People with brain tumours have a range of symptoms that can vary in severity, from headaches to a decline in cognitive function. The symptoms depend on the tumour type and its size, location and growth rate. Understanding what controls the growth rate of brain tumours might therefore lead to the development of therapies that slow cancer progression and improve the quality of life of people who have this type of cancer. Writing in Nature, Venkataramani et al.1, Venkatesh et al.2 and Zeng et al.3 report that, in the brain, neurons and cancer cells form a type of connection between cells called an excitatory synapse, and the formation of this connection boosts tumour growth.
An excitatory synapse is a structure in which two adjacent neurons — termed the presynaptic and postsynaptic neurons — communicate using a neurotransmitter molecule, usually glutamate (Fig. 1). Glutamate release by the presynaptic neuron activates glutamate receptors, known as AMPA receptors and NMDA receptors, on the postsynaptic neuron. Receptor activation causes ion movement across the cell membrane, which produces depolarization — an increase in positive charge inside the postsynaptic neuron that leads to excitation. Certain non-neuronal brain cells called glia surround a synapse and regulate signal transmission across it by removing released neurotransmitter4. Other types of glial cell affect neuronal excitability (the ease with which neurons are depolarized) by regulating extracellular potassium ions5.
Glial cells can give rise to a type of brain tumour called a glioma, which is the leading cause of death from brain cancer in the United States6. One common characteristic among many different types of glioma is that their growth requires the activity of their neighbouring neuronal cells7, but the reason has not been fully understood until now.
Healthy glial cells form interconnected cellular networks. This is because structures on the glial-cell membrane, called gap junctions, enable signalling molecules, such as calcium ions, to move into neighbouring glial cells5. Glioma cells also create interconnected cellular networks by forming gap junctions in what are called tumoural microtubes — long, thin, cell-membrane protrusions that extend from these cells into the surrounding tissue, and which contribute to tumour infiltration and proliferation8.
Using an imaging method called electron microscopy, Venkataramani and colleagues examined tumoural microtubes formed by human gliomas that had been transplanted into mouse brains. They observed that the microtubes had structures characteristic of excitatory synapses, called postsynaptic densities, where glutamate receptors are normally present. Adjacent to these postsynaptic densities, in a nearby neuron, the authors noted clusters of vesicles that store neurotransmitter molecules, which are a feature of a neuronal presynaptic zone. Venkatesh and colleagues made similar observations of synaptic structures arising between glioma cells and neurons.
Venkatesh et al. and Venkataramani et al. provide evidence that genes encoding glutamate receptors and structural components of the postsynaptic region are expressed in a subset of cells in human gliomas, suggesting that glioma cells exploit the same molecular mechanisms used by neurons to establish synapses. To determine whether synapses between tumour cells and neurons function in a similar way to those formed between neurons, both groups transplanted human glioma cells into mouse brains. The stimulation of neurons near the glioma cells produced a rapid depolarizing current in some glioma cells characteristic of excitatory synapses, and this current was mediated by AMPA receptors. Crucially, the type of AMPA receptor expressed in these glioma cells has different pharmacological properties from the AMPA receptors expressed in neuron, making it a promising candidate as a drug target. In some of the other glioma cells, a long-lasting depolarizing current was observed that amplified and spread through gap junctions to the connected network of tumour cells. This prolonged current was not of a synaptic origin — instead, it seemed to come from changes in the extracellular concentration of potassium ions as a result of neuronal activity.
The depolarization of glioma cells induced by neuronal activity caused a transient rise of calcium ions in the cytoplasm, which then spread through the network of glioma cells through their gap junctions. A higher frequency of these calcium signals correlated with increased migration of some of the tumour cells in the network, indicating that synapse formation in a tumour cell altered the properties of other cells in the tumoural network and increased their invasiveness.
To determine the biological importance of the synapses formed between glioma cells and neurons in their model systems, Venkatesh et al. and Venkataramani et al. used either pharmacological tools or genetically engineered glioma cells to block AMPA receptors and thereby prevent depolarizations induced by synaptic activity. These treatments led to an increase in the survival time of animals that had received a transplant of human glioma cells, compared with control animals in which AMPA receptors were not blocked. These manipulations of AMPA receptor function therefore caused a substantial reduction in the effect of synaptic stimulation on the proliferation and invasiveness of glioma cells.
Both groups also engineered human glioma cells to express a light-activatable version of an ion channel that produces cellular depolarization, similar to that obtained by the synaptic-mediated activation of glutamate receptors. Such cells were transplanted into the brains of mice by Venkatesh et al. or grown in vitro by Venkataramani and colleagues. When light was used to depolarize the glioma cells, this promoted tumour proliferation.
Together, the evidence from Venkatesh et al. and Venkataramani et al. indicates that some glioma cells have the capacity to form functional synapses with neurons that are present in their microenvironment. Moreover, these cells form an electrically active tissue that can signal to other glioma cells in the tumour-cell network to promote their migration and growth. The presence of functional synapses between neurons and cancer cells explains why glutamate-mediated neurotransmission is associated with enhanced proliferation, survival and invasiveness of glioma cells9–12.
Zeng and colleagues investigated the role of glutamate-mediated signalling in tumours by examining the expression of glutamate receptors in many different sorts of human cancer. Breast cancer cells, a type of tumour cell that often migrates to the brain, had higher expression of NMDA receptors compared with other types of tumour that they studied. A protein subunit of NMDA receptors called GluN2B — required for synapse formation13 and for changes in the strength of synaptic connections14 — was highly expressed in human and mouse breast cancer cells that the authors determined have a high capacity to migrate to the brain. NMDA receptors allow calcium ions to enter cells and have been implicated in aiding the progression of some human cancers15. Using techniques that included an imaging approach to track calcium levels and electrophysiological recordings, Zeng and colleagues report that NMDA-receptor-mediated signalling occurred in mouse breast cancer cells.
Human breast cancer cells express neuroligin, a protein that aids adhesion between cells and that normally contributes to the formation of synaptic structures between neurons. This suggests that, in a similar way to human glioma cells, human breast cancer cells might exploit the standard mechanisms used by neurons to establish synaptic connections. Indeed, when Zeng et al. used microscopy techniques to study samples of mouse tissue containing human breast cancer cells that had been injected into the animals’ brains, they observed that proteins involved in packing glutamate into vesicles were in close proximity to NMDA receptors, and that synaptic structures had formed between the cancer cells and the neurons. Finger-like cellular processes extended from the cancer cells to reach existing mouse synapses between neurons, adopting the same typical position around a synapse as is adopted by glial cells that remove glutamate from the synaptic cleft. Such an arrangement would allow cancer cells to obtain glutamate to activate their NMDA receptors.
Zeng and colleagues report that, if mice received breast cancer cells engineered to have reduced GluN2B expression, the cells produced smaller brain tumours and the animals had a longer survival time compared with animals that received breast cancer cells with normal GluN2B levels. These results indicate that the signalling mediated by the NMDA receptor promotes the colonization and growth of cancer cells in the brain.
Together, these three studies demonstrate that brain tumours can establish synaptic connections with neurons using the molecular toolkit that forms synapses between neurons. In neurons, synaptic activity enables the depolarization and calcium-ion influx that is necessary for cellular differentiation, proliferation and survival. In cancer cells, these same processes instead support the proliferation of the tumour and contribute to cancer’s eventual lethality. These intriguing findings raise the possibility that approaches targeting specific types of glutamate receptor, postsynaptic signalling processes or the mechanisms needed for synapse formation might provide therapeutic targets for slowing tumour proliferation.
Nature 573, 499-501 (2019)