Increasing Brain Tumor Survival

Nearly 73,000 adults will be diagnosed with a brain tumor in 2016, and for more than one third, their tumor will be declared incurable. Large, collaborative efforts like The Cancer Genome Atlas have helped scientists better understand the genetic changes that define primary tumors, but this information alone is not enough to beat cancer.

Research must also be conducted to understand how tumors respond to treatment, what genetic changes occur with treatment, what type of environment the tumors grow in, how they interact with other cells in the body, and what cell of origin the tumor grows from. All of these problems are currently under investigation in the Brain Tumor Research Center at UCSF.

Oligodendroglioma is a type of primary adult brain tumor. The most common types of adult glioma are comprised of largely oligodendrocytes or astrocytes, called oligodendroglioma and astrocytoma, respectively. Oligodendrogliomas comprise 10% of gliomas.

Normal oligodendrocytes wrap around the axon of a neuron to help speed communication between cells, much like insulation on electrical wiring. Normal astrocytes provide structural support in the brain and help other brain cells get the nutrients they need.

The World Health Organization (WHO) describes adult brain tumors as: 1) low-grade tumors or grade II, or 2) high-grade tumors or grade III or IV. Grade II tumors are less aggressive with better patient outcomes.

It is well established that tumors are heterogeneous, meaning that not all of the cells in the tumor have the same genetic mutations. This heterogeneity can lead to treatment failure; if a treatment is given that only kills a portion of the cells, the remaining cells will repopulate the tumor, and in some cases, make the tumor worse than before treatment. One hypothesis in the field is that stem cells may contribute to the heterogeneity of solid tumors.

Adult stem cells are less differentiated cells that reside in many tissues in the body. When these cells divide, they simultaneously make a copy of themselves and a daughter cell that goes on to differentiate. For example, neural stem cells (NSCs) are primarily active during development and differentiate into neurons, astrocytes, and oligodendrocytes. NSCs are also present in the adult brain but in this context primarily produce neurons.

The presence of stem cells in tumors, termed cancer stem cells (CSCs), is debated. CSCs may contribute to the heterogeneity of tumors, but they have not been definitively defined in solid tumors.

Identifying CSCs in glioma would lead to a better understanding of the origin and development of glioma, and potentially lead to targeted therapies for the disease. If CSCs are responsible for the majority of tumor growth, and have common mutations, using drugs that kill cells with these mutations may halt tumor growth or lead to tumor death.

If we revisit the Central Dogma, our genes, which are comprised of DNA, are transcribed into mRNA, which acts as a template for the creation of proteins, which are the workhorse molecules in a cell. Sequencing of mRNA in tissues tells scientists what genes are expressed at a given time.

Many efforts have been made to profile what types of mRNA are produced by healthy and diseased cells and tissues. The snapshot of what mRNAs are present at a given time, called gene expression profiles, are well established for some stem cells and their differentiated progeny.

mRNA sequencing of tumors can provide novel expression profiles that can inform which drugs may be most effective against a tumor. Additionally, scientists can compare the tumor expression profile to known cellular profiles to determine what types of cells the tumor is comprised of.

For rare cells, like CSCs are proposed to be, the mRNA signal from the stem cells is obscured by the greater volumes of expression information from other tumor cells. This problem can be circumvented by looking at the gene expression profile of individual cells using a technique called single-cell mRNA sequencing or scRNA-seq.

Researchers from the Broad Institute and Massachusetts General Hospital used scRNA-seq to determine if CSCs are present in treatment-naïve grade II oligodendroglioma. Their results were published this month in Nature.

The authors performed scRNA-seq on six tumors.

First, the authors performed a statistical transformation on the scRNA-seq data, called a principal component analysis (PCA). PCA transforms the complex data (expression levels for thousands of genes in each cell) and simplifies it down to a few variables that can best describe the cells.

The PCA determined that the tumor cells could be separated into two major groups, those that express markers of oligodendrocytes, and those that express markers of astrocytes. If the cells expressed oligodendrocyte markers like OLIG1, OLIG2, and OMG, they had low expression of astrocyte genes like APOE, ALDOC, and SOX9, and vice versa.

Cells that had intermediate expression of oligodendrocyte and astrocyte genes preferentially expressed “stemness” genes, or genes that are associated with characteristics of stem cells, like self-renewal. Stemness gene expression is normally found in the prenatal brain and drops after birth, highlighting the abnormality of the stemness genes being expressed in the adult brain.

Additionally, while oligodendrogliomas were initially thought to arise from oligodendrocyte precursor cells, the CSCs identified in this study expressed genes associated with neuronal cells, suggesting an NSC may be the cell of origin of oligodendroglioma.

Next, the authors performed scRNA-seq on known populations of NSCs and neuronal precursor cells (NPCs). The stemness program of the NSCs and NPCs was very similar to the some of the putative oligodendroglioma stem cells identified from the tumor scRNA-seq.

This confirms that the expression profiles of oligodendrogliomas can be broken down into three main groups: oligodendrocyte differentiation, astrocyte differentiation, and a neurodevelopment stemness/progenitor program.

Next, the authors determined the proportion of proliferating cells by looking for high expression of genes associated with an active cell cycle. Interestingly, proliferating cells were in the stem-like state, suggesting that CSCs drive growth of oligodendroglioma.

Finally, the authors wanted to analyze the heterogeneity of the tumors. They used the scRNA-seq dataset to define subclones, or or a small population of cells harboring mutations that are different from the major population of cells in the tumor, based on the number of chromosomes in the cells. Exome sequencing, or sequencing of the protein-coding parts of the genome, was performed on bulk tumor samples. Variations in point mutations in CIC, a gene frequently mutated in oligodendroglioma, were used to confirm subclone presence.

Two tumors were determined to have subclones. Subclones were present in oligodendrocyte-like cells, astrocyte-like cells, and stem-like cells. This suggests that the subclonal mutations were not responsible for the different types of cells present in the tumor.

The authors conclude that CSCs do exist in grade II oligodendroglioma, and that these CSCs lead to the presence of oligodendrocyte-like and astrocyte-like cells that differentiate from a stem-like population. The CSCs are actively growing cells, which means they may be the major driver of glioma.

If these cells can be successfully targeted by therapy, a diagnosis of oligodendroglioma may no longer be a death sentence.