From mapping the inner workings of individual neurons, to improving our understanding of cancer biology – and with deepening integration into multi-omics research – the single-cell revolution continues apace!
Following an illuminating overview from Jonathan Sweedler, we were keen to learn more about the field – and the key figures working within it – in a little more detail. So, as part of a new series, we reach out to several single-cell experts to highlight the landmark accomplishments, most exciting research avenues and applications, as well as the big challenges facing the field today.
Here, Andy Ewing, Professor in the Department of Chemistry and Molecular Biology at the University of Gothenburg, argues that we need to isolate and measure individual organelles – in real time – if we want to make real progress in biology and precision medicine.
Looking at the past decade or so, are there any key developments that you think have catalyzed the evolution of the single-cell analysis field?
One of the biggest developments in the last decade has been advances in imaging techniques, particularly the use of stimulated emission depletion (STED) microscopy and mass spectrometry imaging. We've been doing single-cell analysis for a long time with techniques such as confocal microscopy, but the integration of chemical analysis – especially fluorescence-based methods – has been a major leap forward. Interestingly, while there have been four Nobel Prizes awarded for fluorescence-related techniques, there's never been one for the actual discovery of laser-induced fluorescence. A major milestone came in 2017 when we first imaged inside vesicles using nanoSIMS. Before that, STED microscopy had taken off with some remarkable collaborations – such as those with Steven Hell's group at the Max Planck Institute – which pushed the boundaries of what’s possible. These developments have catalyzed some really exciting science.
Another key area has been protein modeling. Researchers like Silvio Rizzoli and others have done some beautiful work simulating what happens inside a cell or even at a synapse, down to the level of the synaptic bouton. Then, of course, there's all the high-throughput work and massive data analysis, which is pushing the field into areas we haven’t explored before.
Looking ahead, some emerging techniques are particularly exciting, one of which is single-cell NMR spectroscopy. This involves using small diamond crystals with nitrogen vacancies, which originally used fluorescence for readout. Jonathan Sweedler was working on capillary NMR for small volumes, but we're now talking about pushing this down to the single-cell level.
Techniques like these are crucial because they bring us closer to real-time analysis. Right now, a lot of STED imaging is still done on fixed samples, as though live imaging is possible, it remains very challenging – especially for single-cell work. And, of course, mass spectrometry is still primarily limited to fixed or frozen samples.
One area that’s been a major development over the last decade is electrochemical analysis of vesicles. We actually invented a technique ten years ago to measure vesicle content inside cells using electrochemistry, if the contents were electroactive. About five years later, a major advancement came with the development of glutamate sensors. Some of the best work in this area was done by Wei-Hua Huang at Wuhan University, in collaboration with Krantz and Amatore. They published a beautiful paper around 2021 that strongly supported the idea of partial neurotransmitter release during exocytosis – a key aspect of understanding cellular signaling. We also published some work on electroactive compounds that further validated these findings.
Another exciting development is expansion microscopy, which involves physically expanding biological samples to improve resolution. One of the challenges with this is that proteins must be broken up to prevent distortions during expansion, but despite that limitation, scientists such as Ed Boyden at MIT have managed to obtain some stunning data.
Can you tell me a little bit about your current research?
Right now, we're really getting into communication between organelles. I don’t think we’ve fully solved the problem of communication between cells yet as there’s still a lot to explore there, but our focus has shifted somewhat as we’ve gotten better at measuring things in smaller areas. This brings me on to another area that’s going to be a major topic to watch: protein turnover.
When discussing sample preparation, there are two main approaches: you can analyze live samples, or you can work with fixed (dead) samples and take snapshots at different time points to study processes, like protein turnover. For example, we examined protein turnover in stress granules – protein-RNA complexes that form when cells experience stress – thought to play a role in cancer. You can induce their formation using various methods, such as heat shock, arsenite (a toxin), or other stressors. We’re particularly interested in stress granules due to the interactions they have with other organelles. Our research has shown that when stress granules are abundant, they can cause vesicles to undergo homotypic fusion, altering their properties – this is a key reason for our focus.
To study these processes, researchers typically use isotopic labeling or fluorescent labeling of proteins. This allows analysis using STED microscopy or SIMS, but for structures as small as stress granules, NanoSIMS provides better resolution. One effective approach is to analyze protein turnover at different time points after deliberately altering the cells, often using a stop-flow device to precisely control these changes. However, optimized sample preparation is absolutely critical – without it you won’t be able to properly visualize the vesicles, stress granules, or whatever structures you’re investigating.
We do a lot of work with fixed samples, though we also use frozen samples when possible. Frozen samples yield better results, but are significantly harder to work with. The choice of fixation method therefore depends on the specific application. With expansion microscopy, if you want to stretch out the structures for better visualization, you have to break apart larger molecules before fixation.
When working with non-live STED microscopy, mass spectrometry imaging, or expansion microscopy, sample preparation remains one of the biggest challenges. It’s a critical but extremely difficult aspect of these studies, and there’s still a lot of work left to be done in this space.
What are the major challenges in single-cell research today?
Sensitivity. I think it was Thomas Jorgensen that showed that every time you reduce the size by a factor of 10, you need to be 100 times more sensitive. If you're looking at a surface, the signal scales with radius squared – or if it’s a square pixel, it’s side squared. This is a huge challenge to overcome.
Another is spatial resolution, but of course, the two of those are intertwined – if your sensitivity doesn’t improve, you can’t achieve better resolution. In 2009, we published a paper where we used Poisson distribution to show how resolution depends on signal strength. At the time, a lot of people in MALDI were using pixel size as a measure of resolution, but that’s just wrong. Resolution depends on signal strength, which in turn depends on how much material you have and how well your system responds.
Ultimately, we need better responses. With SIMS and MALDI, only a fraction of the material makes it through the mass spectrometer, so there’s still a lot of room for improvement. The same applies to fluorescence techniques, as fluorescence still has a lot of inefficiencies. This is the reason electrochemistry is so sensitive – when you use an electrode to oxidize something in solution, it acts like a little sink. As you oxidize and generate redox species, everything migrates toward the electrode. Diffusion is always from high to low concentration, so you can effectively use up almost all of the available material.
It’s worth mentioning that although electrochemistry is far more sensitive than other techniques, it does have its own limitations: you can only detect a limited range of molecules, unless you use enzymatic reactions to broaden the scope – which is where things like glutamate and acetylcholine detection come in.
Are there any emerging trends in single-cell analysis that you’ve got your eye on?
One thing that I think will catalyze the field is nanopore analysis. Nanopores have already revolutionized how we sequence DNA, which is a huge breakthrough. But now, researchers are using nanopores to analyze proteins, detecting incredibly small changes as proteins pass through. They’re doing this with alpha-hemolysin pores at the moment, and the level of detail they can measure is incredible.
And this ties right back to what we were discussing: nano-entities. If you take the same nanopore technology that we use for DNA sequencing and apply it to nano-entities – whether that’s proteins, organelles, or other biomolecules – you’ve got something big. This is going to be a major area of development.
What big problems could single-cell analysis solve or answer?
Personalized medicine seems like an obvious area where this research will be particularly important. Right now, we’re still at the sampling stage – we take a sample, then analyze it – but in the future, we’ll be able to analyze without sampling. The goal is to reach a point where analysis happens in real time, at the cellular level, no matter the application.
Single-cell analysis is valuable when dealing with heterogeneity or small samples, but there’s another aspect that people often overlook. After almost 40 years in this field, I’ve never fully convinced people of this: if you want to study individual organelles and intracellular processes, you have to work at the single-cell level. If you group cells together, you create a mess, which makes it much more difficult to interpret the data. Biology has made remarkable progress in spite of this challenge, but if we can isolate and measure individual organelles, that’s where the real breakthroughs will happen.
I believe we’re moving toward a future where we’ll advance the resolution of all sorts of spectroscopic techniques. Imagine a scenario where you place your finger on a device, and it performs a spectroscopic analysis – not just of the cells in your finger, but of specific types of cells within it – and provides a precise diagnosis. I truly believe this is the direction we’re headed towards, but to get there, we need a deeper understanding of mitochondria and other organelles. What we need is organelle-level analysis, rather than just broad chemical analysis – or ideally, a combination of both.
We’re making progress, but there’s still plenty of work to be done.
Andrew G. Ewing is Professor of Chemistry and Molecular Biology in the Department of Chemistry & Molecular Biology at the University of Gothenburg, Sweden
I’m really into basic science, but I also work with a lot of medical professionals now. In fact, I just finished a paper in review where we had 179 physicians evaluating statements about long COVID. Much of the discussion revolves around cellular and subcellular effects. This is what people care about, and this is what we need to understand. Medicine isn’t just about infectious diseases; there are also cancers, neurodegenerative diseases, and a whole range of conditions that stem from cellular dysfunction.
That being said, we’re finding more and more evidence that many of these conditions may be triggered by viruses. We had some indications before the pandemic, but now, with all the research that has gone into virology, the evidence is really piling up. Take COVID, for example: I wouldn’t say it causes cancer, but there’s growing evidence that it increases the likelihood of developing certain cancers. We already know it’s linked to an increased risk of neurodegenerative diseases. And this isn’t new – consider mononucleosis (Epstein-Barr virus): we now know that every single person diagnosed with ALS has had Epstein-Barr virus at some point. That doesn’t mean everyone who’s had mono will get ALS, but the fact that there’s a 100 percent correlation strongly suggests a causal relationship.
We know that in some cases, other factors are likely to play a role, too. A friend of mine in Linköping was the one who discovered that spike proteins cause amyloid proteins to aggregate – a truly terrifying discovery. Another collaborator, Resia Pretorius in South Africa, discovered that COVID-19 causes microclots, and that these clots are actually amyloid aggregates in the blood. We have no idea what fraction of people might end up with early-onset Alzheimer’s or other neurodegenerative conditions as a result of this. We know that it takes around a decade for HIV to develop into AIDS, but we’ve only had five years with COVID-19, so we’re still in the early stages of understanding the long-term effects.
We don’t know what we’re going to discover, but it’s these questions which are the kind that cellular analysis will help us answer. We need to expand our focus beyond just single-cell analysis. We should be studying small biological entities in general – that means organelles, protein aggregates, and other subcellular structures. Clearly, protein aggregation is a massive area of research, and I think this is where the next major breakthroughs will come from.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
