All in the family: How 59,000 volunteers and a team of scientists helped uncover six inherited gene mutations for medulloblastoma

Although medulloblastoma is the most common malignant childhood brain tumor, it’s still rare in the big scheme of things. Only 250 to 500 children in the U.S. are diagnosed with this tumor each year. For many years, scientists believed most cases of medulloblastoma did not run in families.

And yet...

At St. Jude, I met parents who had lost two kids to medulloblastoma within a year. How could that be, if the tumor occurs by chance and is rarely hereditary?

When I met that couple, I had already been investigating this question for about five years. Our meeting motivated me to redouble my efforts on behalf of medulloblastoma patients and their parents around the world. Our conversation also provided additional evidence that some cases of this brain tumor have a hereditary component.

To truly understand medulloblastoma, it’s crucial to join forces with other institutions worldwide. That’s the only way to get sufficient numbers to uncover the molecular nuances of a disease that affects such a small group of children.

My colleagues and I partnered with scientists at the European Molecular Biology Laboratory and German Cancer Research Center in Heidelberg, and the Hospital for Sick Children in Toronto. We wanted to find out whether there were inherited gene mutations that could lead to medulloblastoma.

Thousands of volunteers

In our earlier studies of the disease, we had focused on the tumors themselves. We described their molecular landscape, the genes that were mutated, the pathways that were deregulated. But in order to figure out whether the disease could be inherited, we turned our attention to the germline—the normal DNA content present in every cell of a person’s body.

We gathered material from the blood or saliva of 1,022 children with medulloblastoma and existing sequence data from 58,000 healthy individuals. We decided to analyze the genomic sequence of each patient's normal DNA.

We were looking for what's known as genetic predisposition or hereditary predisposition—something that's present in every cell of an individual but might increase the risk of that person developing medulloblastoma at some point in their lifetime—or perhaps another type of cancer, or various cancers, depending on the gene.

How did our study differ from other medulloblastoma studies?

Prior studies on genetic predisposition to medulloblastoma had focused on select clinical cases and pedigrees, or families where multiple generations or siblings had developed medulloblastoma. Often, those studies focused on case reports. A few predisposition genes were already known to be responsible for rare familial tumor syndromes. They had been implicated, but on a small scale.

Not only were we looking for genetic predisposition, but we were doing it with an awareness of how those patients fared in terms of their outcome, or how they were categorized according to their disease subtype.

Medulloblastoma is not a single disease but a collection of distinct disorders. There are four subgroups, named WNT, sonic hedgehog, Group 3 and Group 4. When patients come in with medulloblastoma, they're classified according to their molecular subgroups.

And so we decided to study germline predisposition as it related to these subgroups. We would look for germline mutations that were specific to each subgroup. It was the largest systematic, unbiased analysis of predisposition to medulloblastoma in a massive multi-institutional international cohort.

From 22,000 genes, 110 suspects

We restricted our analysis to only a small group of the 20,000-plus genes in the human genome. We focused our examination on 110 known cancer predisposition genes.

Many of us—meaning you and I, our siblings and our friends—could be carriers of mutations of unknown significance in genes that might cause cancer. It’s really hard to prove whether those mutations or variants are going to actually predispose us to develop cancer.

So we focused only on the very strong predisposition genes—ones where there's already plenty of evidence in the literature that mutations in these genes are known to cause cancer.

From there, we compared potential mutations in our medulloblastoma population to a normal control cohort of 58,000 people. By using biostatistics and computational approaches, we compared the incidence of such mutations in these 110 genes in our 1,022 children with medulloblastoma to that of normal, healthy individuals.

When we did the analysis and computed the statistics, we identified six genes that are significantly and statistically more frequently mutated in medulloblastoma patients than in the normal controls. Those genes are APC, PTCH1, SUFU, TP53, BRCA2 and PALB2.

What we discovered

When we broke the results down according to subgroup, we saw clearly different risks according to subgroups. For instance, nearly one in five patients with sonic hedgehog medulloblastoma have a hereditary predisposition that carries significant implications for their treatment, their future surveillance, their siblings, their parents—for everybody associated with that patient.

Our findings help clinicians who are treating medulloblastoma understand what they should be looking for as they consider possible cancer predisposition in their patients. We’ve proposed guidelines for identifying these patients. We've broken these down according to the subgroup to which each patient matches and have indicated guidelines that we think should be followed.

Based on our findings, we can now pinpoint which patients should be monitored closely for specific types of cancer they could develop, sometimes long after they've survived medulloblastoma.

The patients might assume they’re free and clear because they've survived their brain cancer, but they’re certainly at risk for other cancers. And they also have to be careful with the treatments they're given for their medulloblastoma, because that could then increase their risk further for developing what's called a secondary malignancy.

If a mutation in one of these six genes is identified in their germline, then it will alter how a patient is monitored for development of a secondary malignancy. It might not change directly how they're treated in terms of standard of care for their medulloblastoma, but certainly how they're monitored and watched in months and years following treatment.

The guidelines are already in use at St. Jude. Our collaborators around the globe are encouraging the World Health Organization to incorporate the advice into its guidelines for managing and diagnosing brain tumors.

Patients who have these mutations should be watched very closely, and their families need to be tested. For instance, the siblings I mentioned earlier who died of medulloblastoma both had a germline TP53 mutation. If you're in a family of two or three kids, and one develops medulloblastoma, and it turns out to be a sonic hedgehog subtype with a TP53 germline mutation, you definitely want to test the brothers and sisters.

So, what’s next? Well, now we're going to continue to mine our data and see if there are other genes that could be contributing to hereditary predisposition, either across medulloblastoma or in a particular subgroup. I think there could be more out there that we’ve not yet seen based on our current approach. So the search continues.