Workshop on perinatal exposure to dioxin-like compounds. I. Summary.

An international workshop reviewed 20 ongoing or recently completed studies of the effects of perinatal exposures to dioxins, dibenzofurans, and PCBs on the reproductive, endocrine, neurodevelopmental, and immune systems. Many of the observed effects are consistent with these compounds acting as "environmental hormones" or endocrine disrupters. This report summarizes the conclusions and future directions described at the workshop.

current, recently completed, or planned studies of the effects of perinatal exposures to these compounds on the newborn.
The phrase "dioxin-like compounds" (DLCs) used in the workshop title refers to agents that bind to the Ah receptor. These include representatives of the coplanar, halogen-substituted multiring structures such as the polychlorinated dibenzodioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and some of the coplanar congeners in the polychlorinated biphenyl (PCB) mixtures. The most toxic and well studied of these is 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD).
The objectives of the workshop were: to review the current status of research in the field through presentations by researchers of ongoing, completed or planned studies; to provide a multidisciplinary forum to share insights on the different effects of exposures to DLCs, and to review the use of biomarkers in perinatal studies; and to strengthen future research by establishing contacts and encouraging collaborations between research groups and by setting priorities for future studies.

Orgalnization
The effects of perinatal exposures on the reproductive, endocrine, neurodevelopmental, and immune systems of the offspring were discussed in consecutive sessions, and are summarized in this report. Appendix 1 presents a summary, in chronological order, of all of the studies presented in the workshop sessions, identifying for each the investigator, study design, measures of effect, measures of exposure (compounds under study, source of exposure, matrix/tissue analyzed for dose level), and study conclusions. Fuller descriptions of these studies are given in the accompanying articles, authored by the co-chairs of each session.
In the opening session, speakers gave background information and outlined the task of the workshop. A. Gilman emphasized that organochlorine residues (OCs) continue to persist in the environment. Although the levels of many OCs in western nations have declined over the past decade, populations in different parts of the world show significant deviations from these trends. Increased global use of some persistent OCs, e.g., DDT, toxaphene, PCBs, which are banned or severely restricted in western nations, has led to increased levels of OCs in the Arctic ecosystem. Total global use of DDT may, in fact, be greater in this decade than in the 1970s, when it was banned in North America. U. Ahlborg stated that concern over widespread distribution of these compounds, especially the dioxins, dibenzofurans, and PCB mixtures, and the health consequences of perinatal exposures, have led to the initiation of risk assessments by the World Health Organization (WHO).
The next two speakers contrasted exposures of infants and adults. L. Goldman described several factors that place infants at higher risk than adults from exposures to OCs: physiologic factors (lower barriers to absorption through skin, GI tract, and lungs; lower levels of detoxifying enzymes at birth); nutritional factors (breast milk as major source of nutrition; higher caloric intake per body weight); and behavioral factors (closer dermal contact with the outside and household environments; hand-tomouth exploratory behavior). G. Lindstrom indicated that nursing infants retain almost all of the 2,3,7,8-substituted dioxins and furans that they ingest from breast milk. Nursing infants, on a bodyweight basis, have a dietary intake of TCDD and its equivalents (I-TEQs) that is 100-fold greater than adults; uncorrected for body weight, infant dietary intake is 10fold higher than in adults. Exposure of the fetus is also significant, even though transfer of DLCs from the placenta to the fetus Environmental Health Perspectives is incomplete (i.e., levels in the mother are higher than levels in the fetus).
W. Rogan and J. Jacobson described their published work which identified developmental (growth retardation and ectodermal abnormalities) and neurobehavioral (cognitive and motor functions) deficits in children whose mothers consumed PCBcontaminated cooking oil or large amounts of OC-contaminated fish during pregnancy (14,15,(17)(18)(19)23,24). The possibility that these exposures cause delayed or long-term effects, as suggested by animal studies, has prompted followup, second-generation reproduction studies of exposed cohorts in Seveso, Italy and in Taiwan by P. Mocarelli and W. Rogan, respectively.

Conclusions and Future Directions
Overall conclusions from the workshop, as well as recommendations for future research, are listed below. Following these, conclusions and research needs are given for each study area covered by a workshop session (reproductive, endocrine, neurodevelopmental, and immune toxicity).

General Conclusions
The current dominant paradigm is that the biological activities ofTCDD and the DLCs arise from their binding to the Ah receptor, and the biological effects that occur are believed to be Ah receptor-mediated. Some components of PCB mixtures (DLCs) bind tightly to the Ah receptor and are believed to operate via Ah receptordependent mechanismss. Other components do not, and must act through mechanisms that are independent of the Ah receptor (including phenobarbital-type induction of drug-metabolizing enzymes, binding to other (e.g., estrogen or thyroid hormone) receptors, or altering biogenic amine concentrations). Other PCB components may act via both Ah receptor-dependent and receptor-independent mechanisms.
There is a similarity in response in animals and humans to TCDD and DLCs. For every end point examined, human response is replicated in some species of laboratory animal: growth retardation (monkeys, rats), ectodermal effects (monkeys), spatial learning/memory deficits (monkeys, rats), and changes in lymphocyte subpopulations (monkeys). For wellstudied human end points such as chloracne and enzyme induction, homology exists with several animal species.
A working hypothesis for the mechanism of action of TCDD and DLCs is that they act as endocrine disrupters or "environmental hormones" in perinatal systems. In this way, they provide inter-or intra-cellular signals that alter growth, differentiation and function of cells in a tissue-, stage-, or cellspecific manner.
Consistent with this hypothesis, TCDD and DLCs have been demonstrated to act as multisystem effectors, affecting the developing immune, neurobehavioral, endocrine, and reproductive systems. Acting as hormones or endocrine disruptors, they may cause neoplasia by altering patterns of differentiation/proliferation of specific tissues in the developing or adult organism.

Research Needs
More interaction is needed among laboratory researchers, epidemiologists, and biostatisticians. Collaboration between laboratory researchers and epidemiologists enables animal data to generate hypotheses for human studies, and vice versa. Human and animal studies need biostatistical support so that data are useful for hazard identification, risk assessment, and public health intervention.
Summaries are needed of human and animal studies (male and female) of each organ system. These summaries would describe exposure or dose levels, nature of effects, strength of evidence, and confidence in the studies.
Better data are required for exposure assessment.
The prenatal period is a sensitive period (see animal and human studies) and needs further study.
Research is needed in human populations to investigate the potential for delayed effects in the reproductive, neurobehavioral, and neuroendocrine systems, especially among postpubescent cohorts exposed perinatally.

Reproductive Effects
Conclusions PCB mixtures cause different reproductive effects than DLCs. Lactational exposures to PCB mixtures (Aroclor 1254) cause infertility in male rat offspring without affecting their sperm count (37). TCDD decreases sperm count but does not affect fertility (4,5). Components of PCB mixtures may act like phenobarbital-type inducers, estrogens, or effectors of thyroid status or dopamine levels, while DLCs presumably operate through the Ah receptormediated events.
TCDD reduced fertility in female monkeys (1) and may exacerbate endometriosis (38). In utero and lactational exposures to TCDD caused changes in sexual differentiation in rat pups: feminization of males (reduced anogenital distance, sperm count, and accessory sex glands; abnormal mounting behavior) and urogenital abnormalities in females (absent vaginal openings, cleft clitoris) (2,4,5).
PCB congeners produced growth retardation in offspring of dosed female rats (3); chronic PCB exposures in female monkeys produced growth retardation in offspring (1); and infants born to mothers exposed to PCB/PCDF-contaminated rice oil or contaminated fish had low birth weights (14-23).
Birth size among male infants (Inuits) was inversely related to PCB concentration in breast milk of the mother (7,29).

Research Needs
The effects of perinatal exposures to TCDD and DLCs on the reproductive system of adult males are reasonably welldescribed; information is needed on mechanisms (e.g., effects on spermatogenesis and sexual differentiation).
Research is needed on the effects of these compounds on nonpregnant females (e.g., age at menarche, cyclicity, time to conception, and endometriosis).
Animal studies indicate delayed effects of perinatal exposures on reproductive systems. Perinatally exposed cohorts need to be followed beyond puberty and examined for delayed effects such as age at onset of puberty, age at menarche, reduced fertility, abnormal cyclicity, decreased sperm count, and premature menopause.

Endocrine Effects
Conclusions Exposures of pregnant female rats to small amounts of TCDD caused changes in indices of androgenic status in male offspring: spermatogenesis was inhibited; sexual behavior and the pattern of LH secretion was less masculine and more feminine (2,4,5).
Other specific PCB congeners (#77, #47) caused age-dependent changes in biogenic amine neurotransmitter levels in rats: in adults dopamine levels were reduced by ortho, but not coplanar, PCBs; dopamine levels were raised by perinatal exposures to either (9,12).
TCDD reduced responsiveness of ventral prostate to testosterone in male offspring of dosed female rats without affecting responsiveness of seminal vesicles (4).
TCDD also inhibited sexual differentiation in the CNS without altering sexual dimorphisms in estrogen-receptor concentrations or volumes of brain nuclei (5).

Researh Needs
The mechanistic relationships between DLCs, T4 and the neurodevelopmental system need further study.
Studies should focus on measurements of T4, TSH, and TRH, especially at different levels in the neuroendocrine axis.
Levels of both free and bound T3 should be examined.
More studies on binding of PCBs, metabolites, and DLC congeners to TBG and TTR are needed. Binding to TTR in rodents is important; the significance of the small amount of DLC-binding to TTR in humans should be explored.
Estrogen, testosterone, and corticosteroid levels should be measured in younger cohorts.
Perinatal exposures of monkeys to PCB mixtures, and of rats to specific PCB congeners produced spatial learning/memory deficits (10-13). Cognitive functioning was impaired among children exposed in utero to mixtures of OCs, including PCBs (14,15). PCBs produced motor deficits that persisted to age 2 among children exposed in utero (17-21,23).

Research Needs
Studies employing specific PCB congeners are needed to investigate the associations between perinatal DLC exposures, spatial learning/memory deficits, and levels of dopamine and T4 in specific regions of the brain.
Test batteries for neurobehavioral effects should include standardized, narrow-band, and challenge or stress tests.
Studies of neurobehavioral effects of chemicals are difficult because known environmental factors can overwhelm any effect of the chemical. In rats, sexual differentiation of the CNS is affected by perinatal exposures to TCDD (2,4,5). In humans, our ability to examine this association between DLCs and sexual differentiation awaits the development of measures that are less affected by confounding social/"environmental" factors.
Reliable measures of disease end points are required (e.g., health care providers in clinic networks should be trained to rate neurobehavioral deficits by consistent criteria).

Immunologic Effects Conclusions
A working hypothesis is that perinatal exposures to DLCs and PCBs alter the pattern of differentiation of cells of the immune system and, as a consequence, change the responses of immune cells. The extent and magnitude of these changes depend upon when in the development of these cells the exposures occur: exposures early in development will affect the primordial stem cells, while later exposures affect cells in a more developed system. The thymic atrophy caused by TCDD may arise in part from the depletion by TCDD of prolymphocytes in the bone marrow. TCDD-induced thymic atrophy is accompanied by decreases in lymphocyte stem cell markers which are present only in bone marrow prothymocytes (terminal deoxynucleotidyl transferase (TdT) and recombinase activating gene (RAG) (27,28). In irradiated mice, TCDD-treated prothymocytes are unable to repopulate the thymus.
DLCs may affect the primary antibody response: in humans (Inuits -PCBs and dioxins) (29) and rodents (mice-TCDD) (27,28,34,35) prenatal exposure decreases the ratio of T helper to T s uppressor cells in the thymus.
Perinatal exposures to PCBs affect primary antibody response, as suggested by a 20-fold higher incidence of infectious diseases (e.g. meningitis, measles) and ear infections (otitis media among 1-year-old Inuits with high PCB exposures) than among lesser exposed controls (7,29); by a low immunization take rate among Inuits compared to controls (7,29); and by changes in ratios of lymphocyte sub populations in animals and humans (7,26-30,34,35).

Researh Needs
Biomarkers of immunotoxicity continue to be developed; the effect of TCDD and PCBs on lymphocyte subset ratios (e.g., CD4/CD8) is still unclear in animals and humans.
Studies are needed on the primary antibody response in children with high PCB exposures, possibly employing vaccine challenge tests.
sure on responsiveness of the male rat reproductive system to testosterone stimulation in adulthood. Toxicol Appl Pharmacol