José Pedro Castro on Inflammation and Aging

For the last several years, Dr. José Pedro Castro, a Gladyshev Lab alumnus, who is now Assistant Researcher and Project Principal Investigator at the Institute for Research and Innovation in Health at the University of Porto, has been studying inflammation and its crucial role in aging and disease. Earlier this month, he received the Rising Star Award at Longevity Summit Dublin.

At the summit, Dr. Castro presented his yet unpublished concept of “inflammatory fidelity”: maintaining a healthy, balanced inflammatory homeostasis, which gets disrupted with age. We discussed the idea that inflammation underlies both important functions and many processes of aging, and how future therapies might help us keep up this elusive youthful inflammatory profile.

What is your personal story of getting into the longevity field?

Unfortunately, I don’t have one of those amazing stories about why I joined the longevity field. I finished my degree in microbiology and then started a master’s project using blood from centenarians. We were looking for oxidative changes in proteins and asking whether those proteins were better preserved than in controls.

That’s how I got excited about aging. It’s one of those fields where you start working on one thing and suddenly find yourself in a completely different system. During my master’s, I began with T cells and immunosenescence. At the time, we knew much less about it, and most of the discussion was about oxidative stress and redox biology.

From there, I moved into muscle and adipose tissue and started seeing how closely those tissues are connected. I also began to think that oxidative stress alone couldn’t explain aging. I finished my master’s, did a Ph.D. on T-cell aging and protein aggregation, and then joined Tilman Grune’s group in Germany for my postdoc. We looked more and more at metabolism and aging – for example, how oxidative stress might mimic aging in fat or muscle.

Near the end of that postdoc, I became interested in Vadim Gladyshev’s lab. I read his paper introducing the “deleteriome,” the buildup of many different harmful changes with age. I couldn’t sleep for almost two days after reading it. I kept thinking: aging isn’t just oxidative stress or DNA damage. It’s a combination of many things, including some we may not even know about.

Vadim was taking a systems approach, which was new to me: everything is connected. There may be important hubs, of course, but no single pathway explains the whole process. I wrote to him with ideas for experiments, we connected, I got a grant from Germany, and I joined his lab. Suddenly, I was working across many systems – gene expression, DNA methylation, different tissues, different species.

That completely shifted my interests. If you want to understand aging, you need tools like multi-omics that show how the whole system changes, not just one protein or pathway. So my journey was from one protein and one model to trying to look at as many connected processes as possible. It’s difficult, but it gives you a much better sense of how complex aging really is.

Does that immense complexity ever frustrate you? Do you think we will understand enough to start reversing aging in the foreseeable future?

I don’t know whether we’ll reverse aging completely, because there are so many things we don’t even know to look for, but I think we can slow it meaningfully and perhaps reverse some parts of it.

What’s both exciting and exhausting is that an amazing paper now comes out almost every day, often with powerful new tools and huge datasets. You want to use everything. You see a result and immediately wonder how your model fits into it, but then it becomes very hard to coordinate all of that and still tell a clear story or make a specific contribution.

That’s one reason I decided to focus more narrowly on inflammation. Once I really started digging into inflammation and biological age, I found that despite the enormous literature, some important questions had barely been studied.

Inflammation has been studied for decades and is recognized as a hallmark of aging. At the same time, there is a growing sense that its role may be even broader – that it underlies many aging processes and could be one of the most universal targets for longevity therapies. How should we think about inflammation in aging today?

The term “inflammaging” did something very important: it made people aware that chronic inflammation probably contributes to aging and underlies many chronic diseases. But, there’s a bias. If you search for harmful inflammation and disease, the number of papers has risen almost exponentially. Research on adaptive, regenerative, or otherwise helpful inflammation has stayed much flatter.

Basically, people have focused almost entirely on the bad side, but we need inflammation to fight infections and repair tissues. I think we’ve overlooked that side of it.

That led us to a concept we call “inflammatory fidelity,” which I discussed in Dublin. We need to find inflammatory circuits linked to aging and accelerated biological age, but also different circuits linked to development, adaptation, and regeneration.

Embryos regenerate very quickly. The neonatal heart can still regenerate too, but it loses that ability soon after birth, and inflammation is involved. Using transcriptomic data from embryonic development through late life, we found two broad inflammatory circuits. One is very active during development and early life and then drops sharply. The other starts low and rises with age.

We combine their expression into a fidelity score – very simply, beneficial inflammation divided by harmful inflammation. The score falls in aging, chronic disease, and fibrosis. It rises in regeneration and in signatures linked to longer lifespan.

So, the score seems to track the state of a cell or organism. It could add something to biological-age measures, which may or may not capture inflammation. If you were running a clinical trial, for example, it might not be enough to measure IL-6 or C-reactive protein. Someone could have high pro-inflammatory markers but still have strong regenerative or adaptive pathways that buffer the damage.

Centenarians are a good example. Some studies find high IL-6 and other inflammatory markers in centenarians, yet they stay resilient until very late in life. That suggests they have compensatory mechanisms.

Context matters enormously too. When we activate one of the inflammatory circuits we found, the effect changes with age. In a middle-aged mouse, we see harmful cytokines, inflammatory macrophages, and changes in the immune system, but the mouse can still produce many regulatory T cells that try to hold the response back.

So the question isn’t simply whether inflammation is there. When did it start? What’s the context? IL-6, for example, can be pro-inflammatory in one setting and support regeneration in another.

Or when IL-6 rises acutely in response to exercise, right?

Yes, exactly. In the liver, for example, macrophages need to secrete IL-6 to promote regeneration. The outcome depends on timing, dose, the cell type, and the combination of signals.

Tell me if I understand this correctly: inflammatory fidelity is something like maintaining a healthy inflammatory homeostasis. We have a youthful blueprint in which the system is balanced, and with age, the harmful side begins to dominate. The idea is to bring it back to that more youthful state.

Exactly. That’s the next step for us. We now have candidate transcription factors that seem to control these circuits. We want to test them in different cell types: boost the factor controlling beneficial inflammation, reduce the one controlling harmful inflammation, and see which age-related features improve.

Take macrophages. Young macrophages can help the heart regenerate, while old macrophages promote fibrosis and damage. Can we restore inflammatory fidelity in old macrophages? If we reprogram them for cell therapy, or change them directly in vivo, can we improve regeneration after injury in the heart, liver, or other tissues? And would aging clocks show a younger state? We don’t know yet.

Another big gap is causality. We know inflammation tracks with aging and disease. We give rapamycin, and IL-6 falls; we try another intervention, and some marker goes up or down. But which inflammatory pathways actually make biological age move faster or slower?

In a separate project, we used transcriptomic clocks and their inflammatory modules to predict genes linked to accelerated transcriptomic aging. We chose the strongest, most consistent candidate across datasets such as Tabula Muris Senis. When we treated cells with an agonist for the receptor we found, their transcriptome shifted toward an older state. We’re also testing it in mice.

The main point is that it’s not enough to say inflammation rises with age or falls after a geroprotective treatment. We need to find the specific circuits that actually push biological age up or down.

The inflammatory-fidelity work is still unpublished, but you recently posted a preprint on an inflammation-based epigenetic clock. Tell me more about it.

The fidelity project came later and is still at an early stage. We have a lot of in silico evidence, but we’re only starting the lab validation.

The clock project came from that inflammatory circuit linked to accelerated transcriptomic age. We had a set of genes connected to that state. I contacted Csaba Kerepesi, a former member of Vadim’s lab and a friend, and we asked: if this signature is tied to transcriptomic aging and is strongly enriched for inflammation, can we use it to build an epigenetic clock?

Most epigenetic clocks are black boxes. They can predict very well, but you often don’t know what biology the selected CpG sites represent. Is it inflammation, DNA repair, or something else?

So, you went the opposite way – from the functionality.

Exactly. We started from the biology. We already had a little over 100 inflammation-related genes with a clear function. We knew this group of genes was heavily enriched for inflammation. So we thought: if it really captures inflammation, perhaps a clock built from these genes should pick up diseases in which inflammation plays a major role.

We selected CpG sites in the promoters of those genes and built the clock. It tracks epigenetic age and age acceleration in several chronic conditions, including immune and cardiovascular disorders and many cancers. One interesting analysis compared normal tissue, tissue next to a tumor, and the tumor itself. The predicted age rose from normal to adjacent tissue and then rose again in the tumor.

We also asked whether it could pick up rejuvenation. In OSKM partial-reprogramming data, the predicted biological age went down. To me, that was exciting. We wanted to show that if you start with a meaningful pathway linked to accelerated aging, you can build a clock that still detects age acceleration and deceleration.

OSKM validation is really interesting, but on the other hand it’s all epigenetic reprogramming. You could argue that the factors do something to the epigenome, including somehow recalibrating those genes – and that’s what you see in the readout.

Yes. That doesn’t prove inflammation itself was rejuvenated. OSKM directly remodels the epigenome, so it may simply be changing those promoters, but at least it suggests that this inflammation-related signal is involved.

Have you looked then at other validations, other therapies?

We want to. That work is underway. We’d like to test rapamycin and whatever other intervention datasets are available.

Did you test your inflammation-based clock against more established epigenetic clocks?

Yes. We compared it with established clocks, including Horvath, Hannum, GrimAge, and GrimAge2. They don’t agree perfectly in every dataset, but overall we saw similar trends. That mattered because our clock didn’t come from a search across the whole methylome. We restricted it to a small, biologically chosen gene set and still picked up many of the same signals.

Did you find any organ-specific inflammatory signatures?

Not yet. The current clock was built from blood data. We want to see whether signatures trained in different organs give better or different predictions and respond differently to interventions. Only one graduate student is working on it right now, so progress is slower than we’d like, but organ-specific clocks are an important next step.

Inflammation is clearly more multifaceted than the simple division into pro- and anti-inflammatory factors suggests. Your March preprint examines IL-10, normally considered an anti-inflammatory cytokine, and suggests that chronic IL-10 exposure can instead promote inflammaging and tissue senescence. How does a protective signal become harmful, and how relevant might that be to normal human aging?

I should be careful because I’m a co-author, not the first or senior author, and the paper is still being revised. But, the basic idea is that cytokines probably have an optimal range. You need IL-10 to resolve inflammation, but if it stays too high for too long, it can turn harmful.

The work began in Margarida Saraiva’s lab, which mainly studies tuberculosis. In mice with sustained IL-10 overexpression in CD4+ and CD8+ T cells, those cells were reprogrammed. They became highly inflammatory, entered several tissues, and were linked to tissue dysfunction and signs of aging.

So, we need to understand the right range, not just whether a cytokine is “good” or “bad.” We’re not going to solve this one cytokine at a time, and we can’t simply suppress the immune system because it still has to fight infections.

That’s why I’m cautious about spectacular intervention studies. IL-11 inhibition extends lifespan in lab mice, but what happens in a natural environment? If you shut down a major pathway, perhaps the animals become worse at fighting infection.

The message isn’t that IL-10 is bad. It’s that too much for too long can be bad, and the same may be true for other “anti-inflammatory” cytokines.

In UK Biobank data, higher blood IL-10 was linked to a small increase in mortality risk, although the result was variable and shouldn’t be overinterpreted. So, dose, timing, and context matter more than the simple pro- versus anti-inflammatory label.

That seems like a recurring problem in aging research. We arrive with a simple intervention: this factor is elevated, let’s suppress it; that one is reduced, let’s increase it. Only later do we begin to understand the network. The mouse environment may also be a major confounder, especially for inflammation and immune aging. Lab mice live in unusually protected conditions. Rapamycin can suppress inflammation and some immune responses, and the mice live longer, but how well does that translate to humans who are exposed to many more pathogens?

I agree. I use mice, so I have to work within those limits, but it’s an important problem. One interesting experiment would be to give young lab mice defined inflammatory challenges – viruses, bacteria, LPS, TNF-α – and then see how those exposures change their later trajectory. We’ve started smaller in vitro proof-of-concept work.

My hypothesis is that if young cells get the right amount and pattern of inflammatory stress, they may become more resilient rather than older. It’s still very early and speculative, but some preliminary observations make us want to pursue it, at least in immune cells.

I’m also becoming more interested in non-immune cells. We see inflammatory signatures rise quite strongly in many of these cells with age.

We usually focus on two sources of age-related inflammation: immune cells and senescent cells. You are suggesting that apparently non-senescent, non-immune cells can also participate actively in inflammatory signaling.

Yes. The circuit linked to accelerated transcriptomic age is overexpressed not only in immune cells but also in non-immune cells. We see this clearly in the kidney, where several non-immune cell types become strongly pro-inflammatory with age even without the classical signs of senescence.

It makes sense. When a cell is in danger, it has to tell the immune system and nearby cells that something is wrong. Virus-infected cells do this through interferons. We see a similar pattern with MIF, macrophage migration inhibitory factor. Many cell types seem to express and release MIF when they’re under stress, whether from DNA damage, loss of proteostasis, or something else.

The idea is that non-immune cells start releasing inflammatory signals in response to the damage they build up over time. Those signals could amplify inflammation locally and perhaps eventually across the body.

We see it in the tissue too. In liver and kidney histology, many non-immune cells stain strongly for the marker linked to accelerated inflammatory aging. So, inflammaging isn’t simply an old immune system attacking passive tissues. The tissues themselves may join in.

I also wanted to discuss your 2024 paper on age-associated clonal B cells. I have long been fascinated by clonal expansion, including CHIP, and its relationship to aging and malignancy. What did you find?

I love that paper. It’s one of my favorites, and I want to stress that Anastasia Shindyapina and I were co-first authors.

We hadn’t planned the project when I joined Vadim’s lab. Anastasia and I became obsessed with it during a different intervention study. Some old control mice had very enlarged spleens, and at first we didn’t know whether B cells, T cells, or something else were responsible.

Around the same time, two relevant studies came out. One, from the late Angelika Amon’s group at MIT, linked larger cell size to senescence and aging. Another cross-species study connected larger cells in one tissue with shorter lifespan. So cell enlargement looked like a possible sign of accelerated aging.

We also knew that old C57BL/6 mice often develop B-cell lymphoma. We asked a simple question: are B cells from old mice larger than those from young mice? That same day, with only two young and two old animals, we saw a huge difference. The old B cells were much larger. We thought they might enlarge with age, become dysfunctional, and grow more vulnerable to cancer.

We then compared young mice, old controls, and old mice with B-cell lymphoma, usually identified after their spleens became greatly enlarged. With age, B-cell receptor diversity fell and clonality rose, especially in lymphoma. We then asked how a normal B cell might move toward cancer.

Age-associated B cells, or ABCs, were already known to build up with age and appear in autoimmunity and chronic infections. We thought they might give rise to a clonally expanded population we called age-associated clonal B cells, or ACBCs.

Using public single-cell data, we found markers separating follicular B cells, ABCs, and the clonal cells in lymphoma, and then confirmed them in the lab. ABCs were common in old mice without lymphoma, while ACBCs were abundant in lymphoma. That suggested a path from follicular B cells to ABCs and then to ACBCs.

CellChat pointed to CD22 signaling as one possible driver of the increase in cell size. Follicular B cells grew larger when exposed to ABCs, and blocking CD22 reduced that effect. So, communication between B-cell populations may help push cells toward the ACBC state.

The clones also had internal changes: somatic mutations, epigenetic changes including promoter hypermethylation, and c-Myc activation. These could give some clones an advantage and let them expand. The cells were IgM-positive and didn’t follow the usual germinal-center route.

So, it isn’t one event. The aged environment changes, B-cell diversity shrinks, some cells acquire mutations and epigenetic changes, and particular clones expand.

The big question was whether this mattered in humans. Mouse and human B-cell lymphomas aren’t identical, and the reviewers pushed us to be careful, but the mouse ACBC signature overlapped with human follicular lymphoma and diffuse large B-cell lymphoma. In human data, we also saw B-cell receptor diversity fall and clonality rise with age and found clonal B cells carrying the ACBC signature in people over 50.

So, ABCs may be a useful sign of age-related B-cell problems. They may lead toward autoimmunity, infection-related dysfunction, or cancer, and we still don’t know what decides the path, but they appear in all of those settings.

We also found that inhibiting mTOR or c-Myc in old mice reduced premalignant B-cell changes, so the process may not be irreversible, and those pathways could be targets for prevention.

The implications may extend beyond individual diseases. I have a pet hypothesis that clonal immune aging could be a limiting factor in extreme longevity. Clonal expansions are common in centenarians and supercentenarians, and perhaps they help impose a ceiling on maximum lifespan.

I agree it’s worth exploring. Clonal expansion may be manageable for a long time and then eventually start limiting resilience, but we need much more evidence.

Looking ahead, what is the roadmap for longevity therapies that target inflammation? In an optimistic, almost science-fiction scenario, how much could they affect human lifespan?

Potentially a lot, although I’m less sure that targeting inflammation alone can slow biological aging itself. Most chronic diseases have an inflammatory component, and many are linked to accelerated biological age. So, if we learn to control the right switches – keeping helpful inflammation while reducing harmful inflammation – we could probably extend lifespan by delaying chronic disease.

Whether that changes the aging process itself is harder to say, because many other things are happening, but even if it mainly separates healthy aging from disease and compresses morbidity, the effect could still be big.

We’re still early. As this conversation shows, there are many basic things about inflammation that we still don’t understand.

What are you personally planning to do to move the field in that direction?

I see three main directions. First, I want to find biomarkers that predict a resilient inflammatory state in humans. That would be a dream. We have a cohort of patients with chronic kidney disease that may be a good place to start.

Second, we want to reprogram immune cells so they support regeneration and perhaps lower biological age. This connects directly to inflammatory fidelity: finding the right switches and restoring a younger balance between helpful and harmful inflammatory programs.

The third direction is more speculative. I want to expose young animals and young cells to carefully controlled stimuli and ask whether the right kind of priming – even with things we normally consider damaging – can make them more resilient or biologically younger. It’s related to hormesis, but focused on inflammatory conditioning. The project is still embryonic, but I’m very excited about it.