Long Read - September 2019

brain repair with 3D-printed neural tissues

Brain damage through injury or disease is devastating for the patient, their families and society. Furthermore, the global economic consequences are known to be expanding exponentially; Alzheimer’s Disease alone is projected to affect 92 million people by 2050 when the costs associated with dementia could reach $1.1 trillion a year. So far, existing pharmaceutical and biotherapeutic treatments have been insufficient and clinical trials have repeatedly failed, but in Oxford a team of scientists are trialling a radical new approach.

September 2020

Even five years ago 3D printing brain implants generated from human stem cells would have been considered science fiction. Now we have the means to make it a reality, and to generate a low-cost medical technology to address the growing global catastrophe of brain damage through trauma and disease.

Leading this endeavour at Oxford are Zoltán Molnár and Francis Szele of the University's Department of Physiology, Anatomy and Genetic (DPAG), in collaboration with Hagan Bayley of the Department of Chemistry, and Oxford Martin Fellow Dr Linna Zhou. Through printing a variety of neurons and supporting cells, the team will develop layered ‘pre-organised’ printed neural tissue in vitro that will mimic the basic structure of cerebral cortex columns. They will then transplant the printed layered cortical tissue into animal models of traumatic brain injury, a condition which affects 5.3 million people globally.

Why don’t current treatments for brain damage work?


Not only are all the brain cells and brain areas interconnected in very complex ways, but their position vis-a-vis one another is also very complex. There are probably several hundred different kinds of cells in the brain and the neurons, which are the work horses of the brain, have to be positioned in very particular ways. There are six layers of each brain region; each layer of the cerebral cortex is populated by a different kind of neuron. If that layering gets disrupted, you get problems such as epilepsy.

To repair the adult brain, not only does it have to generate the neurons, but it has to segregate and position them appropriately. So far that has not been achieved through either natural repair mechanisms or by human intervention.

To repair the adult brain, not only does it have to generate the neurons, but it has to segregate and position them appropriately.

Another major reason brain disease is difficult to treat is that drugs do not penetrate the brain easily because it is encased by the blood-brain barrier. A whole class of pharmacological research is working on getting past the blood-brain barrier. Both the penetration of drugs and the very intricate anatomical positioning of the different cell types are very important but unsolved issues in CNS treatment.


One problem is that most of the cells in your nervous system are born in utero and subsequently there is very little turnover. You are trapped with the same brain cells for the rest of your life. Your skin cells are replaced every 2-4 weeks. Your liver regenerates every 5-6 months. But your brain, with some exceptions, is stuck with the same numbers of cells from birth. In fact, your cerebral cortex is the thickest at its peak when you are around 11 years old, then you lose nerve cells and the thickness decreases steadily.

A second problem is that your nervous system is built from millions of over 200 different types of neurons which interconnect with trillions of connections. Our cerebral cortex has a uniform layered structure. The lower layers tend to send projections to distant targets outside the cortex to the brainstem and spinal cord, while the upper layers interconnect with the other cortical areas. Neurons change their connections as the result of complex interaction with the environment. In fact, your brain is reflecting your entire life experience; who you meet, learning to read, playing tennis, etc. How can you regenerate something so delicate and sophisticated? You cannot replace the knowledge of these complex behaviours! For instance, think about language areas in your brain. If you speak Russian, this ability is maintained through enormous neuronal networks that are interconnected in a very specific manner to subserve these complex functions. If you damage a piece of your cortex where some aspects of language are represented in the temporal-parietal lobes of your left hemisphere, then in theory you can replace the neurons with similar neurons, but they have to re-learn these abilities. Our dream is to build some pre-assembled cortical modules that reflect the general structure of the cortex with its basic layers that can integrate into the cortex and perhaps help to re-learn the functions.

What exactly are these 3D printed cells?


Primarily, we will use human induced pluripotent stem cells (hiPSC) that are generating young neurons of the cerebral cortex. The neurons will be 3D printed in droplets of lipid biolayers to give us clusters of segregated neurons.

Each droplet has anywhere from a dozen to several thousand neurons of the same or different types, and we will control which kind of neuron goes into each individual droplet and at what concentration, and then control the position of the droplet to the other droplets. These little droplets also contain the extracellular matrix (ECM), which is the glue outside of the cell holding it all together, and we can also control what kind of ECM is in these droplets.

We end up with columns of tissue, ranging in size according to what we need, that will hopefully replicate the fundamental processing unit of the brain, i.e. the cortical column. We will then implant them into animal models to start with, and eventually into humans for brain disease recovery.


When the brain develops in nature, a single layer of cells starts to produce neurons according to a pre-set choreography; progenitors are next. The subsequently generated neurons bypass them, giving you an inside first outside last pattern in the development of the six layers in the brain. What’s interesting is if you take a fibroblast (the most common cells of connective tissue) from somebody’s skin, induce a stem cell from the skin cell, nurture and culture it, you eventually start generating the cells in the same sequence, even sticking to the same timing, to how it happens in the human brain.

This ‘mini brain’ or organoid can be produced surprisingly well, but it’s not as organised, complex and laminated as our brains even after several months of cultivation. So far, in my 30 years of studying in vivo brain development, and while Francis has been studying how stem cells produce neurons in a particular sequence, we haven’t had the means to influence how these stem cells organise themselves in the dish. We are beginning to understand how to influence the developmental programme of these progenitors, but they follow their own programme and timetable.

What we want to do now, is pre-organise these cells and print them to produce these functional units of the brain and speed things up. After we pre-organise them into the major functional cell types according to layers, we hope the self-organisation will finish the job. Additionally, the uniform six layers in different cortical areas have variation in thickness and composition of cells according to the circuits they form, according to the kind of functions they have to perform. We want to produce modules containing all the cells of the six layers, change the proportions according to the location we want them to integrate, put the modules into the cortex in an organised fashion and we hope that it should integrate much better than if you add disorganised stem cells!

Image 1

Human brain progenitor cells differentiate into neurons and astrocytes 13 days after 3-D printing. [Adapted from Zhou et al., Advanced Materials, 2020]

How did you come up with the idea?


The original ‘aha’ moment for me was when I ran into Hagan Bayley at a 3D printing workshop at Oxford. We started talking over beer while we waited for them to get the projector to work!

I was there to 3D print neurons out of plastic to show my students their size, shape and structure and he was looking for someone to print live 3D neurons with. I was studying subventricular zone cells (SVZ), which are stem cells that continue to make more neurons in the adult brain, one of the exceptions Zoltan referred to earlier. I told him my SVZs were very robust in vitro and could be collected and manipulated, so very soon after we started a few pilot experiments and got a John Fell grant. I’ve been developing the technique to 3D print viable human neural stem cells with the Hagan Lab for a number of years now and we recently published a paper discussing how they yield important insights into human cerebral cortex development - “Lipid-Bilayer-Supported 3D Printing of Human Cerebral Cortex Cells Reveals Developmental Interactions” in Advanced Materials.

However, SVZ cells migrate to the ‘olfactory bulb’ in the brain, which is an ancient part of the brain involved in the sense of smell. I reasoned that we should instead look at the cerebral cortex, which is functionally the most important part of the human brain for regulating who we are as individuals. If you kill your cerebral cortex, you lose your ability to organise your life and to remember, but nobody at the time was recreating the layers of the cerebral cortex. Others have grown cerebral organoids or ‘mini brains’, as Zoltan said, but the cells are all higgledy-piggledy and do not refelect the layered cortical columns necessary for proper cortical functioning.

I realised that we could actually print human neurons from hiPSC in these columnar layered organisations and then use them for two things. One, to better understand human cortical development, as we know a great deal about how the mouse brain develops but relatively little about how human brain cells develop, and make these layered columns. Two, and I have to give a lot of credit to Hagan Bayley for this realisation, was to understand whether we could use these layered cortical columns for functional recovery in conditions such as Alzheimer’s disease, traumatic brain injury or stroke. I also realised that DPAG happens to have one of the world’s experts in the development of the cerebral cortex and that’s Zoltán Molnár!


For the last 30 years I’ve been studying how cerebral cortical neurons are born, migrate and interconnect and how you affect their circuitry if you change their position. I’m interested in the understanding you can generate with this printing: how different proportions of cells self-organise into functional units. Even better if we can then implant these and find they work better than normal unorganised stem cell mixture.

I was initially sceptical, but there are so many limitations to the conventional methods growing organoids, or ‘mini brains’, you have to give another approach a shot, no matter how futuristic.

Why hasn’t this been attempted before?


In the 20th century, a huge amount of developmental biology was based on taking relatively large chunks of tissue and putting them in different organisations in the developing embryo and showing that these transpositioned tissues would be able to induce different developmental events. But we were never able to do that at the cellular level to ask finer grain developmental questions. 3D printing allowed that to happen and 3D printing of live tissues was a quantum leap.

The conceptual breakthrough came alongside a paper by Hagan Bayley, revealing he could print functionalised tissues that have ion channels in the membranes, and that these can be controlled in a variety of ways. We then took the next step and asked: “Can we pepper these tissues with particular neurons at particular locations?”

It’s incredible to think that if we can print in 3D the actual cells, then we can achieve experimentally what nobody else has been able to do, which is at the cellular level control how cells are positioned next to each other.


The other breakthrough was being able to generate, pre-differentiate and then print the different cell types. Through new technology, the field has identified the master regulatory genes to push a stem cell in a particular direction. It means we can influence the differentiation of a cell and produce basic cell types such as a layer 5 projection neuron, or a layer 2 or 3 cortical neuron.

How to produce the basic functional components of the brain, and integrate them meaningfully, is one of the biggest questions in cortical development and in neuroscience.

– Professor Zoltán Molnár

What are your initial goals for the project?


Initially, we will start with basic two layers and build it up further, perhaps eventually to six. Our first goal is to have two layers of two types of neurons: first, long range projection neurons found in the lower layers (layers 5 and 6) of the mammalian brain, second, a type of cell in the upper 2-4 layers that project locally within the cortex.

We worked on this switch, publishing in 2008 and 2015, with Victor Tarabykin’s group. Right now, using key transcription factors which determine the fate of these stem cells, our teams are producing induced pluripotent progenitors for printing upper layers and lower layers for a small piece of printed “cortical” tissue. Then we will implant these two layered cortices in young mice and see how these printed prearranged columns with upper and lower layers integrate into the mouse.

How challenging will the implantation be?


10 years ago, I would have said that implantation of any kind of brain cell into the brain of an adult animal or human is a pipe dream. Zoltán and I invited Professor Afsaneh Gaillard to Oxford, a world expert at implanting human cells into mouse brains and getting functional recovery to a certain extent through anatomical integration. We just didn’t believe it was possible, but she convinced us and then taught me the technique in her lab in France. Fundamentally, it is difficult: when others have tried it by putting the cells all mixed together, they’ve recovered a little function, but not that much, and so we are collaborating with Afsaneh and Professor Pierre Vanderhaeghen as two of the world’s leaders.

While it is risky, we’re hoping that with our technique it will be anatomically and developmentally correct so the recovery of function will be better.


In the past, Pierre has successfully induced one type of cell, or led them on their own to integrate, and he will help us to develop methods for how we put these integrated units into the host brains.

What will be your measure of success?


Oftentimes these stem cells don’t mature fully and because of this they don’t acquire fully mature functions. Firstly, we think that because we’re printing them in anatomically correct patterns that they’ll mature more fully, both in vitro and in vivo. Secondly, we will check for appropriate physiological activity: these neurons are firing away little electrical signals known as action potentials in particular patterns, and our goal is to show with calcium imaging that neural firing is working appropriately and as you would expect for those cortical columns.

For this, we’re collaborating with Professor Owen Ko, a world expert in how cortical columns develop and start to send signals to each other and within themselves. Thirdly, we will know we’ve achieved anatomical success if the cells integrate into the host brain circuitry; they have to receive inputs from and to send outputs to the host brain. Ultimately, we will look for functional recovery: the betterment of defects in the animal models based on the implants.

We will know we’ve achieved anatomical success if the cells integrate into the host brain circuitry. Ultimately, we will look for functional recovery.

We will know we’ve achieved anatomical success if the cells integrate into the host brain circuitry. Ultimately, we will look for functional recovery.

– Professor Francis Szele

Are there any other challenges you expect to face?


We also need to print, study in vitro and implant different ratios of cells. It’s important to get the ratios of excitatory neurons, inhibitory neurons and support cells right and the neurons have to wire up correctly in order to work together.

In addition to our in vivo work we will extensively study the development and function of these 3D printed cortical columns in vitro. Also, extremely importantly, the brain is filled with an extensive vascular network composed of blood vessel cells. Blood vessels are everywhere because the brain demands so much energy. A big goal will be to understand how we’re going to get blood vessel formation in these cortical columns. If you don’t have enough nutrients infiltrating into our 3D printed columns of tissue, they’re not going to survive. Achieving proper irrigation is an essential goal!

You are beginning with traumatic brain injury – will it be possible to translate your results to other types of brain damage?


There are similarities and differences in brain injuries and diseases. Most brain diseases have an inflammatory component, which we will pay attention to. Nevertheless, certainly from my group’s work from the past 20-25 years, every model of disease and injury has subtly to grossly different effects on the regenerative capacity in the brain. One injury does not equate to the other, and the timing of the injury, even the way you make the injury, can have a huge effect on functional outcomes. I predict, therefore, that this would affect the way these stem cells integrate into the brain. We might eventually discover that this kind of 3D printed tissue transplantation is best for a certain type of disease like a traumatic brain injury rather than for a stroke or degenerative disease.


Exactly, degenerative disease is probably the most difficult because you have a complex pathology that is gradually advancing and influencing a large part of your brain. You are losing your capacity, and then you expect a newly transplanted partially preorganised tissue to adopt all these functions in the declining brain. We will need to think about how to integrate the implants in degenerative conditions, perhaps by stimulating some of the cells with gene-targeted drugs, but initially we will start with something simple.


We can monitor these cells upon implantation and in vitro in a dish with two-photon and three-photon microscopy. We’re developing three-photon imaging techniques with DPAG’s Dr Adam Packer to see deep into the brain and look at the function of the neurons in real time in live animals. We might then be able to manipulate the neurons before and after implantation by application of drugs to stimulate their function.


And since we have control over the generation and differentiation of these cells before and after we print them, we could potentially genetically engineer these properties into a subset or for all of these cells. Imposing patterned activity into the grafts could be a game changer for integration and functional recovery.

What is your overarching ambition?


We have two interlaced goals. First, to develop a robust strategy to really understand human functional development of the cerebral cortex for the first time. Second, to create an affordable cost-effective therapy for brain repair.

You can’t do one without the other. You need the in vitro work to understand the fundamental mechanisms of the neurons and then how they behave when they’re put into a diseased person’s brain.


All mammalian brains in any species have a uniform cortical arrangement with six-layers, so we have a prototype circuit across the whole brain, and the basic circuits follow the same principles everywhere. To understand how these circuits are put together and how you produce the normal functional unit of the mammalian cerebral cortex would be already a huge step in my opinion. But, you really have to think about the big questions of how to tackle some of the major brain diseases such as traumatic brain injury or neurodegenerative conditions like Alzheimer’s disease.

The Oxford Martin School encouraged us to think what the biggest contribution to this field would be. We now have the opportunity to understand how to produce the basic functional components of our brain and how to integrate them meaningfully to have a functional implication. This is one of the biggest questions in cortical development and in neuroscience.