Year: 2006
Type: article
Abstract: It is believed that many phthalates affect sperm quality. In this discussion we shall first consider the reaction of the laboratory experimentalists to the epidemiological data and how to improve the epidemiological studies; then, secondly, we will examine how the exposure assessment may be improved by analysis of urine and milk samples for the sum of phthalates; and, thirdly, the epidemiologists should consider how we can produce better models in the laboratory to reflect the real world situation. Two key issues reflecting the potential effects of phthalates on humans, and the observed effects in rodents, are, first, are humans exposed to the dose levels which produce an effect in rodents?; and, secondly, are similar effects seen in humans as occur in rodent models, which is a more difficult question to address? Our dose–response data in rats looking at endpoints in adulthood, namely the incidence of infertility, cryptorchidism and hypospadias, show that significant effects on these parameters are only seen at the highest dose (500 mg/kg/day) levels, although lower doses do produce cases of infertility. In contrast, the fetal testis is more sensitive to the effects of dibutylphthalate (DBP). For example, effects on the endpoint of sex cords containing multinucleated gonocytes is detected consistently at lower doses (100 and 20 mg/kg/day). Other fetal endpoints may be even more sensitive. We have examined in vitro production of testosterone (T) in human fetal explants from abortuses to determine if effects seen in rodents are reproduced in humans. T output is reduced by ketoconazole and stimulated by 22R-hydroxycholesterol, but monobutylphthalate (MBP) and DBP have no significant effect on basal or stimulated T production. MBP and DBP similarly do not affect T production in rat fetal explants in vitro. However, rat fetal testes in vivo show almost complete suppression of T production by MBP or DBP exposure indicating that the in vivo effect is not reproduced in vitro, and the lack of effect of phthalates on human fetal testes in vitro does not signify that there is no effect in vivo (Sharpe, unpublished data). The possibility that the effects produced in vivo are due to metabolites of phthalates other than the monoester may also be considered, but I am not aware of any data to support this directly. Data presented on the association between hormone levels in neonates and human infant anogenital distance (AGD) (Swan et al., 2005; Lottrup et al., 2006) are new and remarkable. The interpretation of the biological data is plausible based on the effects seen in rodents. The dimorphism of sexual AGD is presumably regulated by androgens causing the male AGD to be twice that of females in both rats and humans which shows homology between the two species. This androgen effect presumably occurs at the prenatal or perinatal period. Phthalate exposure causes similar androgen suppression at puberty and in utero. The human data are astounding but it is not surprising that cryptorchidism has not been shown to be related to phthalate exposure because there is not a high prevalence of cryptorchidism in rats at low dose levels. However, the AGD in neonatal rat pups is a very sensitive indicator of androgen status in animals. This may be applicable to humans considering the similar sexual dimorphism, and AGD may be a good indicator of exposure to relatively small doses of endocrine disrupters. Although biologically plausible, this is not proof of cause and effect. Although we have heard about available human data from epidemiologists, toxicologists working with pesticides and industrial chemicals normally do not have direct human data and attempt to extrapolate animal mechanisms into plausible human effects. The mode of action of multiple different phthalates shows consistency across many laboratories. The range of reproductive tract malformations and other abnormalities in animals is highly consistent with reproductive deficits in humans including the testicular dysgenesis syndrome. The mechanisms of inhibition of testosterone biosynthesis and INSL3 downregulation are probably as applicable to human development and testicular descent as in rodents. We know that humans are exposed to critical phthalate metabolites in the general population, and in children via the amniotic fluid, but are we exposed to comparable levels that can produce adverse reproductive effects in rodents, and is there a sufficient margin of exposure between animals and humans to be adequately protective of health? As we look at lower doses in animals, we are recognizing new syndromes which may have relevance to humans. The level of acceptable exposure becomes a political question and policy decisions are not in the control of scientists. We have performed an unpublished multigeneration gross observational Reproductive Assessment by Continuous Breeding study looking at the effect of dietary DEHP at levels of 1.5–7500 p.p.m. (approximately equivalent to 0.1–400 mg/kg/day). Because of the low incidence of reproductive malformation caused by low levels of DBP, we retained an extra three F1 and F2 males where possible in each litter allowing them to reach adulthood before necropsy, but unfortunately without histopathology. Epididymal changes are the most robust indicators of phthalate effects, but were not identified at the lowest dose levels. A few animals had identifiable testicular and epididymal changes when the dose levels reached 15 mg/kg/day (4/64 animals) and 50 mg/kg/day (5/58 animals), but these did not reach statistical significance. A significant number of F1 and F2 animals (21/50) had changes at the 400 mg/kg/day dose level (abstract available on line at http://ntp.niehs.nih.gov/). If the effects seen in low-dose animals are recorded as ‘any reproductive tract malformation’, and these compared to controls where no abnormalities are detected, the observations in the 15 and 50 mg/kg/day animals reach statistical significance (p < 0.05). Our results of dietary DEHP are similar to those of Biro et al. (2006) who used gavage in oil at the same dose levels. Although these are low incidence phenomena, the results appear robust. The abnormalities are not detected in all animals which is similar to the sporadic occurrence in humans. You raise the important epidemiological issue of assessing time of outcome in relation to time of exposure. When analysing human exposure we must consider the temporal relationship, i.e. when the exposure happened, and determine if this was possibly causing a programming effect, or possibly only a short-term acute or subacute effect. We need guidance from experimental models. Epidemiologists could measure prenatal exposure levels and look for remote outcome effects such as time of puberty; however, prospective studies should preferably also follow subjects for meaningful, shorter time intervals. Animal experiments in 1991 showed an effect of phthalates on the testis at a dose level of 500 μg/kg body weight, but recently we have seen effects at much lower doses, even down to 15 μg/kg. How do these levels relate to human exposure, and can rat doses be translated to humans? Susceptible populations such as neonates in intensive care units are receiving very high doses of a single phthalate comparable to, or higher than, the doses which produce effects in rats. Rather than comparing absolute doses in μg/kg, a simpler and more meaningful measure of neonatal exposure is the level of metabolites in the amniotic fluid. I emphasize that the no observed abnormal effect level (NOAEL) of a single compound results in a reduction of the NOAEL of other compounds when there is exposure to a mixture of chemicals, and the cumulative effect is not only due to a mixture of phthalates, but there is also interaction with other substances which disrupt the androgen signalling pathway. The real estimate of damage to human health must include the effect of all the chemicals acting together. Exposure assessment is a key issue and clearly timing is important because we may be seeing a remote programming effect. We may be measuring exposure at the wrong time which would cause increased imprecision of the measurements. This would lead to greater bias towards the null hypothesis, and if any association could be demonstrated epidemiologically, the true association would be much greater. This underestimate of the effect would invalidate the assessment of exposure. Some of these assessments cannot be performed epidemiologically and rodent models are required. For example, the AGD in neonates can be related to the levels of chemicals in breast milk and maternal urine, but the development of testicular cancer at the age of 35 years is the result of a programming in utero developmental effect. If phthalates cause testicular cancer, there is a 35-year latent period. Such long-term surveys are seldom funded. Duty et al. (2003a,b) studied the effect of current exposure to phthalates of adult men who were attending an infertility clinic. Exposure was assessed by analysing the urine, and there was a significant effect on sperm morphology and motility, and there was DNA damage demonstrated by the COMET assay. In addition to looking at the cumulative effect of a mixture of chemicals, we must also consider the cumulative effect of exposure over a prolonged time interval, and perhaps lifelong exposure from in utero to adulthood including exposure in the neonatal period and during lactation. We assessed the level of phthalate exposure per kg body weight in infants and compared these levels with the upper acceptable doses for pregnant women and newborn infants recommended by the German study (Becker et al., 2004). We studied children in the general population and not just the premature and ill neonates in the intensive care units. The levels of phthalates in milk were analysed, and the daily exposure to the babies calculated assuming they received 120 ml milk/kg/day, which is a conservative estimate (Main et al., 2005). The most abundant phthalates in human milk were MBP, MEHP and MINP, and the sum of these three were found to give a median daily dose of 124 μg/kg/day (range 36–11.300) to babies in Denmark, and 115 μg/kg/day (62–1553) to Finnish babies. This median sum exposure was five times the recommended daily dose. Two promising matrices for assessing prenatal exposure to phthalates are meconium and amniotic fluid. Meconium is easily collected non-invasively and is normally discarded. Amniotic fluid can be obtained during routine amniocentesis. However, because amniocentesis involves some potential health risks, surrogate matrices of exposure, such as urine, may be valuable. Before urinary concentrations can be used as a surrogate for concentrations in amniotic fluid, amniotic fluid and urine should be collected simultaneously to determine whether their respective concentrations correlate. The biological matrix is important when assessing levels of phthalate metabolites because all matrices have some degree of esterase activity which may affect the measured amounts of phthalate metabolites. The plastic materials that may be used to collect the samples may contain phthalates and leaching of certain phthalates into the sample may occur. This would also apply to archived samples which may be stored for years in phthalate-containing plastic containers because it is possible that an unknown proportion of the phthalate content of the sample will be due to contamination. The sampling procedures must follow well-established standardized methods. Health effects of humans exposed to current levels of phthalates are not restricted to reproductive abnormalities because there appears to be a correlation with wheezing and asthma in children (Oie et al., 1997; Jaakkola et al., 1999; Bornehag et al., 2004). Phthalates are present in house dust, and children sleeping in bedrooms with PVC wall coverings have a higher incidence of wheezing and asthma which is thought to be related to the level of DEHP in house dust. In animal experiments, intradermal injection of DEHP to mice does not induce an IgE response; however, in other rodents, DEHP is used as an adjuvant with ovalbumen to produce a hyperallergic response (Larsen et al., 2001). Although there is no definite evidence that phthalate exposure is the cause of wheezing and asthma, this is another endpoint to consider which may be a more sensitive endpoint than endocrine disruption because perhaps we are seeing an effect in the population at current exposure levels. As a retired environmental chemist I am impressed by the disturbing and overwhelming facts discussed at this Workshop. Results from the Danish Food Institute and pharmaceutical chemists 25 years ago demonstrated that phthalates migrated from various materials into human blood and liquids. Phthalates were found in North Sea sediments 15 years ago confirming that we are exposed to these chemicals which are present ubiquitously around us. Is it acceptable that all humans, adult men and women, sick and healthy, children and neonates, carry a body load of phthalates, and if acceptable, what level is acceptable? We have examined dose–response in relation to exposure, but we must discuss the level of exposure acceptance, and the level of body load acceptance. We have evidence from toxicologists over many years, and now is the time to decide based on our wisdom, or lack of wisdom, whether we should all be exposed to phthalates and carry them in our bodies. Although we know that something is probably happening as a result of phthalate exposure, we have no definite proof of cause and effect. Our recent book (Late lessons from early warnings: the precautionary principle 1896–2000. Environmental issue report No. 22, EEA 2001) reviewed the history of diethylstilboestrol, polychlorinated biphenyls and tributyltin whose endocrine disruptive effects were recognized and action was taken to regulate their use, but up to 100 years too late. The reality of toxic substances is so complex that we must not get paralysed by over analysis, and we may still not know all the answers to significant questions even after a decade of further research, because the more research we do, the more we realize what we do not know. Studies of climatic changes are complex with much uncertainty and ignorance, but it is concluded that there is a medium likelihood of mankind disturbing climatic conditions, which is defined as being at level 3 of a 5-point scale of level of proof. On a similar scale, what is the probability that phthalates are having significant adverse effects? Is the overall weight of evidence from all sources sufficient to justify a regulatory ban on these substances, whether timed or not depending on the availability of alternatives. It is difficult to decide if certain phthalates should be banned, and it is not my role as an epidemiologist to make such decisions. However, those who choose to use products containing phthalates must be given the appropriate information. Such a choice cannot be made if manufacturers add phthalates secretly to products as so-called ‘inert’ substances without specifying their nature. The first step before banning should be to compel manufacturers to list all the substances in commercial products so that we know what we are being exposed to. I work for the National Institute of Environmental Health Sciences (NIEHS) which is a non-regulatory US Government agency, but we provide data on toxicity to be used by other agencies that may regulate exposures to agents. In general, the weight of evidence must be considered by policy makers and politicians, and manufacturers should be looking for replacement substances. Adipates are alternative plasticizers available for blood bags and other plastic tubing used for medical products. They do not have the same reproductive effects in animals as phthalates; however, they are more expensive than DEHP. We do know of some agents with proven significant toxic effects and these are disappearing from the marketplace because of consumer pressure, more than direct regulatory control. We must complete the circle before instigating a ban on substances. We have several pieces of fairly convincing evidence that phthalates have endocrine disruptive effects and epidemiological surveys confirm the biological plausibility of that evidence. The final step in completing the circle would be to show that the same mechanisms via which phthalates affect the rat fetal testis are operative in humans, and this would give a stronger case for their regulation. The experiments ethically cannot be performed on humans, and results from primate studies could give stronger evidence for human effects than work on rodents. The easy way of using human tissue in vitro unfortunately does not work because we cannot reproduce the in vivo effects in rodents using in vitro rodent explant cultures. We now look at the epidemiological approach to the problem of phthalate exposure in humans and ask if there are any follow-up studies in children known to have been exposed to phthalates as neonates to see if there are any long-term effects at puberty and adulthood which can be traced to this exposure early in life. Toxicological investigation of substances should not require support from epidemiological data before undertaking risk assessment analysis and regulation if harmful effects are demonstrated. This would delay regulatory action if we had to wait to demonstrate that exposure caused effects. Although I work for a regulatory agency, I have no regulatory powers and my function is to introduce scientific data into the process. Ideally, toxic products should be regulated before human exposure and this necessitates validated animal testing to predict ahead of time if substances should reach the marketplace. It is known that toxins other than phthalates were not thoroughly tested before becoming high production volume chemicals. I have the impression that there is a fear of regulation, and I am not optimistic about meaningful regulation being imposed in USA in the current political climate. I have been working with manufacturers of medical devices for several years and have highlighted our concerns about phthalate exposure in infants. The largest manufacturers of medical devices, Baxter and Abbott, are looking for replacements and alternative plasticizers, and they are testing and about to introduce new polymers. It is our collective duty to inform the general population about the risks involved so that there will be insistence on safer materials and chemicals being introduced. We must be concerned about the total body burden of chemicals in an entire population. As citizens of our respective countries we should look at new ways of moving the regulatory process forward, but this is highly dependent on the political climate at the time.
Authors: P. Grandjean, J. Toppari
Language: English
Subfield: Health, Toxicology and Mutagenesis
Field: Environmental Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W1966982759