Retha R. Newbold, Developmental Endocrinology and Endocrine Disruptor Section, Mail-Drop E4-02, National Institute of Environmental Health Sciences, Research Triangle Park, NC 27709, USA. E-mail: email@example.com
Environmental chemicals with hormone-like activity can disrupt the programming of endocrine signalling pathways that are established during perinatal life and result in adverse consequences that may not be apparent until much later in life. Increasing evidence implicates developmental exposure to environmental hormone mimics with a growing list of adverse health consequences in both males and females. Most recently, obesity has been proposed to be yet another adverse health effect of exposure to endocrine disrupting chemicals (EDCs) during critical stages of development. Obesity is quickly becoming a significant human health crisis because it is reaching epidemic proportions worldwide, and is associated with chronic illnesses such as diabetes and cardiovascular disease. In this review, we summarize the literature reporting an association of EDCs and the development of obesity, and further describe an animal model of exposure to diethylstilbestrol that has proven useful in studying mechanisms involved in abnormal programming of various oestrogen target tissues during differentiation. Together, these data suggest new targets (i.e. adipocyte differentiation and mechanisms involved in weight homeostasis) of abnormal programming by EDCs, and provide evidence that support the scientific term ‘the developmental origins of adult disease’. The emerging idea of an association of EDCs and obesity expands the focus on obesity from intervention and treatment to include prevention and avoidance of these chemical modifiers.
Growing scientific evidence supports the hypothesis that numerous environmental chemicals can interfere with complex endocrine signalling pathways and cause adverse consequences (McLachlan, 1985, 1995; Colborn & Clement, 1992; Colborn et al., 1993, 1996). Initial concern focused on chemicals with oestrogen-like activity, but it has become increasingly apparent that estrogenicity is not the only important endocrine mode of action; chemicals that mimic or interfere with the actions of all endocrine hormones including oestrogens, androgens, progestins, thyroid, hypothalamic and pituitary hormones may also be involved. These chemicals are now collectively referred to as ‘endocrine disruptors’, a term that encompasses their multiple hormone-like activities. Likewise, concern first focused only on reproductive parameters and potential carcinogenic effects; however, we now know that multiple organ systems may be affected by endocrine disrupting chemicals (EDCs) including the cardiovascular and neuro-endocrine systems. More recently, studies from our laboratory and others have proposed an association of EDCs with the development of obesity (Baillie-Hamilton, 2002; Heindel, 2003; Heindel & Levin, 2005; Newbold et al., 2005, 2006, 2007a,b). Furthermore, a new study points out the effects of environmental chemicals and their disruption of normal development and homeostatic control over adipogenesis and early energy balance (Grun & Blumberg, 2006). Although uncertainties remain about the full extent of health consequences that may follow exposure to EDCs, we are just beginning to realize that the complexities and interactions of endocrine signalling mechanisms include adipocytes and weight-controlling mechanisms.
The obesity epidemic
Obesity defined as excessive body fat (>25% men; >30% women) is quickly becoming a significant human health crisis worldwide (Oken & Gillman, 2003). The prevalence of obesity has risen dramatically in wealthy industrialized countries over the last 2–3 decades, but it is also on the rise in poorer nations. In the US, obesity has reached epidemic proportions with more than 20% of adults being obese and an additional 30% overweight; the prevalence of obesity is similar in men and women (DHHS, 2005). Obesity and overweight are known to adversely affect health and to impact the risks and prognosis for a number of serious medical conditions such as type 2 diabetes, hyperinsulinemia, insulin resistance, coronary heart disease, high blood pressure, stroke, gout, liver disease, asthma and pulmonary problems, gall bladder disease, kidney disease, reproductive problems, osteoarthritis and some forms of cancer (Mokdad et al., 1999, 2003; Collins, 2005).
Obesity is thought to be caused by a complex interaction between genetic and environmental factors. Commonly held causes of obesity are overeating and a sedentary lifestyle which is imposed on a background of genetic pre-disposition for the disease. Although much interest has focused on these factors including the need to incorporate healthy foods in our diets and more exercise into our lifestyle, the exact cause of obesity is unclear. Further, it remains puzzling why some people are more successful in dieting and loosing weight than others. As obesity is so very difficult to treat, prevention of this disease is essential.
Until the 1990s, adipocytes were considered just storage depots for excess metabolic fuel. However, the discovery of leptin as an adipocyte-derived hormone that communicates energy reserve information from adipocytes to other organs of the body including the central nervous system, lead to a new appreciation that these ‘fat storage cells’ are actually endocrine organs (Collins, 2005). Since then, evidence has shown that adipocytes secrete a large number of cytokines and growth factors that play significant roles in growth and differentiation, as well as in the feedback of information to other endocrine organs. Considering the newly identified endocrine function of adipocytes and knowing that their endocrine signalling pathways are established during perinatal development, it became interesting to investigate whether exposure to EDCs during critical period of differentiation was related to obesity or any of its associated diseases.
The developmental basis of adult disease
It is well accepted that the developing foetus or neonate is extremely sensitive to perturbation by chemicals with hormone-like activity. In fact, Professor Howard Bern described this unique sensitivity in a chapter titled ‘The Fragile Fetus’ a number of years ago (Bern, 1992). Numerous examples now document that developmental exposure to certain chemicals during critical periods of differentiation can lead to adverse effects, some of which may not be apparent until adult life.
The scientific hypothesis that adult health and disease may have an aetiology that arises in foetal or early neonatal life is not unique to the field of endocrine disruption. Reports of poor foetal nutrition have garnered much interest in the past by suggesting that the foetal environment, as a reflection of low birth weight and poor nutrition, is related to increased risk of non-communicable diseases, coronary heart disease, type 2 diabetes, osteoporosis and metabolic dysfunction later in adult life (Barker et al., 2002). These findings lead to the ‘developmental origins of health and disease (DOHaD)’ paradigm in which a substantial research effort has focused on perinatal influences and subsequent chronic disease (Gluckman & Hanson, 2004; Hanson et al., 2004).
Thus, both the fields of endocrine disruption and maternal nutrition have provided examples supporting the idea that perinatal factors can significantly alter the developing organism and cause long-term effects in the adult. As further demonstration of DOHaD effects, exposure to diethylstilbestrol (DES), a well-known perinatal carcinogen, also provides an example of the profound effects of oestrogenic chemicals on the developing foetus and neonate (for review, see Herbst & Bern, 1981; NIH, 1992).
The developmental exposed DES animal model to study obesity
The role of environmental chemicals in the development of obesity is an emerging area of research which is focusing on the identification of obesogens, their possible molecular targets and potential cellular mechanisms through which they might act. Thus, to determine if environmental chemicals with hormone-like activity are playing a role in the development of obesity and, further, study potential mechanisms involved, we used an experimental mouse model of perinatal DES exposure which was developed and characterized in our laboratory to study altered developmental programming of the reproductive tract which is well known to result in disease and dysfunction (Newbold, 1995).
To study the mechanisms involved in DES toxicity, we developed an animal model using outbred CD-1 mice treated with DES by subcutaneous injections on days 9–16 of gestation (the period of major organogenesis in the mouse) (McLachlan et al., 1980) or days 1–5 of neonatal life (Newbold et al., 1990) (a period of cellular differentiation of the reproductive tract and a critical period of immune, behavioral and adipocyte differentiation). The developmental DES animal model has been used to successfully duplicate, and in some cases, predict, many of the alterations (structural, function, cellular and molecular) observed in similarly DES-exposed humans (Newbold, 1995; Newbold et al., 2006).
Although our major focus has been on reproductive tract abnormalities and subfertility/infertility, we also examined the effects of DES on body weight. Treatment of female mice with DES on days 1–5 of neonatal life using a low dose of 0.001 mg/day did not affect body weight during treatment but was associated with a significant increase in body weight as adults (Newbold et al., 2005). Further, data indicated that the increase in body weight in these DES-exposed mice was associated with an increase in the percent of body fat as determined by mouse densitometry (Lunar PIXIMUS; GE Healthcare, Waukesha, WI, USA). Figure 1A schematically shows the weight patterns of low dosed DES-treated mice and controls.
Unlike the low dose of DES (0.001 mg/kg), the high neonatal DES dose of 1 mg/day caused a significant decrease in body weight of female mice during treatment on days 1–5 but this was followed by a ‘catch up’ period lasting until about 2 months of age and then finally resulting in a significant increase in body weight of DES-treated mice as compared with controls (Newbold et al., 2007a,b); Figure 1B schematically shows the body weight comparison of control and DES (1 mg/kg; high dose) treated mice during the time of treatment and continuing into adulthood. The high body weight in all DES-treated mice, both low and high doses, was maintained throughout adulthood. These data suggest that numerous pathways are probably involved in programming for obesity because DES at different doses resulted in obesity whether or not pups were under weight during treatment.
As the densitometry images suggested DES-treated mice had excessive abdominal fat which has been previously reported to be associated with cardiovascular disease and diabetes (Gillum, 1987), weights of various fat pads were measured to determine if specific fat pads were affected by DES treatment or whether it was a generalized effect throughout the mouse. At 6–8 months of age, fat pad weights were compared in DES-treated mice (1 mg/kg) and controls; inguinal, parametrial, gonadal and retroperitoneal fat pads were all increased in DES-treated mice as compared with controls; however, brown fat weights were not significantly different between DES and controls (Newbold et al., 2007a,b).
A recent study describes a role for developmental genes in the origins of obesity and body fat distribution in mice and humans (Gesta et al., 2006). Therefore, exposure to environmental chemicals with hormonal activity may be altering gene expression involved in programming adipocytes. Several genes have been implicated in altering adipocyte distribution and function such as Hoxa5, Gpc4 and Tbx15 and fat cell distribution such as Thbd, Nr2f1 and Sfrp2. We investigated changes in gene expression by microarray analysis in uterine samples from DES-treated mice (1 mg/kg) compared with controls at 19 days of age. In these samples, genes involved in adipocyte distribution were not altered in the uterus following neonatal DES exposure, however, genes involved in fat distribution were. Thbd and Nr2f1 were significantly downregulated and Sfrp2 was significantly upregulated in DES-treated uteri compared with controls (data taken from study published by Newbold et al., 2007a,b). These findings support the idea that environmental oestrogens may play a role in regulating the expression of obesity-related genes in development.
Although DES-treated mice were similar in weight to controls at 2 months, high dose DES (1 mg/kg) mice exhibited elevated serum levels of leptin, adiponectin, IL-6 and triglycerides before overweight and obesity developed suggesting these endpoints may be important early markers of subsequent adult disease (Newbold et al., 2007a,b). The elevated levels of leptin and adiponectin may indicate insensitivity to these hormones and/or a loss of the negative feedback mechanisms that regulate adipogenesis. Further, at 6 months of age, insulin and all of the serum markers except triglycerides were found to be significantly elevated as compared with controls (Newbold et al., 2007a,b).
As, the balance of activity levels and food intake are known contributors to obesity, activity was measured in DES (1 mg/day) and control mice at 2 months of age before a difference in body weight could be detected. Individual mice were placed in an Opto-Max motor activity chamber (Columbus Instruments, Columbus, OH, USA) and their ambulatory activity measured. Overall, there was no statistical difference in this parameter between the two groups although the DES group showed less movement as compared with controls as the experiment progressed. This difference, however, was not sufficient to explain the enhanced weight gain in DES mice as they aged (Newbold et al., 2007a,b). Additional measures of activity will be necessary before decrease in activity can be ruled out as a contributing factor to the development of obesity in these mice.
Feed consumption was also measured over a 2-week period for control and DES-treated mice (1 mg/kg). Although DES-treated mice ate more than controls over the course of the experiment (approximately 3 g more), the amounts were not statistically different from controls (Newbold et al., 2007a,b). Taken into account, both the marginal decrease in activity and the increase in food intake in DES-treated mice as compared with controls, it is unlikely these two measurements can solely explain the development of obesity in DES-treated mice.
Glucose levels were also measured in DES (1mg/kg) and control female mice at 2 months of age prior to the development of obesity (Newbold et al., 2007a,b). Twenty-five percent of the DES-treated mice had significantly higher glucose levels than controls; these mice also showed a slower clearance rate of glucose from the blood because higher levels were seen throughout the experiment (Newbold et al., 2007a,b). It is important to note that altered glucose levels were observed in these mice before they developed excessive weight. Perhaps additional glucose measurements in older mice may help determine if a higher percentage of mice are affected with age, and if higher and sustained levels of glucose can be demonstrated. To date, however, our data suggest that overweight and obesity observed in perinatal DES-treated mice will be associated with the development of diabetes, similar to the association of obesity with diabetes in humans. Interestingly, earlier studies from our laboratory have shown a high prevalence of islet cell hyperplasia in the pancreas of DES-treated mice supporting the idea that these mice have abnormal glucose metabolism (Newbold, R. R., unpublished data).
Although female mice exposed to DES at low or high doses developed obesity as adults, these phenomena were not apparent in similarly exposed males (Figure 2A and B for a comparison of body weight at 2 months for males and females). In fact, DES-treated males were smaller than corresponding controls and the decrease weight was dose dependent. This sex-specific effect is not surprising because organizational effects of sex steroids during critical developmental periods have been well known for years to cause changes in the long-term anatomy and function of the hypothalamic nuclei which regulates reproduction as well as body weight. Certainly, additional studies are necessary to investigate the effects of EDCs on male weight homeostasis and adipocytes because it is not clear whether this is a compound-specific effect, an oestrogenic effect or typical of EDCs with other hormonal activities. Again, our findings point out the complexities of the mechanisms associated with the development of obesity.
Other EDCs and obesity
An interesting review published in 2002 postulated a role for chemical toxins in the aetiology of obesity and summarized data showing that the obesity epidemic coincided with the marked increase in use of industrial chemicals in the environment over the past 40 years (Baillie-Hamilton, 2002). In this article, data presented showed that the current obesity epidemic cannot be explained solely by alterations in food intake and/or decrease in exercise. The author cited numerous studies where chemicals including pesticides, organophosphates, polychlorinated biphenyls, polybrominated biphenyls, phthalates, bisphenol A, heavy metals and solvents caused weight gain possibly by interfering with weight homeostasis such as alterations in weight-controlling hormones, altered sensitivity to neurotransmitters or altered activity of the sympathetic nervous system (Baillie-Hamilton, 2002). Most of the experimental studies cited in the review were conducted to detect toxicity of a specific chemical as determined by decreased weight; thus, a chemical was concluded to be non-toxic at certain levels if it caused no weight loss. Determination of weight gain was not an endpoint determined in the original design of these studies and, if noted, was considered to be a sign of ‘no toxicity’ rather than an adverse effect. It is interesting that in a few cited studies, the chemicals were actually designed to have growth-promoting properties such as with DES which was widely used by the livestock industry (Raun & Preston, 2002).
Since the Baille-Hamilton review (Baillie-Hamilton, 2002), an increasing number of studies has been specifically designed to address the effects of environmental chemical exposure on weight gain and loss. Numerous studies have shown that exposure to numerous EDCs during critical periods of differentiation, at low environmentally relevant doses, can alter developmental programming resulting in obesity. One of the most novel and interesting of these studies involves a class of chemicals called organotins which includes persistent organic pollutants with endocrine-disrupting properties. Tributyl tin chloride and triphenyl tin chloride have been identified as nanomolar agonist ligands for retinoid X receptor (RXR) and peroxisome proliferator-activated receptor γ, nuclear receptors that play important roles in lipid homeostasis and adipogenesis; tributyl tin (TBT) was shown to disrupt normal development and homeostatic controls over adipogenesis and energy balance, resulting in obesity (Grun & Blumberg, 2006; Tabb & Blumberg, 2006). Interestingly, TBT was shown to cause permanent physiological changes in both male and female mice exposed during prenatal life resulting in a pre-disposition to weight gain. The mechanisms describing how these chemicals actually program for the development of obesity was eloquently described (Tabb & Blumberg, 2006). Blumberg and colleagues appropriately called these chemicals ‘environmental obesogens’.
Phytoestrogens, contained in various food and food supplements, in particular soy products, are another class of chemicals that are receiving attention. Genistein and daidzein are two of the most abundant phytoestrogens in the human diet and, genistein, because of its oestrogenic activity, has been proposed to have a role in the maintenance of health by regulating lipid and carbohydrate homeostasis (Park et al., 2005). However, a recent study showed that genistein at pharmacologically high doses did indeed inhibit adipose deposition but, at low doses similar to that found in western and eastern diets, in soy milk or in food supplements containing soy, it induced adipose tissue deposition especially in males. Further, this increase in adipose tissue deposition by genistein was correlated with mild peripheral insulin resistance. Interestingly, like our findings with DES, genistein did not significantly affect food consumption (Penza et al., 2006) suggesting an abnormal programming of factors involved in weight homeostasis.
The data summarized in this review support the idea that brief exposure early in life to environmental endocrine disrupting chemicals, especially those with oestrogenic activity like DES, increases body weight with age. Whether the results in experimental models can be extrapolated to health hazards in humans as the reproductive abnormalities from the DES mouse model did, remain to be determined but epidemiology studies suggest a link between exposure to environmental chemicals such as polychlorinated biphenyls (PCBs), DDE and persistent organic pollutants with obesity. Further, use of soy-based infant formula containing the oestrogenic component genistein, has also been associated with obesity later in life (Stettler et al., 2005). Use of the DES animal model to study ‘obesogens’ and mechanisms involved in altered weight homeostasis (direct and/or endocrine feedback loops, i.e. ghrelin, leptin, etc.) by environmental endocrine disrupting chemicals is an important basic research tool that may shed light on areas of prevention. Public health risks can no longer be based on the assumption that overweight and obesity are just personal choices involving the quantity and kind of foods we eat combined with inactivity, but rather that complex events including exposure to environmental chemicals during development may be contributing to the obesity epidemic.
This research was supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences. The authors gratefully acknowledge Dr. Terry Phillips, Division of Bioengineering and Physical Sciences, Office of Research Services, Office of the Director, NIH, DHHS, Bethesda, MD for the serum measurements. The authors would also like to thank Ms Sherry Grissom for the microarray analysis of gene expression changes in the uterus following neonatal DES treatment.
F. vom Saal
At a recent meeting of DOHAD (Developmental Origin of Health and Disease) it was stated that babies in India are born with an altered ratio of fat to muscle. We should use body composition rather than body weight as a biomarker, and this is possible with the advent of micro CT technology, which gives very accurate measurements. This is especially relevant to rodent studies where body weight is a less reliable measure than the ratio of fat to muscle, and dramatic differences in this ratio can be predictive of the development of diabetes type II.
What is the role of the contraceptive ethinyl oestradiol (EE2) in this context?
I have no information on its effect in obesity.
Dr R Newbold described adult onset obesity in mice exposed to diethylstilboestrol (DES) neonatally. Was the metabolic profile of the liver examined and considered as a potential contributory factor? Neonatal DES is likely to have a dramatic effect on the regulation of liver metabolic function in the adult by acting on the hypothalamo-pituitary-growth hormone axis with effects on the sexually dimorphic patterns of GH secretion. Presumably the expression patterns of enzymes in the liver of adult mice who were exposed to DES neonatally are significanttly altered.
That is an important question but this analysis has not yet been performed.
There are 2 types of fat in the body, yellow fat, and a lesser quantity of brown fat which is thought to burn off energy by thermogenics thereby helping to protect against obesity. Was the relative proportions of yellow fat and brown fat assessed in DES-treated animals compared to controls?
This was not examined in the experiments I presented.
F. vom Saal
We examined the distribution of fat in obese animals and found an increase in epididymal fat, perirenal fat and brown fat. Both white fat and brown fat were increased.
The use of casein in diets must be considered. Rodents and other mammals are normally only exposed to casein early in life when suckling, at a time when steroid hormone levels are usually at very low levels. We must be very careful about diet outside of this unique physiological condition and casein is probably not a good protein to be in the diet throughout life.
Considering the differences between milk based diets and soy based diets, is there enough evidence to recommend not using milk in the diet of mothers and children in an attempt to prevent obesity.
That is an interesting concept, but we are not prepared to make such a recommendation at present because the data are still too preliminary.
There are centres in the hypothalamus which affect body weight by control of food intake and thermogenics. Have you looked at other sites in the brain which may be involved in the neonatal or perinatal programming of body weight?
I am not aware of any such study. We are at a very early stage in our investigations and we have proposed a hypothesis which has to be studied in detail, and hopefully will stimulate other groups to look into similar aspects.
Your hypothesis is a very good suggestion and has stimulated many questions which should be addressed.
There is an interaction between puberty and body weight. We have seen that there are animals exposed tocertain endocrine disrupters during development who are underweight before puberty but overweight after puberty.
The role of activation of endogenous sex steroids production which are increased at puberty is important. In humans, there are some types of puberty which only become apparent after puberty and this is due to the interaction of early and late exposure to endogenous factors.
F. vom Saal
There is a dramatic change in animals immediately after weaning and there are regulatory events occurring during nursing which are not apparent until the pups are removed from their mother. Within 1 week after puberty we see effects of feeding which are not expressed in the same way during the suckling period.
There is a difference in pre- and post-pubertal weight gain. There is not much sex difference in body composition pre-pubertally, but at puberty there is a different response with androgens stimulating muscle gain in boys and oestrogens causing increased fat in girls.