Hemoglobin acts like a rusting agent in the brain
Among the most vulnerable lives on earth — infants born before the 28th week of pregnancy — a silent catastrophe unfolds when fragile blood vessels rupture inside the developing brain, flooding neural tissue with toxic hemoglobin and derailing the architecture of a mind still forming. Researchers across four Swedish institutions have now built a three-dimensional human brain tissue model that recreates this injury at the cellular level, allowing science to watch the damage happen rather than only measure its aftermath. With no established treatment currently existing for preterm cerebral hemorrhage, this model represents a rare opening: a reproducible, human-cell-based window into a wound that has, until now, resisted understanding.
- Cerebral hemorrhage strikes up to one in five extremely preterm infants, leaving many with cerebral palsy, cognitive impairment, or no life at all — and medicine has had no targeted treatment to offer them.
- The injury's cruelty lies in its chemistry: hemoglobin escaping ruptured vessels acts as a cellular poison, oxidizing the very neural stem cells responsible for building the infant brain.
- Previous research was trapped between two imperfect tools — animal models that don't fully mirror human biology, and patient fluid samples that reveal outcomes but not causes.
- A team spanning Malmö University, Lund University, Karolinska Institutet, and KTH engineered a 3D model of the subventricular zone using human cells, making the injury reproducible and observable in real time.
- When the model's findings were checked against samples from actual preterm infants, they aligned — confirming this is not just a laboratory construct but a clinically relevant map of real harm.
- The path forward is now visible: test interventions inside this model, isolate the molecular triggers, and build toward treatments that could protect developing neural stem cells before damage becomes permanent.
Every year, roughly 15 million babies are born before their bodies are ready. For the most fragile among them — those arriving before week 28 — delicate blood vessels in the developing brain can rupture without warning, spilling blood into spaces where cerebrospinal fluid should flow. The damage that follows is not only physical. Hemoglobin, the oxygen-carrying protein in red blood cells, becomes a powerful oxidant when it leaks into brain tissue, poisoning the neural stem cells responsible for generating new neurons. Up to one in five extremely preterm infants experience this kind of cerebral hemorrhage; in the worst cases, it results in cerebral palsy, severe cognitive and motor impairment, or death.
The injury strikes in a region called the subventricular zone — a thin layer lining the brain's fluid-filled cavities where immature blood vessels and neural stem cells sit in close, dangerous proximity. Understanding exactly how hemorrhage translates into lasting neurological harm has long been constrained by the tools available: animal models that imperfectly reflect human biology, and fluid samples from sick infants that describe outcomes without explaining causes.
Magnus Gram of Malmö University, working with collaborators at Lund University, Karolinska Institutet, and KTH Royal Institute of Technology, took a different path. The team engineered a three-dimensional tissue model using human cells, reconstructing the subventricular zone with enough architectural fidelity to observe neural stem cells responding to hemorrhage's toxic byproducts in real time. The model is reproducible — researchers can run identical experiments repeatedly, isolating single variables to distinguish cause from effect — and when its results were compared against samples from actual preterm infants, they aligned.
There is currently no established treatment for preterm cerebral hemorrhage. Care remains supportive: keep the child alive, manage complications, and hope the brain adapts. This model changes the terms of that equation. By revealing the precise molecular sequence through which a hemorrhage becomes a lifetime of disability, it creates a platform for testing interventions before they ever reach a clinical trial — and opens, for the first time, a credible path toward protecting the developing brain at its most vulnerable moment.
Every year, roughly 15 million babies are born before their bodies are ready. For some of the most fragile among them—those arriving before week 28 of pregnancy—a particular danger lurks in the developing brain: blood vessels so delicate they can rupture without warning, spilling blood into the spaces where cerebrospinal fluid should flow. When this happens, the damage cascades through neural tissue still learning how to be a brain. Now, researchers working across Swedish universities have built a three-dimensional model of human brain tissue that allows them to watch this injury unfold in real time, mapping the molecular mechanisms that turn a hemorrhage into lasting neurological harm.
Cerebral hemorrhage strikes up to one in five extremely preterm infants. In the worst cases, it leaves children with cerebral palsy, cognitive impairment, motor difficulties, or death. The injury happens in a region called the subventricular zone, a thin area lining the brain's fluid-filled cavities where fragile, immature blood vessels sit alongside neural stem cells—the cells responsible for generating new neurons throughout development. When those vessels break, blood floods into the ventricles. What makes this catastrophic is not just the physical trauma but the chemistry: hemoglobin, the protein that carries oxygen in red blood cells, is a powerful oxidant. When it leaks into brain tissue where it doesn't belong, it acts like a rusting agent, poisoning the very cells that should be building the child's neural architecture.
Until now, studying this process has meant either analyzing fluid samples from sick infants or inducing hemorrhages in laboratory animals—both approaches carrying significant limitations. Animal models don't perfectly mirror human biology. Fluid samples tell researchers what happened, not why. Magnus Gram, an associate professor of biomedicine at Malmö University, and his collaborators from Lund University, Karolinska Institutet, and KTH Royal Institute of Technology took a different approach. They engineered a three-dimensional tissue model using human cells, recreating the subventricular zone's architecture with enough fidelity to watch neural stem cells respond to the toxic byproducts of hemorrhage.
The advantage is substantial. Unlike animal models, this is human tissue. Unlike patient samples, it is reproducible—researchers can run the same experiment dozens of times, varying single factors to isolate cause from effect. They can introduce hemoglobin and watch what happens to the neural stem cells. They can test potential interventions before they ever reach a clinical trial. When the team compared their findings to samples from actual preterm infants, the results aligned. The model was not just clever; it was relevant.
Gram describes his focus as understanding what happens when hemoglobin ends up in the wrong place in the body. That question, pursued through this new model, opens a door that has been closed. There is currently no established treatment for preterm infants who suffer cerebral hemorrhage. The standard of care is supportive—keeping the child alive, managing complications, hoping the brain can compensate. This research suggests a different future: one where understanding the precise molecular injuries might lead to interventions that protect neural stem cells, preserve their capacity to generate new neurons, and prevent the cascade of damage that turns a hemorrhage into a lifetime of disability. The model is a tool. What it reveals could be transformative.
Notable Quotes
We have managed to create a model where we can look at how these injuries develop following cerebral hemorrhage and thus map the mechanisms and molecules that influence the process.— Magnus Gram, associate professor of biomedicine at Malmö University
The fact that we are seeing relevant responses in both simulated conditions and patient samples is very important, as there is currently no established treatment for these patients.— Anna Herland, professor at the AIMES research centre at KTH and Karolinska Institutet
The Hearth Conversation Another angle on the story
Why does this particular injury happen so early in life? What makes preterm infants uniquely vulnerable?
The brain is still under construction. The blood vessels in the subventricular zone—where new neurons are being made—haven't matured yet. They're fragile, almost permeable. In a full-term infant, those vessels have strengthened. In a preterm baby, they can rupture from the slightest pressure change.
And when they rupture, it's not just the bleeding itself that causes damage?
No. The bleeding is the trigger, but the real injury comes after. Hemoglobin leaks out and oxidizes the surrounding tissue. It's like rust spreading through the brain. The neural stem cells that should be multiplying, creating new neurons, get poisoned by this process.
Why couldn't researchers study this before?
Animal models don't fully replicate human biology. And you can only learn so much from analyzing fluid samples taken from sick babies—you're looking at the aftermath, not the mechanism. This 3D model lets you watch it happen in human cells, over and over, controlling each variable.
What does controlling variables actually mean in this context?
You can add hemoglobin and nothing else, and see exactly how the neural stem cells respond. You can test a potential drug and see if it blocks that response. You can't do that with a patient sample. You can't do it ethically with a living infant.
So this is a stepping stone to treatment?
It has to be. Right now there's no treatment at all. Doctors manage the bleeding, hope the brain adapts. This model could identify which molecules are doing the damage, which ones might be blocked, which interventions might actually work before they're tested on children.
How close are we to that?
This is the foundation. The researchers have shown the model works, that it matches what happens in real infants. Now comes the harder part: using it to find something that actually helps.