BIONEWS
Lyon researchers used 3D printed parts and a perfusion system (a fluid pumping culture chamber) to grow an 8 cm² piece of lab grown tissue for 21 days. Real flow paths inside the tissue were messier than the design.
A team from Université Claude Bernard Lyon 1 showed at the International Society for Cell & Gene Therapy (ISCT) 2026 meeting in Dublin last week that a single 8-cm² piece of lab-grown tissue can be perfused continuously for 21 days. The harder engineering work is not the print itself but the pumps, tubing, and MRI monitoring around it that turn a fragile prototype into a reproducible research tool.
That distinction is what doctoral student Elliot Cowles brought to Dublin. The platform he described, summarized in a GEN News writeup of the ISCT presentation and detailed in Cytotherapy, combines custom 3D-printed components with a Sartorius Ambr250 perfusion bioreactor. The custom parts were validated as autoclavable and leak-proof, and the system pumped culture medium through the tissue at 1.5 mL/min while the bioreactor regulated dissolved oxygen, temperature, and pH. Growing a piece of tissue larger than a fingernail in a controlled way has been the field's bottleneck: a bioprinted tissue is a fragile object made mostly of cells and scaffold, and without a way to perfuse it the way blood perfuses a real organ, and to watch what is happening inside, you cannot scale production, repeat an experiment, or hand the system to a non-specialist lab.
The first attempt at running the platform did not behave the way the design said it would. The team used high-resolution magnetic resonance imaging to reconstruct the tissue's internal 3D structure and to track where fluid actually flowed, non-destructively, over the run. The result: the fluid took pathways inside the tissue that diverged from the computer-aided design predictions. The print is a maze, and the fluid finds its own route through it.
The same monitoring work picked up a second surprise. Oxygen readings inside the tissue chamber ran consistently below the level set in the regulation vessel, even though the controller thought the system was at setpoint. The team traced the gap to oxygen ingress through the flexible tubing connecting the two vessels: a chemical leak in a piece of plumbing, not a sensor or controller fault. Closed-loop control depends on the closed loop being as closed as the model assumes.
The platform's near-term job is to make the kind of surprise it just measured reproducible enough to act on. Next steps include better quantitative flow measurement and the development of computational fluid dynamics models aimed at the local microenvironments inside the tissue. The related Lyon preprint lays out multiple monitoring modalities for the same goal. Three mesenchymal-stem-cell-based tissues were successfully perfused with minimal culture medium during the run, the closest thing to a throughput milestone the team is willing to claim, and the path to larger and more reproducible maturation is the one the field will be watching.
For drug testing and regenerative-medicine research, a piece of lab-grown tissue that can be perfused and monitored for three weeks, and whose internal fluid paths and oxygen levels can be measured honestly, is a tool that can be used in another lab. That is the test of whether this platform becomes infrastructure for the field rather than a one-off demonstration. The 3d.FAB platform at Lyon 1 and the joint Sartorius lab are the institutional pieces making that hand-off possible.
The interesting milestone from Dublin is therefore not the 21-day tissue. It is the platform's willingness to publish its own surprises. Fluid paths that do not match the CAD file, and oxygen that seeps in through flexible tubing, are the kind of finding a closed-loop system is supposed to surface. Whether the field uses that information to converge on reproducible, larger lab-grown tissue is the open question, and the one the next preprint, not the next press release, will answer.