Human Neurons Wire Into Rat Spinal Cords and Bolster Breathing

A week after a controlled blow crushed one side of a rat’s spinal cord high in the neck, a team from Gladstone Institutes and Drexel University injected a million human neurons into the cavity the injury had left.

Those neurons had been grown in a dish from stem cells and engineered to fire whenever blue light hit them. Two months on, with the animal anesthetized and the cord opened a second time, the researchers aimed a blue lamp at the transplant and asked whether anything downstream was still listening. Three minutes of light, pulsed once a second. The light went off, they waited. About five minutes later, the stretch of diaphragm wired to nerves below the injury was pulling harder than it had before.

Which was actually the point of the exercise. Not a tissue graft, not a scatter of stem cells left to become whatever the injury site suggested, but one defined cell type, picked because the spinal cord already leans on that type to route traffic across circuits.

Spinal cords are replete with interneurons, the short-haul cells that sit between other neurons and route traffic between them. One family, the V2a cells, keeps turning up in recovery research: excitatory relays that get pulled into breathing and walking circuits when the usual routes go down. Lyandysha Zholudeva at Gladstone Institutes in San Francisco and her colleagues worked out how to push human induced pluripotent stem cells into becoming that particular cell type, settling on a 15-day recipe tuned with a Wnt-pathway drug so the cells took on a neck-level identity rather than a brain one, and Gladstone president Deepak Srivastava, the study’s senior author, reckons it took “a year and a half of trial and error” to get right. The finished cells freeze into vials and thaw on the morning of surgery, which sounds like housekeeping and isn’t, because an off-the-shelf product is a far easier thing to carry into a clinical trial than a fresh batch grown to order.

Why test any of this on breathing? Because spinal cord injuries affect somewhere between 15 and 20 million people worldwide, on the World Health Organization’s count, neck injuries are among the commonest, and the phrenic circuit that drives the diaphragm is about as well mapped as a motor pathway gets.

A Virus Traced the Wiring Back to the Diaphragm

The injury sat at the third and fourth cervical vertebrae, and the dose was the most the cord would take (any more and the volume itself does damage). Histology at two months showed the transplants had survived the hostile injury site and filled 75 to 90 percent of the cavity, with no sign of donor cells wandering into the thoracic or lumbar cord in the regions examined. To find out whether any of it was actually plumbed in, the researchers put a pseudorabies tracer into the diaphragm, a virus that crosses only genuine synapses, and turned up labeled human cells in 15 of 18 animals.

A fair few caveats travel with that. Every rat was on cyclosporine A from three days before surgery through to the end, the cells went in a week after injury rather than months or years later, and when the team stimulated the transplant and looked for a diaphragm response, they got one in nine animals out of fourteen.

Ordinary Breathing Hid the Difference

Breathing room air, the two groups looked much the same, which is roughly what you’d expect, since a contusion on one side spares enough of the phrenic pool to cover a resting animal. So the researchers leaned on the system. Rats breathed 10% oxygen, and separately 7% carbon dioxide (hypoxia and hypercapnia, in the paper’s language), both of which force the diaphragm to work harder, while electrodes recorded from the dorsal hemidiaphragm, the portion served by motor neurons below the injury. Most vehicle-treated animals answered that demand by contracting less, a paradoxical fade the authors read as fatigue and the kind of respiratory insufficiency that can end in arrest. Among transplant recipients the numbers ran the other way: 73% increased dorsal diaphragm activity under carbon dioxide and 82% did so under low oxygen, against 33% of controls in both challenges.

Zholudeva puts the stakes as the gap between “an injured person who gets a cold and ends up back on a ventilator” and someone with enough reserve to ride it out. Extra capacity when demand spikes, in other words, rather than a different resting life.

Gladstone’s press release renders the result as three-quarters of treated rats passing the challenges without difficulty, and describes most untreated rats as showing signs of respiratory failure. Neither is quite what was measured: the electrodes recorded which way a muscle signal moved, and no animal in the paper is reported to have gone into failure or to have been scored on whether it sailed through.

Not Every Donor Neuron Joined In

Sequencing the transplants at two months turned up something the team had not gone looking for. Of the 37,358 human nuclei recovered, nearly all the cells the tracer had labeled sat in one small transcriptional cluster, marked out by adhesion and axon-guidance genes that in adult mice tag mature V2a neurons. Across the five sequenced animals (three that had responded and two that had not), the share of those integrated neurons correlated with how strongly the diaphragm answered a challenge, which is suggestive and, at that sample size, not a great deal more. The authors are careful about it, and careful too about a structural question their data cannot settle, since they cannot separate a direct brainstem synapse onto a donor cell from a longer route through several interneurons, though they argue a polysynaptic route would do fine for repair.

Bigger animals come next, and with them the harder question of whether cells laid into a cord scarred for years behave anything like cells laid into one injured last week. And then the circuit that people with neck injuries actually rank first, which is not breathing at all but the hand.

• For further reading see: What Is Neuroplasticity? The Science of How the Brain Rewires Itself

Frequently Asked Questions