The fifth session of the workshop was devoted to a panel session on research needs. Roundtable member Hal Zenick of the U.S. Environmental Protection Agency (EPA) introduced the four discussants, who discussed the field’s research needs from various perspectives, and then presided over a wide-ranging discussion that involved the discussants, Roundtable members, and workshop participants.
The first discussant was Linda Birnbaum, director of the National Institute of Environmental Health Sciences (NIEHS) and a Roundtable member. She began with a brief review of some of NIEHS’s efforts in the area of the environment and obesity.
NIEHS held its first symposium on the fetal origins of obesity in 2004. A number of years later it announced a program to explore the connections among environmental chemicals, obesity, and diabetes. It began funding grants for that program in 2011 and currently has funding for 57 grants related to obesity, 32 of which involve human cohorts, 20 of which are studies with animal models, and 5 of which are mechanistic studies.
The institute’s work with the National Toxicology Program has shown, she said, that it will be important to think more broadly about which chemicals to consider. Given that it is practically impossible to test all the 80,000 chemicals that are currently in commerce, it will be important to get a better sense of how to prioritize which chemicals to screen.
An important need is to identify environmental chemicals that cause weight gain or changes in body composition, she said. Another important need is to understand the windows of susceptibility. It is likely that children are much more susceptible to the effects of environmental chemicals than adults, but exactly when are they the most susceptible?
In examining the effects of environmental chemicals, it will be important to look past their effects on adipocytes and think also about their effects on other tissues, including the gastrointestinal tract, the pancreas, the liver, muscle tissue, and the brain. “We need to remember that the brain is controlling much of our behavior, both conscious and unconscious, and we need to focus on that,” she said. There are satiety centers in the brain, for instance, but the brain also controls circadian rhythms, and they can play a role in overeating. Toxicologists know that they may get very different dose–response curves if they dose their animals at different points in the circadian cycles.
In doing toxicological studies, it would be valuable to maintain the lab animals long enough to see effects, particularly when the research is looking at developmental effects. Researchers often hold animals just to weaning, Birnbaum noted, or perhaps until puberty, but to see the sorts of effects on weight and metabolism that are of interest, it will be important to keep the animals at least until they are 1 year of age.
Another question that researchers should ask themselves is if they are using the right animal model. Effects can vary from one species to another, from one strain to another, and from one sex to another. For example, some strains of mice gain only a small amount of weight on a high-fat diet, while others grow to gargantuan proportions.
In humans, researchers studying obesity should remember to take into account such factors as race/ethnicity, socioeconomic status, and even birth order. One interesting difference between Europe and the United States is that new mothers in Europe are more likely to nurse their children for a significant amount of time. That can affect obesity. Gestational diabetes is another factor that may play a role, Birnbaum said, but not enough is known about it to say for sure.
She commented that obesogens are likely to affect more than the risk of obesity; they have various other health effects as well. A chemical that is very bioactive is probably going to do different things in different tissues, she said, particularly in context-dependent endocrine-mediated processes.
In studying obesity in both animals and humans, body weight gain may not be the best metric to use, Birnbaum said. In particular, body composition is extremely important. So researchers should probably consider more than just body mass index.
The microbiome is likely to play an important role in obesity, she said, and learning more about the microbiome in general and about its role in obesity in particular should be a focus of future research.
Concerning epidemiological studies of the effects of various environmental factors on obesity, Birnbaum said that not only are prospective longitudinal studies needed, but also researchers should think about using randomized controlled intervention trials. Such trials can be ethically done in an environmental context, and there is certainly potential for them in the future, she said.
Researchers should be shifting their focus from solely on obesity to paying more attention to the metabolic syndrome and diabetes, Birnbaum suggested, because it is not obesity per se that is the major health risk but, rather, diabetes and metabolic issues.
It would be valuable to coordinate the epidemiological studies with the animal studies, she said. “The toxicologists want to think that they are the ones finding a thing and then the epidemiologists go and look,” she said. “I think it may be more the other way around, that frequently it’s the epidemiologist who sees an association that nobody has ever considered before and then the toxicologists need to see if there is biological plausibility there.”
Econometrics studies will also be important, Birnbaum said. By putting a dollar value on the costs of obesity, diabetes, and the metabolic syndrome, it should be possible to drive policy changes.
The current paradigm for obesity in the United States is one of intervention and treatment, Birnbaum noted. Certainly it will be important to continue to improve on ways to intervene and treat, but it would be much better to learn how to prevent obesity in the first place. “That is going to have to start not only in children but actually prenatally and maybe even preconception,” she said. “I am not giving anyone a bye on exercise and diet, but are we making it harder for people to control their weight? I think that is the question.”
The next discussant was Suzette Kimball, the acting director of the U.S. Geological Survey (USGS) and a member of the Roundtable. The issue of the environment and obesity intersects with work that is occurring at USGS, she said, and her presentation was from that perspective. In particular, she said, the body of research on environmental exposures and the role that they play in increasing the risk of obesity and other health challenges that is emerging is related to the research done at USGS, which focuses on environmental conditions and factors.
Recognition that chemical exposures may play a role in the development of obesity can lead to the development of new avenues for
risk reduction and prevention, she said, but this requires determination of which particular toxicants are increasing the risks of obesity and other health issues.
The research gaps that need to be addressed in this area are not unique to obesity research, Kimball noted. Research on the mechanisms by which environmental chemicals cause adverse health effects and the way that these chemicals may interact in the environment to provide cumulative effects is needed. New and better methods to measure the presence of these chemicals in the environment are also needed.
Kimball then spoke about some of the challenges she sees in the area. They include understanding the potential health effects of chronic exposures to extremely low concentrations—that is, less than 1 part in 1 billion—and gaining greater knowledge about the potential increased vulnerability of various populations, including the elderly, the very young, and those who have other complicating health factors. Another challenge is figuring out how exposures in early life stages have effects much later in life or even in subsequent generations.
From the perspective of USGS, she said, a major research challenge is identifying and characterizing the environmental drivers of exposure. The fate and transport of environmental chemicals—as well as human exposure to such chemicals—are affected by both natural and anthropogenic changes in the environment, she said. The human-driven changes in the environment vary across the world, depending on societal demands for land and natural resources and the changes in resource consumption that are driven by economic prosperity. Those changes are compounded by natural Earth processes, climate trends, and related climatic events.
Scientists at USGS specialize in understanding the factors at the interface of the environment and health, Kimball said. That is, they work to characterize the interactions among the physical environment, the living environment, and people to understand how various processes affect human, wildlife, and ecological exposures to environmental disease agents.
A recent focus at the agency has been endocrine-disrupting chemicals (EDCs). USGS spends approximately $5 million per year on research related to these chemicals, Kimball said. However, it also collaborates on such research with many other agencies, such as EPA, the U.S. Department of Agriculture, the U.S. Army Corps of Engineers, and various state public health and agricultural agencies, nongovernmental organizations, and universities. “So the actual figure that is dedicated to this research is much larger than that $5 million,” she said.
In 2014, USGS developed a national strategy for research on EDCs, Kimball said, and that strategy focuses on the science related to the occurrence, exposure pathways, and effects of these chemicals on natural resources, including both terrestrial and aquatic wildlife. Some of the research challenges and research directions identified in that strategy have a great deal in common with the work on these chemicals that the public health sector is undertaking.
The research needs identified in the USGS strategy include the systematic evaluation of sources; the distribution and state of EDCs in the environment; identification of the routes of exposure to EDCs of fish, wildlife, and humans; studies of how EDCs accumulate in animal and human tissues; research into the effects of simultaneous exposures to mixtures of chemicals; investigations into the mechanisms of action of EDCs and the mechanisms by which EDCs cause adverse effects; improvements in analytical techniques, laboratory methods, and biological assays for identifying EDCs in the environment; investigations into less studied endocrine pathways involving metabolism, behavior, fat storage, bone development, and immunity; and the identification of new potential EDCs. “Those challenges that we have identified as a research strategy specifically for EDCs can, as I said, connect very well to the kinds of challenges that I heard articulated over the last day and a half,” she said.
The third discussant was John Rogers, the director of the Toxicity Assessment Division of the National Health and Environmental Effects Research Laboratory in the Office of Research and Development at EPA.
He began by saying that one thing that struck him about the topics discussed at the workshop was that the standard regulatory developmental toxicity tests at EPA would miss most of what was being talked about. For one thing, when EPA scientists do teratology studies, they take the fetuses from the mother just before they would be born, so they do not see effects that appear later in life. And even in multigenerational studies, he said, they do not follow the offspring for very long. Furthermore, things like insulin resistance, elevated blood pressure, or changes in body composition that did not affect body weight would not be detected at all in the experiments. He also noted that even when there was elevated body weight in some of the studies, the extra weight was not seen as an adverse effect. In short, he said, we could be missing some of the biggest problems we are seeing now in children, like obesity and diabetes.
In research on developmental exposures, Rogers said, a key research need is to learn about critical periods. Do the critical periods all come before birth? And even if the critical periods are prenatal in humans, a lot of the development—especially neurobehavioral development—that occurs in utero in humans happens after birth in rodents. This raises the question of what sorts of animal models are needed to model the human situation as best as possible.
Many different questions arise concerning the appropriate animal models, he said. A variety of inconsistencies appear among studies, and some of those may be due to differences in the animal models used in the different studies. Given that there are important differences among species—between the mouse and rat, for example—and even among strains of the same species, how do you choose which is the most appropriate? Not enough is known about those differences yet. Other issues are the following: What kind of diet should be used for animal models? Do the animals need to be put on a high-fat diet to stress their systems and demonstrate that they respond differently if they have a high caloric intake? What should be the exposure period and the dose of test agents?
Dose can be a particularly tricky issue because of nonmonotonic dose–responses. For example, a low dose of a substance might stimulate some pathway that leads to obesity, but increasing the dose might trigger toxic mechanisms that start to drive the weight down, so at some point the response changes from weight gain to weight loss. In particular, if the dose at the start is too high, it might seem that there is no effect or even that the substance causes weight loss when at a lower dose it can lead to obesity.
One of the most important things for researchers to do with animal models, Rogers said, is to understand the mechanisms that explain how certain environmental exposures lead to obesity. One particular focus, he said, should be how insulin affects other endocrine systems.
Researchers should be looking for animal models for some of the things that affect the body composition of offspring in humans, such as maternal obesity, preexisting obesity prior to pregnancy, high maternal weight gain during pregnancy, and gestational diabetes. With such models, researchers could start to ask how these various factors make the offspring more likely to be affected by the mother’s exposure to certain environmental chemicals.
Many researchers working with animals models look only at selected endpoints and do not observe other outcomes, Rogers said. It would be
useful for researchers to start doing more complete evaluations and to measure the various aspects of the metabolic syndrome—not just obesity, but insulin resistance, blood pressure, and blood lipids—as well as changes in the stress response.
Because animal tests take longer to perform if the animals have to be kept for up to 1 year and tested at the end of that time, it would be useful to have biomarkers that would provide early indications of later effects. That could considerably shorten the time that it takes to run such tests.
Researchers should recognize that body weight is not a particularly good endpoint for work with either animal models or humans. Body composition, in particular, measures of fat content and location, is a much better indication of what is going on.
The fourth discussant was Nik Dhurandhar from Texas Tech University, who said that his comments would be grounded mainly in the obesity perspective and what he sees as research in that area.
If one thinks of obesity as the result of an energy surplus, he said, the obvious solution is to reduce that energy surplus by decreasing energy intake or increasing energy expenditure, which in turn is done either by increasing activity or by increasing metabolic need. Clinicians have tried to attack obesity in this way by asking people to eat less or to move more or by using drugs or surgery to reduce intakes or increase expenditures.
However, Dhurandhar said, one should also ask what is upstream of increased energy intake or decreased activity. The answer, he said, is a dysregulation of energy balance. People who are not obese are able to maintain their body weight, but it is not done consciously. People cannot say at the end of the day exactly how many calories that they have consumed or expended, yet they have balanced the two. That means that they are regulating their body weight at some subconscious level, which in turn implies that in those people who are not able to maintain their body weight, there is a dysregulation of energy balance, which is what leads to the energy surplus. What is interesting to him, Dhurandhar said, is the question of what causes this dysregulation of energy balance and does so only in some individuals.
He then showed a slide that he had borrowed from Claude Bouchard at the Pennington Biomedical Research Center at Louisiana State University and that he had modified slightly. It was a partial list of dozens of putative contributors to weight gain in four categories: the social environment, the physical environment, behavior, and biology.
“So this is a reality,” Dhurandhar said. “In one individual, several of these factors may be operational, and in another individual, it could be a completely different set of factors that may be determining that person’s body weight or obesity.” The point is, he said, that if people focus on just one or a few factors in trying to determine why people are obese, it will likely be inadequate.
To explain the problem, he offered a metaphor: suppose there is a buffet and you are concerned about the calorie consumption of people enjoying this buffet. You decide that you are going to focus on one of those dishes for its caloric contribution to that person’s caloric intake, and perhaps you reduce the amount of that dish that a person is allowed to eat or you completely remove that dish from the buffet. But if you remove a particular product that is contributing to caloric intake, there are still many others on the table that can make up the difference. “That is what I think of as digging a hole in water,” he said. “It is really going to get filled by something else.” The moral is this: a focused attempt on just one aspect that contributes to obesity may not yield the desired results.
Therefore, Dhurandhar said, it makes sense to refer to “obesities” rather than just “obesity” to make it clear that multiple conditions that may have similar symptoms but that are really different types of conditions are involved. He compared the situation to jaundice, which describes a symptom but not what has caused it, such as viral hepatitis, cirrhosis, cancer, or some other condition.
Therefore, he said, it is important that researchers identify the true causes of obesity—the truly upstream causes, not the midstream causes. “As a physician,” he said, “I always like to say that when somebody presents with a cold, I don’t treat the nose; I treat what caused that cold. Maybe there’s an allergy; maybe there is an infection. I think that is a strategy that requires a multifactorial approach to a multifactorial disease.”
Linda Birnbaum began the discussion by noting that the collection of environmental chemicals that had been talked about at the workshop was much smaller than the complete list of chemicals of concern. There was no mention of pesticides, for instance, and little talk of nicotine and tobacco smoke, even though there are some striking data connecting maternal smoking with both obesity and type 2 diabetes in offspring. Many of the environmental chemicals of concern work through endocrine-mediated mechanisms, she noted, and much of the work looking at
endocrine disruptions has focused on estrogen, androgen, and thyroid hormone, but there are many other endocrine systems that can be disturbed.
Jerry Heindel of NIEHS asked the speakers to comment on next steps to advance the field of environmental exposures and obesity.
Nik Dhurandhar answered that one of the first things that needs to be done is to identify what is not known, and he said that the workshop had been a good start to that end. One thing to keep in mind, he said, is the difference in clinical treatment and research. While clinicians today need to offer the best tools in the toolbox for the treatment of obesity, it is the job of researchers to move things forward and to develop strategies and treatments that are truly effective in producing meaningful and durable weight loss for the majority of people suffering from obesity. For instance, research is needed to understand better the various factors that contribute to obesity and to determine which of those are preventable and which are not.
Finally, he commented that while it may seem simple for people to simply eat less, it is not. Although eating is under volitional control in the sense that you can decide whether to put food in your mouth at any given moment, that does not mean that it is easy for someone to control their eating patterns over the months and years that it takes to lose weight and keep it off. He emphasized the importance of treatments that work not only in clinical trials, at the hands of talented researchers, but also in the real world, at the population level.
Paul Sandifer from the College of Charleston and Hollings Marine Lab said that he had been struck during the workshop by the paucity of animal models used in studies of obesity. In particular, he noted, only one study mentioned an aquatic vertebrate, which was the zebrafish. He suggested that many organisms may provide useful experimental models for this work. Because so many of the EDCs are waterborne or water mediated, he said, it might offer some additional insights to deal more with fish, which live in a water environment all the time.
Suzette Kimball responded to that comment by saying that there is a fairly robust body of ongoing research concerning bioaccumulation, looking at how these chemicals of interest reach higher and higher levels as you move up the food chain, and fish are the species of choice in that work. Because fish are also a human food source, it is important to understand the role that bioaccumulation might play in exposing people to high levels of obesogens.
Birnbaum added that it would also be important to study bioaccumulation in terrestrial animals as well because some of the patterns of bioaccumulation appear to be different between aquatic and terrestrial species.
Kristina Rother of the National Institutes of Health (NIH) suggested that it would be useful if different researchers could combine resources in a way that allowed each of them to focus on what they do best. As an example, she offered a hypothetical scenario: “I’m a clinician; I don’t have easy access to adipocytes. Barbara wants to know the concentrations of artificial sweeteners that occur under [a] normal daily intake situation (information I have), and somebody else needs a liver enzyme experimental setup to test something else. Even within just this group, we could potentially help each other so that I don’t have to do adipocyte experiments but can focus on my breast milk analyses.”
A member of the webcast audience suggested that research on obesity and diabetes could be helped with the establishment of a major research initiative, something that would be similar to the human genome project but that would also be multinational.
Birnbaum responded that NIH is in the process of launching a major research initiative on precision medicine that will eventually involve more than 1 million subjects compiled mainly from existing studies. “The idea,” she explained, “is to have access to the electronic health records and expand that with additional questionnaires, so you would have both biomedical exams and extensive questionnaire data and potentially take biological specimens and so on.” The project might offer the opportunity to include some environmental factors as well, she said. She also said that while a large international trial would be very useful, it is extremely difficult to start studies with these very large cohorts. Finally, she said that what researchers really need to examine in these trials is early life exposures. There are a number of large birth cohorts in numerous countries around the world in which such early life exposures are now being studied, and there are similar, albeit much smaller, studies being carried out in the United States.
Barbara Corkey of Boston University said that it will be important to find better and more effective ways to share data and information among researchers. By sharing data and information, scientists and other workers can focus on what they do best. “For example,” she said, “handling the data is a very special skill set that, for example, basic scientists rarely possess, but it’s a very important aspect of interpreting
data. So we should work together to achieve these things rather than each trying to learn them.”
John Rogers suggested that a similar thing is true in terms of the levels at which different scientists work. By coordinating in vitro work with animal studies and with human studies, it would be possible to get a much stronger, more cohesive picture. That would not be difficult to do if the different research approaches were designed together.
One audience member asked what sort of funding proposals researchers should be submitting, given that everyone agrees that understanding obesity will require understanding multiple factors and how they interact but that the funding agencies still seem to respond best to very focused proposals. Are the funding agencies ready for some interdisciplinary research with experts from various fields coming together and trying to answer questions about obesity?
Birnbaum responded that NIH tries to find the proper balance between individual research grants and program funding that brings together a variety of different types of research. The more that is spent on the one, the less there is to spend on the other. Furthermore, NIEHS has developed some approaches to increasing interdisciplinary work. One example is the Victor Program, in which one investigator receives an individual RO1 grant and as many as two other investigators can be added on to do additional studies that were not in the original grant but that are related to it. Still, Birnbaum said, it often is the case that to get a grant funded, an investigator must write the proposal in a very narrow way, which, given the nature of obesity, risks missing the forest for the trees. It is an issue that funding agencies and researchers must work further on.
Lynn Goldman of George Washington University turned the discussion to the broader public health perspective on obesity. She noted that chronic diseases associated with obesity cost the economy billions of dollars every year and that the prevalence of obesity is twice what it was in the 1960s. Even though obesity rates have been leveling off in the youngest children, they are not going back down to where they used to be. So the nation faces an enormous future cost associated with obesity as well.
It is important, she noted, to look at the interventions that are happening to identify and disseminate those that are actually working, because some interventions are clearly more effective than others. Right now there is a very scattershot approach to obesity prevention, with government agencies, states, foundations, and companies all making
efforts. From a public health perspective, it would be very useful, Goldman said, to be able to select from all of these effects those that are actually working and then to promote them, while at the same time doing the necessary research to understand the problem fundamentally. “We can’t really afford 10 or 20 or 30 years of understanding the entire biology of a system before we start taking action to protect people,” she said.
The same thing is true for treatment, she said, because so many people are already obese and are struggling with such diseases and diabetes and the metabolic syndrome.