How the HIV Drug Maraviroc Rejuvenates the Aging Immune System & Boosts Strength and Coordination

RapAdmin · September 23, 2026, 5:43pm

As mammals age, the immune system becomes increasingly biased toward producing myeloid cells over lymphoid cells, driving chronic systemic inflammation and elevating the neutrophil to lymphocyte ratio. This paper demonstrates that bone marrow infiltrating CD4+ T cells acquire mitochondrial dysfunction over time, which activates the cGAS-STING pathway and triggers the secretion of the signaling protein CCL5. This protein binds to CCR5 receptors on hematopoietic stem cells, forcing them to overproduce neutrophils. Pharmacological blockade of the CCR5 receptor using the FDA approved antiviral drug maraviroc reversed this myeloid skewing, normalized circulating immune profiles, and significantly improved physical strength and motor coordination in aged mice.

The neutrophil-to-lymphocyte ratio is one of the least glamorous numbers on a standard blood panel. It costs nothing extra, falls out of any complete blood count, and predicts death from nearly every age-associated disease studied. It also climbs steadily with age. Until now the reason it climbs has been largely a matter of description rather than mechanism.

A team led by María Mittelbrunn and Enrique Gabandé-Rodríguez at the Centro de Biología Molecular Severo Ochoa in Madrid, working with collaborators across Spain, Italy, Switzerland, Germany, and the US National Institute on Aging, has now traced a specific causal chain. The culprit is a type of T cell that should not really exist in large numbers. CD4+ T cells are conventionally the coordinators of immunity, not the assassins. But with age a subset acquires the killing machinery of CD8+ cells: perforin, granzymes, the transcription factor EOMES. These cytotoxic CD4+ cells have been noted in aged humans and in supercentenarians, and their significance has been argued about for years.

The Madrid group found these cells piling up specifically in bone marrow. When they transferred T cells from old mice into young T-cell-deficient recipients, myeloid output jumped within seven days. Transfer young T cells instead and nothing much happens. The bone marrow, it turns out, is where these cells preferentially accumulate, more than spleen, liver, fat, or colon.

The mechanism runs through the mitochondria. Persistent antigenic stimulation wrecks mitochondrial function in these cells. Damaged mitochondria spill DNA into the cytosol, where the cGAS-STING pathway reads it as an infection signal and switches on CCL5 production. Mice engineered to have broken mitochondria only in their T cells developed the same picture at eight months of age that normal mice reach at twenty-four. That is a compression of roughly two-thirds of the mouse lifespan into a single genetic lesion, which is a strong argument that the mitochondrial defect is upstream rather than incidental.

The receiving end matters too. Aged hematopoietic stem cells and granulocyte-monocyte progenitors upregulate CCR5. In elegant competitive chimera experiments where half the marrow could see CCL5 and half could not, only the CCR5-competent progenitors expanded when old T cells arrived. The signal is being received, not merely broadcast.

The translational hook is that a CCR5 blocker already sits on pharmacy shelves. Maraviroc has been approved for HIV since 2007. One month of daily injections in old mice pushed the NLR back toward youthful levels, reduced neutrophils infiltrating lung and liver, and improved a spread of physiological markers. Young mice given the same drug showed essentially no change, which suggests the drug corrects a pathological signal rather than broadly suppressing immunity.

Actionable Insights

For individuals focused on longevity therapeutics, this paper provides highly practical value regarding the neutrophil to lymphocyte ratio (NLR). This ratio is easily obtained from standard complete blood count panels and serves as a highly sensitive, low cost proxy for systemic aging and bone marrow niche health. In large human cohorts a rising NLR ratio tracks with mortality risk across cardiovascular disease, cancer, and infection. This study supplies a mechanistic reason to take that number seriously rather than treating it as noise, because in mice the ratio is not a passive readout but a consequence of an active signaling loop that can be interrupted.

The application of a CCR5 inhibitor successfully lowered this ratio and reversed associated physical decline. In the study, 23 month old mice treated with maraviroc experienced an approximate 39 percent reduction in their neutrophil to lymphocyte ratio, dropping from a median of 0.33 to 0.20. This intervention generated physical improvements with substantial effect sizes. Grip strength increased by roughly 25 percent, representing an absolute gain of approximately 15 grams of force compared to vehicle treated controls. Furthermore, motor coordination measured by rotarod latency improved dramatically, with treated mice remaining on the apparatus for roughly 100 to 150 seconds longer than untreated age matched controls.

Maraviroc carries a black box warning for liver toxicity in humans. While maraviroc is currently available, it carries immunosuppressive risks that require careful clinical management.

Context and Source

Open Access Paper: Cytotoxic CD4+ T cells induce age-associated myelopoiesis through CCL5–CCR5 signaling
Primary institution: Centro de Biología Molecular Severo Ochoa (CSIC-Universidad Autónoma de Madrid), Madrid, Spain. Corresponding authors Enrique Gabandé-Rodríguez and María Mittelbrunn.
Country: Spain (lead), with international consortium.
Journal: Nature Aging (Springer Nature)
Dates: Published online 26 August 2026.
Impact evaluation: The impact score of this journal is 25.0 (2025 two-year Journal Impact Factor; five-year JIF 26.0), evaluated against a typical high-end range of 0–60+ for top general science, therefore this is a High impact journal.

Short answer

No. The paper does not establish that, and reading it that way conflates two different things: the absence of a pathological signaling axis with the absence of one downstream readout of it.

Why the young-mouse null does not transfer

The young-mouse arm is not a test of “normal NLR predicts non-response.” It is a test of “an animal lacking the entire upstream axis does not respond to a drug targeting that axis.” Young mice differ from old mice in at least four ways relevant here, and normal NLR is the last of them, not the first:

Young bone marrow has essentially no CCL5+ CD4 CTLs. The reported figure is 1.64 percent of CD4 cells in young versus 26.78 percent in old. There is almost no ligand being produced.
Young HSCs and GMPs have lower CCR5. 15.5 versus 21.9 percent on HSCs, 9.4 versus 24.0 percent on GMPs. There is less receptor to block.
Young T cells do not show the mitochondrial depolarization, cytosolic mtDNA, or STING activation that drives CCL5 in the first place.
Young mice have a normal NLR.

Maraviroc failed in young mice because items 1 through 3 were absent. Item 4 was a consequence of 1 through 3, not an independent cause. Blocking a receptor that is not being engaged does nothing regardless of what the blood count says. [Confidence: High]

The question the paper never asked

The design compares young versus old versus old-treated. It does not stratify old mice by baseline NLR and ask whether high-NLR animals responded better than low-NLR animals. That responder analysis is absent, and it is the analysis that would actually answer your question.

Notably, the data would have supported attempting it. The individual data points in the old vehicle NLR panel are widely dispersed, with some aged animals sitting close to young values and others roughly three times higher. With n of 7 the analysis would have been badly underpowered, but the variance is visibly there and goes unexamined. [Confidence: Medium, based on figure inspection rather than source data]

NLR is a downstream aggregate, not the target

This is the substantive reason to resist the inference. The drug target is CCR5 on myeloid progenitors. The proximate pharmacodynamic markers are CCR5 occupancy, GMP frequency, and LS−K frequency. NLR sits several steps downstream and is influenced by a long list of inputs that have nothing to do with this axis: acute infection, subclinical inflammation, corticosteroid exposure, smoking, recent exercise, circadian timing, and in humans the benign ethnic neutropenia associated with the Duffy-null phenotype.

An aged individual could therefore plausibly have a normal NLR while still carrying an activated CD4 CTL to CCL5 to CCR5 axis, with the neutrophil expansion masked or offset by something else. Conversely, an elevated NLR can be driven by an entirely unrelated process on which maraviroc would have no purchase. NLR is a screening signal for this biology, not a measure of it. [Confidence: High]

The paper’s own data argues mildly against your inference

Two of the maraviroc effects in old mice were not proportional to an NLR deficit.

Serum glucose was essentially identical between young and old vehicle mice, roughly 60 versus 62 mg/dl, meaning there was no aging gap at all. Maraviroc nonetheless pushed glucose down to approximately 48, below both. That is a pharmacological effect in a domain with no baseline abnormality, which is exactly the situation your question posits.

The respiratory exchange ratio improvements in both light and dark phases are also not obviously downstream of neutrophil count, and the authors do not attempt to link them mechanistically to NLR.

Neither observation is strong evidence, and the glucose finding is as likely an off-target effect as a benefit, but both show that the drug does things in aged animals that the NLR framing does not predict. [Confidence: Low to Medium]

What would actually be needed

To answer your question properly, someone would need to treat aged mice stratified by baseline NLR into tertiles, with adequate power, and test for a treatment-by-baseline interaction. That study does not exist. Until it does, the honest position is that the young-mouse null tells you the drug is not broadly immunosuppressive in a healthy animal, which is reassuring for safety, and tells you nothing about response prediction in aged animals.

The better candidate predictive biomarker, on this paper’s own logic, would be CCR5 expression on circulating myeloid progenitors or CD4 CTL burden, neither of which is a routine clinical measurement. NLR is being used here as a convenient proxy because it is cheap, not because it was validated as a response predictor. [Confidence: High]

Rapamycin Longevity News