Electrical and synaptic integration of glioma into neural circuits

Welcome to our Monthly Journal Club! Each month I post a paper or two that I have read and find interesting. I use this as a forum for open discussion about the paper in question. Anyone can participate in the journal club, and provide comments/critiques on the paper by leaving a comment below. I picked this month’s paper because it describes a scary phenomenon (brain tumors interlocking with neural circuits!) that could have far-reaching consequences for how we treat brain cancer. This paper, in combination with two others in the same issue, will become classics in the new field of ‘cancer neuroscience’. The paper we are discussing is titled Electrical and synaptic integration of glioma into neural circuits” (click the hyperlink to see the paper) by Michelle Monje & colleagues at Stanford University. Two other great papers outlining integration of cancer and neural circuits can be found at these links:

Synaptic proximity enables NMDAR signalling to promote brain metastasis

Glutamatergic synaptic input to glioma cells drives brain tumour progression

Together, these studies provide evidence that cancer cells synaptically communicate with neurons in the brain, and this communication boosts tumor growth! (see the figure below)

Neural activity promotes tumor growth and progression. Glutamate signaling depolarizes (activates) tumor cells that are ‘listening in’ on normal neural communication. (Credit:    Barria, 2019   ).

Neural activity promotes tumor growth and progression. Glutamate signaling depolarizes (activates) tumor cells that are ‘listening in’ on normal neural communication. (Credit: Barria, 2019).

Below, check out a video of brain cancer cells in a mouse brain expressing a fluorescent activity indicator (GCaMP6s). Spontaneous activity of the cancer cells can be see as waves of green propagating throughout the network. Cell nuclei are labeled red. (Credit: Venkatesh et al., 2019).

High grade gliomas are the most prevalent and deadly of brain cancers in adults and children. Due to their intimate interaction with normal brain tissue, it is extremely hard to eliminate this cancer without destroying the surrounding cells. Significant focus has been on understanding the ‘intrinsic’ mechanisms within the cancer cells that regulate the tumor’s growth and progression. More recently, the role the ‘microenvironment’ plays has come center stage. This includes the cells and extracellular material around and bathing the tumor itself. Michelle Monje’s group demonstrated that this microenvironment is very important for tumor growth as neuronal release of neuroligin-3 is required for glioma growth (click the hyperlink to see the paper). Building on these findings from a few years ago, her group was interested in whether neuronal activity directly influences brain tumor growth and progression. This would require functional synaptic connections to form between neurons and tumor cells.

To examine this, they started to hunt for signs of synaptic connections between these two cell types in primary human tumor samples and mouse models of brain cancer, namely ‘diffuse intrinsic pontine glioma’ (DIPG; see Figure 1 below)

Figure 1:  Evidence for functional synapses between neurons and brain tumor cells. In (a) we can see that the expression levels of synapse-related genes (GRIN1, GRIA1,2,3, GRIK2, DLG4, NLGN3, HOMER1) are highly enriched in malignant vs. non-malignant tissues from cancer patient samples. In (b) the data are arranged to see the lineage (x axis) and stemness (y axis) of cells from primary patient samples. In (c) we can see physical signs of synapses using electron microscopy in a human (left) and mouse (right) brain tumor. In (e) and (f) the researchers found signs of synaptic transmission in glioma by labeling the protein post-synaptic density-95 (PSD-95) and synapsin. (Credit: Venkatesh et al., 2019).

Figure 1: Evidence for functional synapses between neurons and brain tumor cells. In (a) we can see that the expression levels of synapse-related genes (GRIN1, GRIA1,2,3, GRIK2, DLG4, NLGN3, HOMER1) are highly enriched in malignant vs. non-malignant tissues from cancer patient samples. In (b) the data are arranged to see the lineage (x axis) and stemness (y axis) of cells from primary patient samples. In (c) we can see physical signs of synapses using electron microscopy in a human (left) and mouse (right) brain tumor. In (e) and (f) the researchers found signs of synaptic transmission in glioma by labeling the protein post-synaptic density-95 (PSD-95) and synapsin. (Credit: Venkatesh et al., 2019).

Using a variety of methods including transcriptomic profiling, electron microscopy, and immunohistochemistry, the researchers were able to demonstrate signs of the synapses between glioma cells and neurons! This was a great first step…but what are those synapses doing? Are they functional? Do glioma cells propagate neural signals or does the signal ‘die’ when it hits the tumor?

To investigate the functionality of these tumor synaptic connections, Dr. Monje’s group transplanted tumor cells into a mouse's brain and then recorded from these cells during stimulation of a specific neural pathway (Schaffer Collateral). By doing this, they can tell whether the tumor formed functional connections with neurons that innervate that brain area (See Figure 2).

Figure 2 : Glioma cells form functional glutamatergic synapses. Mice were transplanted with DIPG tumor cells into the hippocampus (shown in a schematic in (a)). Following some time to allow the tumor to integrate into the tissue, the researchers stimulated a pathway known as the “Schaffer collateral” pathway in the hippocampal CA1 region, which is a very well defined neural pathway in the brain. They recorded any responses to this stimulation in the tumor cells using a recording electrode. In (c-i) the authors demonstrate that stimulating this pathway causes depolarization (change in voltage) in the tumor cells. In (k,l,m) they demonstrate that when they use GCaMP recordings instead of electrophysiology, they can detect large changes in activity in tumor cells following stimulation! (Credit: Venkatesh et al., 2019).

Figure 2: Glioma cells form functional glutamatergic synapses. Mice were transplanted with DIPG tumor cells into the hippocampus (shown in a schematic in (a)). Following some time to allow the tumor to integrate into the tissue, the researchers stimulated a pathway known as the “Schaffer collateral” pathway in the hippocampal CA1 region, which is a very well defined neural pathway in the brain. They recorded any responses to this stimulation in the tumor cells using a recording electrode. In (c-i) the authors demonstrate that stimulating this pathway causes depolarization (change in voltage) in the tumor cells. In (k,l,m) they demonstrate that when they use GCaMP recordings instead of electrophysiology, they can detect large changes in activity in tumor cells following stimulation! (Credit: Venkatesh et al., 2019).

When they stimulated this pathway, they observed large depolarizations (change in voltage that causes action potentials to fire) in the connected tumor cells! This was repeated in several different DIPG models, and was blocked by administration of the drug NBQX, which blocks AMPA receptor signaling. AMPA receptors are fast ionotropic glutamate receptors that play a major role in neural communication throughout the brain. Additionally, when the switched out a recording electrode for a fluorescent indicator of activity (GCaMP6), they observed large increases in tumor fluorescence following stimulation of the Schaffer collateral pathway. This strongly indicates that tumors can form functional connections with neurons, and they talk to each other through classical glutamate AMPAR-mediated signaling! They went on to further characterize ionic communication between neurons and tumor cells using similar techniques as above.

The final and most important question for them to answer was: Does this matter? Does neuron-tumor communication actually influence the lethality or progression of brain cancer? To test this, they implanted tumors cells into the mouse brain (as above) and stimulated neurons in the surrounding area using optogenetics. Then, they analyzed the ‘proliferation index’ of tumor cells in that area, reasoning that if stimulating the neurons in the area promoted tumor growth, then mice that received stimulation would have faster growing tumors than mice that did not receive stimulation. Using the proliferation marker Ki67, they demonstrated that neural activity drives brain cancer progression (see Figure 3)!

Figure 3 : Neural activity drives glioma progression. (a,b,c,d) optogenetic stimulation of neurons in the tumor microenvironment promotes tumor growth/proliferation. (e,f,g,h) Over-expressing the AMPA receptor sub-unit GluA2 accelerates brain tumor lethality, and inhibiting GluA2 expression using a dominant negative approach suppresses brain tumor lethality! (i) shows us the big difference between mice with normal levels of GluA2 and those with a non-functioning dominant negative version. (j) Indeed, the tumor burden of mice with reduced AMPA receptor signaling (GluA2-DN-GFP) had a much lower tumor burden than their counterparts with normal levels. This effect was mirrored when AMPA receptor antagonists were used (e.g., Perampanel) instead of the transgenic approach. (Credit: Venkatesh et al., 2019).

Figure 3: Neural activity drives glioma progression. (a,b,c,d) optogenetic stimulation of neurons in the tumor microenvironment promotes tumor growth/proliferation. (e,f,g,h) Over-expressing the AMPA receptor sub-unit GluA2 accelerates brain tumor lethality, and inhibiting GluA2 expression using a dominant negative approach suppresses brain tumor lethality! (i) shows us the big difference between mice with normal levels of GluA2 and those with a non-functioning dominant negative version. (j) Indeed, the tumor burden of mice with reduced AMPA receptor signaling (GluA2-DN-GFP) had a much lower tumor burden than their counterparts with normal levels. This effect was mirrored when AMPA receptor antagonists were used (e.g., Perampanel) instead of the transgenic approach. (Credit: Venkatesh et al., 2019).

To provide further evidence that AMPA receptors are important for neural activity-induced tumor progression, the researchers over-expressed a subunit of the AMPA receptor GluA2 in tumor cells that were transplanted into mouse brains. Then mice were followed to see how long they survived. Mice that had high GluA2 expression (and presumably more AMPA signaling between neurons and tumors) died more rapidly from brain cancer than mice with normal levels of this protein. In the opposite experiment, they replaced GluA2 with a non-functioning version using a dominant-negative approach. This time, mice with non-functioning GluA2 survived significantly longer than their counterparts with normal levels. This suggests that neuron-to-tumor communication through AMPA receptors drives tumor progression and lethality!

The authors moved on to see if there was any evidence that this occurs in living human patients with brain cancer. Indeed, they demonstrated that neurons in brain areas with tumor infiltrations were ‘hyper-excitable’ (Figure 4). This suggests that enhanced brain-to-tumor signaling is a defining feature of glioma infiltrating healthy parts of the brain!

Figure 4 : Neurons are extra-excitable in the glioma-infiltrated human brain! (Credit: Venkatesh et al., 2019).

Figure 4: Neurons are extra-excitable in the glioma-infiltrated human brain! (Credit: Venkatesh et al., 2019).

This paper, along with other published in the same issue, form the foundation for what is to become a vibrant field linking cancer and neuroscience! There are a ton of unknowns here…which makes the area ripe for potentially life-saving discoveries! I had a lot of fun reading these crazy papers…but as this is a journal club, let me know what you think by leaving a comment below! Until next time…STAY CURIOUS! - JCB