Shapeshifters in the Brain

Posted date

New Single-cell Technologies Are Revealing How Aggressive Brain Tumors Evade Treatment

From Paths of Progress 2020

By Nicole Davis

brain653x516.jpg

Glioblastoma is the worst of the worst. A form of brain cancer that affects adults and children, it is both aggressive and difficult to treat. Tragically, most patients die within two years of diagnosis. But new research is cracking open the biology of the disease, helping explain the tumor's tenacity even when faced with targeted therapies. And those insights are at last giving scientists a critical toehold in the hunt for new, more effective glioblastoma treatments.

In a study published last summer, a team from Dana-Farber, Boston Children's Hospital, and other leading research organizations used single-cell techniques, which can examine the genetic information within individual cells, one at a time. The researchers scrutinized more than 23,000 cells from both pediatric and adult tumors. They uncovered some surprising similarities among the cells that drive these cancers, insights that could help spur new drugs aimed at glioblastoma.

mariella-filbin250x275.jpg
Mariella Filbin, MD, PhD

"We've kind of come full circle in how we think about adult and pediatric glioblastoma," said Mariella Filbin, MD, PhD, a lead author of the recent paper, which appeared in Cell, and a pediatric neuro-oncologist at Dana-Farber/Boston Children's Cancer and Blood Disorders Center. "Up until about a decade ago, we thought they were essentially the same, so we treated them that way. But more recently, studies of the tumors' genomes revealed distinct genetic mutations — and that got us thinking that glioblastoma in children is quite different than in adults."

Now, Filbin and her colleagues are finding that the two cancers may actually be more alike than not. They determined that the cells that drive both adult and pediatric glioblastoma exist in four distinct states, which are defined by the constellation of genes that are switched on or off — a kind of molecular signature. These signatures closely resemble those seen in the assortment of normal cell types that make up the developing brain.

Importantly, Filbin and her team revealed that these driver cells are not static and can readily switch from one state to another, like a game of whack-a-mole. That discovery helps explain a sobering clinical reality: why existing, single-target drugs for glioblastoma fail to curb the cancer's growth.

Opening Brain Cancer's Bag of Tricks

Like other forms of brain cancer, glioblastoma is alarming because of where it originates. As the epicenter of the mind and body, the brain can endure only a limited amount of surgery and radiation. It is also separated from the bloodstream by a special structure, the blood-brain barrier, which restricts the flow of drugs and other chemicals to brain tissue, further complicating treatment.

Yet glioblastoma also has other tricks up its sleeve. The tumor becomes interwoven into the fabric of the brain, mingling with normal structures and making itself inseparable from them.

"It's actually like trying to remove a virus with surgery," said Keith Ligon, MD, PhD, one of Filbin's collaborators and a neuropathologist at Dana-Farber/Boston Children's.

Although Filbin understood this dismal picture, she initially set out to study a different cancer, a rare brain tumor called diffuse intrinsic pontine glioma (DIPG), which affects mostly young children and, sadly, also has a grim prognosis.

She was struck by the power of new single-cell technologies, specifically single-cell RNA sequencing, which provides a cell-by-cell snapshot of which genes are active and which are not — the so-called "transcriptome." That view goes beyond what can be gleaned simply by looking at the cells' DNA, which only reveals which genes are missing or broken, not how those changes impact cell function. Because these techniques are painstakingly applied to individual cells, they can reveal subtle differences within a tumor that may be missed by cruder, bulk profiling methods.

"DIPG was the very first cancer I wanted to look at because these kids have such poor survival rates," said Filbin. "We need to do better and we need to do something different to understand that disease."

So she and her colleagues harnessed single-cell RNA sequencing to examine nearly 2,500 cells from six DIPG patients. In a paper published in Science in 2018, they revealed that the majority of tumor cells are developmentally stuck — instead of growing up and becoming more specialized, as normal cells do, DIPG cells are trapped in an immature, stem-cell like state.

ligon-and-chow653x447.jpg
Keith Ligon, MD, PhD, and Kin-Hoe Chow, PhD, associate director of the Center for Patient Derived Models.

Filbin and her team also noticed something else. Prior to any treatment, a small fraction of tumor cells can overcome this developmental roadblock. They simply grow up or "differentiate" — and, at the same time, lose their tumor-forming potential.

"That is super exciting," said Filbin. "We now have the first evidence that a pediatric brain cancer can differentiate on its own."

She and her colleagues are now exploring whether they can exploit this capability therapeutically. Can a drug be identified that coaxes more tumor cells to differentiate, thereby rendering the cancer harmless? This concept, known as differentiation therapy, has been around for decades in pediatric oncology and currently forms the basis for treating a form of pediatric leukemia as well as another pediatric tumor called neuroblastoma.

A Cell-By-Cell View of Glioblastoma

Filbin was so impressed with the idea of differentiation therapy for pediatric brain cancers, she wondered if it might apply to other aggressive forms, like glioblastoma. Her colleagues at Massachusetts General Hospital were already using single-cell techniques to study adult forms of the disease, so the researchers banded together to conduct a broader analysis of pediatric as well as adult tumors. They examined tumors from 20 adult and eight pediatric patients.

Prior to the advent of single-cell methods, scientists delved into the molecular makeup of patients' tumors using a bundled approach. They would isolate a chunk of tumor tissue, grind it up, and perform experiments on that cellular "soup" — effectively averaging across all of the cells present in the mix. And yet tumors don't grow, spread, or die as an average unit. They do so cell by cell.

"A good analogy here is your friendships," said Ligon. "If you averaged all of your friends' personalities, that won't really tell you anything about what each person is like."

With single-cell technologies, Filbin, Ligon, and their colleagues were able to glimpse for the first time a detailed picture of glioblastoma on a scale that matches its biology.

The team discovered that tumor growth is fueled by four distinct cell types or "states," which exist at varying degrees in all glioblastoma patients, both children and adults. These states resemble different cells normally found in the developing brain, including astrocytes and oligodendrocyte progenitor cells, both types of support cells called glia; neural progenitors, which can give rise to both neurons and glia; and mesenchymal cells, which give rise to connective tissue.

These glioblastoma cell states are highly plastic and can readily shift from one to another. When Filbin and her colleagues injected just one of the four cell types into mice, tumors formed that contained all four types, helping to explain why existing single-target therapies fail to curtail glioblastoma growth.

Unfortunately, the data do not give credence to differentiation therapy as an effective treatment for glioblastoma, though, as Filbin's earlier work indicates, it remains a promising approach for DIPG. Nevertheless, Filbin and her team believe their findings offer a roadmap for identifying new glioblastoma therapies, and they have launched early-stage research projects to pursue them.

"In every glioblastoma we looked at, whether it was from a kid or an adult, it had only those four cell types, regardless of its genetic makeup," said Filbin. "That gives us hope that if we can find a way to block all four at the same time, we can kill the tumor."

Such a four-pronged attack would signify a groundbreaking approach to glioblastoma treatment — one that could deliver long awaited hope to patients.

Other Feature Stories from Paths of Progress 2020

To see the entire publication, please download the PDF version.

Request a Publication

Receive by mail the current issue of a Dana-Farber publication by completing this request form.

Request a Publication

Media Contacts

If you are a journalist and have a question about any of our stories or need more information, email media@dfci.harvard.edu or call 617-632-4090 and ask to speak to a member of the media team.

The Media Team cannot respond to patient inquiries. For more information on contacting Dana-Farber, please see Contact Us.

Disclosure of Advisory Roles and Certain Industry Support

See an overview of relevant board memberships, outside support, and significant advisory roles as reported by several physicians and researchers featured in the 2020 issue of Dana-Farber's Paths of Progress magazine.

VISIT OVERVIEW