We've updated our Privacy Policy to make it clearer how we use your personal data. We use cookies to provide you with a better experience. You can read our Cookie Policy here.


Understanding Cell-to-Cell Communication in Cancer

Four attached pink cancer cells.
Credit: iStock
Listen with
Register for free to listen to this article
Thank you. Listen to this article using the player above.

Want to listen to this article for FREE?

Complete the form below to unlock access to ALL audio articles.

Read time: 7 minutes

For many years, tumors have been considered as masses of rogue cells growing uncontrollably, brought about by aberrant cell signaling pathways. But as we learn more about the disease, it appears that the communication pathways that cancer cells use between themselves, and with cells in their environment, are much more intelligent. Here, we explore how by targeting the language of cancer cell communication, we can find new ways to stop tumor growth and prevent drug resistance.

Cancer cells are a community with their own language


“There’s probably two broad themes that have emerged in recent years that have led us to think differently about cancer,” says Professor Chris Bakal at The Institute of Cancer Research, London. “The first is the idea that cancer cells are not just communicating with cells in their environment and between themselves, but are communicating with normal cells in the body, such as cancer-associated fibroblasts or neurons. The second idea is that cancer cells aren’t just growing uncontrollably and oblivious to what’s around them, but are talking to each other in a more strategic fashion, coordinating actions among themselves as a community.”

Bakal’s lab focuses on how cancer cells change their shape and how shape affects their function. They specialize in using imaging technologies and developing machine learning methods to quantify cell shape in different settings, whether it’s the lab or the clinic. In a recent study,1 Bakal collaborates with Dr. Amanda Foust and Professor Mustafa Djamgoz at Imperial College London. Foust had developed a method for visualizing voltage fluctuation in cells, and the Bakal lab found a way of processing this complex information so it could be quantified and compared. Together, they observed that breast cancer cells were making different fluctuations in voltage compared to healthy cells, and these were not random – they were very predictable and looked in many ways like a signal.

Bakal and Foust are now exploring this observation further with melanoma cells. “A lot of the cells we work on change their shape to look a lot like neurons, developing synapses and other functions that neurons have,” he explains. “Melanoma cells come from neural crest progenitors so it’s not that much of a stretch to imagine that they can adapt in this way. Neurons have this incredible property, changing their shape to make long extensions that ultimately form a network, so we initially thought the changes in melanoma cell shape were to enable them to invade surrounding tissue. But, although we still think this is important, we’re also starting to consider that this change in shape might be a way for cancer cells to propagate information like neurons.”

Unlike the relatively slow, soluble signal produced by a ligand being released from a cell towards a receptor over a short distance, these electrical pulses offer a rapid and distant transmission system, perhaps like the quorum sensing seen in bacteria and electrical signals used by fungi. “It’s not inconceivable that there are similar quorum sensing systems in communities of cancer cells, allowing them to tell each other where nutrients are and coordinate their metabolism,” says Bakal. “If you look across many different cancer cell types, there’s a limited number of distinct electrical patterns. Once you start to get predictable classes of signal, you can begin to see the building blocks of a code or language, where different patterns might be received differently by different cells. If we can stop these signals, it would dysregulate the community, rather like having an army without its general. The challenge is, we might understand the language, but we don’t yet know what the cellular response is.”

Brain tumor cells communicating with healthy neurons


In the Hervey-Jumper lab at University of California, San Francisco, Dr. Saritha Krishna has been exploring how glioblastoma cells influence neural activity in distal sites within the brain.2 “We know from preclinical studies that there’s a strong positive feedback loop between neurons and glioblastoma cells in the microenvironment,” says Krishna. “We wanted to ask what this feedback loop means for patients in terms of their cognitive function.”

The Hervey-Jumper team recruited volunteers awaiting surgery whose tumors had infiltrated the brain region controlling speech and used intra-operative mapping during surgery to capture brain activity during different language tasks. In addition to finding neural activity in the specific brain areas responsible for language processing, there was activation in broader regions of the brain unrelated to these cognitive functions.

“This told us that the gliomas were remodeling wider neural circuits, so we set out to understand why,” said Krishna. They collected tumor biopsies from regions with high functional connectivity (HFC) to the rest of the brain and regions with low functional connectivity (LFC) and ran a series of experiments looking at gene and protein expression. The HFC regions were enriched for genes that form synapses with other cells, including thrombospondin 1 (TSP1).

“When we co-cultured these TSP-expressing HFC cells with neurons, they started to rapidly proliferate and, in a 3D culture system, the tumor cells integrated extensively in the neuron organoids. By contrast, the LFC cells seemed to care less about being near neurons and showed minimal neuron integration,” Krishna added. When TSP1 was exogenously added to the LFC cells, they started to behave like the HFC glioma cells, suggesting that TSP1 is critical for this neuron–glioma interaction.

Having established that glioblastoma cells had these neuronotrophic properties, Krishna and her colleagues looked at whether this resulted in the increased hyperactivity they saw in the brains of glioblastoma patients: “Again, we found that in the presence of the HFC cells, the neurons were hyperexcitable. We think that these HFC glioblastoma cells are literally integrating into neural circuits.”

Based on these findings, the team has already tested a cheap and readily available anti-convulsant drug, gabapentin, and was able to successfully disrupt the glioblastoma–neuronal communication in preclinical models. “We are still in the early stages, but this gives us hope that this drug, which is already used in brain tumor patients to control seizures, might also limit tumor expansion. In the future, it might also be possible to use neuromodulation techniques to target the hyperexcitable neurons in the tumor microenvironment and spare the healthy neurons. I’m hopeful that this finding could be beneficial to patients.”

Drug-resistant tumor cells communicate directly with immune cells


It’s not only healthy surrounding tissues that cancer cells are influencing with communication tactics. There is growing evidence that tumors can gain the ability to signal directly to non-cancer cells within the tumor microenvironment. Dr. Jason Griffiths, an associate professor at Department of Medical Oncology and Therapeutics Research, City of Hope National Medical Center and the University of Utah, Salt Lake City, studies the ecology of cancer and the phenotypic evolution that we see within tumors. His lab uses single-cell molecular data to look within the tumors of patients enrolled on clinical trials before, during and after treatment. “We know that as cancer cells evolve and acquire resistance to therapy, one way in which they do this involves altering communication with non-malignant cells in the tumor microenvironment (TME),” said Griffiths. “However, the specific interactions between malignant and non-malignant cells that cause this drug resistance remain widely unknown.”

In a recent study they profiled more than 400,000 single cells from serial biopsies of estrogen receptor-positive breast cancer tumors in patients being treated with either endocrine therapy alone or in combination with ribociclib, a cell cycle blocking drug.3 They used single-cell RNA sequencing to determine the gene expression profile of each cell and an immune cell classifier to understand the immune composition of the different tumor biopsy samples.

Once they had identified the cell types, they used an extended version of the expression profiling method to measure cell interactions. “We looked at ligand and receptor gene expression in each cell, and then applied a statistical analysis method to tell us how much each of these different subpopulations of cells are signaling to one another. This produced an overall network map of how strongly these cell types are communicating,” Griffiths explained. Next, they compared the networks of communication between resistant and responsive tumor types. “What we were able to see from this network analysis is that the resistant tumor cells lacked normal communication with immune cells but were sending lots of immunosuppressive signals to macrophages,” said Griffiths. “We know cancer cells can alter the phenotypes of macrophages and myeloid cells, and there's a lot of talk about the role of cancer-associated fibroblasts and endothelial cells that can support the growth of the cancer by producing networks of capillaries that provides resources. What was unclear before doing this study was which of these cell types play these roles.”


Griffiths is now looking more closely into the mechanism of immune suppression in tumors treated with ribociclib: “We’ve been using in vitro systems where we can co-culture tumor cells and macrophages to understand more about the social communication aspect, and we’re moving into spatially resolved analyses, where we try to preserve the spatial orientation of the cells and look at how those cell interactions play out in a tissue. But the most important thing we discovered is that tumors that don’t respond to ribociclib lack activated T cells, which suggests ribociclib has a dual role: stopping proliferation by blocking the cell cycle, but also having a negative effect on anti-cancer T cells. This tells us that patients who have not responded well to ribociclib might benefit from additional therapy to reactivate their immune system.”


1.     Quicke P, Sun Y, Arias-Garcia M, et al. Voltage imaging reveals the dynamic electrical signatures of human breast cancer cells. Commun Biol. 2022;5(1):1178.  doi: 10.1038/s42003-022-04077-2

2.    Krishna S, Choudhury A, Keough MB, et al. Glioblastoma remodelling of human neural circuits decreases survival. Nature. 2023;617(7961):599-607. doi: 10.1038/s41586-023-06036-1

3.       Griffiths JI, Cosgrove PA, Castaneda EM, et al. Cancer cells communicate with macrophages to prevent T cell activation during development of cell cycle therapy resistance. bioRxiv 2022.09.14.507931; doi: 10.1101/2022.09.14.507931. This article is a preprint and has not been certified by peer review.