On a recent press trip to Oxford, Gustav Ceder visited John Todd, PhD, at the Centre for Human Genetics to discuss his latest research and reflect on the path that brought him there. Ceder having followed Todd’s work for years, has always been fascinated by the way Todd pairs early adoption of new technologies with a clear translational purpose, and maintains a sustained focus on the type 1 diabetes community and more recently, on the field of neuroinflammation.
Q: We’re speaking today at the Wellcome Centre–what has this place meant to you personally and briefly, what influence has this center had on the scientific community?
John Todd: The Wellcome Centre for Human Genetics started in 1994, when a colleague of mine, Sir John Bell and I with others wrote a grant to the Wellcome Trust to put this building up. The doors opened in 1999. I actually left to go back to Cambridge in 1998 and spent 18 years there. Then nine years ago I returned to the University of Oxford and the Wellcome Centre for Human Genetics (now the Centre for Human Genetics), because I wanted to run clinical trials. I wanted to take the knowledge we’d gained over the last 30 years and translate it into randomized placebo-controlled, trials. And Oxford was the place to do that. The Centre for Human Genetics offers all the facilities [and] expertise. It’s a fantastic community and I was director for five years from 2019. We have diversity [and] equality, and we’re having fun and doing great things.
Q: What drew you to science as a general field, and then what got you into the areas of diabetes, inflammation, and autoimmune diseases?
Todd: My story goes back to secondary school or grammar school. I grew up in Northern Ireland and my biology teacher, whose nickname was Wee Hick, was a total inspiration. He inspired me in biology because there was a course running at the time called the Nuffield Biology Course. They would give you some problems and questions and clues and you had to work out what was going on in this biological system. It might be about Daphnia swimming around in a pond, and there was something about that learning approach that I really liked.
University of Oxford
From school, I was able to get into Edinburgh University. At Edinburgh, there was a lecture on immunology. I thought, “Wow, what is this immunology stuff about? It’s fascinating!” That was a seed. At this stage I was still a microbiologist. I was getting interested in protein biochemistry and combining it with genetics, [and] had an idea. I took that to the University of Cambridge, the biochemistry department there. My brilliant PhD supervisor, David Ellar, PhD, said, “Oh, I like that idea, John. If you get a fellowship, come and do your PhD in my lab in biochemistry.” I did that in 1980, and the idea worked out.
I got my first Nature paper in 1982, and I stayed in the lab to learn recombinant DNA technology. I was fortunate enough to get into the unbelievably mighty Medical Research Council Laboratory of Molecular Biology for a year to work with Greg Winter, PhD, and César Milstein, PhD, both Nobel Prize winners. I learned how to clone and mutate DNA and lots of other experimental tricks. I learned a lot in that year because what I really wanted to do, instead of the genetics and biochemistry of bacteria, was genetics and biochemistry of human disease.
I looked around all the labs in the world that did that and there were only two labs. One was Leroy Hood’s lab in Caltech and the other was Hugh McDevitt’s lab in Stanford. The McDevitt lab was doing human work and I got a letter from Hugh McDevitt saying, “If you get a fellowship, John, you can come and join the lab.” I joined Hugh’s lab in 1985. I met John Bell, who was a senior postdoc there.
The breakthrough was, at that time, PCR (polymerase chain reaction). One of the guys that really got it working was Randy Saiki, who was a Cetus employee. I used to go trout fishing with Randy. Randy would say to me on every expedition to the northern lakes in California, “You got to try this PCR stuff.” John and I eventually did in March ’86, and it worked the first time. John and I began to realize that we could sequence DNA from patients in a matter of days almost, rather than weeks in the traditional way of doing molecular genetics.
That led to some major discoveries, including in type 1 diabetes. A major determinant of the disease is a complex region of the human genome called the HLA. That paper got into Nature in 1987 and is one of our most cited papers. Recently, we’ve been working out the mechanism of that association. It tells you how long it can take to do research and get to the nuts and bolts of mechanism.
Because John had lots of connection with Oxford, I started looking for jobs at the end of my Stanford postdoc in 1985. One of those trips was to Barbara Davis Center for Diabetes (Colorado). I walked into reception and I tripped over a teddy bear. I looked around and in the reception of this research clinical center, with parents, children, staff scientists, physicians, and I went, “Wow, type 1 diabetes is a really serious and common disease.” Children are affected and adults. We’ve just made this major genetic discovery about type 1 diabetes susceptibility. That was the moment that I thought this is the specific disease area where I want to combine genetics, biochemistry, molecular biology, whatever it takes to make a difference.
In 1985 I took up my first group leader position, at the University of Oxford and I stayed there for 10 years before leaving for Cambridge. Since 1986 I’ve been working on type 1 diabetes, understanding the mechanisms, taking that forward to the clinic.
Q: You’ve been around and seen a lot of technological advances happening quite rapidly. Can you elaborate on some key advances?
Todd: Technologies are fundamentally important and enabling. As Sydney Brenner, DPhil, said, “You can have ideas, you can make discoveries, but unless you actually have the technology available to answer the question, you won’t be able to answer the question and make the discoveries.” PCR was a total revolution, it changed everything—how we did human genetics, how we analyze and sequence the human genome. The current revolution, and we’re living through one at the minute, is obviously AI and all sorts of technological advances: targeted proteomics, untargeted proteomics, mass spectrometry. We live in an age that is 100 times more revolutionary than PCR in 1986 because our capabilities have really exploded in the last 10 years. Combining all that gives us an ability that we’ve never seen before.
Another good example of enabling technology: we wanted to map all the other genes apart from HLA in type 1 diabetes. We started in the mouse model and had to create our own genetic map of the mouse. We realized that type 1 diabetes was polygenic, it wasn’t a simple recessive disease. We thought that if we could collect enough families and cases and controls, we could do that in human type 1 diabetes. But the technology wasn’t there to do enough polymorphisms across the human genome at sufficient density to capture the information in thousands of people.
Without exaggeration, I had to wait 13 years before Affymetrix and Illumina developed single nucleotide polymorphism (SNP) arrays to do something we coined genome wide association studies (GWAS), in which hundreds of thousands of SNPs could be analyzed in tens of thousands of people. In the early ’90s I published a paper about a technique called TaqMan. In 1995 there was no way that could be scaled up, but in 2015/2016, both Affymetrix and Illumina made advances and we used those technologies. My most cited paper is with a group called the Wellcome Trust Case Control Consortium. Again, published in Nature, which many people think is a milestone in genome wide association studies. That was one of the first applications of the early Affymetrix SNP arrays. Of course, now SNP arrays are almost obsolete because DNA sequencing costs have come down, once more due to advances in sequencing technologies.
High throughput sequencing has many implications and one of them is reading out proteomic analyses. I’m interested in very rare molecules in human blood, such as interleukin-2 (IL-2), and sensitivity and dynamic range of detection are crucial. We’ve done a lot of work using the Olink platform. However, more recently, over the last 18 months, we’ve been using NULISA™ (Alamar Biosciences). It’s been up and running smoothly now for months. We’re very interested in neuroinflammation and we’ve been applying the central nervous system neurology panel. It offers sensitivity, which is crucial, modified forms of phosphorylated tau, and now, phosphorylated tau 217. This is an FDA approved biomarker of Alzheimer’s disease. That’s a breakthrough because trials or early experimental medicine studies can be directed at lowering the amount of that protein very early in the disease process. You want the most sensitive and the most specific assay, which the NULISA platform offers.
I can see these technologies, with their accuracy, sensitivity, and quantitation, ending up at the bedside as routine predictive tools for molecularly defined disease. When drug companies such as GSK (GlaxoSmithKline) want to figure out how to intervene in a pathway, they’re going to be intervening in the causative mechanism before the disease is entrenched and it’s too late to treat or reverse the disease.
I think the singleplex or three or four plex is a very important development. We need to keep bringing costs down. We also need to keep expanding the number of proteins detected so that when the literature tells us of new important biomarkers, an assay can immediately be designed and deployed.
Q: Alzheimer’s disease is often referred to as type 3 diabetes. Is this term relevant to talk about today?
Todd: We’re actively working with GSK on neurodegenerative disease and neuroinflammation, primarily the motor neuron disease, ALS (amyotrophic lateral sclerosis), Parkinson’s disease, and Alzheimer’s. All three have a causative inflammatory component. Some authors do refer to Alzheimer’s as type 3 diabetes. One reason for that is an clear co-association in populations of Parkinson’s and Alzheimer’s and with a diagnosis of type 2 diabetes. If you look at a cohort of Parkinson’s patients, the frequency of type 2 diabetes in that cohort will be two or three times that of the population level.
What does that mean? I looked at the genetic results for Alzheimer’s and other diseases, and the genetic susceptibility results for type 2 diabetes. Are any of the top hits for Alzheimer’s in common with type 2 diabetes? It’s not obvious, but I think the explanation for the concordance of diabetes with Alzheimer’s and Parkinson’s is related to the body-wide effects of aging. As we age, our immune systems become less regulated. There’s a phenomenon called chronic inflammation. Chronic inflammation may exacerbate the accumulation of certain proteins behind these neurodegenerative diseases: alpha synuclein, TDP-43, beta amyloid, and, of course, tau and its phosphorylation. Modifications of these proteins and formation of fibrils and plaques accumulate in neuronal cells as you get older. But it’s not only neurons that are affected but also other cell types that express higher than average levels of these proteins. For example, we have found that pancreatic beta cells, which produce the body’s insulin, also express relatively high levels of tau, which could be a partial explanation why type 2 diabetes is more common in Alzheimer’s and Parkinson’s patients than in the general population. All these proteins have normal healthy functions but with a combination of ageing, genetic variants and environmental exposures (such as alcohol, smoking, hearing loss), they become modified and end up forming cell-damaging plaques.
To take an example of an environmental exposure, if you were a very keen boxer or footballer or rugby player, you might have a history of head impact. We know that induces phosphorylation of tau in the brain, which is quite clearly an early causative factor in neuronal damage and neurodegenerative disease.
Three years ago there was a very good paper in one of the Nature journals, where they took all the data from GWAS publications and integrated it and asked, “Which … disease? is most closely associated with Alzheimer’s disease?” Type 1 diabetes was the answer. Parkinson’s, Alzheimer’s, and type 1 diabetes have HLA associations that overlap directly, and also with multiple sclerosis. Therefore, we know there’s an inflammatory component, but probably there’s an autoimmune component in all three neurodegenerative diseases, and of course that classic autoimmune disease, type 1 diabetes.
So, a lot of research to be done. What are the autoantigens? Last week there was a Nature paper saying one of the key proteins in ALS, C9orf72, there’s an autoimmune response to it. Equally interesting, a few months ago in the journal Lancet, there was trial reporting that in early ALS, they could reduce the death rate by giving low-dose IL2 to promote regulation of the immune system. We’re beginning to tease apart the inflammatory and probably autoimmune components of these common and often fatal neurodegenerative diseases.
Why is that a particularly exciting opportunity? Well, immunotherapy has really come on. There are more than a dozen drugs to treat autoimmune inflammatory diseases. IL-2 and its redirection towards ALS is just one of the several examples of these proven drugs that protect and treat inflammatory diseases such as rheumatoid arthritis and type 1 diabetes. They could be redeployed into individuals with early signs of neurodegenerative diseases, and in particular, individuals showing the inflammatory biomarkers that can be accurately measured with NULISA.
Gustav Ceder is a freelance science writer and photographer.
