus/dpm per ml plasma. P-Glycoprotein
    was found in all placentas from fetuses with a positive allele but not
    in those from from fetuses with only negative alleles, in the one
    litter for which data were reported (Lankas et al., 1997b).

    (c)  Placental P-glycoprotein levels in CF-1 mice of the +/+ genotype

        Ten female CF-1 mice homozygous (+/+) for P-glycoprotein were
    cohabited with males of the same species, strain, and genotype. Day 0
    of gestation was designated as the day a vaginal plug was found. One
    or two females were killed on day 9, 11, 13, or 15 of gestation, and
    their placentas were removed and placed in dry ice. The placentas from
    a single litter on gestation days 9, 11, and 13 were pooled, and those
    from days 15 and 17 were assayed individually (the placentas from
    gestation day 17 were obtained from a study of fetal uptake reported
    elsewhere in this addendum.) Placentas were assayed for P-glycoprotein
    by western blotting. P-Glycoprotein was present as early as day 9.
    Relative to total placental protein, P-glycoprotein expression
    increased with duration of gestation, the highest levels being found
    on day 17 (Lankas et al., 1997c).

    (d)  P-Glycoprotein in immature rats

        The levels of P-glycoprotein were measured in the brain and small
    intestine of six young (six weeks), sexually immature rats (strain
    unspecified) of each sex by immunohistochemistry. P-Glycoprotein
    staining was found in cerebral and cerebellar capillary endothelial
    cells and in brush-border epithelial cells of the jejunum. It was
    reported that rats of this age are insensitive to the toxicity of
    avermectin (Lankas & Cartwright, 1995a).

    (e)  Development of P-glycoprotein in rat fetuses

        The development of P-glycoprotein was investigated in
    Sprague-Dawley Crl:CD(SD)BR rats by immunohistochemical and western
    blot analyses of brain and intestine of female rats, of their fetuses
    on day 20 of gestation, and of pups on days 2, 5, 8, 11, 14, 17, and
    20  post partum. The 40 F0 females, 10 weeks old (weighing 200-300
    g) at the start of the study, were paired 1:1, and mating was
    confirmed by the presence of a vaginal plug. (Whether this was
    considered day 0 or day 1 of gestation is not stated, although in
    previous studies from this laboratory the day on which a vaginal plug
    was found was considered to be day 0 of gestation.) Four unmated
    females were also studied. Deaths among the pregnant females were
    checked daily. One fetus of each sex per litter of four F0 females
    was killed on day 20 of gestation by hypothermia, and the brain and
    jejunum were snap frozen and analysed for P-glycoprotein. The brain,
    jejunum, and uterus of each of the four dams were also analysed for
    P-glycoprotein. On postnatal day 0, the litters were culled to five
    pups of each sex, with fostering as required. On days 2, 5, 8, 11, 14,
    17, and 20, one pup of each sex from each of four litters was killed,

    and the brain and jejunum were processed for immunohistochemical and,
    in some cases, western blot analyses. The uteri of the four
    non-pregnant females were examined by the same procedures as used for
    the pregnant animals. 

        No deaths were observed, and fetal weights were within the normal
    range. The litter size of the four rats sacrificed before parturition
    ranged from 13-17. No external malformations were seen in any fetuses
    or pups. P-Glycoprotein was found in F0 females in a diffuse pattern
    on the endothelial-cell surface of the capillaries of the cerebrum and
    cerebellum and on the brush border of jejunal epithelial cells;
    moderate amounts were present on the luminal surface of the uterine
    epithelium. No P-glycoprotein was observed in the uteri of
    non-pregnant rats. P-Glycoprotein staining was seen in the brains of
    all fetuses and pups; however, the intensity of staining was much
    weaker in fetuses on day 20 of gestation and in younger than in older
    pups. Virtually no immunohistochemical staining for P-glycoprotein was
    observed in the jejunum on gestation day 20 or on days 2 and 5  post 
     partum; minimal staining was seen on day 8, increasing in intensity
    with time. Western blotting of the brain and scanning dosimetry of
    P-glycoprotein showed that the levels of P-glycoprotein, expressed as
    the percentage of adult levels, were 11, 6.5, 5.7, 4.4, 6.9, 19, 37,
    and 89% on gestation day 20 and on days 2, 5, 8, 11, 14, 17, and 20
     post partum, respectively. Thus, P-glycoprotein levels are much
    lower in the brain and jejunum in the early postnatal phase, and,
    since the presence of P-glycoprotein in the tissue barrier reduces
    penetration of abamectin, the lower levels in the brain and jejunum
    are likely to permit increased penetration of abamectin. Once a
    critical level of abamectin is achieved in the brain, the toxicity to
    pups seen in the multi-generation studies is observed (Cukierski et
    al., 1996).

    (f)  Immunohistochemistry of P-glycoprotein in rhesus monkeys

        The immunohistochemistry of P-glycoprotein was investigated in the
    endothelial-cell surface of capillaries in the cerebellum and
    cerebrum, in the brush border of jejunal epithelial cells, and in the
    canaliculi of the liver of four untreated, one- to two-year-old rhesus
    monkeys of each sex. The animals were killed by pentobarbital
    injection. P-Glycoprotein staining was observed in a diffuse pattern
    in the tissues examined, the staining being least intense in the
    jejunal cells and most intense in the liver; staining was
    approximately the same in all brain sections (Schinkel et al., 1994).
    Lankas and Cartwright (1995a) state that 'It is known that juvenile
    monkeys are relatively insensitive to avermectin-related toxicity',
    probably due to a well-developed P-glycoprotein export system in the
    brain.

    (g)  P-Glycoprotein in monkey fetuses

        In another exploratory study on the immunohistochemistry of
    P-glycoprotein, five male and four female fetal monkeys were killed by
    phenobarbital injection, and sections of brain and jejunum from the
    fetuses and biopsy samples of placenta and endometrium from the
    maternal animal were frozen and analysed. A diffuse pattern of
    P-glycoprotein staining was seen on the endothelial-cell surface of
    capillaries in the cerebellum, cerebrum, cerebellar peduncle, and pons
    of the fetal monkeys, with comparable intensity of staining in all
    areas of the brain. Diffuse staining was also seen in the syncytial
    trophoblasts lining the chorionic villi, and trophoblasts in the
    marginal zone of the placenta were also weakly stained. Patchy
    P-glycoprotein staining was also seen on the endothelial surface of
    maternal uterine blood vessels, the staining being most intense in the
    trophoblasts of the chorionic villi. No staining was found in the
    jejunum of the fetal monkeys. The intensity of staining in the brains
    of fetal monkeys was comparable to that in juvenile monkeys; however,
    jejunal P-glycoprotein was observed in juvenile monkeys. Infant
    monkeys were stated to be insensitive to avermectin-induced toxicity
    (Lankas & Cartwright, 1995b).

    (h)  P-Glycoprotein in human fetuses

        Bohr and Mollgärd (1974) showed that tight junctions composed of
    strands are present in human fetuses of a crown-rump length of 60-150
    mm. Transepithelial permeability is believed to depend on the presence
    of such strands, which develop early in the human fetus. Betz and
    Goldstein (1981) investigated changes in the metabolism and transport
    properties of capillaries isolated from rat brain 1-45 days  post 
     partum and concluded that various aspects of brain capillary
    function show distinct developmental patterns which may be related to
    changes in blood-brain barrier permeability during development. The
    blood-brain barrier in rats is thus incompletely developed at birth,
    in contrast to the early development of tight junctions in the human
    fetus. 

        Van Kalken et al. (1992) examined the expression of P-glycoprotein
    in human fetuses at 7-28 weeks of gestation by immunohistochemical
    staining. P-Glycoprotein mRNA was detected in seven-week-old fetuses,
    but P-glycoprotein expression in the brain was not detected until week
    28 of gestation, when the level was comparable to that in adults.
    MacFarland et al. (1994) examined the distribution of P-glycoprotein
    in the human placenta. In the first trimester, P-glycoprotein was
    present in the syncytiotrophoblast microvillus border, and some
    placental macrophages showed weak staining. At term, no staining was
    seen in the trophoblasts but strong staining was seen in placental
    macrophages. These findings indicate that P-glycoprotein is present in
    the human fetal brain and in the placenta, but is minimal or absent in
    the brains of fetal rats.

    2.  Toxicological studies

    (a)  Acute toxicity

    (i)  Abamectin

        Female CF-1 mice weighing 23-29 g, were genotyped by RFLP with
    radiolabelled DNA probes to identify mice homozygous (+/+) and
    heterozygous (+/-) for P-glycoprotein expression. No homozygous (-/-)
    mice were used. Abamectin (purity not specified) in sesame oil was
    administered orally to groups of five female mice at doses of 0, 10,
    20, 40, or 80 mg/kg bw. All mice homozygous for P-glycoprotein that
    were given 40 mg/kg bw died within 3-6 days, and all those at 80 mg/kg
    bw within 1-3 days; all mice heterozygous for P-glycoprotein that were
    given 20 mg/kg bw died within 2-4 days and all those at 40 and 80
    mg/kg bw died within 1-2 days. The physical signs in all treated
    animals were tremors, decreased activity, and bradypnoea within hours
    of treatment and persisting for 1-4 days. Lateral recumbency was
    generally seen before death. The oral LD50 values for abamectin were
    28 mg/kg bw for homozygous (+/+) mice and 14 mg/kg bw for heterozygous
    (+/-) mice. In previous studies, the oral LD50 values for mice of the
    -/- genotype were 0.3-0.4 mg/kg bw (Hall, 1997; Lankas et al., 1997d).

    (ii)  8,9-Z Isomer of abamectin

        Groups of three randomly selected CD-1 female mice (7-8 weeks old,
    weighing 22-28 g) were given doses of 50, 90, 162, 292, or 525 mg/kg
    bw, and groups of five randomly selected CF-1 male mice (14 weeks old,
    weighing 33-39 g) were given doses of 10, 20, or 30 mg/kg bw of the
    8,9- Z isomer orally in sesame oil, and observed for seven days. CD-1
    mice at 292 mg/kg bw and above showed decreased activity and
    bradypnoea, and some animals had tremors before death; all of these
    animals were moribund within 5-6 h of treatment. No deaths occurred at
    lower doses, but mice at 90 mg/kg bw and above appeared unkempt, and
    those at 162 mg/kg bw had a hunched posture; however, all of these
    mice were normal on day 3 (at 90 mg/kg bw) or day 6 (at 162 mg/kg bw).
    Of the CF-1 mice, 1/5 at 10 mg/kg bw , 3/5 at 20 mg/kg bw, and 2/5 at
    30 mg/kg bw died, all on day 1 after treatment. Surviving animals were
    normal by day 2. The signs of toxicity included ptosis (at 10 mg/kg bw
    only), decreased activity, and bradypnoea. The acute oral LD50 of the
    8,9- Z isomer of abamectin was 217 mg/kg bw in female CD-1 mice and
    about 20 mg/kg bw in male CF-1 mice. No sex differences in toxicity
    have been reported with any of the avermectins (Lynch, 1996). 

    (b)  Short-term toxicity

        Abamectin (purity 96% by weight based on combined analysis for
    avermectin B1a and B1b) was administered to groups of 49 male and 50
    female CF-1 mice and to groups of five male and five female CD-1 mice
    at a dose of 0.8 mg/kg bw per day for four days. Severe tremors and/or
    ataxia were seen in 12 female and five male moribund CF-1 mice after
    the first dose. These animals were killed 3-4 h after treatment,
    together with two female and one male CF-1 controls; 20 of the

    remaining (insensitive) CF-1 mice and all of the remaining CF-1
    controls were killed on study day 4, when the brain and small
    intestine were removed. The cerebral cortex, cerebellum, and jejunum
    were then cut into 3-mm sections, flash-frozen, and processed
    immunohistochemically for P-glycoprotein (Schinkel et al., 1994) with
    a primary antibody specific for the intermembrane binding site for
    P-glycoprotein (Kartner et al., 1985). Samples of cerebellum,
    cerebrum, and small intestine were also processed for assessment of
    P-glycoprotein by western immunoblotting after crude extraction of
    lysed cells to obtain cell membranes and other cytoplasmic components.
    The protein content of the samples was measured by solubilizing them,
    electrophoretic transfer to nitrocellulose and incubation overnight at
    4 ¡C with C219, a primary antibody for P-glycoprotein, then reaction
    for 1 h at room temperature with horseradish peroxidase-linked
    anti-mouse immunoglobulin A. Immunoreactive P-glycoprotein was then
    detected with a chemiluminescent substrate. After completion of the
    biochemical studies, five male and five female insensitive CF-1 mice
    and a random sample of five male and five female or 10 female CD-1
    mice were given single doses of 1-10 mg/kg bw abamectin per day.

        P-Glycoprotein was not detected in the brain or small intestine of
    11/12 female or 5/5 male  CF-1 mice killed after the first dose
    because of moribundity; the one remaining female had very small
    amounts of P-glycoprotein in the brain. The surviving CF-1 mice and
    all CD-1 mice had very slight to intense P-glycoprotein staining.
    Staining for P-glycoprotein of variable intensity was seen in vehicle
    control CF-1 mice. Western immunoblots indicate intense banding of
    P-glycoprotein in all three tissues from CD-1 mice and
    abamectin-insensitive CF-1 mice but virtually no P-glycoprotein in
    sensitive CF-1 mice.

        Challenge with higher doses (1, 2.5, 5, or 10 mg/kg bw) of
    abamectin resulted in mild signs of toxicity in 6/10 insensitive CF-1
    mice at 5 mg/kg bw and in 8/10 at 10 mg/kg bw. The signs included
    slight tremors and ataxia 5-6 h after treatment, with recovery within
    24 h except for one mouse at 10 mg/kg bw which still showed slight
    signs of toxicity but recovered by 48 h. CD-1 mice at 10 mg/kg bw
    showed no signs of toxicity (Lankas et al., 1994, 1997d).

    (c)  Reproductive toxicity

    (i)  Multigeneration reproductive toxicity

           In a multigeneration study of reproductive toxicity, groups of
    20 female and 10 male Crl:CD(SD)BR rats were given ivermectin (purity,
    97.7%) in sesame oil orally at doses of 0, 0.4, 1.2, or 3.6 mg/kg bw
    per day. No effects were seen on parental body weights, either before
    treatment (70 days) or during mating and pregnancy. The duration of
    gestation was increased in animals at 3.6 mg/kg bw per day. The
    incidence of live offspring per litter on day 1  post partum was
    slightly increased at 1.2 mg/kg bw per day in F1a litters but not at
    other doses or in F1b litters. Deaths of pups by day 7 were increased

    at 1.2 and 3.6 mg/kg bw per day in F1a litters and at all doses in
    F1b and F2a litters, and the dose of 3.6 mg/kg bw per day was
    terminated because of the high mortality rate: 53% by day 14 of
    lactation. The weights of pups of the F1a, F1b, and F2a generations
    were generally comparable at all doses on day 1 but were decreased by
    day 7 and thereafter in F1a litters at 3.6 mg/kg bw per day and in
    F1b and F2a litters at 0.4 and 1.2 mg/kg bw per day. Delayed incisor
    eruption was also seen at 3.6 mg/kg bw per day. Since treatment-
    related increases in pup mortality early in lactation and reduced pup
    weight gain were seen at all doses, the study was terminated early,
    and a second study with doses of 0, 0.05, 0.1, 0.2, and 0.4 mg/kg bw
    per day was initiated. No effects were reported, except for an
    increase in mortality in F3a litters at 0.4 mg/kg bw per day between
    days 1 and 7.

        In order to assess the postnatal toxicity of ivermectin,
    cross-fostering was studied. Forty female rats were given 2.5 mg/kg bw
    per day for 61 days and then mated with untreated males but continued
    to receive ivermectin throughout mating and gestation. On day 1 
     post partum, the litters of 40 control female rats treated with
    sesame oil and those of the treated rats were culled to four pups of
    each sex, which were cross-fostered in four groups: control parents ×
    control pups, control parents × treated pups, treated parents ×
    treated pups, and treated parents × control pups. Postnatal survival,
    growth, and development were monitored until weaning on day 25  post 
     partum. Increased mortality and decreased pup weights were seen in
    treated parents × treated pups and treated parents × control pups, but
    not in the other crosses. These results indicate that neonatal
    toxicity is a function of postnatal exposure, since the progeny
    affected were those of dams continuously exposed to ivermectin.

        Metabolic studies were also performed in eight-week-old female
    rats given tritiated ivermectin orally at 2.5 mg/kg bw per day
    (specific activity, 0.2 mCi/mg). Six rats in group 1 were given the
    compound for 61 days and then throughout mating and gestation to day 9
     post partum, and six in group 2 received the compound on days 1-9
     post partum. Blood was collected from the orbital sinus of two rats
    in group 1 on days 5, 10, 15, 20, 40, and 60 of exposure and from two
    dams in each group on days 1, 4, and 6  post partum. Two pups were
    taken from each group 1-5 h and 1, 4, and 6 days  post partum, and
    radiolabel was measured in blood and in pup brain and carcass. All
    animals were killed 24 h after the last dose on day 10  post partum, 
    and blood, liver, brain, and carcass were collected. Milk samples were
    obtained from two dams in each group on days 4, 6, and 10  post 
     partum.

        The levels of radiolabel in plasma in animals in group 1 increased
    and plateaued after 10-15 days. The plasma levels during lactation
    increased by three to four times between day 60 (before mating) and
    day 1  post partum and then gradually diminished to levels comparable
    to those before mating by day 10  post partum. In group 2, the plasma
    levels were low on day 1  post partum but increased to levels about
    equal to those in group 1 by day 10. The levels of radiolabel in milk

    were consistently two to three times higher than those in plasma in
    both groups. The levels in brain, liver, and carcass on day 10 were
    comparable in the two groups. Pups from group 1 had very low plasma
    levels on day 1  post partum, which increased rapidly, however, so
    that and on days 6 and 10 the level was two to three times greater
    than that in the dam. In pups from group 2, no radiolabel was detected
    in plasma on day 1; by days 4 and 6, the level was about half that
    seen in pups, and by day 10 the levels were comparable to those in
    pups from group 1. The residues in pup liver, brain, and carcass on
    day 10  post partum exceeded those in adults by two to three times.
    The plasma:brain ratios of radiolabel in offspring of the two groups
    were about 1 on days 1 and 4  post partum and 2-3 on days 6 and 10.
    These results can be interpreted in terms of the development of the
    blood-brain barrier in rats and of maternal fat mobilization and
    consequent release of lipophilic ivermectin. They are consistent with
    development of the blood-brain barrier on days 6-10  post partum, as
    also indicated by the mortality rate among pups. In humans, the
    blood-brain barrier develops prenatally (Lankas et al., 1989).

    (ii)  Developmental toxicity

        Groups of 25 mated Crl:CF-1 BR mice, 10 weeks old and weighing
    19-29 g, were given the 8,9- Z isomer of avermectin B1 (purity, 99%)
    in sesame oil at doses of 0, 0.015, 0.03, or 0.06 mg/kg bw per day on
    days 6-15 of gestation by gavage. The presence of a vaginal plug was
    considered to be day 0 of gestation. Physical signs were reported
    daily; body weights were recorded on days 0, 6, 8, 10, 12, 14, 16, and
    17 of gestation; and mice were killed on day 17. The uteri were
    examined, and implants were counted and reported as resorptions or
    dead or live fetuses. Maternal animals were subjected to gross
    necropsy, and the fetuses were weighed, sexed, and examined
    externally. Every third fetus per litter was dissected, as were all
    externally malformed fetuses. The head of every third fetus was fixed
    in Bouin's fixative. All fetuses were processed for alizarin staining
    and skeletal examination.

        No maternal deaths were reported. A single incidence of abortion,
    probably occurring on gestation days 10-12 and unlikely to be related
    to treatment, was observed in an animal at 0.06 mg/kg bw per day. No
    treatment-related physical signs were observed, and no effects were
    seen on body weight or body-weight changes between days 6-16 and 16-17
    of gestation. Food intake was not significantly affected at any dose.
    No gross organ changes were observed in dams, and no compound-related
    effects were seen on the incidence of implantations, fetal deaths, or
    resorptions, the sex ratio of fetuses, or fetal body weight. The
    external malformations reported were exencephaly in three fetuses in
    three litters at 0.03 and in three fetuses in two litters at 0.06
    mg/kg bw per day. One of the fetuses at 0.03 mg/kg bw per day also had
    a ventricle septal defect in the heart. A single fetus at 0.015 mg/kg
    bw per day group had a cleft palate. Undated data on historical
    control CF-1 mice gave the incidence of exencephaly as 68/25 037
    fetuses (i.e. 0.27%), with a maximum of 7/428 (1.6%). In the study
    reported, the incidence was 1.3% at 0.03 mg/kg bw per day and 1.4% at

    0.06 mg/kg bw per day, i.e. within the upper limits of those of
    historical controls. Similarly, the incidence of cleft palate in this
    study is well within the range in historical controls. Lack of a
    dose-response relationship also mitigates against compound-induced
    effects. The NOAEL for both maternal and developmental toxicity was
    0.06 mg/kg bw per day, the highest dose tested (Gordon et al., 1985a).

        Five groups of 25 mated female Crl:CF-1 BR mice, 9.5 weeks of age
    and weighing 21-28 g were given the 8,9- Z isomer of avermectin B1
    (purity, 99%) in sesame oil at doses of 0, 0.015, 0.03, 0.1, or 0.5
    mg/kg bw per day by gavage on days 6-15 of gestation. The presence of
    a vaginal plug was considered to be day 0 of gestation. Physical signs
    were observed daily, and body weights recorded on days 0, 6, 8, 10,
    12, 14, 16, and 17 of gestation. At sacrifice on day 17 of gestation,
    uterine implants were counted and classified as resorptions or live or
    dead fetuses. Complete gross necropsies were performed. All fetuses
    were weighed, sexed, and examined externally. Every third fetus per
    litter was dissected, as were all externally abnormal fetuses. The
    head of every third fetus was fixed in Bouin's fixative for
    sectioning, and all fetuses were processed for alizarin staining and
    skeletal examination.

        Physical signs of toxicity (chromodacryorrhoea and lethargy) were
    seen in one female at 0.5 mg/kg bw per day on day 9 of gestation,
    which continued until sacrifice when moribund on day 12. The effects
    were considered to be due to treatment. No effects on body weight,
    weight gain, or food intake were seen in other animals. No gross
    macroscopic changes were reported. The number of implantations per
    pregnant female was slightly (not statistically significantly) reduced
    at 0.5 mg/kg bw per day, and the incidence of resorptions was
    increased at 0.03 and 0.5 mg/kg bw per day; however, the latter was
    not dose related and was statistically significant only at 0.03 mg/kg
    bw per day. Fetal weights were not affected, but were slightly
    increased (not statistically significantly) at 0.5 mg/kg bw per day,
    probably due to the reduced number of live pups arising from the
    reduced implantation rate and slightly increased resorption rate.

        The external malformations included exencephaly in one fetus in
    one control litter, two fetuses in two litters at 0.015 mg/kg bw, five
    fetuses in two litters at 0.03 mg/kg bw, no fetuses at 0.1 mg/kg bw,
    and two fetuses in two litters at 0.5 mg/kg bw. Cleft palate was found
    in none of the controls, in one fetus in one litter at 0.015 mg/kg bw,
    in one fetus in one litter at 0.03 mg/kg bw, in six fetuses in one
    litter at 0.1 mg/kg bw, and in 23 fetuses in six litters at 0.5 mg/kg
    bw. The percentage incidences of fetuses with exencephaly were 0.3,
    0.7, 2.1, 0, and 0.8% and of affected litters, 0.4, 0.8, 0.87, 0, and
    0.87%, and the percentage incidences of fetuses with cleft palate were
    0, 0.35, 0.42, 2.1, and 9.9%, and of litters, 0, 4, 4.3, 4.2, and 26%.
    The clumping of incidence of cleft palate within litters at 0.1 and
    0.5 mg/kg bw per day is very clear. Visceral and skeletal examinations
    revealed a number of minor malformations (cervical ribs, lumbar ribs,
    and sternebral ossification defects), which were not considered to be

    related to treatment. The NOAEL for maternal toxicity was 0.1 mg/kg bw
    per day, and that for teratogenicity was 0.03 mg/kg bw per day (Gordon
    et al., 1985b).

        A group of 276 female Crl:CF-1 BR mice, 9-10 weeks old and
    weighing 22-32 g, was given abamectin (purity, 97.1%) at 0.4 mg/kg bw
    in sesame oil in order to distinguish those that were sensitive and
    insensitive to this drug. Mice that had tremors or recumbency were
    considered to be sensitive. Insensitive mice were further tested by
    giving them an additional oral dose of 0.8 mg/kg bw. Sensitive mice
    comprised 27% of the tested population, but 69% of the sensitive mice
    died during the selection process and only 23 survived. Sensitive and
    insensitive mice were mated 1:1 with CF-1 males of unknown sensitivity
    two to three weeks after treatment. Four groups of 25 mated,
    insensitive animals were given the 8,9- Z isomer (purity, 79.8%) in
    sesame oil at 0, 0.5, 1, or 1.5 mg/kg bw per day on days 6-15 of
    gestation, day 0 of gestation being considered the day of mating. A
    control group of four mated, sensitive female CF-1 mice was treated
    similarly with sesame oil. Other mated, sensitive CF-1 female mice
    were given one to three doses of 0.2 mg/kg bw per day, followed by two
    doses (to all but one female which had not reached gestation day 6) of
    0.3 mg/kg bw per day. The dose was increased further to 0.5 mg/kg bw
    per day for one day and then to 1 mg/kg bw per day for a further day.
    After this dose, 17/18 females showed decreased activity or were
    recumbent, and the eighteenth female (last mated) was recumbent after
    two doses of 1 mg/kg bw per day. Treatment was therefore withheld for
    two days, after which 6/18 females appeared normal and were given one
    to three doses of 0.75 mg/kg bw per day. One mouse on day 9, two mice
    on day 12, two mice on day 13, three mice on day 14, four mice on day
    15, and two mice on day 17 of gestation died or were killed when
    moribund. Of the four survivors, three were recumbent for two days
    before and on the day of terminal sacrifice (day 18 of gestation).
    Animals that died before term were checked for pregnancy and
    discarded.

        All mice were observed for physical signs of toxicity daily, with
    an additional examination 1-5 h after treatment. Body weights were
    recorded on gestation days 0, 6, 8, 10, 12, 14, 16, and 18, and food
    intake was measured between gestation days 3-5, 6-8, 9-11, 12-14, and
    15-17. At term all female mice were killed, internal organs were
    examined for gross changes, pregnancy status was determined, and
    corpora lutea were counted. The cerebrum and cerebellum of the four
    surviving sensitive mice, seven insensitive controls, and 11
    insensitive animals receiving 1.5 mg 8,9- Z isomer were processed for
    P-glycoprotein immunochemistry. The uterine contents were examined for
    the numbers of implants, live fetuses, dead fetuses, and resorption
    sites. Fetuses were examined externally after weighing, and then half
    of the fetuses per litter (including externally malformed and/or dead
    fetuses) were dissected for visceral examination. The heads of live
    fetuses were sectioned. All fetuses were processed for skeletal
    staining and examination.

        Histopathological examination of the brains of CF-1 females at
    termination of the study showed diffuse P-glycoprotein staining on the
    endothelial-cell surface of capillaries of the cerebrum and cerebellum
    of insensitive mice and the absence of such staining in sensitive
    mice.

        No adverse physical signs were observed, except in the sensitive
    females given increasing doses. No effects were seen in this group at
    doses of 0.2 or 0.3 mg/kg bw per day, but at 0.5 mg/kg bw per day
    closed eyes and decreased activity were reported, and decreased
    activity, recumbency, and ptosis were also observed at 1.0 mg/kg bw
    per day. These signs persisted when the dose was reduced to 0.75 mg/kg
    bw per day. The body weight, weight gain, and food consumption of
    insensitive mice were unaffected. In sensitive mice, body weight and
    weight gain were reduced from about day 10 of gestation, and food
    intake declined throughout the study, the decrease increasing with
    time. The incidence of pregnancy, number of corpora lutea, incidence
    of resorptions, sex ratio, pup weight, and placental structures were
    comparable in all insensitive groups. The four sensitive controls, one
    of which had only three implantations, of which one was resorbed
    early, had comparable numbers of corpora lutea; the increased
    preimplantation loss was due to one mouse that had only three implants
    (13 corpora lutea). The placental structure was normal. In the treated
    sensitive mice, the numbers of corpora lutea per pregnant female were
    comparable to those in controls. All females became pregnant, but
    three of the four surviving litters comprised only dead pups, probably
    due to immobility of the females during the last two or more days of
    gestation. The body weights of the 11 live pups in one litter were
    comparable to those of controls. 

        An increasing incidence of cleft palates was seen in the
    insensitive animals, with 2.4% in controls, 4.4% in those at 0.5 mg/kg
    bw per day, 6.9% at 1.0 mg/kg bw per day, and 20% at 1.5 mg/kg bw per
    day, and also in sensitive mice, with 0% in controls and 45% in the
    single surviving live litter of treated animals. Other malformations,
    including exencephaly, open eyelids, brachydactyly, and hind-limb
    rotation, were observed in insensitive mice, but the highest incidence
    of these malformations was in the control group. Malformations found
    only in treated animals were limited to insensitive mice and comprised
    anencephaly (1/295 fetuses at 0.5 mg/kg bw per day) and macrophthalmia
    (1/307 fetuses), micrognathia (1/307 fetuses), and tail malformation
    (2/307 fetuses) at 1.5 mg/kg bw per day. Vertebral malformations were
    seen in all groups except the sensitive controls, the incidence being
    similar and low. Variants included a dose-related increase in the
    incidence of pseudopolydactyly in insensitive mice. The occurrence of
    extra cervical ribs, supernumary ribs, and sternal ossification
    defects did not appear to be associated with treatment. The apparently
    dose-related increase in the incidence of cleft palate in the treated
    insensitive mice is not totally unexpected, since the expression of
    cleft palate is dependent on fetal genotype. Insensitive parental mice
    would include both the +/+ and +/- genotypes, and hence pups could
    exhibit any of the three genotypes; thus, the sensitivity of the pups
    within each test group will vary. In the absence of information on

    fetal genotype, NOAELs cannot be determined. There was no NOAEL for
    induction of malformations in sensitive or insensitive mice. The NOAEL
    for maternal toxicity in insensitive CF-1 mice was 1.5 mg/kg bw per
    day, the highest dose tested; that in sensitive mice was much lower
    but could not be determined because of the changing doses (Wise et
    al., 1996a).

        Four groups of 22 mated female Crl:CD-1 BR mice, aged 12 weeks and
    weighing 24-32 g, were given the 8,9- Z photoisomer of abamectin
    (purity, 98.1%) in sesame oil at doses of 0, 0.75, 1.5, or 3 mg/kg bw
    per day by gavage on days 6-15 of gestation. They were cohabited 1:1
    with males; the day a copulatory plug was found was considered to be
    gestation day 0. The test solution was shown to be of adequate
    stability, with 95-100% of the nominal concentrations. Physical signs
    were observed daily; body weight was recorded on days 0, 6, 8, 10, 12,
    14, 16, and 18, and food consumption on days 3-5, 6-8, 9-11, 12-14,
    and 15-17. All mice were killed on day 18 and subjected to gross
    necropsy. Implants were counted and classified as resorptions or live
    or dead fetuses. Placentas were examined for gross abnormalities.
    Fetuses were weighed and examined for gross abnormalities. One-half of
    the fetuses and all dead or abnormal fetuses were dissected. The heads
    of live fetuses were fixed in Bouin's fluid for subsequent sectioning,
    and all fetuses were processed by alizarin staining for skeletal
    examination. Distal tail segments were collected from all fetuses and
    mating partners for analysis of the P-glycoprotein genotype.

        No adverse physical signs or deaths were reported. The mean
    body-weight gain of treated animals was slightly (not statistically
    significantly) increased, but the effect was not dose related and was
    not considered to be toxicologically significant. Food intake was
    generally comparable in all groups. No treatment-related findings were
    observed on gross necropsy. All of the adult animals examined (12 of
    each sex at 3 mg/kg bw per day, 10 males and nine females at 1.5 mg/kg
    bw per day, 11 of each sex at 0.75 mg/kg bw per day, and 11 male and
    12 female controls) showed the +/+ genotype for P-glycoprotein. The
    numbers of preimplantation losses were higher in controls than in
    treated groups, mainly due to high losses in 4/20 females, which
    resulted in a slightly lower control implantation rate (12) than in
    treated animals (> 13). The incidence of resorptions was comparable
    in all groups. The birth of 12 stillborn pups to dams at 0.75 mg/kg bw
    per day before sacrifice resulted in an increased incidence of dead
    offspring. The sex ratios were normal. 

        The malformations observed included cleft palate, in none of the
    controls, two pups (0.73%) at 0.75 mg/kg bw, one (0.31%) at 1.5 mg./kg
    bw, and four (1.4%) at 3 mg/kg bw per day; one tail malformation at
    0.75 mg/kg bw per day; and hind-limb extension in two fetuses in each
    group. The incidence of cleft palates at 3 mg/kg bw per day was four
    in three litters, which is within the incidence of historical controls
    of this strain drawn from contemporary studies (1995-96) performed
    within the same laboratory. The incidences in 14 control groups were
    used, but 10 of these were from range-finding studies in which the
    mice were injected intravenously with sodium bicarbonate in saline

    (the carrier in these studies). The percentage of affected fetuses
    ranged from 0.32 to 3.7% (1/312 to 11/296). In the present study, the
    percentage incidence of cleft palate at the high dose was 1.5%
    (4/272). Two of the control groups exceeded this level (2% and 3.7%).
    Further, it should be noted that the incidences at the three doses
    were 2/254, 1/251, and 4/272, with no strict dose-effect relationship.
    The other variants observed and the limb extension and tail
    malformations were probably not related to treatment. The NOAEL for
    maternal toxicity and for induction of malformations was 3 mg/kg bw
    per day, the highest dose tested (Wise et al., 1996b).

        The genotypes of Crl:CF-1 BR female mice, approximately 11 weeks
    of age and weighing 23-31 g, were determined by RFLP and Southern
    blotting, and groups of 12 female mice were paired according to
    genotype, as follows: sesame oil controls, +/- females × +/- males and
    -/- females × -/- males; treated, +/+ females × +/+ males, +/- females
    × +/+ males, and+/- females × -/- males. After mating, day 0 of
    gestation being considered the finding of a copulatory plug, treated
    females were given the 8,9- Z photoisomer of abamectin (purity,
    94.3%) at a dose of 1.5 mg/kg bw per day in sesame oil on gestation
    days 6-15. Clinical signs were recorded daily before treatment and
    twice daily during treatment. Body weight was measured on gestation
    days 0, 6, 8, 10, 12, 14 16, and 18. Mice were killed on day 18, and a
    tail clip was collected for further genotyping if necessary. After
    confirmation of pregnancy, implantations were counted and classified
    as resorptions or live or dead fetuses, and fetuses were weighed and
    examined externally for malformations. They were then killed and a
    hind limb was collected for possible P-glycoprotein genotyping.
    Samples were also collected of five placentas from a control dam of
    the negative genotype (-/- × -/-) and a control dam of the positive
    genotype (+/+ × +/+) and of all placentas from two control animals
    with the control genotype (+/- × +/-). Heads and placentas were also
    collected from all fetuses of four control dams of the +/- × +/-
    genotype and four dams at 1.5 mg/kg bw per day of the +/- × -/-
    genotype; the heads and placentas were preserved for P-glycoprotein
    immunohistochemistry. Finally, the brain of the control dam with the
    +/+ × +/+ genotype from which the five placentas were collected was
    preserved for immunohistochemistry.

        No deaths or adverse clinical signs were observed during the
    study. The body weights were not statistically significantly affected,
    but the +/- × +/- control group had moderately increased body-weight
    gain, which was considered to be incidental. The incidence of
    pregnancy was slightly lower in both +/- × +/- and -/- × -/- control
    groups (8 and 9/12, respectively) than in the treated groups (12/12 in
    all groups). This finding is probably incidental, since treatment did
    not commence until day 6 of gestation. The number of implants was
    increased slightly in the +/- × +/- cross, resulting in a slightly
    larger litter size (13 ± 0.9) than in the second control (-/- × -/-,
    12 ± 4.4) or treated groups (10 ± 4.0 to 12 ± 2.8). The incidences of
    resorptions were comparable in all groups. No fetal deaths were
    recorded, and fetal weights were comparable in all groups.

        Genotypic analysis showed that all of the offspring of +/+ × +/+
    crosses that were analysed were +/+ and all of those of the -/- × -/-
    crosses were -/-. Matings of +/- control males and females resulted in
    15 +/+, 32 +/-, and 18 -/- offspring (the expected 1:2:1 Mendelian
    ratio). Western blotting of the placentas showed that the amount of
    P-glycoprotein in the placenta depended on the genotype of the fetus,
    the +/+ genotype having the greatest P-glycoprotein content, the +/-
    genotype having less, and the -/- genotype having no P-glycoprotein.
    The incidence of cleft palate was 0 in the treated +/+ × +/+ group,
    12% in the +/- × +/+ treated group, and 58% in the +/- × -/- treated
    group. The incidence was 0.83% in the +/- × +/- control group and 0 in
    the -/- × -/- control group. No +/+ offspring of treated dams showed
    cleft palate, whereas 30/31 fetuses with the -/- genotype and 29/70
    (41%) of fetuses with the +/- genotype had this malformation. One
    control fetus of the -/- genotype had exencephaly. The incidence of
    postaxial pseudopolydactyly was 8.3% of all fetuses for +/- × +/-
    controls and 2.8% for -/- × -/- controls. The incidences were 11% in
    treated +/- × -/- offspring, 0.8% in +/- × +/- offspring, and 1.4% in
    +/+ × +/+ offspring. The results clearly indicate that the fetal
    genotype is closely related to both the incidence of cleft palate and
    the presence of P-glycoprotein (Lankas et al., 1996).

    (d)  Special study: Hypersensitivity of dogs to ivermectin

        Beagles and most others strains of dogs tolerate single doses of
    3-10 mg/kg bw ivermectin; however, some collie dogs have been reported
    to be hypersensitive to doses of 0.1-0.2 mg/kg bw. In order to
    investigate this phenomenon, a colony of collies made up of
    individuals derived from individual breeding pairs was screened for
    sensitivity to ataxia, depression, tremors, and mydriasis at low
    doses. Three sensitive and three insensitive dogs were killed, and
    their livers and brains were analysed immunohistochemically for
    P-glycoprotein. The insensitive dogs showed slight, consistent
    staining of brain capillaries for P-glycoprotein, but the sensitive
    dogs showed no staining. In the liver, P-glycoprotein was strongly
    stained in bile caniculi and hepatocytes of both sensitive and
    insensitive dogs. These results suggest that collies have a rare
    mutation of the  mdr1a gene, which is responsible for brain
    P-glycoprotein expression, but have no such mutation in the  mdr2 
    gene, responsible for liver P-glycoprotein expression.

    3.  Observations in humans

        Ivermectin has been administered to humans for the treatment of
    parasitic diseases. Fifty million doses of 0.2 mg/kg bw have been
    given worldwide. No evidence of toxicity has been reported, even when
    ivermectin was used at much higher doses (up to 1.6 mg/kg bw), and no
    adverse neurological effects have been reported (personal
    communication from Merck Research Laboratories).

    Comments

    Abamectin

        In a study to establish the LD50 of abamectin in CF-1 mice
    genotyped for P-glycoprotein expression, similar signs of toxicity
    were seen in +/+ (homozygous) and +/- (heterozygous) mice. The oral
    LD50 for +/+ mice was 28 mg/kg bw, and that for +/- mice was 14 mg/kg
    bw. Separate studies indicated an LD50 for -/- mice of 0.3-0.4 mg/kg
    bw. Thus, in CF-1 mice, the oral LD50 appears to be related to the
    genotype for P-glycoprotein.

        A four- to five-day study with CF-1 and CD-1 mice given abamectin
    at 0.8 mg/kg bw per day resulted in severe toxicity in 17% of the CF-1
    mice after the first dose. P-Glycoprotein was not detectable in the
    brain or jejunum of these mice, except in one mouse which had minimal
    expression in the brain. Insensitive CF-1 mice showed
    slight-to-intense P-glycoprotein staining, and CD-1 mice showed
    intense P-glycoprotein staining. Oral administration of challenge
    doses (10 mg/kg bw) to groups of insensitive CF-1 mice after five days
    of treatment at 0.8 mg/kg bw per day caused slight toxicity, with
    complete recovery within one to two days. The results indicate severe
    toxicity in mice of the -/- genotype and variable toxicity in those of
    the +/- and +/+ genotypes. No toxicity was seen in CD-1 mice, which
    are probably of the +/+ genotype. These results indicate that the
    genotype of CF-1 mice with respect to P-glycoprotein expression
    governs the toxicity of abamectin. There is no evidence that mutations
    of this genotype occur in CD-1 mice.

    8,9-Z Isomer

        Studies of tissue distribution in genotyped CF-1 mice after
    administration of radiolabelled 8,9- Z isomer indicated marked
    differences according to genotype. In the brain, the levels in -/-
    mice of each sex were about 60 times greater than those in +/+ mice.
    By 24 h, the difference was even greater, since clearance occurred in
    +/+ mice but not in -/- mice. A similar pattern was seen in the
    testes, with the highest levels in -/- mice. At 24 h, the testicular
    levels in +/- and +/+ genotype mice were in equilibrium with those in
    plasma. In plasma, the level of radioalabel was highest in -/- mice,
    but the differences between the +/+ and -/- genotypes were much less
    than in the organ systems.

        A single oral dose to CD-1 and CF-1 mice resulted in LD50 values
    of 220 mg/kg bw for female CD-1 mice and about 20 mg/kg bw for male
    CF-1 mice. Data on other avermectins indicate no sex difference for
    acute toxicity. Since the signs fit the pattern for neurotoxicity, it
    is probable that the low LD50 value (i.e. increased susceptibility)
    in CF-1 mice is related to the accessibility of the target organ to
    the test material and hence to the presence or absence of
    P-glycoprotein expression.

        A number of studies of developmental toxicity were performed in
    CF-1 mice. In the first study, the NOAEL for both maternal and
    developmental toxicity was 0.06 mg/kg bw per day, the highest dose
    tested. Cleft palate and exencephaly were observed, but the incidence
    was within historical control limits. A second study showed an NOAEL
    for maternal toxicity of 0.1 mg/kg bw per day and an LOAEL, based on
    signs of toxicity, of 0.5 mg/kg bw per day. The NOAEL for
    teratogenicity was 0.03 mg/kg bw per day, the incidences of cleft
    palate at doses of 0.1 mg/kg bw per day and above being greater than
    those in historical controls; however, at 0.1 mg/kg bw per day, cleft
    palates were seen in only one litter; at 0.5 mg/kg bw per day, the
    incidence of cleft palate showed clumping within litters. A third
    study with the 8,9- Z isomer was performed in female CF-1 mice which
    had been screened for sensitivity to abamectin before the start of the
    study. Sensitive and insensitive female mice were paired with males of
    unknown sensitivity, and the doses given to sensitive female mice were
    varied during exposure. Marked effects on sensitive mice occurred at
    doses above 0.5 mg/kg bw, only 4/18 animals surviving to term. Of
    these, only one mouse had a live litter. No effects were seen on
    insensitive mice at doses up to 1.5 mg/kg bw per day; however, cleft
    palates were observed at all doses between 0.05 and 1.5 mg/kg bw per
    day. This study demonstrates that the incidence of malformations is
    not related to maternal toxicity.

        Yet another study of developmental toxicity was performed using
    parental mice of known genotype for the  mdr-1 gene, which encodes
    P-glycoprotein expression. This study indicated a relationship between
    the parental genotype and the incidence of cleft palate: at 1.5 mg/kg
    bw per day, cleft palate was observed in none of the offspring of +/+
    × +/+ crosses, in 12% of those of +/+ male × +/- female crosses, and
    58% of those of -/- male × +/- female crosses; one cleft palate
    occurred in the control +/- × +/- cross and none in the -/- × -/-
    cross. Genotypic analysis of the fetuses from treated females showed
    no cleft palates in +/+ mice, 41% in +/- mice, and 97% in -/- mice.
    Analyses of the placentae for P-glycoprotein showed a correlation with
    the genotype of the fetus, the levels being highest in +/+ fetuses and
    absent in -/- fetuses; the +/- control matings yielded a Mendelian
    distribution of 15 +/+, 32 +/-, and 18 -/- pups. A close relationship
    between fetal genotype and the presence of P-glycoprotein in the
    placenta would be expected, since much of the placenta is formed from
    fetal tissue. The relationship between the incidence of cleft palate
    and genotype appears to be a reflection of prevention by
    P-glycoprotein expression of penetration of the test material through
    placental membranes.

        The presence of placental P-glycoprotein was investigated in +/+
    male and +/+ female CF-1 mice. Western blotting of the placentae
    indicated the presence of P-glycoprotein on day 9 of gestation, the
    levels increasing with the duration of gestation. Levels present at
    the time of palatal closure (about day 15) would be sufficient to
    hinder placental transfer of the 8,9- Z isomer of abamectin. Passage
    of the radiolabelled isomer across the placenta was investigated after
    administration to +/- female mice on day 17 of gestation. The maternal

    plasma levels of radiolabel were variable but maximal 8 h after
    treatment. The levels of radiolabel in the fetus depended on the fetal
    genotype, being lowest in +/+ fetuses and highest in -/- fetuses,
    indicating that blockage of placental transfer depends on genetically
    controlled expression of placental P-glycoprotein.

        A study of developmental toxicity in CD-1 mice treated by gavage
    with doses of 0, 0.75, 1.5, or 3 mg/kg bw per day of the 8,9- Z 
    isomer showed no adverse effects on the maternal mice. The incidences
    of cleft palate were 0, 2, 1, and 4 (or 0, 0.73, 0.31, and 1.4%) at 0,
    0.75, 1.5, and 3 mg/kg bw per day, respectively. These incidences were
    not dose-related and fell within control incidences seen after oral or
    intravenous administration of vehicles in the same laboratory. The
    NOAEL for maternal, embryo-and fetal toxicity was 3 mg/kg bw per day.

    Ivermectin

        A multigeneration study in rats given ivermectin at doses of 0.4
    mg/kg bw per day and greater resulted in early mortality of pups 
     post partum and reduced pup body-weight gain. The study was
    terminated early. A further study at doses of 0.05-0.4 mg/kg bw per
    day showed no effects, except increased pup mortality in the F3a
    litters at 0.4 mg/kg bw per day between days 1 and 7  post partum. 
    Postnatal toxicity was assessed in a series of cross-fosterings of
    newborn pups and was shown to be due to postnatal, not in-utero,
    exposure. The postnatal toxicity was further investigated with
    radiolabelled ivermectin given either before or throughout mating and
    gestation. In the rats dosed  post partum, the levels of radiolabel
    in plasma were initially low but were comparable to those observed
    after long-term exposure by day 9  post partum. The levels of
    radiolabel in milk were consistently three to four times the plasma
    levels in both groups, which may reflect mobilization of ivermectin
    from fatty tissues. In the offspring of parents treated  post 
     partum, radioalabel was not detected in plasma on day 1  post 
     partum, but was half that in the offspring treated by long-term
    exposure on days 4 and 6, and equivalent on day 9. The tissue levels
    in the pups on day 9 were two to three times those in the parents. The
    plasma:brain ratios of radiolabel in the offspring of both groups were
    1 on days 1 and 4  post partum and 2-3 on days 6-9. These findings
    can be interpreted to indicate that the development of the blood-brain
    barrier in rat offspring is delayed, occurring some time after
    parturition. The postnatal toxicity observed in rats may be a function
    of the accessibility of the target organ to the toxin, owing to the
    late formation of the blood-brain barrier and to possible mobilization
    of ivermectin from adult fatty tissues.

    P-Glycoprotein distribution in adult and young animals

        In the immature rat (about six weeks old), P-glycoprotein is
    present in the brain and in the brush-border epithelial cells of the
    jejunum. In fetal animals (day 20 of gestation), however, minimal
    P-glyco-protein was detected in the brain, the levels being less than
    10% of that in adult animals up to day 14 and then increasing rapidly.

    No P-glycoprotein was detected in the jejunum of fetal rats or in rats
    on days 2 or 5  post partum; P-glycoprotein was detectable by day 8
     post partum, and the levels increased with time thereafter. These
    data indicate late expression of P-glycoprotein, occurring some 10-15
    days  post partum. In non-pregnant adult rats, P-glycoprotein was not
    observed in the uterus; it was present, however, on the luminal
    surface of the uterine epithelium in pregnant rats.

        P-Glycoprotein was present on the endothelial surface of
    capillaries in the cerebrum, cerebellum, cerebellar peduncle, and pons
    of rhesus monkey fetuses. The intensity of staining was comparable in
    all areas of the brain. P-Glycoprotein was also present in the
    placenta, but none was detected in fetal jejunum. The brain levels of
    P-glycoprotein in monkey fetuses were comparable to those in the
    brains of one- to two-year-old rhesus monkeys examined in another
    study.

        In humans, P-glycoprotein was detected in the brain capillaries of
    fetuses aborted at 28 weeks, but not at earlier gestational ages. The
    levels found were comparable to that in the adult brain. In human
    placenta, P-glycoprotein was found in the syncytiotrophoblast
    microvillus border and in some placental macrophages in the first
    trimester, but mainly in the placental macrophages at term.

    Species sensitivity to avermectins

        Increased sensitivity has been seen in CF-1 mice and in Collie
    dogs. In no other species or strain of animal has increased
    sensitivity to avermectins been observed. In humans, 50 000 000 doses
    of  0.2 mg/kg bw ivermectin have been administered for treatment of
    parasitic diseases, with no report of toxicity directly attributable
    to the drug. Higher doses (1.6 mg/kg bw) have also not resulted in
    toxicity. Treatment of humans is not usually required more often than
    yearly.

        The data reviewed on the effects of the 8,9- Z isomer on
    developmental toxicity in the CF-1 mouse clearly indicate a strong
    relationship between the increased incidence of cleft palate and
    reduced expression of P-glycoprotein in this strain of mouse. Since
    the phenomenon is seen only in these animals, the Meeting considered
    that use of the results of studies with this strain is not appropriate
    in establishing an ADI. The NOAEL for teratogenic activity in the CD-1
    mouse was 3 mg/kg bw per day.

        Data on the avermectins have been used in the overall review of
    abamectin. In the multigeneration studies of reproductive toxicity of
    ivermectin and abamectin in rats, the critical adverse effects were
    those on pups during early lactation. The studies with ivermectin
    indicate that the pup mortality and reduced body-weight gains seen
    early in lactation may be associated with delayed development of
    P-glycoprotein expression. Reduced glycoprotein expression in early
    lactation correlates with the mortality of pups. Young pups are not
    only more susceptible to abamectin because they lack P-glyco-protein

    expression, but they are also exposed to levels of abamectin in milk
    that are two to three times those in maternal plasma. P-Glycoprotein
    expression in humans is fully developed by week 28 of gestation. The
    appropriateness of the multigeneration study of reproductive toxicity
    of abamectin as the basis for the ADI is, therefore, questionable.

        Because of the hypersusceptibility of rats postnatally, the
    Meeting determined that a reduced interspecies safety factor would be
    appropriate for establishing an ADI. A safety factor of 50 was
    therefore applied to the NOAEL from the multigeneration study in rats
    (0.12 mg/kg bw per day) to give an ADI of 0-0.002 mg/kg bw, which is
    supported by the NOAEL of 0.24 mg/kg bw per day in the one-year study
    in dogs, applying a safety factor of 100. A single ADI for both
    abamectin and its 8,9- Z isomer was deemed to be appropriate, since
    the potential teratogenicity of the isomer has been satisfactorily
    explained.

    Toxicological evaluation

    Levels that cause no toxicological effect

     Abamectin

        Mouse:     4 mg/kg bw per day (two-year study of toxicity and
                   carcinogenicity)

        Rat:       1.5 mg/kg bw per day (two-year study of toxicity and
                   carcinogenicity)
                   0.12 mg/kg bw per day (two-generation study of
                   reproductive toxicity)

        Dog:       0.25 mg/kg bw per day (one-year study of toxicity)

     8,9-Z isomer

        Mouse:     3 mg/kg bw per day (study of developmental toxicity in
                   CD-1 mice)

     Estimate of acceptable daily intake for humans (sum of abamectin 
     and 8,9-Z isomer)

           0-0.002 mg/kg bw

     Studies that would provide information useful for continued 
     evaluation of the compound

        Further observations in humans, possibly involving repeated
        exposure

    References

    Betz, A.L. & Goldstein, G.W. (1981) Developmental changes in
    metabolism and transport properties of capillaries isolated from rat
    brain.  J. Physiol., 312, 365-376.

    Bohr, V. & Mollgärd, K. (1974) Tight junctions in human choroid plexus
    visualized by freeze-etching.  Brain Res., 81, 314-318.

    Cukierski, M.A., Lankas, G.R., Umbenhauer, D.R. & Cartwright, M.E.
    (1996) Exploratory study of P-glycoprotein development in rat fetuses.
    Unpublished report TT#94-739-0, submitted to WHO by Merck Research
    Laboratories, Three Bridges, NJ, USA.

    Gordon, L.R., Clark, R.L., Nickell, B.E., Collevechio, K. & Landes,
    D.K. (1985a) 8,9 Isomer of avermectin B1: Oral teratology study in
    mice. Unpublished report TT#85-710-0, submitted to WHO by Merck
    Research Laboratories, Three Bridges, NJ, USA.

    Gordon, L.R., Clark, R.L., Nickell, B.E., Collevechio, K. & Geiger,
    J.E. (1985b) 8,9 Isomer of avermectin B1: Oral teratology study in
    mice. Unpublished report TT#85-710-1, submitted to WHO by Merck
    Research Laboratories, Three Bridges, NJ, USA.

    Hall, S. (1997) Abamectin: Exploratory acute oral toxicity study in
    mice. Unpublished report TT#96-2727, submitted to WHO by Merck
    Research Laboratories, Three Bridges, NJ, USA.

    Kartner, N., Evemden-Porelle, D., Bradley, G. & Ling, V. (1985)
    Detection of P-glycoprotein in multidrug resistant cell lines by
    monoclonal antibodies.  Nature, 316, 820-823.

    Lankas, G.R. & Cartwright, M.E. (1995a) Exploratory P-glycoprotein
    immunohistochemistry in rhesus monkeys. Unpublished report (project ID
    # 618-936-PGRM), dated 14 March 1995, submitted to WHO by Merck
    Research Laboratories, West Point, PA, USA.

    Lankas, G.R. & Cartwright, M.E. (1995b) Exploratory P-glycoprotein
    immunohistochemistry in fetal rhesus monkeys. Unpublished report,
    submitted to WHO by Merck Research Laboratories, Three Bridges, NJ,
    USA.

    Lankas, G.R. & Cartwright, M.E., (1995c) Exploratory study of
    P-glycoprotein in non-pregnant rats. Unpublished report, submitted to
    WHO by Merck Research Laboratories, Three Bridges, NJ, USA.

    Lankas, G.R., Minsker, D.H. & Robertson, R.T. (1989). Effects of
    ivermectin on reproduction and neonatal toxicity in rats.  Food 
     Chem. Toxicol., 27, 523-529.

    Lankas, G.R., Cartwright, M.E. & Umbenhauer, D. (1994) Exploratory
    5-day oral toxicity study comparing abamectin sensitivity and
    P-glycoprotein levels in CF-1 and CD-1 mice. Unpublished report
    TT#94-2775, dated 12 September 1994, submitted to WHO by Merck
    Research Laboratories, Three Bridges, NJ, USA.

    Lankas, G.R., Wise, L.D., Cartwright, M.E., Umbenhauer, D.R.,
    Cukierski, M.A., Nickell, B.E., Collevechio, K., Powzaniuk, W.C.,
    Kulp, S.H., Jefferson, T., Overs, M.G. & Aversa, D. (1996) L-652,280:
    Exploratory oral developmental toxicity study in CF-1 mice of known
    P-glycoprotein genotype. Unpublished report TT#96-721-0, submitted to
    WHO by Merck Research Laboratories, Three Bridges, NJ, USA.

    Lankas, G.R., Hall, S.J. & Umbenhauer, D.R. (1997a) L-652,820:
    Exploratory acute radiolabelled oral toxicity study in CF-1 mice.
    Unpublished report TT#97-2578, submitted to WHO by Merck Research
    Laboratories, Three Bridges, NJ, USA.

    Lankas, G., Umbenhauer, D. & Wise, L.D. (1997b) L-652,280: Exploratory
    fetal uptake study in CF-1 mice. Unpublished report TT#97-2538,
    submitted to WHO by Merck Research Laboratories, Three Bridges, NJ,
    USA.

    Lankas, G.R., Wise, L.D. & Umbenhauer, D. (1997c) Exploratory acute
    breeding study in CF-1 mice. Unpublished report TT#97-2568, submitted
    to WHO by Merck Research Laboratories, Three Bridges, NJ, USA.

    Lankas, G.R., Cartwright, M.E. & Umbenhauer, D. (1997d) P-Glycoprotein
    deficiency in a sub-population of CF-1 mice enhances
    avermectin-induced neurotoxicity.  Toxicol. Appl. Pharmacol., 142,
    357-365.

    Lynch, D. (1996) L652,280: Exploratory acute oral toxicity study in
    mice. Unpublished report TT#95-2754, submitted to WHO by Merck
    Research Laboratories, Three Bridges, NJ, USA.

    MacFarland, A., Abramovich, D.R., Ewen, S.W.B. & Pearson, C.K. (1994)
    Stage-specific distribution of P-glycoprotein in first trimester and
    full-term human placenta.  Histochem. J., 26, 417-423.

    Schinkel, A.M., Smit, J.J.K., van Tellingen, O., Beijnen, J.M.,
    Wagenaar, E., van Deemter, L., Mol, C.A.A.M., van der Valk, M.A.,
    Robanus-Maandag, E.C., te Riele, H.P.J., Berns, A.J.M. & Borst, P.
    (1994) Disruption of the mouse  mdr1a P-glycoprotein gene leads to a
    deficiency in the blood brain barrier and to increased sensitivity to
    drugs.  Cell, 77, 492-502.

    Van Kalken, C.K., Giaccone, G., van der Valk, P., Kuiper, C.M.,
    Hadisaputero, M.M.N., Bosma, S.A.A., Scheper, R.J., Meijer, C.J.L.M. &
    Pinedo, H.M. (1992) Multidrug resistance gene (P-glycoprotein)
    expression in the human fetus.  Am. J. Pathol., 141, 1063-1072.

    Wise, L.D., Lankas, C.R., Cartwright, M.E., Cukierski, M.A., Nickell,
    B.E., Collevechio, K., Powzaniuk, W.C., Kulp, J.H., Jefferson, T.,
    Hipple, A.L. & Duchai, D.M. (1996a) L-652,280: Oral developmental
    toxicity study in CF-1 mice. Unpublished report TT#95-741-0, submitted
    to WHO by Merck Research Laboratories, Three Bridges, NJ, USA.

    Wise, L.D., Lankas, G.R., Umbenhauer, D.R., Stabinski, L.G.,
    Cukierski, M.A., Nickell, B.E., Collevechio, K. & McMahon, M.G.
    (1996b) Oral developmental toxicity study in CD-1 mice. Unpublished
    report TT#96-732-0, submitted to WHO by Merck Research Laboratories,
    Three Bridges, NJ, USA.
    


    See Also:
       Toxicological Abbreviations
       Abamectin (Pesticide residues in food: 1992 evaluations Part II Toxicology)