Blog – The Pros and Cons of using iPSCs in Dementia Research

One of the big flashy trends in dementia research over the past decade has been the use of induced pluripotent stem cells, or iPSCs, to try and understand the fundamental molecular and cellular biology underpinning disease. At the start of my PhD, a little over three years ago, I thought this was tremendously exciting, and joined the trend, embarking on a project with a lab specialised in iPSC models for neurodegenerative disease. After three years of working with stem cells and stem cell derived neurons, I figured it would be useful to reflect, from experience, on the advantages and disadvantages of using this model, from both a scientific standpoint but also a personal one.

But first – what are iSPCs and why might we want to use them in the first place? Stem cells are cells that are highly proliferative – they multiply quickly – and have the potential to become more specific kinds of cells. All cells in the body derive from one original stem cell – a totipotent stem cell, because it has the potential to become every other cell in the body. The totipotent stem cell divides into pluripotent stem cells in the early embryo, that have the ability to differentiate into more and more specific cell and stem cell types, such as hair follicle stem cells that will turn into hair, or skin stem cells that produce skin cells, neural precurors that produce brain cells, etc. The way a cell changes identity is by changing the genes its expressing. A brain cell expresses genes that make it a brain cell, likewise a heart cell expresses genes that make it a heart cell, and same for stem cells. Biochemical signals in the environment, or exogenous factors, can cause gene expression to change, allowing a stem cell to turn into a specific cell type. It stands to reason though, that since all cells contain the same set of DNA and only use a subset of genes, you should in theory be able to get a cell to stop expressing the genes that make it, for example, a skin cell, and start expressing genes that make it a stem cell.

Shinya Yamanaka and his team in Japan first published “Induction of Pluripotent Stem Cells from Mouse Embryonic and Adult Fibroblast Cultures by Defined Factors” in 2006. Here, they outline their discovery that using only four essential transcription factors, proteins that regulate the expression of particular sets of genes, they could induce skin cells to return to a pluripotent stem cell state. This landmark discovery earned the 2012 Nobel Prize in Physiology or Medicine, and is important because of the huge downstream potential of this technology. Now that the stem cell state has been induced, the cell can be guided, either by coaxing using various signalling molecules, or forced with viral gene delivery, to turn into any other kind of cell in the body. This allows the functioning of live cells – any kind of cells – to be examined in a dish in the lab.

Shinya Yamanaka won the Nobel Prize for iPSC discovery in 2012.

Why are iPSCs Useful for Dementia Research?

For dementia research, this is especially momentous. One of the simple but nevertheless overwhelming reasons that research into diseases like Alzheimer’s and Parkinson’s is so hard, and that after over a century of work, we still have limited understanding of what actually causes and goes on in these diseases, is because the brain is hard to study. Like, really hard, especially at the cellular and molecular level, where it seems these diseases manifest.

For the most part, if we want to study what happens to cells in the brains of patients with dementia, we have to wait until the patient has died. This has been extremely useful and valuable to our understanding of dementia related pathology, but presents two important limitations. First is that we cannot see the live functioning (and dysfunction) of the cells as they progress from healthy to diseased. The skull presents a dense layer between cellular functions and any hope of observation without highly invasive and dangerous surgery. Techniques like MRIs and PET scans enable observations of large scale changes in the live brain, but lack the resolution to see precise circuit or cellular changes, let alone subcellular molecular changes. Furthermore, post mortem studies only allow us to see the end stage of disease – we see cellular grave yards, and cannot glean their initial cause of death, with may have occurred as many as 10 or 20 years earlier. Studying how live cells change in the context of dementia is essential to understanding disease.

For reasons that should be obvious, it’s not easy or ethical to acquire live brain tissue from a human patient. It is not impossible – sometimes biopsies are necessary in patients with suspected brain cancers, and in some diseases including brain cancers or epilepsy, it may be necessary to remove brain tissue from a live patient, which can then be used, with consent, for research. However, this is rare, and has all the additional considerations of the patient, who may or may not have dementia, having different diseases that would confound results.

This has lead to a heavy reliance on animal models in dementia research. Studying the brains of rats, mice, worms, fish, and flies, is a staple of dementia research. They allow us to examine the functioning and progressive dysfunction of cells in the context of a whole brain at different stages of disease in a highly controlled manner. The problem with animals, however, is that now we are examining a human disease in non-human cells. Notably, none of these animals actually naturally get dementia, so it has to be induced artificially. While these models have been useful for understanding the effects of different pathologies and genetic mutations on cellular functioning, they have a poor translational track record, and remain extremely limited in our ability to study human brain disease in a human cellular context.

The Advantages of the iPSC Model

This extensive preamble is to underscore what a gamechanger iPSC technology is. Pluripotent stem cells can be induced from tissue acquired from patients and healthy donors in a safe, simple, one off, not particularly invasive procedure that doesn’t touch their brain at all. From these stem cells, neurons can be grown in the lab in vitro that enable the comparison of cells from patients with disease to those without disease, and enables us to study the effect of toxins, drugs, therapies, disease related proteins, genetic mutations, and any other kind of change on the live functioning of human cells. Since an inherent property of stem cells is that they divide rapidly (and since they are typically induced from skin fibroblasts, which are not only easy to access but also divide rapidly), it means that once a patient has donated a cell sample, vast quantities of cells can be grown and used for years on end.

Given that these are in vitro studies, this gives us as experimenters a substantial level of control in experiments, allowing us to start to ask causative rather than simply correlative questions, which is what post-mortem analyses are limited to, and truly begin to interrogate the earliest stages and fundamental biochemical drivers of disease. In short, we can now study the live human brain cells of patients with Alzheimer’s and Parkinson’s disease, without harming them or ever having to access their actual brain. Moreover, where animals models have largely failed to translate clinical results to humans, stem cells bear great promise for high throughput screening of candidate drugs or other interventions with a view to better understanding therapeutic targets and mechanisms, in the context of cells with human genetics.

Not only is this nothing short of miraculous, but it is also just really cool. An underrated advantage of working with iSPCs is that you get to say you grow human brain cells in a dish at parties. This is, objectively, extremely cool. Not only is this conceptually lucrative though, but aesthetically, working with stem cells is also superb. The reality is that neurons just look really pretty, and when you get to watch them grow and move around, and when you stain them with fluorescent markers, you get stunning presentation images to wow any audience, spectacular desktop wallpapers, great material for image competitions, and impressive pictures for your parents to show to their friends (pour one out though for all the poor people who do purely gene expression or protein analysis type work, and just have gels, heat maps, and PCR graphs to work with).

iPSCs were first created from mouse cells in 2006, and human iPSCs followed in 2007. 

Caveats of working with iPSCs

Given all of the above, you would be forgiven for assuming that the process of making stem cell neurons is simple and easy. Rest assured; it is not. It is, in fact, extremely onerous, time consuming, and expensive. Going from skin fibroblast to iPSC involves at least a month’s worth of work, and after all of that, iPSCs are a fickle cell type to work with, requiring daily attention, and sensitive to all manner of things like temperature and media changes. In short, they need a lot of high maintenance care, attention to detail, and unless you have some kind of extremely expensive automated culturing and feeding system (or minions), you can say goodbye to most of your weekends. And that’s just the iPSCs.

Differentiating iPSCs into neurons is an adventure in its own right. While conceptually amazing, practically its extremely tedious (as most lab work is, in fairness), but also highly time consuming. Depending on the cell type you are making and the protocol you are using, it may take between 2-3 weeks to grow neural precurors, early and immature forms of the cells you’re interested in. To get to properly ‘mature’ neurons, which have appropriately polarised membranes, express neuronal subtype specific markers, have functional synapses, complex dendritic arbours, and reliably fire spontaneous action potentials, can take as much as 2-3 months. While this is still shorter than many rodent experiments, cell cultures tend to require substantially more care and maintenance along the way, and differentiation protocols can involve having to use a complicated variety of different chemical or genetic factors to achieve the cell types you’re interested in, with frequent if not daily attention.

As an Australian studying in the UK, friends at home often quip how excellent it must be situated near Europe, with so much exciting travel on the doorstep. The reality is that working with stem cell derived neurons means it’s often tricky to get away from the lab for more than a couple of days at a time, and that for better or worse, your calendar has to be planned months in advance, depending on your cell and experimental time lines. Running well controlled experiments where you do interventions or tests on the same day or time point means that you lose some of the flexibility in your life you may have if you’re working with, for example, frozen tissue, and it can just stay in the freezer until tomorrow if something unexpected crops up. It turns out that the advantage of being able to work with live cells is also its own disadvantage – you have to keep it alive (and hopefully you don’t get to day 80 of growing your cells and suddenly find them contaminated, or accidentally feed them the wrong thing, or discover that for reasons beyond your comprehension, they’ve all suddenly died). As well as being psychologically costly (especially when your cells die), long experimental time lines also means that not only are labour costs high, but so are equipment and reagent costs. Media to grow cells, especially the highly customised media for maintaining iPSCs, does not come cheap, nor do the various recombinant growth factors used to differentiate cells into neurons.

Beyond the lifestyle and financial inconveniences though, there are other scientific caveats to working with stem cell derived neurons. One is that even if you do grow cells out for 2-3 months (and growing cells like astrocytes can take even longer, over 100 days), these cells aren’t ‘mature’ in the strict sense. In fact, the process of inducing cells back to pluripotency ‘de-ages’ the cells. iPSCs, in an epigenetic sense, look like embryonic cells in terms of their biological age. They’re no longer the cells of a 40-something year old patient, but of a developing embryo, and 100 day old ‘’mature”neurons are just that – a mere 100 days old. It’s not entirely clear what this means in a biological sense – how different is a mature functioning cell in an adult brain from a ‘mature’ functioning cell in a stem cell derived culture? While conceptually fascinating that we can take adult cells and rejuvenate them, this is somewhat counterproductive when studying a disease of aging. One obvious concern for Alzheimer’s research in stem cell neurons is that a key feature of developing neurons is high levels of tau phosphorylation – a confounding issue when it comes to studying tau pathology and dynamics in these cell models. While some groups have developed protocols for direct conversion of cells from fibroblasts to neurons, preserving the biological age of the cell, these ‘induced neurons’ are not nearly as scalable as iPSCs are.

Another inherent limitation of working with cell cultures is that we have now stripped a complex brain disease, likely to have interplay between different neuronal and glial cell types, vascular factors, a role for the gut and other bodily organs, and immune function, into a highly simplified 2-dimensional model. Gone is the complex context of the brain and nuances necessary for understanding the full interrelations of the disease. In one sense, this is an advantage of the model – it allows us to isolate the cell autonomous factors driving disease, and examine specific interactions with careful cocultures of particular cell types – for example, growing neurons alongside microglia. Recent developments have also enabled the growing of 3D “organoids’’, assembling neurons into structures that resemble cortical layers. As these technologies develop, they will likely come to include vascularisation and enable complex modelling of blood brain barrier function too, another aspect of these diseases that iPSCs are already facilitating.

iPSCs, in part because of the way they are made, also tend to be highly variable. The transcription factors introduced to induce their pluripotency can introduce all kinds of unexpected genetic alterations and genomic instability. iPSCs need careful and regular quality control and karyotyping (all of which is more work), and even with this, one differentiation is never quite the same as the next, and often the same cells grown in parallel by two different researchers can end up looking somewhat different, and experiments are hard to reproduce, especially when methods have insufficient detail, which is almost always the case. This means you need many differentiations and ideally many cell lines to overcome the high variability between cell lines and differentiations, but this is, again, more work than can be reasonably expected of anything less than a small team, more time, and more expense. N numbers are something to watch out for in papers using iPSCs – they’re typically dire.

Even with all this work, while you might end up with nicely functioning neurons, they may not be precisely the cells you want. Depending on the differentiation protocol, you may end up with a very generic cell type such as “cortical neurons” which in reality in the brain encompasses a huge array of different cell types across vast swathes of neural real estate, or you may end up with not quite the cell type you were hoping for. For example, a popular protocol for generating midbrain dopaminergic neurons is often used in studies of Parkinson’s disease, where midbrain dopaminergic cells in the substantia nigra are highly vulnerable to degeneration. The problem is that most of the cells generated by this differentiation protocol seem to reflect the identity of midbrain dopaminergic neurons in the ventral tegmental area, rather than the substantia nigra.

All this being said, iPSCs are still an incredibly new technology, and improving at a rapid pace. Differentiation procotols are being improved and developed to generate more cell types, more accurately, and more quickly. As cost barriers to iPSCs come down and some parts of the protocol can be automated, their use can hopefully become more widespread, and methods developed to understand how to even model aging in these developing cells. In the end, all methods and models have their pros and cons, and the same is true for iPSCs. While they can be a pain to work with, they are, undeniably, extremely cool, and potentially a huge scientific breakthrough for our ability to test therapies for and understand the fundamental biochemistry of diseases in the brain and beyond.

Ajantha Abey

Author

Ajantha Abey is a PhD student in the Kavli Institute at University of Oxford. He is interested in the cellular mechanisms of Alzheimer’s, Parkinson’s, and other diseases of the ageing brain. Previously, having previoulsy explored neuropathology in dogs with dementia and potential stem cell replacement therapies. He now uses induced pluripotent stem cell derived neurons to try and model selective neuronal vulnerability: the phenomenon where some cells die but others remain resilient to neurodegenerative diseases.

Follow @ajanthaabey