Recovery of function after complete spinal cord transection
Our clinical target is contusion — the crush and compression injury behind almost every human case. We began from the harder end instead: a cord cut clean through, where nothing recovers on its own and there is nothing to mistake for spontaneous improvement. Three species, three peer-reviewed studies. Every treated group recovered function. No control animal recovered anywhere in the programme.
10
Human cadaver operations developing the contusion protocol
430
Animals across the programme to date
3
Peer-reviewed publications
0
Control animals that recovered, in any study
Model
Design and outcome
Record
Published
20 Rabbits10 treated / 10 control
T9 · 30 daysMovement, sensation and sphincter control partially restored in the treated group. Controls stayed paraplegic and anaesthetic — and 4 of 10 controls died against none of the treated.
T7–T9 · 60 daysFirst large-animal test and the study where the protocol took its current shape — vertebral fixation, cord cooling and systemic infusion added to the gel. All treated animals walked on four limbs by day 60; DiI tracing crossed the lesion.
T8 · 60 daysThe pilot repeated under a stricter design. Sensation by day 2, bladder control by day 5, walking on four limbs in all three treated animals by day 60. FluoroGold tracing and NF-200 histology confirm axons crossing; controls show cysts and glial scar. p < 0.001 on all three scales.
OngoingRats are where the compound itself gets decided: conjugate variants compared head to head to find the best-performing formulation, at group sizes that are impossible in swine. Also where the move to contusion is being worked out.
Preprint in progress
10 CadaversHuman anatomy
OngoingTen operations on human cadavers developing the surgical protocol for contusion injury — resection to healthy tissue, shortening and fixation, worked out on the anatomy the procedure will actually meet.
Ongoing
Pipeline
60 SwinePowered · contusion
NextThe winning conjugate carried back into the large-animal model at a group size that settles the small-n objection — and, for the first time, against contusion rather than a clean cut.
—
02 — Evidence, model by model
Increasingly better results, three times over — and the compound is now being readied for real surgical application.
Each study is a harder test than the one before it. Rabbits showed the effect exists. The swine pilot showed it holds in an anatomy close to ours and gave the protocol its shape. The blinded repeat added the tissue evidence to match the behaviour. Rats are now deciding the compound itself.
Rabbits — the effect exists
n = 20 · 10 / 10 · T9 · 30 days Surg Neurol Int 2023;14:423
Twenty rabbits, complete transection at T9, evenly split. The treated group received an intraoperative injection of photo-crosslinked chitosan mixed with PEG. By day 30 movement, sensation and sphincter control had partially returned. Controls stayed paraplegic and anaesthetic throughout — and four of the ten controls died, against none in the treated group. First signs of recovery appeared within 24–48 hours — the same window that recurs in both swine studies.
Chart slotmBBB scores over 30 days, treated vs controlOpen article ↗
Modified BBB scaleTreated animals separate from control within the first week and keep climbing.
Image slotVideo storyboard, treated rabbit at day 30Open article ↗
Day 30A short jump and increased hind limb range of motion in a treated animal.
Established — fusogenic repair produces functional recovery after complete transection
Survival signal — 0 of 10 treated died vs 4 of 10 controls
Formulation — a physical mixture, not yet the covalent conjugate
Open — no tracing, no large-animal anatomy, 30 days only
Swine, pilot — the effect scales
n = 5 · 3 / 2 · T7–T9 · 60 days Surg Neurol Int 2023;14:424
Five Hungarian Mangalica pigs, complete thoracic transection. This is where the protocol took its current shape: local Neuro-PEG gel, intravenous PEG intra- and postoperatively, local cord cooling for neuroprotection, and vertebral stabilisation — used here for the first time in any fusogen study — to hold the stumps apposed. Scored on ILMS, PTIBS and modified BBB, with sensory and sphincter assessment. DiI tracing showed dye crossing the lesion.
Chart slotILMS, PTIBS and mBBB across 60 daysOpen article ↗
Three independent scalesTreated animals rise on all three; controls stay at floor.
Image slotDiI tracing through the transection siteOpen article ↗
In vivo tracingLabelled fibres running through the repaired gap.
Established — recovery in an anatomy comparable to human
Protocol — fixation, cooling and systemic infusion added
Endpoint — independent ambulation on four limbs at 60 days
Open — small groups, unblinded, no vehicle arm
Swine, confirmatory — the effect repeats
n = 5 · 3 / 2 · T8 · 60 days · blinded PLOS One 2026;21(6):e0349579
The pilot run again under a stricter design — randomised allocation and blinded neurological scoring. Sensation returned on day 2, bladder control on day 5, standing with support by day 18; by day 60 all three treated animals walked on four limbs with full pelvic control. Controls never changed. Differences on all three scales reached p < 0.001, with inter-scale correlation of r = 0.95–1.0. Retrograde FluoroGold tracing and NF-200 immunofluorescence show axons physically crossing the lesion; controls show cysts, retraction bulbs and glial scar.
Chart slot · Fig 3Neurological scores over 60 days, three scalesDownload figure ↗
mBBB · PTIBS · ILMSTreated curves climb steadily; control curves plateau at the floor for the full 60 days.
Image slot · Fig 7FluoroGold retrograde tracing across the lesionDownload figure ↗
In vivo tracingDye injected 5 cm below the cut travels up living axons and crosses it. Continuity is anatomical, not inferred.
Established — result holds under blinding, with morphological confirmation
Statistics — p < 0.001 on all three scales; r = 0.95–1.0 between them
Tissue — axonal bridges present in treated, absent in control
Open — no vehicle arm, no electrophysiology, 60 days
Rats — choosing the compound
n = 400 · complete transection Preprint in progress
Every study above shares one weakness: groups too small to compare anything. Rats are where that is solved. Four hundred animals let conjugate variants be screened against each other rather than against nothing — which is how we settle the best way to make Neuro-PEG, instead of assuming the first formulation that worked is the right one. The same cohort is where the move from clean transection to contusion is being worked out. Arms, endpoints and readout date to be added.
Decides — which conjugate goes into the 60-swine study
Answers — statistical power, the standing objection to the swine data
Opens — the contusion model, the actual clinical target
Status — running; preprint in preparation
Theseus · preclinical stage, not an offer of treatmentandrei@dowell.bio
