The challenge of treating childhood brain cancers

When Mariella G. Filbin was doing her residency in pediatrics in 2011, she was struck by how little she had to offer her patients with diffuse intrinsic pontine glioma, a type of childhood brain cancer also known as DIPG. There were no drugs for these kids. Most died within a year. This dearth of treatments inspired Filbin, who has a PhD in biochemistry and molecular biology in addition to her MD, to devote her career to studying deadly pediatric brain cancers with the aim of giving kids more options.

More than ten years later, the treatment landscape hasn’t changed much, says Filbin, a pediatric neuro-oncologist at Dana-Farber/Boston Children’s Cancer and Blood Disorders Center. All her patients with DIPG die, despite doctors’ best efforts. “To be honest, it’s very depressing to be a clinician in this field. And I can only do it because I want to change it,” she says.

While results in the clinic are still sobering, during the past decade, advances in chemical biology have given doctors and scientists deeper insight into the molecular mechanisms behind childhood brain cancers. And researchers are hoping they can eventually translate that knowledge into therapeutics.

But tackling brain tumors is tough. Medulloblastoma, juvenile pilocytic astrocytoma, ependymoma, low-grade glioma—these are a just few of the more than 100 types and subtypes of pediatric brain cancers for which researchers are trying to develop treatments.

Even when scientists understand the underlying biology, they don’t always know what the best molecular targets are. And when they do have a target, chemists creating a drug that can hit it face an additional degree of difficulty: to kill brain cancer cells, a drug must be able to reach them. Sometimes a tumor has made the brain permeable to drugs. But more often, chemotherapy for brain tumors must cross the blood-brain barrier and avoid the brain’s mechanisms for flushing out foreign molecules. This is where insights in medicinal chemistry can move the field forward.

A tough problem

In 2020, more than 30,000 children and adolescents around the world were diagnosed with a brain or spinal cord tumor, according to Globocan 2020, a database of global cancer statistics. This makes those diseases the second-most-common childhood cancers, behind leukemia. But brain tumors kill more kids than leukemia. They are the deadliest form of childhood cancer, according to the US National Center for Health Statistics.

Not all pediatric brain tumors are fatal. The 5-year survival rate for children diagnosed with brain tumors varies from 0 to 90%, depending on tumor type, according to the Central Brain Tumor Registry of the US. But even children who survive their cancer have a difficult path because the treatment for brain tumors often causes serious side effects.

Why are pediatric brain cancers so difficult to treat? “Partly because the tools we have are very blunt,” says Robert Wechsler-Reya, a professor at Columbia University Irving Medical Center who studies medulloblastoma. Those blunt tools are surgery, radiation, and chemotherapy. Surgeons remove as much of a tumor as they can, and then doctors use radiation and drugs to eliminate what’s left.

With medulloblastoma, Wechsler-Reya says, “the kids who do survive end up having very severe long-term side effects from the radiation, because we’re blasting their developing brains.” The same is true for kids with other types of tumors.

Chemotherapy is also more challenging in children than adults, says Sidharth Mahapatra, a pediatrician at the University of Nebraska Medical Center who studies medulloblastoma. “Kids can’t get maximal chemotherapy; they’re just too fragile,” he says.

But doctors and scientists say they have reason to believe that better treatments are on the horizon. “In the last decade, there has been an explosion in understanding the genomics of cancer,” Mahapatra says.

Genetic sequencing can reveal which mutations drive tumor growth, and as researchers learn more about a cancer’s molecular drivers, they can design targeted treatments and perhaps avoid using radiation and untargeted chemotherapy, which cause damaging side effects. “We’ve made a lot of progress in understanding these diseases, and we’re beginning to see some of that science turn into therapies,” Wechsler-Reya says.

Amar Gajjar, a pediatric neuro-oncologist at St. Jude Children’s Research Hospital, says doctors used to think that what works for tumors in adults should also work for kids. “Now we know that’s a flawed way of thinking,” he says.

It has been only in the past decade that researchers have been able to get molecular insight into brain cancers. They learned that childhood and adult brain cancers often have different molecular mechanisms—​​distinct mutations that drive a cancer’s growth.

For example, under the microscope, cells from adult and pediatric high-grade glioma tumors look the same, but when you analyze their molecular genetics, their mutations and developmental drivers turn out to be completely different. “So all this time, we were sort of operating in the dark, treating just the histology, but we didn’t know the underlying biology,” Gajjar says.

Cancer’s molecular drivers

To better grasp what drives diffuse midline gliomas in children and adults, Dana-Farber’s Filbin recently used a molecular profiling method called spatial single-cell transcriptomics. She studied these cancer cells’ structures in people across a range of ages who had tumors in various locations in the brain. Filbin’s team focused on tumors driven by a genetic mutation called histone H3-K27M, wherein a lysine in histone 3 has been replaced by a methionine.

Filbin and her colleagues found that in kids, these tumors have more immature cells that are prone to proliferate quickly because they are similar to stem cells. In adults, the cells are more like the ones that heal damaged tissue. But there is no genetic difference between the cells in children and adults. Instead, the researchers suspect that the immune microenvironment around the tumors might be what makes them different. That view suggests that treatments that work for adults might not work for kids (Nat. Genet. 2022, DOI: 10.1038/s41588-022-01236-3).

Brain tumors are connected to normal brain development in many ways, Filbin says. The mechanisms that help a brain develop can also feed tumor growth. “How to disentangle what’s bad while preserving what’s good in the normal brain is just really difficult,” she says.

A good understanding of the underlying biology is what drove researchers to develop the combination therapy of dabrafenib and trametinib. In March, the US Food and Drug Administration approved this drug combination, which is marketed by Novartis as Tafinlar and Mekinist, to treat low-grade glioma in children who are 1 or older and have what’s known as the BRAF V600E mutation.

This mutation in the BRAF gene causes a valine at the 600 position in the protein B-Raf to be replaced with a glutamic acid. It’s known to drive a signaling pathway that leads to uncontrollable cell growth and therefore cancer.

The mutation occurs in 15–20% of children with low-grade glioma, but it isn’t specific to that cancer. In fact, the dabrafenib-and-trametinib combination had previously been approved to treat several other cancers with this mutation, including melanoma, thyroid cancer, and lung cancer.

The type of cancer wasn’t important to deciding to use the dabrafenib-and-​trametinib combination, but the BRAF V600E mutation was, explains Jeff Legos, head of oncology drug development at Novartis. “We were selecting patients on the basis of the mutation rather than the organ where the cancer had originated,” he says.

Dabrafenib, one half of the treatment, is a highly selective inhibitor of the mutated B-Raf protein. Chemists at GSK reported the molecule in 2013 (ACS Med. Chem. Lett., DOI: 10.1021/ml4000063). Scientists working on the project realized that cancers would be able to develop resistance to dabrafenib, which is why they began combining it with trametinib early in clinical trials.

“We worked very hard to understand the mechanisms of resistance very early on in the development program,” Legos says.

Trametinib blocks the mitogen-activated protein kinase kinase, which is also known as MEK or MAP2K and is downstream of B-Raf in a signaling pathway. The drug was first reported by scientists at GSK and Japan Tobacco in 2011 (ACS Med. Chem. Lett., DOI: 10.1021/ml200004g). Novartis acquired both drugs in a 2015 deal.

Low-grade glioma likely disrupts the blood-brain barrier, and preclinical data suggested that either the two drugs or their active metabolites could cross the blood-brain barrier, Legos says. Whatever the mechanism, the fact that these drugs could enter the brain got researchers thinking the compounds could benefit people with brain tumors with the BRAF V600E mutation as well as other cancers with the mutation that had spread to the brain.

Realizing that the BRAF V600E mutation plays a role in some forms of pediatric low-grade glioma, the scientists working on the project got the formulation team involved early, Legos says. That team developed an oral suspension of the drugs that could be given to young children.

Before this year’s approval of dabrafenib and trametinib to treat pediatric low-grade glioma, “almost 4 decades have gone by where these pediatric patients just continued to receive the same age-old chemotherapy,” Legos says. “It’s incredibly important that we continue to do more and to do better for these pediatric patients.”

Past the blood-brain barrier

The team that developed dabrafenib and trametinib wasn’t necessarily aiming for drugs that could travel into the brain. But that type of challenging medicinal chemistry design work has become important to drugmakers working on brain cancer treatments.

“What gives me optimism for what is to come in the future is that there is a deeper appreciation for how to assess brain penetration,” says Timothy P. Heffron, vice president of discovery chemistry at Genentech. He points to the development of paxalisib as an example.

Paxalisib was designed to block phosphoinositide 3-kinase (PI3K). Aberrant PI3K signaling is one way cancer cells proliferate. Several FDA-approved drugs pursue this target, but none of them cross the blood-brain barrier. This property, Heffron says, was probably by design. Drugmakers didn’t want to worry about off-target effects in the central nervous system when they were trying to treat cancers elsewhere in the body.

Genentech was developing PI3K inhibitors when Heffron and his coworkers realized that by avoiding the brain, they might be missing an opportunity to create therapeutics for brain cancers, both in kids and adults.

The PI3K pathway is implicated in 80% of glioblastoma multiforme tumors. It’s also a driver for many cancers that can metastasize and travel to the brain, where they can evade chemotherapies that can’t cross the blood-brain barrier. So the chemists decided to develop a PI3K inhibitor that could cross into the brain.

The chemists’ molecular starting point—a compound that was in clinical development for peripheral cancers driven by PI3K—was too polar to cross the blood-brain barrier.

But the researchers made the key realization that they could remove an indazole group on that starting molecule and replace it with an aminopyrimidine. In doing so, they created an additional contact between their molecule and PI3K. With that additional contact, the chemists reaped an added benefit: they could eliminate a part of the molecule that had significant polarity while still retaining the molecule’s potency.

The change also reduced the number of hydrogen-bond donors in the molecule, which was key to getting it across the blood-brain barrier and avoiding the mechanisms that shuttle molecules back out of the brain—a process known as efflux.

During Genentech’s medicinal chemistry effort, scientists at Pfizer published a series of papers describing an algorithm that assessed central nervous system penetration, Heffron recalls. The algorithm did a good job at picking out which molecules would be affected by efflux.

“We decided to apply this algorithm to see if that would help us to more efficiently make molecules that have the desirable attributes that we’re after,” Heffron says. After the researchers applied the algorithm, many more of the molecules they made avoided efflux. Further medicinal chemistry efforts to ensure metabolic stability brought the chemists to paxalisib (ACS Med. Chem. Lett. 2016, DOI: 10.1021/acsmedchemlett.6b00005).

Paxalisib is in clinical trials for several types of brain cancers. It received an orphan drug designation from the FDA in 2020 for treating DIPG and in 2022 for atypical rhabdoid or teratoid tumors, a rare and aggressive childhood brain cancer. Kazia Therapeutics licensed paxalisib from Genentech in 2016.

Other medicinal chemistry teams have sought to build cancer-fighting molecules that can slip beyond the blood-brain barrier. For example, Pfizer’s lorlatinib, which the FDA granted accelerated approval to in 2018 and is marketed as Lorbrena, was designed to get into the brain and avoid efflux.

Lorlatinib inhibits anaplastic lymphoma kinase (ALK), which can drive cancer in several ways. In designing lorlatinib, Pfizer medicinal chemists were challenged to make key hydrogen-bond contacts to ALK but also limit the number of hydrogen-​bond donors to ensure the compound penetrated the brain.

They had previously developed crizotinib, marketed as Xalkori, and knew that an aminopyridine group was critical to ALK binding. Making lorlatinib a macrocyclic compound that contained that aminopyridine-binding motif and several other structural features of crizotinib turned out to be critical to success. Macrocyclic molecules created in this medicinal chemistry campaign were better than acyclic molecules at avoiding efflux.

Also, the macrocycles have fewer rotatable bonds and are forced to adopt a more compact shape than their molecular weight would suggest. These features help lorlatinib cross the blood-brain barrier (J. Med. Chem. 2014, DOI: 10.1021/jm500261q).

Lorlatinib recently replaced crizotinib in a Phase 3 clinical trial of children with high-risk neuroblastoma whose tumors have a genetic alteration in the ALK gene (Nat. Med. 2023, DOI: 10.1038/s41591-023-02297-5).

Reducing the number of hydrogen-bond donors was also important for the development of vorasidenib. The compound is in clinical trials for patients 12 years and older with low-grade gliomas that have mutations in the genes that code for isocitrate dehydrogenase 1 and 2 (IDH1 and IDH2). When the enzymes have specific mutations, they can produce a metabolite associated with low-grade gliomas. Vorasidenib blocks those enzymes. Agios Pharmaceuticals created the compound, which is being developed by Servier.

Chemists at Agios modeled vorasidenib after another molecule they developed, called enasidenib. Enasidenib, which is marketed as Idhifa by Celgene, has a triazole core but doesn’t penetrate the brain, most likely because it has three hydrogen-bond donors. To create vorasidenib, Agios chemists kept the triazole core but changed the groups around the periphery to limit the number of hydrogen-bond donors in the molecule to two (ACS Med. Chem. Lett. 2020, DOI: 10.1021/acsmedchemlett.9b00509).

More work to do

Even as chemists create cancer-fighting molecules that can cross into the brain, there’s still a long way to go in bringing therapies to kids with brain cancer. In addition to understanding what’s driving brain tumors in children and finding targets that won’t affect developing brains, doctors say multiple therapies will most likely be needed to prevent these cancers from developing resistance.

The work is not for the faint of heart, says Dana-Farber’s Filbin. But scientists have an opportunity to make a tremendous impact.

“For people who really want to make a big splash in their lifetime with their skills, I think this is the field to go in,” she says. “A lot of things will fail because it’s tough. But if we don’t do it, who will?”