How Tumours Use the Nervous System: Cancer Neuroscience, Drug Repurposing, and the Evidence Big Pharma Ignores

What nerves and tumours are really doing

The easiest way to see how far the old picture has collapsed is to start at the most disturbing end. In several brain cancers, tumour cells do not just sit near neurons. They behave like them. They receive electrical input, fire currents and use that activity as fuel for growth.

In work that has now become canonical, researchers took slices of human glioma, placed them in culture with healthy brain tissue and stimulated the surrounding neurons. The tumour responded like part of the circuit. Stimulation made the cancer cells electrically active. That activity correlated with more aggressive behaviour and poorer survival in patients whose tumours showed this kind of coupling.

Subsequent studies went further. They revealed that these cancers form conventional glutamatergic synapses with neurons. Under a microscope, the contacts look like ordinary synapses. At the molecular level they use the same receptors a neurologist would expect to find at a normal brain connection. It is the partner that is different. One side of the synapse is now a cancer cell.

Outside the brain the story bends again. In models of stomach cancer, tumours attract sensory neurons that ordinarily convey pain and visceral signals. Those nerves then release peptides that drive tumour growth and help set up distant spread. In breast cancer models, nerves transfer mitochondria directly into malignant cells, increasing their metabolic capacity and helping them survive the stress of metastasis.

At this point it is no longer credible to treat nerves as scenery. Tumours are behaving as counterfeit organs that wire themselves into the host circuitry. They do not just coexist with the nervous system. They annex it.

The overlooked histology that saw this coming

The irony is that the field did not need synapse recordings to see that something was wrong with the old picture. The evidence was already in the wax blocks and glass slides of routine pathology.

For generations pathologists have documented a pattern called perineural invasion. Under the microscope, cancer cells cluster around nerves, wrap them and often travel along the nerve sheath into adjacent tissue. Clinicians learned that when this pattern appears, outcomes worsen. Yet the finding was treated as a marker rather than a mechanism, as if the tumour simply enjoyed the view along the nerve rather than deriving anything from contact.

In the late nineteen nineties a small number of investigators asked the obvious question. What happens if you put tumour cells and neurons together in a dish and watch them interact? The answer was unambiguous. Cancer cells and neurons grew towards each other, established structured contacts and proliferated faster once the connection formed. Later work in human prostate tissue showed that malignant glands contained far more nerves than normal tissue, and that higher nerve density associated with worse prognosis.

Population scale data added another clue. Men with spinal cord injuries that interrupt neural signals from the pelvis and lower body have roughly half the risk of prostate cancer compared with men of similar age without such injuries. Something about disrupted wiring appears to protect against the disease.

These are not delicate signals at the edge of detection. They are coarse patterns that persisted across experiments and cohorts. Yet they sat on the margins of the story because they did not fit the gene centred narrative that dominated cancer biology and cancer funding for decades.

Cutting the wires in animals

Once you accept that nerves are part of the machinery, the next step is brutal but simple. What happens if you cut them?

In prostate cancer models, investigators removed autonomic nerves that serve the gland or destroyed them chemically. Tumours that would normally appear and progress quite readily now struggled to establish themselves. The denervated prostate became hostile soil.

In gastric cancer models, the same principle played out in a different organ. By severing branches of the vagus nerve or injecting botulinum toxin into the stomach wall to silence local nerve terminals, researchers could suppress tumour initiation and growth. When they combined this with chemotherapy, the denervated regions of stomach responded far better than the control side.

Other work has teased apart the branches of the autonomic and sensory systems. Sympathetic fibres that drive the classical stress response tend to accelerate cancer growth and spread in organs such as the breast, ovary and prostate. Parasympathetic fibres can either restrain or promote tumours depending on context and location. Sensory nerves that carry pain can be co opted to provide growth signals in stomach and other cancers.

The message from the animal work is not that all nerves are bad. It is that the nervous system is a major regulator of tumour behaviour and that its influence is specific to organ, fibre type and timing. That complexity is a scientific challenge. It is not an excuse for pretending the system is irrelevant.

When cheap drugs hit expensive questions

Cutting nerves in patients is crude and often impossible. For most clinicians the more realistic idea is to dampen their signals chemically. Here the most interesting candidates are not futuristic molecules from discovery pipelines. They are drugs that cardiologists and neurologists have used for years.

Beta blockers are among the most obvious examples. They blunt the effects of adrenaline and related stress mediators on the heart and blood vessels. If sympathetic signalling helps tumours grow and spread, then beta blockade should in theory weaken that effect.

Retrospective analyses of health records have repeatedly reported that patients who happened to be on beta blockers for heart disease at the time of a cancer diagnosis often fare somewhat better than similar patients who were not. In several breast cancer cohorts, women already taking agents such as propranolol or carvedilol showed lower rates of recurrence or death over the following years, although the findings differ between drugs and studies.

Prospective trials, though small, suggest that the effect is not just a quirk of record keeping. In one randomised study, women with newly diagnosed breast cancer received either a week of propranolol before surgery or placebo. Tumour samples removed at operation from the propranolol group looked less aggressive and more heavily infiltrated by immune cells. In metastatic melanoma, a phase one trial that combined propranolol with an immune checkpoint inhibitor reported response rates markedly higher than expected from immunotherapy alone.

In the brain, the synapse story naturally leads to drugs that block the receptors those synapses use. Perampanel, an anti epileptic drug that antagonises AMPA type glutamate receptors, has shown the ability to reduce neuron to glioma electrical coupling and slow tumour growth in mouse models. Early trials are now exploring whether this translates into a survival benefit in people with glioblastoma.

These are not miracle cures. They are plausible adjuncts, grounded in mechanism and supported by early data. The most striking feature is not that they may work, but that they are so cheap and familiar. Propranolol and similar agents are off patent. Perampanel is hardly a speculative molecule. Botulinum toxin is already ubiquitous in clinical practice. None of them requires a discovery platform or a new factory.

Why the system stalled

This is where biology collides with economics. Drug development in oncology has been built around a simple pattern. A company discovers a new target, designs a proprietary molecule, runs expensive trials and then recoups those costs through years of high priced monopoly sales. Everyone involved, from investors to regulators to hospital payers, knows how that story works.

Repurposed drugs break the story. If a cheap generic agent turns out to cut recurrence risk when given around surgery or to double the benefit of an immunotherapy, no single company can own the result. The intellectual property resides in historical chemistry rather than in a new structure. Any firm that spends serious money to prove the benefit would mainly be funding competitors who are free to sell the same molecule under a different label.

That is why much of the research in cancer neuroscience has been driven by academic groups working with modest grants. They can show that neural blockade changes tumour biology. They can run early phase safety studies and produce promising signals in small cohorts. What they cannot easily do is the large, long and expensive trials that guideline bodies and insurers demand before changing standard care.

The result is a kind of suspended animation. The science is mature enough to fill review articles and win major prizes. The clinical signals are strong enough to justify serious investigation. Yet the trials that would settle the question are repeatedly delayed, downsized or never funded at all, because they do not fit the prevailing commercial template.

What a rational system would do now

If you strip away the commercial constraints and ask what a rational public health system would do with the current state of cancer neuroscience, the answer looks quite different from actual practice.

You would not hand out beta blockers and anti epileptics to every person with a malignancy. The evidence does not justify that. You would instead treat neural modulation as a strategic frontier and fund trials at a scale only public or philanthropic money can sustain.

That would mean large randomised studies of perioperative beta blockade in cancers where stress pathways are known to matter, with recurrence and survival as primary outcomes. It would mean glioma trials that enrol only those patients whose tumours show electrical coupling on preoperative recordings, then test perampanel against hard endpoints. It would mean deliberate combinations of neural interventions with immunotherapies and chemotherapies rather than leaving them to chance in retrospective analyses.

It would also mean integrating nerve related metrics into staging and risk models, instead of treating them as curiosities. Nerve density, perineural invasion scores, markers of synaptic wiring and circulating stress mediators would sit alongside tumour size, nodal involvement and mutation status.

Above all, a rational system would accept that some of the most powerful additions to modern cancer treatment may never be the basis of blockbuster products. They will be public goods, supported because they save lives and reduce suffering, not because they support a particular balance sheet.

The moral of the story

There is a certain horror in watching a tumour reach out to a nerve, form a synapse, steal a mitochondrion and ride an electrical impulse. It looks like a hostile intelligence. It is easy to dwell on that image and forget what it says about us.

We have uncovered a layer of biology that helps explain why some tumours spread early, why some patients relapse after seemingly successful surgery and why chronic stress appears to correlate with worse outcomes. We have also identified drug classes that can plausibly interrupt those circuits, at prices health systems can bear.

The fact that progress is now constrained less by a lack of knowledge than by a lack of patentable molecules is not a scientific problem. It is a political choice. Cancer neuroscience has revealed how closely malignant cells can entwine themselves with the nervous system. It has also revealed how closely our treatment options are entwined with a particular way of making money from disease.

The nerves inside tumours have been helping cancer for more than a century. The question now is whether we are prepared to stop helping them.