Electroceuticals jolt into the clinic, sparking autoimmune opportunities

In 2016, GlaxoSmithKline and Alphabet life sciences company Verily staked a bold claim on the ‘electroceutical’ landscape. By launching Galvani Bioelectronics — a joint venture, backed by over US$700 million for 7 years — the partners set out to develop a new therapeutic modality that can harness the electrical signals that course through the body to treat various chronic diseases.

Now Galvani, led by Kristoffer Famm, has launched a landmark trial of its electroceutical platform. They are chasing SetPoint Medical, which is also testing an immune-modulating electroceutical device for rheumatoid arthritis. But whereas SetPoint’s device acts on the vagus nerve, which controls the functions of all internal organs, Galvani’s end-organ approach targets only the spleen, via the splenic nerve. For Famm, this selective strategy will be key to the safety of electroceuticals in rheumatoid arthritis, and to the future of these medicines in other indications. Data will have to bear that out.

Other big questions also hang over the field. How do different neural circuits play into disease biology? What is the best intervention point for these? How big a treatment effect is possible? With tools now in hand, Famm expects that answers are on the way.

“It has been a very heavy lift to build the device system,” says Famm. “Now we have a clinical grade end-organ system that can stimulate nerves. This is what makes the next 10 years exciting.”

How did you get involved in bioelectronic medicine development?

This started back 10 or 11 years ago, when GSK wanted to look for a new class of treatments to draw from, in addition to small and large molecules. I had the privilege of orchestrating this search, and we homed in on what we ended up calling bioelectronic medicines or electroceuticals. Our premise was that there's a lot of physiology and pathophysiology that we are interested in that has a neural control component. All the visceral organs central in many chronic diseases are innervated, and the nerves are there solely for control purposes.

Pharmaceutical and biotech firms are not utilizing that axis of control. But there are neuromodulation devices: deep brain stimulation for Parkinson disease and spinal cord stimulation for pain have existed for decades. We hadn't merged those two worlds. For GSK, that was an attractive opportunity, when taking the 10–20-year view. We made a decision to get in at the ground floor.

You outlined your electroceutical vision in Nature in 2013, and provided an R&D roadmap in Nature Reviews Drug Discovery in 2014. How is this going?

That’s right. We deeply believe that this is not something we can do alone; a number of disciplines need to come together, both in academia and industry.

When we set out that roadmap, our goal was to create a foundation to build not one, but multiple, bioelectronic medicines. High level, the fact that we are now bringing our first bioelectronic medicine into the clinic for a chronic disease shows that roadmap has delivered. But if you start going into the components, there are aspects of the roadmap that have not yet delivered.

We called for a focus on creating a visceral nerve atlas — a structural and functional map of the neural connectivity and control of visceral organs. There, progress has mostly come from the public sector. The NIH’s SPARC initiative has been the biggest contributor there. Maps are arising, but they are patchy.

But there are a number of organs where people have studied that sort of anatomical innervation, and, to a lesser degree, the neural signals that go to and from them. And there are definitely therapeutic opportunities emerging from that.

Next on the roadmap was miniaturization of interface technologies.

There has been partial progress here, and more to do. We have focused on building durable neural interfaces, but it has not been a quick process. I would say it's the most time-consuming part of the development effort, because you have to optimize something for chronic reliability and safety that can interface with the soft, moving, irregular anatomy of the human body.

If you look at what we’ve brought into the clinic for splenic nerve stimulation, we have this quite advanced cuff that sits around the splenic neurovascular bundle. You need to hug that bundle in such a way that it can move with every pulsation of the artery, and yet be close enough to the nerves to electrically stimulate them. And it needs to do this year in and year out. It took several iterations to get to a reliable and safe version of that, and the clock ticks on quite quickly. That's been a bottleneck, but it is now a key capability.

What about hopes for ‘closed loop’ interfaces that can record and analyse signalling, not just stimulate nerves?

We are not there, and we're not close.

The third piece of the roadmap was establishing therapeutic feasibility. Why rheumatoid arthritis, via the splenic nerve?

We — and others — have seen that you can quickly get to therapeutic proof of principle in animals. In rodent models, you can stimulate or block a nerve in a disease model, and get impressive treatment effects. But there is a significant translational journey as you scale up to human physiology.

It looks like the neural circuits are quite conserved. If a nerve controls something in mice, it is likely to control it up the species tree. But anatomical access becomes a challenge: we’ve had nerves in mice that look promising, but these are then spread out or hidden in a ligament in humans such that we couldn't translate our finding. In polycystic ovary syndrome, for example, we had beautiful data in rodents only to find that this nerve in humans was not possible to access.

We've had other translational challenges too. There are great preclinical data on electrically blocking nerve signals in the carotid sinus nerves in type 2 diabetes. But in humans you would need to block these nerves 24/7, and this technology was just too challenging — too power-intensive — to actually build. You get attrition in translation.

We knew early on that the splenic nerve was a promising target, and we mentioned it 9 years ago in our manifesto. It's also one of those nerves that translates from small animals, to large animals, to human anatomy, and we've been able to develop a surgical procedure and a device system to access this nerve. I'm convinced from what we've seen that this is not a one-off.

How does splenic stimulation modulate rheumatoid arthritis disease activity?

We know a fair bit about this, but there's much more to be done. A lot of it still needs to be published. But what I can describe is when we stimulate the splenic nerve, we get neurotransmitter release in the spleen. Mostly it is noradrenaline, but also some other neurotransmitters. We know that the neurotransmitters work on immune cells, primarily macrophages and monocytes. And what we see is a quite a broad shift in the properties of these cells — and the cells that they act on — as they go from pro-inflammatory to anti-inflammatory states.

For example, there is a broad downregulation of pro-inflammatory cytokines, and an upregulation of anti-inflammatory cytokines. There is production of pro-resolving mediators that lead to inflammation resolution. And we see a shift in how the impacted macrophages and monocytes affect B and T cells. Exactly the cascade, and the comprehensive map, is going to take years to chart. But the contour is that we can programme circulating immune cells, dialing in the degree of reactivity. We have this lever in the spleen.

Rheumatoid arthritis, our first indication, offers significant unmet need, good endpoints and good precedent for feasibility studies. And in animal models we see a very robust effect on disease, on the same order as seen with anti-TNF antibodies.

But this is not a rheumatoid-arthritis-specific intervention. It's an immunomodulation that is orthogonal to how we treat today in immune-mediated inflammatory disorders. It's not targeting just one cytokine. We stress ‘immune modulation’ as opposed to ‘immune-suppressing’ activity, because we shift the balance. I think it can stand alone as a new class of treatments, but can also be complementary with current drugs.

What are the safety concerns?

There are three classes of safety to watch. The highest risk, in our risk register, is that associated with the implantation procedure. Then we have the safety and reliability of the device systems, that we will explore in the clinic: does the device work reliably, does it have any ‘off-target’ effects on the body. And then there is the safety of the stimulation: do we see any inadvertent adverse effects of the immunomodulation, or consequences of stimulating the splenic nerve?

Like risk of infection?

That's definitely one bucket. Specifically for immunosuppressive effects, we are not that concerned though. What we're seeing preclinically is that we are not fully suppressing any cytokines. We’re getting a shift, reducing TNF levels, for example, by two thirds or so. But that's not what you see with monoclonal antibodies, which mop it all up.

How will ‘dose selection’ — or optimization of the stimulation protocol — work?

We’ve optimized stimulation parameters in large animal studies, thinking about amplitude, pulse width, frequency and pattern. It's quite a complex optimization landscape. We've sampled something that leads to good immunomodulatory effect, without impacting the blood flow or the contraction of the spleen, which are also controlled by the splenic nerve. There's a therapeutic window, and we’ll test that in clinical studies by stepping up to this target dose in patients.

This nerve bundle has some sensory fibres, so we step up to the point where we are activating the nerve, but we don't have an uncomfortable sensation. The clinical studies will show whether people will feel anything or not.

It is hard to predict how much further optimization will be needed after that. We don't know if we are talking about a 10% more potential treatment benefit from further optimization, or a multiplier effect. But future research will look into this.

You told Business Insider that approval could be 4–5 years away. Is that realistic?

It's ambitious, but it is feasible.

When it comes to the regulatory path, these are implantable medical devices. We can do feasibility studies — that's what we are doing now — and if we show safety and have a reason to believe in efficacy, we can go straight into a pivotal study. And just one pivotal study is needed. In rheumatoid arthritis, that could be a single 200–300-patient study. It then comes back to the question of dosing. If we get a good treatment effect in our feasibility study, 4–5 years is in play. But we’ll need to see.

What else is in your pipeline?

There are three ways we want to build out from this foothold.

First, we have our splenic chronic device system that we are now testing in rheumatoid arthritis. There are a number of indications that we build that out to. Pick your inflammatory disorder, and look at the breadth of the anti-inflammatory antibodies.

Then, we have adjacent uses for splenic stimulation. We can acutely stimulate the spleen during some surgeries, and that might improve surgical recovery. There are other potential acute indications too.

And beyond the spleen, our platform system can provide near-organ stimulation to a number of organs. We have not explicitly disclosed what we are working on here, but a good example is pancreatic lymph nodes for type 1 diabetes. We have published together with academic collaborators some very impressive preclinical data, showing that by stimulating the nerve to the pancreatic lymph nodes we can shift the immune cells that drive the self attack on β-cells, protecting these cells and restoring insulin control in mice. This is as black and white an early proof of concept as you can get.

Our rheumatoid arthritis trial is a big milestone for us, but this isn't just about rheumatoid arthritis. This is the first near-organ precision neuromodulation for a chronic disease, and the next couple of years will be the gatekeeper to all the big potential we’ve been discussing.