Human Gene Tied to Rare Iron-deficiency Disorder

After 12 years of clinical and laboratory research, a team of physicians and scientists has identified the genetic defect that causes a rare iron deficiency disorder they have termed iron-refractory iron deficiency anemia (IRIDA).

The journey started in 1996, when a child was referred to Nancy Andrews, then a pediatric hematologist at Children’s Hospital Boston, because of her expertise in iron metabolism. The patient had severe iron deficiency anemia that oral iron supplements did not cure. Even intravenous iron did not fully correct the anemia. Because of the early presentation, Andrews suspected the anemia was congenital. A subsequent sibling showed a similar clinical picture, strengthening her hunch.

One pair of siblings, however, simply was not enough to go on. So Andrews, now dean of Duke University School of Medicine, and co–senior author Mark Fleming, then in the Andrews lab and now an HMS associate professor of pathology at Children’s, took what action they could. They developed and maintained a human research protocol and created a repository of DNA samples from those first patients. Over the ensuing years, they added more samples from similar families.

In the meantime, they continued their lab research into iron metabolism. “We found new genes in mice that led to iron deficiency” and other blood disorders, said Fleming. They checked each new gene to see if it could be implicated in their patients’ illnesses. For IRIDA patients, said Fleming, “we could never find the right gene.”

By 2007, the team had recruited enough families and individual samples to consider launching a genomewide search for the gene or genes underlying the disorder. One family had two children affected with IRIDA; the children’s parents were first cousins and their great, great grandparents were also closely related. This family’s genetic similarity, along with confidence in their phenotype based on tests performed by co-author Matthew Heeney, HMS instructor in pediatrics at CHB, gave the researchers the statistical power they needed to begin the genomewide search in a small cohort of people.

As they were about to begin, a group studying a similarly consanguineous family from Sardinia with multiple cases of IRIDA linked the disorder to chromosome 22q12-13. Knowledge of this locus narrowed the search to about 100 candidate genes.

Just as first author Karin Finberg, a former research fellow in the Andrews lab at CHB who is now at Duke, and co-author Dean Campagna, lead research technologist in the Fleming lab, had begun to sequence this subset of genes in the stored DNA of their IRIDA patients, the investigators got another lucky break.

“We had just started sequencing them when the Beutler group released their abstract for the American Society of Hematology meeting,” said Finberg. This abstract, submitted by Ernest Beutler from the Scripps Research Institute, described a mouse model with a defect in the murine homolog of a gene called TMPRSS6 within the human 22q12-13 region. In terms of IRIDA, said Finberg, “their mouse looked very similar to our patients.” She and Campagna immediately sequenced that gene and saw a variety of mutations. “We realized we’d hit the jackpot.” Their results appear in the May Nature Genetics.

“The real hero in this whole thing was the repository and the slow, gentle persistence of keeping it going,” said Heeney.

This persistent effort involved paperwork and legwork, most of which has become more burdensome over time. According to Fleming, the length of the original 1996 protocol and consent were eight pages and one page, respectively. They run to 40 and eight pages today. Though the regulatory changes driving this increase are in the best interest of research subjects, Fleming said, they also increase the energy required to do clinical research.

In this case, the extra effort paid off. Those with IRIDA were found to have severe TMPRSS6 mutations that likely cause loss of function in the encoded protein. Family members who have at least one normal copy of the TMPRSS6 gene do not have IRIDA. Moreover, mutations found in families with IRIDA were not detected in chromosomes from a large group of control individuals.

Though the team has not yet identified a molecular mechanism by which the serine protease TMPRSS6 acts, they hypothesized that a loss of function would result in inappropriately high levels of hepcidin, a hormone produced by the liver that regulates iron absorption (see figure); normally, iron deficiency anemia results in undetectable hepcidin levels. To test this hypothesis, the team reconnected with their IRIDA patients—in some cases, eight years after collecting their DNA—and tested their urinary hepcidin levels. All of the IRIDA patients had inappropriately elevated levels of hepcidin, which, said Heeney, is “like the thermostat that’s on all the time even though the house is warm.”

In the Andrews lab at Duke, Finberg is working to understand the exact molecular mechanism by which loss of normal Tmprss6 function leads to elevated hepcidin levels and IRIDA in mouse models. Finberg also plans to cross Tmprss6-deficient mice with mouse models of hemochromatosis, a form of iron overload in which there is too little hepcidin. This work will help determine if pharmacological modulation of Tmprss6 could dampen the iron overload of hemochromatosis.

Meanwhile, in the clinic, Heeney and Fleming will explore the possibility that this gene could also be involved in less severe forms of iron deficiency. To investigate this hypothesis, they will continue their clinical protocol of collecting and analyzing data from patients. “Every experienced hematologist will tell you about a patient, ‘Oh yeah, old Mrs. Widgets. She had iron deficiency, and we just couldn’t make it better,’” said Heeney. “It would be great to be able to identify those with a genetic risk for this and, also, it would be even better to be able to do something about it.”

If this broad influence pans out, this discovery could trigger a change in thinking about iron deficiency. The same way clinicians now know that diabetes has a genetic component, there may also be a genetic susceptibility to iron deficiency. “There might be extenuating reasons why you want to do more than the conventional therapy” of supplemental oral iron, said Fleming. “We’re trying to break down the notion that nutritional deficiencies are purely nutritional deficiencies.”