Skip to main content
Premium Trial:

Request an Annual Quote

Nature and Nurture


This article has been updated from a previous version to correct the spelling of Bahar Taneri's name.

For the nearly 4 million people worldwide suffering from a crippling lymphedema, the solution is simple: shoes.

Podoconiosis, also known as mossy foot, is caused by continual exposure to red-clay soil derived from volcanic rock in the highlands of southern Ethiopia and similar regions in some nine other countries. It presents as painful inflammation of the lower extremities, eventually leading to foot disfigurement. Mossy foot is a disease of poverty; wearing shoes prevents exposure to the volcanic dust and thus prevents the disease.

Still, not everyone who goes barefoot on such soil will experience the extreme immune response that leads to the swelling and lesions — only about 1 in 20 of those who are environmentally susceptible develop podoconiosis. Several studies have shown familial clustering, suggesting heritability.

"Podoconiosis is a really classic example of how the environment puts people at risk who have a genetic susceptibility," says Charles Rotimi, director of the National Human Genome Research Institute's Center for Research on Genomics and Global Health. "You can have the genetic susceptibility, but if you wear shoes and therefore are not exposed to this fine-grain sand seen in the high-altitude volcanic regions in Ethiopia and some other countries, then you won't get it — you won't get the disease."

Prevention is the cornerstone of modern public health. Because genes exist within, and are affected by an environment, incorporating gene-environment interactions into genomics research will help researchers paint a more complete picture of human health and disease. And because altering environmental factors is ostensibly more feasible than modifying the genome, incorporating gene-environment interactions into prevention measures seems, to many in the field, the best way forward for public health genomics.


"One of the lessons that we are learning in trying to use genetics or environmental factors all by themselves to try to understand disease is that we always get an incomplete picture," Rotimi says. "What we are learning is that it is that environment that shapes what we see at a genomic level. So, to properly understand why we see certain diseases, especially when we are comparing populations of ethnic or ancestral groups, it is critical for us to put that understanding within the context of how individuals have lived their lives in the environments that they have lived their lives, and their ancestors have lived their lives."

'A simplified model'

Though it is a non-communicable disease, people who develop podoconiosis are highly stigmatized and often shunned by their communities.

"It's a disease that you cannot hide — your leg is swollen. Most of the time, you usually cannot move about, and sometimes, because of the nature of the disease, it actually produces odor, so people really don't want to be around you," Rotimi says. The social stigmas run deep — affected children can be excluded from classrooms, and people are reluctant to marry into families with a history of the disease. "People tend to hide it as much as they can, but it's not something they can really hide," Rotimi says, adding that "some [affected] people actually starve to death" while keeping away from others.

While its environmental cause is abundantly clear, what is not so apparent is why only some families are susceptible to the disease. Clustering of cases within pedigrees suggests that genetic factors play some role in the pathogenesis of podoconiosis, but until recently, researchers were unsure as to just what those factors were.

With funding from the Wellcome Trust, Rotimi and his colleagues in the UK and Ethiopia designed a genome-wide association study of 194 cases and 203 controls from the same broad geologic area covered by red-clay soil containing particles derived from volcanic rock from southern Ethiopia. The researchers chose to perform an agnostic search, "because we didn't understand where the genetic signal [was] in the genome," Rotimi says. They extracted participants' DNA from saliva samples, which they then shipped off to Iceland for genotyping at Decode Genetics. Rotimi's group analyzed the resulting data in his lab at NHGRI. "The challenge initially was really analyzing the data correctly. There's not a whole lot of reference data for Ethiopian genetics out there," Rotimi says.

The team identified a significant association of mossy foot phenotype with a SNP located 5.8 kilobases from the HLA-DQA1 locus as well as suggestive associations with seven additional SNPs in or around HLA-DQA1 and the nearby loci HLA-DQB1 and HLA-DRB1.

Because assembling a case-control replication cohort for this highly stigmatized condition was all but unfeasible, Rotimi and his colleagues used family-based association testing to validate the associations they found.

"When you are designing gene--environment interaction [studies], you want to be very careful as to who you are calling a control. By understanding the epidemiology of the disease, we're able to now select what we consider the optimal controls," he says. For the family-based replication study, control subjects had to have lived in the community, not worn shoes, and been exposed to the soil their entire lives. "Most of our controls are in their 60s or older. Therefore, if they were going to get this disease, they would have gotten it," he adds. "Those are the kind of basic principles you need to be able to do very good gene-environment interaction-type studies."

The team's discovery-phase GWAS results did indeed replicate in its family-based HLA typing study. The researchers confirmed the associations, reporting their results in the New England Journal of Medicine in March. In their paper, the researchers also showed that alleles HLA-DRB1*0701, DQA1*0201, DQB1*0202, and HLA-DRB1*0701-DQB1*0202 haplotype are risk variants for podoconiosis. In NEJM, the team suggested that "podoconiosis may also represent a simplified model of gene-environment interactions, which remain poorly understood in many complex genetic diseases."

For Rotimi, "the basic principle — which we think is going to be applicable to a lot of chronic diseases, like diabetes — is that the first thing you notice is [that] not everybody gets it. Therefore, there is some basic background of genetic susceptibility," he says. "We also know that that background of genetic susceptibility is not sufficient most of the time to express the disease. So there has to be some environmental trigger. … By characterizing that [environmental] component and also characterizing the gene, you are able to fully appreciate who will be exposed."


Researchers had characterized the environmental factors at play in podoconiosis — recently recognized as a neglected tropical disease — long before digging into the genetics. But for many more common, complex diseases, researchers do not yet fully understand the environmental nor genetic triggers, though work toward the latter — primarily through GWAS — has produced many robust associations to date.

Common, complex

Carolyn Hutter from the Fred Hutchinson Cancer Research Center seeks to understand gene-environment interactions in the context of colorectal cancer. For Hutter, who is also a lecturer of epidemiology at the University of Washington School of Public Health, placing genetic variants in context is key to understanding them.

"The main thing that gene-environment interactions allow us to do is to recognize that genetic factors — which we do know play an important role in health, both in terms of genetic disorders and also complex diseases — are impacted by environmental factors," she says. "If we're going to try to claim to understand the genetic factors, we have to study them in the context of environmental factors as well."

Hutter and her colleagues sought to integrate environmental factors with GWAS data for colorectal cancer in a meta-analysis. "We were saying: 'Let's look at genetic variants that have been found from previous GWAS that are known to be common genetic variants that impact susceptibility to colorectal cancer and see whether the impact of any of them is modified by environmental factors,'" she says. "Or you could flip it around, and also say: 'Are we able to identify anywhere where the impact of these known or putative environmental factors are impacted by the genotype?' ... So the idea in that case was taking these known genetic factors and established environmental risk factors and just seeing: Do they interact with one another? Does one of them modify the effect of the other?"

In an April Cancer Research paper, Hutter and her colleagues describe their investigation of potential effect-modifications between known colorectal cancer-associated SNPs and probable or established environmental risk factors for the disease derived from nine cohort and case-control studies. Among the 12 risk factors the researchers assessed — including sex, BMI, aspirin/NSAID use, smoking status, and dietary intake — the strongest statistical evidence they found for gene-environment interactions across all studies involved vegetable consumption and the SNP rs16892766 on chromosome 8q23.2.

Collating gene and environment data for 7,016 cancer cases and 9,723 controls was no simple task. "One of the biggest challenges is actually getting the environmental data, as we call it, harmonized — put together in a way that you can look across studies and examine environmental factors across studies," Hutter says. "Just having data that you can use for these environmental factors that is well-measured and consistently measured in the sample sizes that we need for these types of studies" is perhaps the single greatest challenge, she adds.

Another challenge when working with published studies is that not all will have examined the same environmental factors. "We were interested in post-menopausal hormone use — [it] is a variable that's also thought to have a role in colorectal cancer etiology," Hutter says. "We would have liked to have included that, but we weren't able to come up with a good variable that we could use across all of the studies."

Beyond that, statistical challenges remain. "The statistical approaches are still a little bit in their infancy," Hutter says. "There have been a couple of really good advances recently, but we're still working out that side."

While some researchers continue to test and debate the various approaches to mine population data for potential gene-environment interactions, others are preparing for a different but no less complex problem — dealing with personal data.

Predict, prevent

Bahar Taneri, associate professor of molecular biology and genetics at the Eastern Mediterranean University in North Cyprus, says that because health and disease are moderated by gene-environment interactions, "if you want to address these you have to address the genome-based knowledge as well as envirome-based knowledge." And so, working with her colleagues at Maastricht University in The Netherlands — where she was a visiting researcher for six months in 2011 — Taneri set out to develop a model to incorporate the interplay among genetic and environmental factors into public health genomics.

The resulting model, published in Personalized Medicine in January, integrates personal and population gene-environment data for common, complex diseases. "What is integrated in our model is personal data — a person's genome sequence, and a person's envirome data — and [then] what we do is cross-compare this with [population] data from the literature," Taneri says. "That includes genome data from the literature plus environmental factors listed and validated to contribute to these common complex diseases."

While the researchers validated their model using data for three common behavioral disorders, Taneri says that it is broadly applicable to all common, complex diseases for which there is personal and population data for both genetic and environmental factors. "We want to use this model as a predictive and personalized tool," Taneri says. "Basically, the public health genomics goal when we look at our model is to use a personalized, predictive approach to prevent on a personal level, which will actually translate to population health."

The Hutch's Hutter at has a similar prevention-oriented goal in mind. Because colorectal cancer screening is invasive, researchers might use genetic and environmental factors to identify an individual's screening needs. That has been difficult to do thus far because "you really need to have large sample sizes with very well-measured environmental data, and having both of those simultaneously is a challenge," she says. "We need to be spending time and energy to really try and scale up our measures of environmental factors the same way that we scaled up our measures of genetic factors. We've done a very good job on the genetic side, and the environmental side is sort of lagging," Hutter adds.

Rotimi's group at NHGRI has already seen prevention efforts borne of its public health genomics work put into practice. Working with nonprofit advocacy groups, Rotimi and his colleagues are helping to prioritize the delivery of shoes to families affected by podoconiosis. "We try to give shoes to the younger people ... first, because they will be at higher risk," he says.

The Scan

Follow-Up Data Requests to Biobank Participants Ineffective, Study Finds

An effort to recontact biobank enrollees for additional information reports low participation in a new BMJ Open study.

Study Finds Widespread Transmission of Resistant Bacteria in Vietnam Hospitals

A sequencing study in The Lancet Microbe finds widespread transmission of drug-resistant Escherichia coli, Klebsiella pneumoniae, and Acinetobacter baumannii in two Vietnam ICUs.

Novel Brain Cell Organoids Show Promise for Autism Research

University of Utah researchers report in Nature Communications on their development of brain cell organoids to study SHANK3-related autism.

Study Finds Few FDA Post-Market Regulatory Actions Backed by Research, Public Assessments

A Yale University-led team examines in The BMJ safety signals from the US FDA Adverse Event Reporting System and whether they led to regulatory action.