KLRG1 identifies regulatory T cells with mitochondrial alterations that accumulate with aging

Soto-Heredero, Gonzalo; Gabandé-Rodríguez, Enrique; Carrasco, Elisa; Escrig-Larena, José Ignacio; Gómez de las Heras, Manuel M.; Delgado-Pulido, Sandra; Francos-Quijorna, Isaac; Blanco, Eva M.; Fernández-Almeida, Álvaro; Abia, David; Rodríguez, María Josefa; Fernández-Díaz, Cristina M.; Álvarez-Flores, María Beatriz; Ramírez de Molina, Ana; Jung, Sascha; del Sol, Antonio; Zorita, Virginia; Sánchez-Cabo, Fátima; Torroja, Carlos; Mittelbrunn, María

doi:10.1038/s43587-025-00855-9

Download PDF

Article
Open access
Published: 30 April 2025

KLRG1 identifies regulatory T cells with mitochondrial alterations that accumulate with aging

Nature Aging volume 5, pages 799–815 (2025)Cite this article

9394 Accesses
7 Citations
32 Altmetric
Metrics details

Subjects

Abstract

Recent studies using single-cell RNA sequencing technology have uncovered several subpopulations of CD4⁺ T cells that accumulate with aging. These age-associated T cells are emerging as relevant players in the onset of inflammaging and tissue senescence. Here, based on information provided by single-cell RNA sequencing data, we present a flow cytometry panel that allows the identification of age-associated T cell subsets in systematic larger analysis in mice. We use this panel to evaluate at the single-cell level mitochondrial and senescence marks in the different age-associated CD4⁺ T cell subpopulations. Our analysis identifies a subpopulation of regulatory T (T_reg) cells that is characterized by the extracellular expression of the co-inhibitory molecule killer cell lectin-like receptor subfamily G member 1 (KLRG1) and accumulates with aging in humans and mice. KLRG1-expressing T_reg cells display senescence features such as mitochondrial alterations, increased expression of cell-cycle regulators and genomic DNA damage. Functionally, KLRG1⁺ T_reg cells show a reduced suppressive activity in vivo accompanied by a pro-inflammatory phenotype.

Identification of human exT_reg cells as CD16⁺CD56⁺ cytotoxic CD4⁺ T cells

Article 10 August 2023

Harnessing the biology of regulatory T cells to treat disease

Article 16 December 2024

Telomeric injury by KML001 in human T cells induces mitochondrial dysfunction through the p53-PGC-1α pathway

Article Open access 02 December 2020

Main

With age, the immune system loses the ability to respond to infections, cancer or vaccination. Instead, it engages in autoimmune and pro-inflammatory responses, which favor tissue damage and the acquisition of a low-grade systemic chronic inflammation, known as inflammaging, increasing the risk of many age-related diseases¹. Thus, the decline of immune function during aging represents a major clinical challenge in many disease conditions including autoimmune diseases, infectious diseases, cancer and neurodegenerative and cardiovascular disorders^2,3,4. The use of single-cell RNA sequencing (scRNA-seq) has allowed a deeper characterization of the different subsets of CD4⁺ T cells that accumulate during aging. CD4⁺ age-associated T cells (TAAs) include effector/memory T cells; a population of regulatory T (T_reg) cells expressing activation genes (denoted as activated T_reg cells or aT_reg cells); cells with an exhaustion signature (denoted as exhausted); and cells overexpressing genes associated with cytotoxicity that have previously been described in the context of viral infections and cancer as CD4⁺ cytotoxic T cells (denoted as cytotoxic)⁵.

T_reg cells, a subtype of CD4⁺ T cells characterized by the expression of the forkhead box transcription factor (FOXP3), are required to maintain immune homeostasis and avoid excessive tissue damage preventing inflammatory and autoimmune diseases⁶. Understanding the molecular mechanisms behind age-associated changes in the T_reg compartment is critical to understand the deterioration of the immune system during aging and the consequences for immunosenescence and inflammaging. T_reg cells in aged mice are augmented in lymphoid organs (for example, spleen and lymph nodes)^7,8,9 and other nonlymphoid tissues such as the visceral fat and lungs¹⁰, whereas they are reduced in muscle¹¹. T_reg cells also undergo molecular changes with aging, becoming more dependent on IL-15 signaling owing to a lower expression of the IL-2 receptor CD25, while exhibiting enhanced levels of CD122 (IL-2/IL-15 receptor chain β)^12,13. Moreover, they acquire a memory-like profile during aging, characterized by the downregulation of CD62L and the upregulation of CD44 and CD69 (refs. ^14,15). These aged effector–memory T_reg cells display increased expression of canonical pro-inflammatory transcription factors such as T-BET and RORγT, and pro-inflammatory cytokines such as IFN-γ or IL-17A¹⁶. T_reg cells with pro-inflammatory phenotypes have been denoted as fragile T_reg cells in the context of cancer, although its function in antitumoral responses is still debated^17,18.

Killer cell lectin-like receptor subfamily G, member 1 (KLRG1) is a co-inhibitory receptor expressed by natural killer (NK) cells and antigen-experienced T cells. The expression of KLRG1 increases dramatically with age in human blood samples¹⁹. In mice, KLRG1 has been used to identify memory precursor cells from effector T cells. In acute viral infection models, KLRG1 distinguishes short-lived effector CD8⁺ T cells (KLRG1^hi) and memory precursor effector CD8⁺ T cells (KLRG1^lo)^20,21. Importantly, KLRG1 identifies senescent CD8⁺ T cells in humans and mice, as KLRG1⁺CD8⁺ T cells exhibit reduced cytokine production and senescence characteristics such as a reduced proliferative capacity^19,22,23,24. Regarding CD4⁺ T cells, KLRG1 also identifies terminally differentiated cells²⁵. Specifically in T_reg cells, KLRG1 is considered a marker of effector suppressive T_reg cells that display increased levels of activation markers (CD69, CD44, CD103 and CD39) and enhanced production of IL-10 (ref. ²⁶).

Mitochondria are the key controller of cellular metabolism, but also play a crucial role as signaling hubs for inflammation and cell death. Mitochondrial function declines with age in different cell types and tissues²⁷, including T cells^28,29, affecting the mitochondrial ATP generation as well as reactive oxygen species (ROS) production and cellular signaling³⁰. In addition, mitochondrial dysfunction is sufficient to drive cellular senescence in fibroblasts^31,32. In contrast to conventional T cells, which engage glycolysis during immune response, T_reg cells preferentially use mitochondrial respiration as source of ATP³³ and the genetic induction of T_reg-specific mitochondrial dysfunction leads to premature death by uncontrolled inflammation^34,35. Here, by using multiparametic spectral flow cytometry, we investigate the specific subpopulations of T cells that accumulate mitochondrial alterations during aging. Among them, we identify a population of T_reg cells characterized by the expression of KLRG1 (kT_reg cells) and FOXP3 that displays senescence features such as nuclear DNA damage. Although the suppressive capacity of kT_reg cells is maintained in vitro, it is compromised in vivo. kT_reg cells produce pro-inflammatory cytokines and, importantly, are also increased among human peripheral blood mononuclear cells (PBMCs) with age.

Results

T cells with mitochondrial alterations accumulate with aging

To investigate age-associated mitochondrial dysfunction in T cells, we examined mitochondrial mass and mitochondrial membrane potential (MMP) by flow cytometry in circulating CD4⁺ T cells from C57BL/6 mice at different ages ranging from 2 to 21 months old by combining MitoTracker Green (MtG) and MitoTracker DeepRed (MtDR) staining. As a control, we used oligomycin that, by inhibiting the proton ATPase, increased the MMP and therefore the MtDR signal (Extended Data Fig. 1a). Both the MtG and MtDR signals declined during aging in circulating CD4⁺ T cells (Extended Data Fig. 1b,c), suggesting that the proportion of T cells with mitochondrial alterations increases in aged mice (Fig. 1a,b). We obtained similar results using two additional mitochondrial probes: MitoTracker Red CM-X-ROS (Fig. 1c,d) and tetramethylrhodamine methyl ester (TMRM) (Extended Data Fig. 1d). Interestingly, MtG^loMtDR^lo cells, which accumulate with aging, showed a reduced MtDR/MtG ratio, suggesting mitochondrial defects (Extended Data Fig. 1e). To confirm that MtG^loMtDR^lo cells have altered mitochondria, we sorted MtG^hiMtDR^hi and MtG^loMtDR^lo CD4⁺ T cells from aged mice and performed electron microscopy (Fig. 1e,f). MtG^loMtDR^lo cells showed an increased percentage of mitochondria with altered morphology³⁶ identified as round, small and with unstructured or even lost mitochondria cristae (Fig. 1f,g).

**Fig. 1: CD4⁺ T cells progressively accumulate altered mitochondria during aging.**

Because the T cell subsets dramatically change with age and the most notable changes are the loss of naive T cell subsets and the increase in the effector memory pool³⁷, we assessed mitochondrial probes in naïve (CD62L^hiCD44^lo) and effector–memory (CD62L^loCD44^hi) CD4⁺ T cells in mice at different ages. We observed that, even in old mice, naive T cells were MtG^hiMtDR^hi, while effector–memory cells were characterized by MtG^loMtDR^lo, even in younger mice (Fig. 1h,i).

KLRG1 identifies age-associated regulatory T cells

To assess whether age-associated mitochondrial dysfunction equally affects all TAA subsets, we developed a multiparametric staining panel to identify the different age-associated CD4⁺ T cell subsets. Our panel included surface markers based on gene expression of surface proteins identified by scRNA-seq^5,38, complemented with known markers for immunosenescence, such as KLRG1 (ref. ¹⁹). In previous scRNA-seq experiments, splenic T cells from young mice were mainly classified as naive, naive_ISG15 and rT_reg cells, whereas T cells from old mice were characterized by the accumulation of TAAs such as aT_reg, effector memory T (T_EM) cells, exhausted and cytotoxic T cells⁵. Unbiased clusterization of spectral flow cytometry analysis of splenic CD4⁺ T cells from young (2-month-old) and old (21-month-old) mice resulted in seven subpopulations of T cells (Fig. 2a,b). Naive T cells were characterized by the expression of the adhesion molecule CD62L and the absence of the activation marker CD44. Three subsets of T_reg cells were identified by the expression of CD25 and discriminated by different expression levels of CD62L, CD44 and KLRG1. The other CD4⁺ T cell clusters presented marks of differentiation (CD62L^lo and CD44^hi) and were identified as cytotoxic based on the expression of NK-like markers, such as NKG2A and NKG2D, or as exhausted, based on the expression of classical exhaustion markers such as PD-1 and TIM3 (Fig. 2c,d). Young mice mostly displayed naive and rT_reg cells, while aT_reg, kT_reg, T_EM, cytotoxic and exhausted cells were increased in spleen from old mice (Fig. 2e,f). Importantly, this panel discriminated six out of the seven subpopulations previously defined by scRNA-seq analyses; however, because the panel is based on surface markers, we could not identify the ISG15_naive T cell subpopulation (Extended Data Fig. 2a,b). Our analysis identified a TAA subset characterized by the expression of KLRG1 (Fig. 2b and Extended Data Fig. 2a,b). In fact, circulating KLRG1⁺ CD4⁺ T cells progressively accumulated with aging (Extended Data Fig. 2c). Notably, these cells co-expressed KLRG1 and CD25, suggesting that KLRG1⁺ CD4⁺ T cells are probably a subpopulation of T_reg cells (Fig. 2c,d). To confirm whether these CD25⁺ KLRG1⁺ cells were T_reg cells, we combined the panel with the intracellular marker FOXP3 and we found that most of KLRG1⁺ cells expressed CD25 and FOXP3 (83–87%) (denoted as kT_reg cells), supporting their T_reg identity (Fig. 2g). The percentages of splenic kT_reg cells assessed by spectral flow cytometry gradually increased with age within both the T_reg compartment and CD4⁺ T cells (Extended Data Fig. 2d,e).

**Fig. 2: Multiparametric spectral flow cytometry identifies a new age-associated T_reg subset.**

In addition, we reanalyzed existing and publicly available scRNA-seq data from young and old mice to separate the novel kT_reg cluster⁵. By increasing the depth of the clusterization, we obtained an additional cluster, which accumulated during aging, within the aT_reg subset (Extended Data Fig. 3a,b). Interestingly, the most determinant genes of this cluster were reminiscent of kT_reg cells (Supplementary Table 1). These cells showed increased gene expression of Klrg1 and altered expression of genes related to T_reg function compared with aT_reg cells, such as Icos and Maf (Extended Data Fig. 3c). We also found that kT_reg cells expressed higher levels of inflammatory genes such as those belonging to the s100a family (including S100a4, S100a6, S100a10, S100a11 and S100a13) and inflammation-related proteins such as galectins (Lgals1 and Lgals9), IL-18 receptor (Il18r) and IFN-γ receptor (Ifngr1) (Extended Data Fig. 3d).

Then, we investigated the tissue distribution of kT_reg cells. To map their preferential location in the different tissues, we used the multiparametric flow cytometry panel to analyze T cells from different tissues including spleen, bone marrow, liver, colonic lamina propria, white adipose tissue and Peyer’s patches from young and old mice. kT_reg cells were found in all these tissues (Fig. 3a), and interestingly, the highest percentage of kT_reg cells was observed in the colonic lamina propria (Fig. 3b,c). Importantly, the absolute numbers and the percentage of kT_reg cells were increased in different tissues during aging (Fig. 3d) and their percentage was increased in spleen, bone marrow, liver and Peyer’s patches but not in lamina propria and white adipose tissue (Fig. 3e). Altogether, we have developed a multiparametric flow cytometry method that allows the identification of TAA subsets in both lymphoid and nonlymphoid organs. Using this strategy, we have identified a subset of T_reg cells characterized by the expression of KLRG1 that accumulates with aging.

**Fig. 3: kT_reg cells accumulate peripheral tissues during aging and are preferentially located in the colonic lamina propria.**

KLRG1⁺ T_reg cells show mitochondrial alteration and senescence features

To evaluate if any of the CD4⁺ T cell subsets are more susceptible to the age-associated mitochondrial dysfunction, we combined our spectral flow cytometry panel together with the assessment of mitochondrial probes in young and old mice. Old CD4⁺ T cells showed an increased proportion of MtG^loMtDR^lo cells in all subsets compared with young CD4⁺ T cells (Fig. 4a,b). Strikingly, even within the same mouse, there is a remarkable difference in the percentage of MtG^loMtDR^lo in the distinct T cell subsets. While the naive cluster mainly exhibits low frequencies of MtG^loMtDR^lo cells, T_EM cells displayed the major accumulation of MtG^loMtDR^lo cells in both young and old mice. Among T_reg cells, kT_reg cells showed an increased percentage of MtG^loMtDR^lo cells compared with other T_reg subsets in both young and old mice (Fig. 4a,c), suggesting that kT_reg cells are terminally differentiated T_reg cells with mitochondrial alterations.

**Fig. 4: TAAs have different predisposition to mitochondrial perturbations.**

To better characterize the different T_reg subpopulations, we compared the transcriptomic fingerprint of rT_reg, aT_reg and kT_reg cells isolated from young (2-month-old) and old (26-month-old) mice by bulk RNA-seq. This transcriptomic dataset allowed us to improve the resolution of the kT_reg cell expression profile. The clusterization grouped the samples by their identity rather than the age of the mice (Fig. 5a,b). We then analyzed the differentially expressed genes between the two age-associated T_reg subsets, aT_reg and kT_reg. We found Klrg1, Il1rl1 and Cd200r1 as the most significantly increased genes in kT_reg cells. Other genes upregulated in kT_reg cells included Foxp3, Gata3, Ahr, NK-related genes (Klra8), granzymes (Gzmb, Gzma and Gzmc) and chemokine receptors (Ccr3 and Ccr10) (Fig. 5c). The pathway analysis showed that kT_reg cells exhibit increased expression of cellular senescence genes and P53 target genes, decreased expression of genes related to mitochondrial function and increased expression of inflammatory genes (Fig. 5d).

**Fig. 5: KLRG1 identifies T_reg cells with senescence features.**

Given that a mitochondrial dysfunction is a trigger of cellular senescence³¹, and that the transcriptional fingerprint of kT_reg cells showed increased expression of senescence-related genes, we interrogated whether kT_reg cells display other features of cellular senescence beyond mitochondrial alterations. We analyzed the expression of the cell cycle regulators P16 and P21, which are increased in CD4⁺ T cells during aging (Fig. 5e,f). We compared the three different subsets of T_reg cells and found that kT_reg cells displayed the highest expression of P16 and P21 (Fig. 5g,h). Moreover, as an additional marker of cellular senescence, we analyzed the genomic DNA damage measured as the phosphorylation of the histone H2AX in the serine 139 (γH2AX), a marker of DNA double strand breaks that is increased in CD4⁺ T cells from old mice (Fig. 5i). Interestingly, kT_reg cells also displayed increased DNA damage compared with aT_reg cells (Fig. 5j).

KLRG1⁺ T_reg cells express ST2 and are induced by IL-33 in vivo

In light of the increased expression of Gata3 in kT_reg cells (Fig. 5c), and to further characterize the kT_reg subset, we compared the expression of different transcription factors related to T cell function by flow cytometry. This analysis revealed that kT_reg cells expressed higher levels of FOXP3 compared with the other T_reg subsets (Fig. 6a and Extended Data Fig. 4), whereas the levels of RORγT and T-BET were similar to aT_reg cells. Importantly, kT_reg cells expressed higher levels of GATA3 than aT_reg cells in both young and old mice (Fig. 6a and Extended Data Fig. 4), supporting the RNA-seq results. TOX and MAF expression was higher in kT_reg cells compared with rT_reg and aT_reg cells, but this difference was not observed in old mice (Fig. 6a and Extended Data Fig. 4).

**Fig. 6: KLRG1⁺ T_reg differentiation depends on the IL-33–ST2 axis.**

GATA3 is essential for the regulatory function of T_reg cells^39,40. Given that GATA3-expressing T_reg cells are characterized by the expression of the IL-33 receptor ST2 (refs. ^41,42), we analyzed the expression of ST2 in the different CD4⁺ T cell clusters. We observed that kT_reg cells expressed higher levels of ST2 than the rest of T cell subsets in both young and old mice (Fig. 6b). Based on that, we tried to induce the differentiation of kT_reg cells in vivo by administration of IL-33 in young mice. Before IL-33 administration, we injected IL-2, which induces the expansion of T_reg cells due to their expression of the IL-2 receptor CD25 (ref. ⁴³) (Fig. 6c). Upon this treatment, we found an expansion of T_reg cells in the injected mice (Fig. 6c) and an increase in the percentage and absolute number of kT_reg cells (Fig. 6e,f). Interestingly, this treatment seems to induce the differentiation of aT_reg cells into kT_reg cells (Fig. 6g), suggesting that kT_reg cells rely on ST2–IL-33 signaling for differentiation.

KLRG1⁺ T_reg cells exhibit a pro-inflammatory phenotype

To investigate the function of kT_reg cells, we analyzed the suppressive function of these cells both in vitro and in vivo. First, we isolated T_reg cells from old mice, sorted them on the basis of KLRG1 expression and co-cultured them with naive T cells loaded with CellTrace Violet at a 1:2 ratio (1 T_reg:2 naive). Both KLRG1⁺ and KLGR1⁻ T_reg cells showed a similar suppressive capacity in vitro (Extended Data Fig. 5a,b). Next, we performed an in vivo suppression assay by injecting naive CD45.1 CD4⁺ T cells alone or in combination with CD45.2 KLRG1⁻ or KLRG1⁺ T_reg cells into T cell-deficient mice (CD3ε^−/− CD45.2) (Fig. 7a). While KLRG1⁻ T_reg cells were able to restrain the weight loss induced by naive T cells, mice injected with naive T cells together with KLRG1⁺ T_reg cells lost weight to a similar extent as mice exclusively injected with naive T cells (Fig. 7b). Four months after the adoptive transfer, CD45.1 cells were more abundant in mice injected with KLRG1⁺ T_reg cells (Fig. 7c). In addition, we found that transferred CD45.1 CD4⁺ T cells were more activated when co-transferred with KLRG1⁺ T_reg cells, as shown by their increased expression of CCL5 and PD-1, than when co-transferred with KLRG1⁻ T_reg cells (Fig. 7d). Importantly, we observed an increased prevalence of fibrotic colon and fecal blood in mice injected with naive T cells alone or together with KLRG1⁺ T_reg cells than in mice injected with naive T cells together with KLRG1⁻ T_reg cells (Fig. 7e,f). Altogether, these data suggest that the function and/or viability of KLRG1⁺ T_reg cells in vivo is compromised.

**Fig. 7: Analysis of the suppressive and pro-inflammatory function of KLRG1⁺ T_reg cells.**

To further explore the activity of kT_reg cells, we assessed their capacity to produce cytokines by ex vivo intracellular flow cytometry. We found that kT_reg cells from old mice expressed higher amounts of the anti-inflammatory cytokine IL-10 than the rest of the T_reg subsets (Fig. 7g). We found no differences in the production of IL-17 or IFN-γ by kT_reg cells compared with aT_reg cells, but kT_reg cells produced more TNF (Fig. 7g). Similar results were observed when comparing T_reg subsets from young mice (Extended Data Fig. 5c). Furthermore, we measured the cytokine secretion of the different T_reg subsets. We induced kT_reg expansion in vivo by IL-33 administration and sorted rT_reg, aT_reg and kT_reg cells. After 24 h in vitro, we assessed the cytokine release by multiplex analysis of the culture media and we observed that, in agreement with our flow cytometry data, kT_reg cells secreted higher amounts of IL-10 together with pro-inflammatory cytokines than the other T_reg subsets (Fig. 7h). Notably, kT_reg cells produced higher levels of senescence-associated secretory phenotype (SASP)-related factors including IL-1β, IL-6, IL-13, IFN-γ, GM-CSF, CXCL2, CXCL1, CCL1 and CCL3 than rT_reg or aT_reg cells (Fig. 7i). Noteworthy, upon activation, aT_reg cells secreted similar or even higher levels of most cytokines than kT_reg cells. However, kT_reg cells maintained increased production of IL-10, IL-6, GM-CSF and CXCL2 (Extended Data Fig. 5d,e). Finally, to investigate the function of kT_reg cells in vivo, we adoptively transferred KLRG1⁻ and KLRG1⁺ T_reg cells into CD3ε^−/− mice (Extended Data Fig. 5f). In agreement with the increased production of pro-inflammatory cytokines, we found that mice transferred with KLRG1⁺ T_reg cells showed splenomegaly together with increased percentages of circulating neutrophils and basophils compared with noninjected or mice injected with KLRG1⁻ T_reg cells (Extended Data Fig. 5g,h).

KLRG1⁺ T_reg cells increase in human blood samples during aging

We wondered whether the kT_reg population is also increased in humans during aging. We used flow cytometry to analyze the presence of kT_reg cells in PBMCs from a human cohort of 144 healthy volunteers separated into two groups: young (ranging from 18 to 25 years old) and senior (over 55 years old) (see Extended Data Fig. 6 for the gating strategy). We found an increased frequency of kTregs in the senior population (14.56% of T_reg cells) compared with young individuals (6.35% of T_reg cells) (Fig. 8).

**Fig. 8: KLRG1⁺ T_reg cells are increased in human blood during aging.**

Altogether, by using deep flow cytometry analysis, we have identified a subset of T_reg cells characterized by the expression of KLRG1 and ST2. Besides typical T_reg cytokines such as IL-10, kT_reg cells produce other SASP-related cytokines and chemokines such as IL-1β, IL-6, IL-13, IFN-γ, GM-CSF, CXCL2, CXCL1, CCL1 and CCL3. kT_reg cells also harbor senescence features such as mitochondrial decline, expression of cell cycle inhibitors and DNA damage. These kT_reg cells with senescence features are induced in response to IL-33 and accumulate with aging in different tissues, but they are especially abundant in the gut in mice. Importantly, circulating kT_reg cells are increased in humans during aging.

Discussion

T cells play a critical role in recognizing and depleting damage or infected cells and sustaining tissue homeostasis preventing inflammation and tissue deterioration²¹. However, like every cell type in the body, T cells undergo changes as they age, leading to diminished function and compromising tissue homeostasis. Recently, scRNA-seq technology has enabled the decoding of the heterogeneity of age-associated CD4⁺ and CD8⁺ T cells that accumulate over time^5,38. Using this information, we have developed a multiparametric panel of 20 antibodies, which allows routine discrimination of T cell heterogeneity during aging by spectral flow cytometry. With this strategy, we have mapped the different TAAs that accumulate during aging in different tissues and we have identified KLRG1-expressing T_reg cells as a subset accumulating in blood, lymphoid and nonlymphoid tissues during natural aging and harboring aging features, including mitochondrial alterations.

Functional T_reg cells are essential to restore tissue homeostasis and to prevent autoimmunity and inflammation. Although it is well established that T_reg cells are augmented in several organs during aging in both humans and mice, whether their function is compromised with aging is still a matter of debate. Some authors showed that young and old T_reg cells have equivalent suppressive capacity^44,45, whereas others showed an age-related increase in T_reg function linked to increased risk of malignancies and infections in older adults^7,14. By contrast, there is also evidence of a deterioration of T_reg function during aging. Old T_reg cells are not able to suppress IL-17⁺ T cells upon chronic inflammation⁴⁴, failing to control hypersensitivity⁴⁶ and producing reduced amounts of IL-10 (ref. ⁴⁷).

KLRG1 is a marker of highly differentiated T cells, and it has been considered as a marker of immunosenescence²¹. In addition, a subset of terminally differentiated T_reg cells that express KLRG1 has been identified as short-lived T_reg cells expressing higher levels of suppressive (that is, CTLA-4 and IL-10) but also inflammatory molecules²⁶ located in nonlymphoid barriers, metabolic tissues or at sites of inflammation^9,48. For example, KLRG1⁺ T_reg cells contribute to insulin resistance⁴⁹ and IFN-γ- and IL-10-secreting KLRG1⁺ T_reg cells also infiltrate the central nervous system during experimental autoimmune encephalomyelitis⁵⁰. Using our flow cytometry panel, we mapped the presence of kT_reg cells in different tissues including spleen, bone marrow, liver, colonic lamina propria, white adipose tissue, Peyer’s patches and blood during aging. Our results demonstrate that kT_reg cells increase during aging in most of the analyzed tissues and that the colonic lamina propria is a preferential location for kT_reg cells, reaching 50% of the total T_reg pool. Importantly, T_reg cells are critical for gut homeostasis^51,52, which is essential for healthy aging.

Our results also support that kT_reg cells express the IL-33 receptor ST2 and that their differentiation depends on IL-33. IL-33 is an interleukin-1-like cytokine that plays a critical role in mucosal immunity, favoring Th2 differentiation⁵³. IL-33 is increased in the adipose tissue with aging⁵⁴, and it induces thymic involution and T cell dysfunction⁵⁵. Further research is required to define whether IL-33 levels change in lymphoid tissues or in the circulation during aging or if the differentiation of kT_reg cells depends on other factors.

By combining the multiparametric panel with mitochondrial probes, we assessed which of the different T cell subsets that accumulate during aging show mitochondrial dysfunction. We found that, whereas naive T cells are MtG^hiMtDR^hi, even in old mice, TAAs show increased percentage of cells with low MtG and MtDR stainings. Within T_reg cells, we found that kT_reg cells show higher percentage of MtG^loMtDR^lo, indicating that they harbor reduced mitochondrial mass and MMP. The MMP is an important indicator of mitochondrial integrity and function. During the mitochondrial respiration, electrons are ejected from the mitochondrial matrix to the intermembrane space to generate an electrochemical gradient. An optimal MMP is required for most of the mitochondrial functions including the generation of ATP, the uptake of ions, the synthesis of iron–sulfur clusters and others. When the MMP is not recovered, unpolarized mitochondria are degraded by mitophagy⁵⁶. The simultaneous assessment of mitochondrial mass and MMP by flow cytometry have been extensively used to characterize the health status of mitochondria in different tissues, including T cells, under different conditions such as tumor environment^57,58,59,60. Nevertheless, changes in the molecular composition of mitochondria, particularly in cardiolipins, lead to a reduction in the binding of MitoTracker probes in intraepithelial lymphocytes that, besides their low signal of MtG and MtDR, showed a similar mitochondria number on electron microscopy images⁶¹. Importantly, a reduction in the staining of these probes could also be explained by changes in the plasma membrane potential or by the expression of multidrug resistance transporters that expel the fluorescent probes⁶². To confirm that the changes that we observed in MtG and MDR stating are related to mitochondria, we sorted MtG^hiMtDR^hi and MtG^loMtDR^lo CD4⁺ T cells and assessed mitochondrial morphology by electron microscopy. Electron microscopy analysis shows that MtG^loMtDR^lo cells have mitochondrial changes including signs of cristae remodeling, supporting that MtG^loMtDR^lo T cells have mitochondrial alterations. However, additional research is needed to determine whether the MtG^loMtDR^lo T cell population also has structural or biochemical changes in the composition of mitochondria, in the plasma membrane potential and/or in the expression of multidrug transporters that could explain the dramatic reduction in MtG and MtDR staining. Importantly, beyond the key role of mitochondrial metabolism for the suppressive function of T_reg cells³⁵, a proper mitochondrial metabolism of T_reg cells is essential for maintaining liver homeostasis during aging. Deleting Altre, a long noncoding RNA found in T_reg cells and upregulated with aging, leads to mitochondrial dysfunction, causing liver inflammation, fibrosis and cancer in aged mice⁶³. Therefore, the increase in kT_reg cells with impaired mitochondria could be detrimental with aging. However, the function of kT_reg cells during aging and age-associated diseases requires further exploration.

Our findings demonstrate that, besides mitochondrial alterations, kT_reg cells harbor other signs of senescence. A recent study found that T_reg cells undergo senescence faster than conventional T cells through a ROS-dependent mechanism⁶⁴. In addition, both mitochondrial decline and cellular senescence are tightly interconnected^31,32. An important number of senescent-associated changes are dependent on mitochondrial function, especially the pro-inflammatory phenotype. Therefore, the pro-inflammatory phenotype observed in kT_reg cells could be a consequence of their mitochondrial dysfunction. The mitochondrial damage leads to the release of mitochondrial DNA (mtDNA) to the cytosol. mtDNA fragments trigger the activation of the cGAS–STING pathway, leading to the production of type I interferons, such as IFN-β, and pro-inflammatory cytokines, such as TNF or IL-6 (ref. ⁶⁵), which are part of the SASP⁶⁶. Cytosolic mtDNA, especially the oxidized mtDNA, also activates NLRP3 promoting the assembly of the inflammasome, a structure that leads to the activation of caspase I, therefore leading to the release of active IL-18 and IL-1β by the pyroptotic cell death⁶⁷. As kT_reg cells harbor mitochondrial damage, they could also experience a release of mtDNA that would trigger their pro-inflammatory phenotype. Moreover, the mitochondrial dysfunction also contributes to lysosomal stress favoring a pro-inflammatory phenotype in CD4⁺ T cells⁶⁸.

Surprisingly, there is recent evidence supporting the idea that senescence could be a reversible state. Chemotherapy-induced senescent cells can be released from senescence and reenter into the cell cycle⁶⁹. In addition, tissue senescence can be transitorily induced to promote tissue regeneration in Hydractinia symbiolongicarpus⁷⁰. Regarding CD4⁺ T cells, a recent study demonstrated that the expression of the senescence marker KLRG1 is transient in a subset of resident T_reg cells, despite their maintenance of an activated phenotype⁷¹. Because the specific ablation of mitochondria from senescent cells is sufficient to reverse many features of the senescent phenotype⁷², mitochondria are a promising target for interventions aimed at reducing the harmful effects of senescence in aging tissues. Interestingly, mitochondria dysfunction in exhausted CD8⁺ T cells can be reversed by scavenging ROS or boosting mitochondrial function with nicotinamide riboside^57,73,74, suggesting that mitochondrial stress is reversible and that mitochondria could be key to restoring T cell function. Future research should clarify whether mitochondrial damage and other features of senescence are targetable in kT_reg cells.

Sustaining a proper function of T cells during aging is essential for healthy aging. Unraveling the changes in different T cell compartments will help to better understand the underlying mechanisms of age-associated immunological decline and, thus, to develop therapeutic interventions that enhance immunity in older adults.

Methods

Animal procedures

C57BL/6J HccRsd mice were purchased from Envigo or generated at the Centro de Biología Molecular Severo Ochoa (Madrid, Spain) animal facilities. All mice required for this study were bred and aged in specific-pathogen-free conditions in the animal facility of Centro de Biología Molecular Severo Ochoa (Madrid, Spain). All mice were housed in ventilated cages within animal rooms maintained under a 12 h–12 h light-dark cycle. Animal rooms were temperature and humidity controlled. Standard diet and water were available ad libitum. Both male and female mice were used in this study.

All the procedures with animals were previously evaluated and approved (PROEX 287/16 and PROEX 52.1/23) by the ethics commitee on animal experimentation of the CBMSO, the authorized commitee of the Spanish National Research Council or the Universidad Autónoma de Madrid, and the regional government (Comunidad de Madrid). All mice were checked for any macroscopic abnormalities (according to the Jackson guide ‘AGED C57BL/6J MICE FOR RESEARCH STUDIES’). Mice were used at different ages: young (less than 4 months of age), adult (4–20 months of age) and old (over 20 months of age).

CD3ε^−/− mice (JAX stock no. 004177) were kindly provided by Prof. B. Alarcón (CBM). CD45.1 (JAX stock no. 002014) mice were kindly provided by Dr. C. Cobaleda (CBM). Both strains were in C57BL/7 background.

IL-2 + IL-33 treatment

The IL-2 immune complex was prepared as previously described⁴³. In brief, 1 µg of IL-2 (Peprotech, #212-12) was mixed with 5 µg of α-IL-2 (BioLegend, #503706) per mouse and incubated at room temperature (RT) for 30 min. Each mouse was intraperitoneally injected for three consecutive days. As a control, mice received only 200 μl of NaCl 0.9%. From the third day, IL-2 was combined with IL-33. 1 µg of IL-33 (BioLegend, #580506) was injected in 200 μl of NaCl per mouse for five consecutive days. As a control, mice received only 200 μl of NaCl.

Adoptive transfer of T_reg cells

For the study of the function of kT_reg cells in vivo, KLRG1⁻ and KLRG1⁺ T_reg cells (CD4⁺CD25⁺) were isolated by fluorescence-activated cell sorting (FACS) from IL-33-treated mice and injected into CD3ε^−/− young mice for three consecutive days. Approximately 8 × 10⁵ total T_reg cells were injected per mouse. Mice were weighed and monitored every 3 days for 4 months after the adoptive transfer and then were euthanized and the spleens were analyzed.

Suppression assays

For the in vivo suppression assays, KLRG1⁻ and KLRG1⁺ T_reg cells (CD4⁺CD25⁺) were isolated by FACS from IL-33-treated mice. Naive CD4⁺ T cells were purified by magnetic isolation (Stemcell Technologies, #19765) from young CD45.1 mice. A total of 6 × 10⁵ naive CD4⁺ T cells alone or in combination with 1.5 × 10⁵ KLRG1⁻ or KLRG1⁺ T_reg cells were intravenously injected into CD3ε^−/− young mice. The animals were weighed and monitored every 3 days for 4 months after the adoptive transfer and then were euthanized and the colonic lamina propria was analyzed.

For the in vitro suppression assays, KLRG1⁻ and KLRG1⁺ T_reg cells (CD4⁺CD25⁺) were isolated by FACS from old mice. Naive CD4⁺ T cells were purified by magnetic isolation from young CD45.1 mice. CD45.1⁺ naive CD4⁺ T cells were labeled with 1 µM of CellTrace Violet (Invitrogen, C34571) in PBS for 30 min at 37 °C. A total of 2.5 × 10⁴ naive T_reg cells were incubated with 1.25 T_reg cells for 3 days. Cells were stimulated with Mouse T-Activator Dynabeads (Gibco, 11452D) following the manufacturer’s instructions. The percentage of suppression was determined as follows: 100 − (% of proliferating cells with T_reg cells)/(% of proliferating cells without T_reg cells).

Human samples

Volunteer recruitment was performed through the GENYAL Clinical Trials Platform of IMDEA Alimentación (Madrid, Spain). This study was approved by the institutional Research Ethics Committee (IMDEA Food Foundation, IMD PI-052 and IMD PI-055) and performed in accordance with the principles of research involving human subjects stated in the Declaration of Helsinki (1964). All patients were clearly informed about the study methodology and provided written informed consent.

Peripheral blood samples were collected from a total of 117 volunteers belonging to two different population groups, whose inclusion criteria were as follows:

Young healthy population (n = 42): 14 male and 28 female healthy volunteers aged between 18 and 25 years.
Senior population (n = 75): 28 male and 47 female volunteers over 55 years of age.

Exclusion criteria were common for both study groups and included: decreased cognitive function, pregnancy or breastfeeding, severe chronic health conditions (for example, chronic kidney, liver and heart disease), immunodeficiencies and autoimmune diseases, and immunosuppressive or psychotropic pharmacological treatment. The results shown in the study apply to both sexes. The experimental groups have been defined by the age or the participants, and no sex- and gender-based analyses have been performed. Volunteers received financial compensation after the study.

Tissue processing for flow cytometry

Mice were euthanized with CO₂ followed by perfusion with cold PBS. The indicated tissues were extracted and processed as specified.

Spleen and lymph nodes

Lymph nodes were collected from inguinal, mesenteric, cervical and axillar areas. Spleen and lymph nodes were mashed and filtered through a 70-μm cell strainer. The cell suspension was centrifuged at 300g for 5 min at 4 °C. Red blood cells were removed using 5 ml of erythrocyte lysis buffer (ammonium chloride 0.15 M, sodium bicarbonate 0.01 M and EDTA 0.0001 M) for 5 min. Cells were washed, centrifuged, resuspended and counted.

Blood

Blood was extracted from either the facial vein or the heart in living or euthanized mice, respectively. The cell suspension was centrifuged at 300g for 5 min at 4 °C. Red blood cells were removed resuspending the cells in 5 ml of erythrocyte lysis buffer for 5 min. Cells were washed, centrifuged, resuspended and stained.

Colonic lamina propria

T cells from colonic lamina propria were isolated as previously reported⁷⁵. Colon samples between the cecum and rectum were obtained and cleaned from fat and feces. Tissues were cut longitudinally, washed with cold PBS and then cut transversally into 1-cm-long fragments, mixed in prewarmed 5 mM EDTA, 14 mH HEPES, 10% FBS PBS and incubated under shaking at 180 rpm for 30 min at 37 °C. After washing with PBS, tissue pieces were then minced and mixed in prewarmed 25 mM HEPES, 10% FBS RPMI supplemented with 300 U ml⁻¹ collagenase type VIII (Sigma, C2139) under shaking at 180 rpm for 45 min at 37 °C. The digested tissue was filtered through a 70-μm cell strainer, washed with 5 mM EDTA, 14 mH HEPES, 10% FBS PBS and centrifuged at 650g for 5 min at RT. To further enrich in leukocytes, supernatants were centrifuged in a 40%/70% Percoll gradient (Sigma, GE17-0891-01) at 750g for 20 min at RT with minimum acceleration and without brake. Isolated cells were washed with PBS and resuspended in 2% FBS RPMI for counting.

White adipose tissue

Gonadal white adipose tissue was obtained from the mouse abdominal cavity and mixed in 2 mg ml⁻¹ BSA, 2% FBS RPMI supplemented with 2 mg ml⁻¹ collagenase type II (Sigma, C6885) under shaking at 180 rpm for 40 min at 37 °C. The digested tissue was vertically rested to separate fat from the aqueous phases, which were obtained using an 18 G syringe. Then, the cell suspensions were filtered through a 70-μm cell strainer and washed with 2% FBS RPMI. Finally, erythrocytes were removed by incubation with a lysis buffer for 5 min at 4 °C, washed with 1 mM EDTA PBS and finally resuspended in 1 ml of 1 mM EDTA 2% FBS PBS for counting.

Peyer’s patches

Peyer’s patches were collected from the intestine and mashed into a 70-μm cell strainer. The cell suspension was centrifuged at 400g for 5 min at 4 °C. Finally, the cell pellets were resuspended in 1 ml of 2% FBS RPMI for counting.

Liver

Liver was collected and cut into prewarmed 25 mM HEPES, 10% FBS RPMI supplemented with 0.4 mg ml⁻¹ collagenase type VIII (Sigma, C2139) under shaking at 180 rpm for 45 min at 37 °C. Digested tissue was filtered through a 70-μm cell strainer and centrifuged at 350g for 5 min at 4 °C. Red blood cells were removed using 5 ml of erythrocyte lysis buffer for 5 min. To further enrich in leukocytes, supernatants were centrifuged in a 40%/70% Percoll gradient (Sigma, GE17-0891-01) at 1,250g for 30 min at RT with acceleration on 6 and without brake. Isolated cells were washed with PBS and resuspended in ml 2% FBS RPMI for counting.

Bone marrow

Femurs and tibias were collected and the cells from the bone marrow were obtained by centrifuging the bones at 6,000g for 1 min. Red blood cells were removed using 5 ml of erythrocyte lysis buffer for 5 min. Cells were washed and resuspended in 2% FBS PBS for counting.

Human samples

Human blood samples were collected by venipuncture in an overnight fasting state. Three milliliters of blood were collected in TransFix/EDTA vacuum blood collection tubes (Cytomark) and preserved until the day of staining and cell acquisition.

Flow cytometry

To differentiate between live and dead, the cells were first stained with the Zombie NIR Fixable Viability Kit (BioLegend, 423106, 1:3,000), the Zombie Yellow Fixable Viability Kit (BioLegend, 423104, 1:3,000) or the Ghost Dye Violet 540 (Tonbo Biosciences, 13-0879, 1:3,000) for 20 min at 4 °C. Then, the cells were washed with FACS staining buffer (PBS supplemented with 2% fetal bovine serum and 1 mM EDTA) and incubated with Fc receptor blocker purified rat anti-mouse anti-CD16/CD32 (BD Biosciences, 553142, 1:200) for 20 min at 4 °C. Cells were then incubated with primary antibodies for 20 min at 4 °C and were washed twice with FACS staining buffer. The following antibodies were diluted in Brilliant Stain Buffer (BD Biosciences, 566349) for surface antigen staining:

Antigen	Fluorochrome	Dilution	Clone	Supplier	Catalog
CD38	Pacific Blue	1:500	90	BioLegend	102720
TIM3	BV480	1:200	5D12/TIM-3	BD Biosciences	747618
CD244.2 (2B4)	BV510	1:200	2B4	BD Biosciences	740115
CD44	BV570	1:200	IM7	BioLegend	103037
CD69	BV650	1:100	H1.2F3	BioLegend	104541
CD62L	BV711	1:400	MEL-14	BioLegend	104445
CD95	BV750	1:100	Jo2	BD Biosciences	747413
KLRG1	BV785	1:200	2F1	BioLegend	138429
CD223 (LAG3)	BB515	1:200	C9B7W	BD Biosciences	566210
CD49d	PerCP-Cy5.5	1:100	R1-2	BioLegend	103619
PD-1	PerCP-eFluor710	1:200	J43	ThermoFisher	46-9985-80
ST2	PerCP-eFluor710	1:200	RMST2-33	ThermoFisher	46-9333-82
NKG2A	PE	1:200	16A11	BioLegend	142804
NKG2D	PE-Dazzle594	1:200	CX5	BioLegend	130214
CD25	PE Cy5	1:400	PC61	BioLegend	102007
CD8	PE-Fire700	1:1000	53-6.7	BioLegend	100792
CD28	APC	1:50	E18	BioLegend	122016
CD153	R718	1:200	RM153	BD Biosciences	751871
CD27	APC-Cy7	1:400	LG.3A10	BioLegend	124226
CD4	APC-Fire810	1:1000	Gk1.5	BioLegend	100480

For intracellular staining, after staining for membrane markers, the cells were fixed and permeabilized using the FOXP3/Transcription Factor Staining Kit (eBioscience, 00-5523-00) for 20 min at RT and darkness. To assess cytokine production, cells were stimulated for 4 h with 50 ng ml⁻¹ phorbol 12-myristate 13-acetate (PMA) (ThermoFisher, 356150010) and 1 μg ml⁻¹ ionomycin (Tocris, 1704) in the presence of brefeldin A (eBiosciences, 00-4506-51). Cells were then stained the following intracellular antibodies:

Antibody	Fluorochrome	Dilution	Clone	Supplier	Catalog
IL-10	Pacific Blue	1:100	JES5-16E3	BioLegend	505020
IL-17	PE	1:100	TC11-18H10.1	BioLegend	506904
TNF	APC	1:100	MP6-XT22	BioLegend	506308
IFN-γ	Spark NIR 685	1:100	XMG1.2	BioLegend	505861
MAF	eFluor450	1:100	sym0F1	ThermoFisher	48-9855-41
FOXP3	FITC	1:200	FJK-165	ThermoFisher	11-5773-82
RORγt	PE	1:100	Q31-378	BD Biosciences	562607
T-BET	AlexaFluor594	1:100	4B10	BioLegend	644834
T-BET	APC	1:100	4B10	BioLegend	17-5825-82
TOX	eFluor660	1:200	TXRX10	ThermoFisher	50-6502-80
γH2AX	–	1:300	20E3	CellSignaling	9718
P16	–	1:200	Polyclonal	ThermoFisher	PA5-119712
P21	Alexa Fluor 647	1:400	Polyclonal	Abcam	ab237265
Donkey anti-rabbit	Alexa Fluor 647	1:500	Polyclonal	ThermoFisher	A-31573

All flow cytometry experiments from mouse samples were performed with four-laser (violet, blue, yellow-green and red) or five-laser Aurora analyzers (Cytek Biosciences). Data were analyzed with the FlowJo v10.5.3 software (BD Biosciences). Gating strategies were set on the basis of fluorescence minus one controls, unstained samples and unstimulated samples (when needed). All the samples in the experiment excluded dead cells, clumps and debris.

In human samples, each tube contained at least 2 × 10⁶ human whole blood cells. Cells were labeled by incubation with appropriate fluorescence-conjugated antibodies for 15 min at RT in the dark. Cells were then lysed with 2 ml of FACS lysing solution (BD Biosciences) for 10 min and centrifuged at 500g for 5 min at RT. Then, the cells were washed with 5 ml of PBS. The following antibodies were used for surface antigen staining:

Antibody	Fluorochrome	Dilution	Clone	Supplier	Catalog
CD8	BUV496	1:100	RPA-T8	BD Biosciences	612942
CD25	BUV615	1:100	2A3	BD Biosciences	612996
CD19	BUV661	1:100	1D3	BD Biosciences	612971
CD45	BV510	1:100	HI30	BD Biosciences	563204
CD4	BV786	1:100	SK3	BD Biosciences	664528
CD127	PE	1:100	HIL-7R-M21	BD Biosciences	561028
CD3	APC	1:100	HIT3a	BD Biosciences	555342
KLRG1	APC-Cy7	1:100	2F1/KLRG1	BioLegend	138426

Experiments with human samples were performed in a BD FACSymphony A5 SORP flow cytometer (BD Biosciences). To generate comparable results among patients and over time, the photomultiplier voltages were adjusted to unlabeled lysed whole blood cells to obtain optimal photomultiplier tube voltages for the resolution of dim cell populations. The target values resulting of the optimization were used for subsequent calibrations to maintain instrument standardization. When possible, at least 40,000 events of CD4 population were acquired to reach the maximum T_reg events. Data were analyzed using the FlowJoTM v.10 software.

The gating strategy for identifying kT_reg cells is illustrated in the Extended Data Fig. 6. In brief, first, doublets and debris were excluded in forward scatter (FSC) – side scatter (SSC) dot plots. Then, a region in CD45 and SSC parameters were used to discriminate leukocytes. Next, lymphocytes were gated in a SSC – CD4 dot plot. Lymphoid cells were further cleaned in a SSC – CD45 plot. After that, CD3⁺CD19⁻ T cells were gated and, among them, CD4⁺ T lymphocytes were distinguished. The T_reg cells were identified, within the CD4⁺ T cells, by gating the CD25^hiCD27^lo population. Then, the expression of KLRG1 was analyzed in T_reg cells. To have a negative control to properly gate KLRG1⁺ cells, the gate was established in B cells.

Analysis of mitochondrial fitness

Analysis of mitochondrial mass and MMP was performed by flow cytometry in cells labeled for 30 min with 50 nM MitoTracker Green FM (Invitrogen, M7514), 200 nM MitoTracker DeepRed FM (Invitrogen, M22426), 40 nM Image-iT TMRM (Invitrogen, I34361) or 50 nM MitoTracker Red CMXRos (Invitrogen, M7512) in RPMI medium with no FBS in a 37 °C and 5% CO₂ incubator before extracellular staining. After incubation, cells were washed with FACS staining buffer and incubated with antibodies for extracellular staining.

Dimensional reduction and clustering analysis of flow cytometry data

Dimensional reduction and clustering analysis of flow cytometry data was done using OMIQ (Dotmatics). First, nonlymphocyte cells, doublets and dead cells were excluded on the basis of viability staining and FSC and SSC parameters. Then, 15,000 CD4⁺ cells from each sample were subsampled for further analysis. For dimensional reduction, the Uniform Manifold Approximation and Projection (UMAP) algorithm was applied with the following parameters: neighbors = 15, minimum distance = 0.4, components = 2, learning rate = 1, epochs = 200. For unbiased clustering, the Cluster-X algorithm was applied with alpha = 0.001. Mean fluorescence intensity for each marker projected on the UMAP plots was used to infer the cluster’s identity, and similar clusters were combined.

RNA-seq

rT_reg, aT_reg and kT_reg cells were isolated from young and old mice by FACS. RNA was extracted with the RNeasy Micro Kit (Qiagen, 74004) following manufacturer’s instructions. Sample quality was measured using the Qubit 3.0 fluorimeter for sample concentration and Agilent 5400 for fragment analysis. SMART-Seq V4 Ultra Low Input RNA kit for Sequencing 480 Rxns (Takarabio, 634893) was used for efficient cDNA synthesis and library preparation. After amplification, cDNA was purified using AMPure XP beads (Beckman Coulter, A63882) to remove any contaminants and ensure high-quality cDNA. The library was checked with Qubit and real-time PCR for quantification and bioanalyzer for size distribution detection. Quantified libraries were pooled and sequenced on Illumina Illumina Novaseq X plus using a PE150 (pair-end 150 base pair) strategy to produce 6G of data.Fastq files quality check was performed using FastQC v0.11.9. RNA-seq reads were mapped to the Mus musculus reference genome, GRCm39, using Hisat2 v2.2.1 software. Reads were then preprocessed with SAMtools v1.13 to transform Sequence Alignment/Map files into Binary Alignment/Map files and sorted. The number of reads covered by each gene is calculated by HTSeq-Count v1.99.2. Downstream data analysis was performed with R v4.4.1. Analysis of differentially expressed genes was performed using DESeq2 v1.44.0. Genes with P < 0.05 and |log₂fold change| >log₂(1.5) were determined to show statistically significant differences in group comparison. Over-representation analysis and gene set enrichment analysis were performed using clusterProfiler v4.12.0 package in GO, KEGG, WikiPathways, Reactome and the Hallmarks of the Molecular signatures database. Principal component analysis (PCA) plots and volcano plot were visualized using ggplot2 v3.5.1. Heat maps were visualized using heatmap v1.0.12.

Analysis of cytokine secretion by T_reg subsets using Multiplex

rT_reg, aT_reg or kT_reg cells were isolated from IL-33-treated young mice by FACS. A total of 5.5 × 10⁵ cells were incubated in 200 µl of complete RPMI medium supplemented with 10% FBS in resting conditions or in presence of Mouse T-Activator Dynabeads for 24 h. The cells were centrifuged at 300g for 5 min, and the supernatant was used for cytokine detection with magnetic bead technology (Invitrogen, Cytokine & Chemokine 26-Plex Mouse ProcartaPlex Panel 1).

Electron microscopy

For electron microscopy, splenocytes from old mice were labeled for 30 min with 50 nM MitoTracker Green FM and 200 nM Mitotracker DeepRed FM before extracellular staining. Cells were then labeled with a fluorochrome-conjugated antibody against CD4, and MtG^hiMtDR^hi and MtG^loMtDR^lo alive CD4⁺ T cells were isolated by FACS. After sorting, cells were fixed with 4% paraformaldehyde, 2% glutaraldehyde diluted in 0.1 M phosphate buffer. Cells were postfixed with 1% osmiun tetroxide and 1% potassium ferrocyanide for 60 min. After washing, incubation with 0.15% tannic acid in buffer phosphate 0.1 M for 1 min was achieved. After washing, cells were counterstained with uranyl acetate 2% for 1 h. Then, cells were dehydrated with lowering concentrations of ethanol and embedded in resin EPON. The preparations were examined with a transmission electron microscope (JEM1400 Flash, Jeol), and images were acquired with a CMOS Oneview camera (Gatan).

Reanalysis of scRNA-seq data

scRNA-seq data were extracted from a previously published work⁵ and reanalyzed with the Seurat package v4.2.0 in R v4.1.3. Variable genes were identified with the FindVariablesFeatures function across the range of expression values and used to perform a PCA with the RunPCA function. Clustering was performed with the FindClusters function with the Leiden algorithm and the first 20 principal components. Cluster identification was done with the FindMarkers function in each subset with a minimum log fold change of 0.25 and a P value <10^-3. The kT_reg cluster was separated by using the FindSubCluster function in the aT_reg cluster and then identified with the FindMarkers function.

Statistics and reproducibility analysis

Unless otherwise specified, n represents the number of individual biological replicates and is represented in graphs as one dot per sample. All the data are extracted from one representative experiment (n ≥ 3 per group) of a minimum of two separate experiments. Flow cytometry plots are representative of at least three replicates. No statistical method was used to predetermine sample size, but a minimum of three samples were used per experimental group and condition. Data collection and analysis were not performed blind to the age of the mice.

For statistical analysis, GraphPad Prism (version 9) was used. Data are expressed as mean ± s.e.m., unless otherwise indicated. Outliers were identified by the ROUT method (5%) and removed. Comparisons for two groups were calculated using unpaired two-tailed Student’s t-tests (for two groups meeting the normal distribution criteria) or Mann–Whitney U test (for two groups without normal distribution) according to the Shapiro–Wilk normality test. When comparing different populations within the same mouse, comparisons were calculated using two-tailed paired Student’s t-tests (for two groups meeting the normal distribution criteria) or Wilcoxon test (for two groups without normal distribution) according to the Shapiro–Wilk normality test. Comparisons for more than two groups were calculated using one-way analysis of variance (ANOVA) with Tukey’s correction for multiple comparisons (for three or more groups meeting the normal distribution criteria) or Kruskal–Wallis test with Dunn’s correction for multiple comparisons (for three or more groups without normal distribution) according to the Shapiro–Wilk normality test. When comparing at least three distinct populations within the same mouse, comparisons were calculated using repeated measures (RM) one-way ANOVA with Tukey’s correction for multiple comparisons (for samples meeting the normal distribution criteria) or Friedman test with Dunn’s correction for multiple comparisons (for at least three groups without normal distribution) according to the Shapiro–Wilk normality test.

The statistical significance is represented as follows: *P < 0.05, **0.05 < P < 0.01; ***0.01 < P < 0.001, ****P < 0.001.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data supporting the findings of this study are available within the Article and its Supplementary Information. The scRNA-seq dataset reanalyzed is publicly available (Single Cell Portal; accession number SCP490)⁵. The RNA-seq dataset generated during the current study is available in the Gene Expression Omnibus repository (GSE279926). The rest of the data were originally generated.

References

Franceschi, C. & Campisi, J. Chronic inflammation (inflammaging) and its potential contribution to age-associated diseases. J. Gerontol. A 69, S4–S9 (2014).
Article Google Scholar
Ferrucci, L. & Fabbri, E. Inflammageing: chronic inflammation in ageing, cardiovascular disease, and frailty. Nat. Rev. Cardiol. 15, 505–522 (2018).
Article CAS PubMed PubMed Central Google Scholar
Goronzy, J. J. & Weyand, C. M. Immune aging and autoimmunity. Cell. Mol. Life Sci. 69, 1615–1623 (2012).
Article CAS PubMed PubMed Central Google Scholar
Carrasco, E. et al. The role of T cells in age-related diseases. Nat. Rev. Immunol. 22, 97–111 (2022).
Article CAS PubMed Google Scholar
Elyahu, Y. et al. Aging promotes reorganization of the CD4 T cell landscape toward extreme regulatory and effector phenotypes. Sci. Adv. 5, eaaw8330 (2019).
Article CAS PubMed PubMed Central Google Scholar
Sakaguchi, S., Yamaguchi, T., Nomura, T. & Ono, M. Regulatory T cells and immune tolerance. Cell 133, 775–787 (2008).
Article CAS PubMed Google Scholar
Sharma, S., Dominguez, A. L. & Lustgarten, J. High accumulation of T regulatory cells prevents the activation of immune responses in aged animals. J. Immunol. 177, 8348–8355 (2006).
Article CAS PubMed Google Scholar
Trujillo-Vargas, C. M. et al. Immune phenotype of the CD4⁺ T cells in the aged lymphoid organs and lacrimal glands. GeroScience 44, 2105–2128 (2022).
Article CAS PubMed PubMed Central Google Scholar
Soto-Heredero, G., Gómez de las Heras, M. M., Escrig-Larena, J. I. & Mittelbrunn, M. Extremely differentiated T cell subsets contribute to tissue deterioration during aging. Annu. Rev. Immunol. 41, 181–205 (2023).
Article CAS PubMed Google Scholar
Lorenzo, E. C. et al. Senescence-induced changes in CD4 T cell differentiation can be alleviated by treatment with senolytics. Aging Cell 21, e13525 (2022).
Article CAS PubMed Google Scholar
Kuswanto, W. et al. Poor repair of skeletal muscle in aging mice reflects a defect in local, interleukin-33-dependent accumulation of regulatory T cells. Immunity 44, 355–367 (2016).
Article CAS PubMed PubMed Central Google Scholar
Tsaknaridis, L. et al. Functional assay for human CD4⁺CD25⁺ Treg cells reveals an age-dependent loss of suppressive activity. J. Neurosci. Res. 74, 296–308 (2003).
Article CAS PubMed Google Scholar
Raynor, J. et al. IL-15 fosters age-driven regulatory T cell accrual in the face of declining IL-2 levels. Front. Immunol. 4, 161 (2013).
Article CAS PubMed PubMed Central Google Scholar
Lages, C. S. et al. Functional regulatory T cells accumulate in aged hosts and promote chronic infectious disease reactivation. J. Immunol. 181, 1835–1848 (2008).
Article CAS PubMed Google Scholar
Raynor, J. et al. IL-6 and ICOS antagonize Bim and promote regulatory T cell accrual with age. J. Immunol. 195, 944–952 (2015).
Article CAS PubMed Google Scholar
Morales-Nebreda, L. et al. Aging imparts cell-autonomous dysfunction to regulatory T cells during recovery from influenza pneumonia. JCI Insight 6, e141690 (2021).
Article PubMed PubMed Central Google Scholar
Overacre-Delgoffe, A. E. et al. Interferon-γ drives Treg fragility to promote anti-tumor immunity. Cell 169, 1130–1141 (2017).
Article CAS PubMed PubMed Central Google Scholar
Zagorulya, M. et al. Tissue-specific abundance of interferon-gamma drives regulatory T cells to restrain DC1-mediated priming of cytotoxic T cells against lung cancer. Immunity 56, 386–405 (2023).
Article CAS PubMed PubMed Central Google Scholar
Henson, S. M. et al. KLRG1 signaling induces defective Akt (ser473) phosphorylation and proliferative dysfunction of highly differentiated CD8⁺ T cells. Blood 113, 6619–6628 (2009).
Article CAS PubMed Google Scholar
Joshi, N. S. et al. Inflammation directs memory precursor and short-lived effector CD8⁺ T cell fates via the graded expression of T-bet transcription factor. Immunity 27, 281–295 (2007).
Article CAS PubMed PubMed Central Google Scholar
Henson, S. M. & Akbar, A. N. KLRG1–more than a marker for T cell senescence. Age 31, 285–291 (2009).
Article CAS PubMed PubMed Central Google Scholar
Voehringer, D. et al. Viral infections induce abundant numbers of senescent CD8 T cells. J. Immunol. 167, 4838–4843 (2001).
Article CAS PubMed Google Scholar
Schwartzkopff, S. et al. Tumor-associated E-cadherin mutations affect binding to the killer cell lectin-like receptor G1 in humans. J. Immunol. 179, 1022–1029 (2007).
Article CAS PubMed Google Scholar
Li, L., Wan, S., Tao, K., Wang, G. & Zhao, E. KLRG1 restricts memory T cell antitumor immunity. Oncotarget 7, 61670–61678 (2016).
Article PubMed PubMed Central Google Scholar
Ramello, M. C. et al. Polyfunctional KLRG-1⁺CD57⁺ senescent CD4⁺ T cells infiltrate tumors and are expanded in peripheral blood from breast cancer patients. Front. Immunol. 12, 713132 (2021).
Article CAS PubMed PubMed Central Google Scholar
Cheng, G. et al. IL-2 receptor signaling is essential for the development of Klrg1⁺ terminally differentiated T regulatory cells. J. Immunol. 189, 1780–1791 (2012).
Article CAS PubMed Google Scholar
López-Otín, C., Blasco, M. A., Partridge, L., Serrano, M. & Kroemer, G. The hallmarks of aging. Cell 153, 1194–1217 (2013).
Article PubMed PubMed Central Google Scholar
Ron-Harel, N. et al. Defective respiration and one-carbon metabolism contribute to impaired naïve T cell activation in aged mice. Proc. Natl Acad. Sci. USA 115, 13347–13352 (2018).
Article CAS PubMed PubMed Central Google Scholar
Bektas, A. et al. Age-associated changes in human CD4⁺ T cells point to mitochondrial dysfunction consequent to impaired autophagy. Aging 11, 9234–9263 (2019).
Article CAS PubMed PubMed Central Google Scholar
Steinert, E. M., Vasan, K. & Chandel, N. S. Mitochondrial metabolism regulation of T cell-mediated immunity. Annu. Rev. Immunol. 39, 395–416 (2021).
Article CAS PubMed PubMed Central Google Scholar
Velarde, M. C., Flynn, J. M., Day, N. U., Melov, S. & Campisi, J. Mitochondrial oxidative stress caused by Sod2 deficiency promotes cellular senescence and aging phenotypes in the skin. Aging 4, 3–12 (2012).
Article CAS PubMed PubMed Central Google Scholar
Wiley, C. D. et al. Mitochondrial dysfunction induces senescence with a distinct secretory phenotype. Cell Metab. 23, 303–314 (2016).
Article CAS PubMed Google Scholar
Michalek, R. D. et al. Cutting edge: distinct glycolytic and lipid oxidative metabolic programs are essential for effector and regulatory CD4⁺ T cell subsets. J. Immunol. 186, 3299–3303 (2011).
Article CAS PubMed Google Scholar
Chapman, N. M. et al. mTOR coordinates transcriptional programs and mitochondrial metabolism of activated Treg subsets to protect tissue homeostasis. Nat. Commun. 9, 2095 (2018).
Article PubMed PubMed Central Google Scholar
Weinberg, S. E. et al. Mitochondrial complex III is essential for suppressive function of regulatory T cells. Nature 565, 495–499 (2019).
Article CAS PubMed PubMed Central Google Scholar
Venditti, P., Di Stefano, L. & Di Meo, S. Mitochondrial metabolism of reactive oxygen species. Mitochondrion 13, 71–82 (2013).
Article CAS PubMed Google Scholar
Mittelbrunn, M. & Kroemer, G. Hallmarks of T cell aging. Nat. Immunol. 22, 687–698 (2021).
Article CAS PubMed Google Scholar
Mogilenko, D. A. et al. Comprehensive profiling of an aging immune system reveals clonal GZMK⁺ CD8⁺ T cells as conserved hallmark of inflammaging. Immunity 54, 99–115 (2021).
Article CAS PubMed Google Scholar
Wang, Y., Su, M. A. & Wan, Y. Y. An essential role of the transcription factor GATA-3 for the function of regulatory T cells. Immunity 35, 337–348 (2011).
Article CAS PubMed PubMed Central Google Scholar
Wohlfert, E. A. et al. GATA3 controls Foxp3⁺ regulatory T cell fate during inflammation in mice. J. Clin. Invest. 121, 4503–4515 (2011).
Article CAS PubMed PubMed Central Google Scholar
Schiering, C. et al. The alarmin IL-33 promotes regulatory T-cell function in the intestine. Nature 513, 564–568 (2014).
Article CAS PubMed PubMed Central Google Scholar
Siede, J. et al. IL-33 receptor-expressing regulatory T cells are highly activated, Th2 biased and suppress CD4 T cell proliferation through IL-10 and TGFβ release. PLoS ONE 11, e0161507 (2016).
Article PubMed PubMed Central Google Scholar
Boyman, O., Kovar, M., Rubinstein, M. P., Surh, C. D. & Sprent, J. Selective stimulation of T cell subsets with antibody-cytokine immune complexes. Science 311, 1924–1927 (2006).
Article CAS PubMed Google Scholar
Sun, L. et al. Aged regulatory T cells protect from autoimmune inflammation despite reduced STAT3 activation and decreased constraint of IL-17 producing T cells. Aging Cell 11, 509–519 (2012).
Article CAS PubMed Google Scholar
Garg, S. K. et al. Aging is associated with increased regulatory T-cell function. Aging Cell 13, 441–448 (2014).
Article CAS PubMed PubMed Central Google Scholar
Zhao, L. et al. Changes of CD4⁺CD25⁺Foxp3⁺ regulatory T cells in aged Balb/c mice. J. Leukoc. Biol. 81, 1386–1394 (2007).
Article CAS PubMed Google Scholar
Hwang, K.-A., Kim, H.-R. & Kang, I. Aging and human CD4⁺ regulatory T cells. Mech. Ageing Dev. 130, 509–517 (2009).
Article CAS PubMed PubMed Central Google Scholar
Delacher, M. et al. Precursors for nonlymphoid-tissue Treg cells reside in secondary lymphoid organs and are programmed by the transcription factor BATF. Immunity 52, 295–312 (2020).
Article CAS PubMed PubMed Central Google Scholar
Beppu, L. Y. et al. Tregs facilitate obesity and insulin resistance via a Blimp-1/IL-10 axis. JCI Insight 6, e140644 (2021).
Article PubMed PubMed Central Google Scholar
Tauro, S., Nguyen, P., Li, B. & Geiger, T. L. Diversification and senescence of Foxp3⁺ regulatory T cells during experimental autoimmune encephalomyelitis. Eur. J. Immunol. 43, 1195–1207 (2013).
Article CAS PubMed PubMed Central Google Scholar
Hadis, U. et al. Intestinal tolerance requires gut homing and expansion of FoxP3⁺ regulatory T cells in the lamina propria. Immunity 34, 237–246 (2011).
Article CAS PubMed Google Scholar
Cebula, A. et al. Thymus-derived regulatory T cells contribute to tolerance to commensal microbiota. Nature 497, 258–262 (2013).
Article CAS PubMed PubMed Central Google Scholar
Kaur, H., Kaur, G. & Ali, S. A. IL-33’s role in the gut immune system: a comprehensive review of its crosstalk and regulation. Life Sci. 327, 121868 (2023).
Article CAS PubMed Google Scholar
Goldberg, E. L. et al. IL-33 causes thermogenic failure in aging by expanding dysfunctional adipose ILC2. Cell Metab. 33, 2277–2287 (2021).
Article CAS PubMed PubMed Central Google Scholar
Xu, L. et al. IL-33 induces thymic involution-associated naive T cell aging and impairs host control of severe infection. Nat. Commun. 13, 6881 (2022).
Article CAS PubMed PubMed Central Google Scholar
Monzel, A. S., Enríquez, J. A. & Picard, M. Multifaceted mitochondria: moving mitochondrial science beyond function and dysfunction. Nat. Metab. 5, 546–562 (2023).
Article PubMed PubMed Central Google Scholar
Yu, Y.-R. et al. Disturbed mitochondrial dynamics in CD8⁺ TILs reinforce T cell exhaustion. Nat. Immunol. 21, 1540–1551 (2020).
Article CAS PubMed Google Scholar
Monteiro, L., de, B., Davanzo, G. G., de Aguiar, C. F. & Moraes-Vieira, P. M. M. Using flow cytometry for mitochondrial assays. MethodsX 7, 100938 (2020).
Article CAS PubMed PubMed Central Google Scholar
Montes de Oca, M. et al. IL-27 signalling regulates glycolysis in Th1 cells to limit immunopathology during infection. PLoS Pathog. 16, e1008994 (2020).
Article CAS PubMed PubMed Central Google Scholar
Chen, P.-M. et al. CD38 reduces mitochondrial fitness and cytotoxic T cell response against viral infection in lupus patients by suppressing mitophagy. Sci. Adv. 8, eabo4271 (2022).
Article CAS PubMed PubMed Central Google Scholar
Konjar, Š. et al. Mitochondria maintain controlled activation state of epithelial-resident T lymphocytes. Sci. Immunol. 3, eaan2543 (2018).
Article PubMed PubMed Central Google Scholar
Kowaltowski, A. J. & Abdulkader, F. How and when to measure mitochondrial inner membrane potentials. Biophys. J. 123, 4150–4157 (2024).
Article CAS PubMed Google Scholar
Ding, C. et al. A T_reg-specific long noncoding RNA maintains immune-metabolic homeostasis in aging liver. Nat. Aging 3, 813–828 (2023).
Article PubMed Google Scholar
Guo, Z. et al. DCAF1 regulates Treg senescence via the ROS axis during immunological aging. J. Clin. Invest. 130, 5893–5908 (2020).
Article CAS PubMed PubMed Central Google Scholar
Sun, Z. & Hornung, V. cGAS–STING signaling. Curr. Biol. 32, R730–R734 (2022).
Article PubMed Google Scholar
Dou, Z. et al. Cytoplasmic chromatin triggers inflammation in senescence and cancer. Nature 550, 402–406 (2017).
Article CAS PubMed PubMed Central Google Scholar
Linder, A. & Hornung, V. Inflammasomes in T cells. J. Mol. Biol. 434, 167275 (2022).
Article CAS PubMed Google Scholar
Baixauli, F. et al. Mitochondrial respiration controls lysosomal function during inflammatory T cell responses. Cell Metab. 22, 485–498 (2015).
Article CAS PubMed PubMed Central Google Scholar
Milanovic, M. et al. Senescence-associated reprogramming promotes cancer stemness. Nature 553, 96–100 (2018).
Article CAS PubMed Google Scholar
Salinas-Saavedra, M. et al. Senescence-induced cellular reprogramming drives cnidarian whole-body regeneration. Cell Rep. 42, 112687 (2023).
Article CAS PubMed PubMed Central Google Scholar
Burton, O. T. et al. The tissue-resident regulatory T cell pool is shaped by transient multi-tissue migration and a conserved residency program. Immunity 57, 1586–1602 (2024).
Article CAS PubMed Google Scholar
Correia-Melo, C. et al. Mitochondria are required for pro-ageing features of the senescent phenotype. EMBO J. 35, 724–742 (2016).
Article CAS PubMed PubMed Central Google Scholar
Vardhana, S. A. et al. Impaired mitochondrial oxidative phosphorylation limits the self-renewal of T cells exposed to persistent antigen. Nat. Immunol. 21, 1022–1033 (2020).
Article CAS PubMed PubMed Central Google Scholar
Scharping, N. E. et al. Mitochondrial stress induced by continuous stimulation under hypoxia rapidly drives T cell exhaustion. Nat. Immunol. 22, 205–215 (2021).
Article CAS PubMed PubMed Central Google Scholar
Valle-Noguera, A., Gómez-Sánchez, M. J., Girard-Madoux, M. J. H. & Cruz-Adalia, A. Optimized protocol for characterization of mouse gut innate lymphoid cells. Front. Immunol. 11, 563414 (2020).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

Funded by the European Union. Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. This study was supported by European Research Council grant ERC-2021-CoG 101044248-LetTBe, and the Y2020/BIO-6350 NutriSION-CM synergy grant from Comunidad de Madrid, Ministerio de Ciencia e Innovacion, Spain (grants PID2022-141169OB-I00 and PID2022-138295OB-I00). GENYAL Clinical Trials Platform of IMDEA Alimentación led by A. Ramírez de Molina and R. Ramos Ruiz conducted the volunteer recruitment and human clinical trial. Flow cytometry of human samples was conducted by Flow Cytometry Unit, CNIC, being the cytometer part of the grant EQC2018-005009-P funded by MCIN/AEI/10.13039/501100011033 and by ‘ERDF A way of making Europe’. G.S.-H. is supported by a FPI-UAM grant (Universidad Autónoma de Madrid). E.G.-R. was funded by a Juan de la Cierva grant (IJC2018-036850-I; Universidad Autónoma de Madrid). M.G. and J.I.E.-L. are supported by FPU grants (FPU19/02576 and FPU20/04066, respectively), both from Ministerio de Ciencia, Innovación y Universidades (Spain). E.C. is supported by SI4/PJI/2024-00166 from Comunidad de Madrid and UAM (Spain). S.D.-P. was funded by a PIPF grant (PIPF-2022/SAL-GL-25208), from Comunidad de Madrid (Spain). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Author information

Authors and Affiliations

Departamento de Biología Molecular, Facultad de Ciencias, Centro de Biología Molecular ‘Severo Ochoa’, Universidad Autónoma de Madrid, Madrid, Spain
Gonzalo Soto-Heredero, Enrique Gabandé-Rodríguez, Manuel M. Gómez de las Heras & Isaac Francos-Quijorna
Consejo Superior de Investigaciones Científicas, Centro de Biología Molecular ‘Severo Ochoa’, Universidad Autónoma de Madrid, Madrid, Spain
Gonzalo Soto-Heredero, Enrique Gabandé-Rodríguez, José Ignacio Escrig-Larena, Sandra Delgado-Pulido, Isaac Francos-Quijorna, Eva M. Blanco, Álvaro Fernández-Almeida, David Abia & María Mittelbrunn
Departamento de Biología, Facultad de Ciencias, Centro de Biología Molecular ‘Severo Ochoa’, Universidad Autónoma de Madrid, Madrid, Spain
Elisa Carrasco
Servicio de Microscopía Electrónica, Centro de Biología Molecular ‘Severo Ochoa’, Universidad Autónoma de Madrid, Madrid, Spain
María Josefa Rodríguez
IMDEA Food Institute, CEI UAM+CSIC, Madrid, Spain
Cristina M. Fernández-Díaz & Ana Ramírez de Molina
Centro Nacional de Investigaciones Cardiovasculares, Instituto de Salud Carlos III, Madrid, Spain
María Beatriz Álvarez-Flores, Virginia Zorita, Fátima Sánchez-Cabo & Carlos Torroja
Luxembourg Centre for Systems Biomedicine, University of Luxembourg, Esch sur-Alzette, Luxembourg
Sascha Jung & Antonio del Sol
CIC bioGUNE-BRTA (Basque Research and Technology Alliance), Bizkaia Technology Park, Derio, Spain
Sascha Jung & Antonio del Sol
IKERBASQUE, Basque Foundation for Science, Bilbao, Spain
Sascha Jung & Antonio del Sol

Authors

Gonzalo Soto-Heredero
View author publications
Search author on:PubMed Google Scholar
Enrique Gabandé-Rodríguez
View author publications
Search author on:PubMed Google Scholar
Elisa Carrasco
View author publications
Search author on:PubMed Google Scholar
José Ignacio Escrig-Larena
View author publications
Search author on:PubMed Google Scholar
Manuel M. Gómez de las Heras
View author publications
Search author on:PubMed Google Scholar
Sandra Delgado-Pulido
View author publications
Search author on:PubMed Google Scholar
Isaac Francos-Quijorna
View author publications
Search author on:PubMed Google Scholar
Eva M. Blanco
View author publications
Search author on:PubMed Google Scholar
Álvaro Fernández-Almeida
View author publications
Search author on:PubMed Google Scholar
David Abia
View author publications
Search author on:PubMed Google Scholar
María Josefa Rodríguez
View author publications
Search author on:PubMed Google Scholar
Cristina M. Fernández-Díaz
View author publications
Search author on:PubMed Google Scholar
María Beatriz Álvarez-Flores
View author publications
Search author on:PubMed Google Scholar
Ana Ramírez de Molina
View author publications
Search author on:PubMed Google Scholar
Sascha Jung
View author publications
Search author on:PubMed Google Scholar
Antonio del Sol
View author publications
Search author on:PubMed Google Scholar
Virginia Zorita
View author publications
Search author on:PubMed Google Scholar
Fátima Sánchez-Cabo
View author publications
Search author on:PubMed Google Scholar
Carlos Torroja
View author publications
Search author on:PubMed Google Scholar
María Mittelbrunn
View author publications
Search author on:PubMed Google Scholar

Contributions

G.S.-H., E.G.-R., E.C. and M.M. set up the research project and designed the experiments. G.S.-H., E.G.-R., E.C., J.I.E.-L., M.M.G., S.D.P., I.F.Q. and E.M.B. conducted most of the experiments and acquired the data. M.J.R. performed and analyzed the data in Fig. 1f,g. V.Z. provided technical and scientific support. C.M.F.-D. and M.B.A.-F. performed the experiments and analyzed the data in Fig. 8 under the supervision of A.R. A.F.A., D.A., S.J., A.d.S., F.S.-C. and C.T. performed bioinformatic analysis. G.S.-H. assembled the figures. G.S.-H. and M.M. wrote the manuscript. M.M. obtained financial support and provided supervision.

Corresponding author

Correspondence to María Mittelbrunn.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Aging thanks Jorg Goronzy and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 CD4⁺ T cells progressively lose mitochondrial mass and mitochondrial membrane potential during aging.

(a) Representative plots of mitochondrial mass and mitochondrial membrane potential analyzed by flow cytometry in CD4⁺ T cells treated or not treated with oligomycin. Cells were gated as MtG^hiMtDR^hi, MtG^hiMtDR^lo and MtG^loMtDR^lo. Mitochondrial mass analyzed as gMFI of MtG (b) and mitochondrial membrane potential analyzed as gMFI of MtDR (c) measured by flow cytometry in circulating CD4⁺ T cells from mouse from 2- (n = 5), 9- (n = 5), 13- (n = 5), 17- (n = 5) and 21- (n = 6) month-old mice. (d) Representative flow cytometry plots and quantifications of the simultaneous analysis of MtG and TMRM in splenic CD4⁺ T cells from young (2-month-old, n = 5) and old (21-month-old, n = 4) mice. (e) Quantification of the MtDR/MtG ratio in MtG^hiMtDR^hi and MtG^loMtDR^lo CD4⁺ T cells from mice at different ages (n = 6 mice per group). gMFI: Geometric Mean Fluorescence Intensity; MtG: Mitotracker Green; MtDR: Mitotracker DeepRed. Each dot represents an individual mouse. Data are presented as mean values ± SEM. Statistical analysis was performed using one-way ANOVA with post hoc Tukey’s correction (b), Kruskal-Wallis test with post hoc Dunn’s correction (c), two-tailed Student’s t test [(d): TMRM^hiMtG^hi; (e)] or two-tailed unpaired Welch’s t test [(d): TMRM^loMtG^lo] (b). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Exact P values and additional statistical parameters can be found in the source data.

Source data

Extended Data Fig. 2 Identification of age-associated T cells by scRNAseq and by spectral flow cytometry.

(a) UMAP representation of CD4⁺ T cells from young (left) and old mice (right) analyzed by scRNAseq (top) or spectral flow cytometry (bottom). (b) UMAP with clustering overlay showing the distribution of the different clusters of CD4⁺ T cells identified by scRNAseq or by spectral flow cytometry: naive (blue), naïve_ISG15 (dark blue), rTregs (yellow), aTregs (orange), kTregs (red), TEM (light green), cytotoxic (brown) and exhausted (dark green) (n = 4 mice per group). (c) Percentage of circulating KLRG1⁺ cells in CD4⁺ T cells from 2- (n = 6), 9- (n = 6), 13- (n = 5), 17- (n = 6) and 21- (n = 6) month-old mice assessed by flow cytometry. (d-e) Percentage of splenic KLRG1⁺ cells in CD4⁺FOXP3⁺CD25⁺ T cells (d) or in CD4⁺ T cells (e) from 2- (n = 3), 9- (n = 4), 17- (n = 4) and 21- (n = 4) month-old mice. Each dot represents an individual mouse. Data are presented as mean values ± SEM. Statistical analysis was performed using one-way ANOVA with post hoc Tukey’s correction [(c), (d), (e)]. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Exact P values and additional statistical parameters can be found in the source data.

Source data

Extended Data Fig. 3 Reanalysis of publicly available scRNAseq data identifies kTregs with a pro-inflammatory signature.

(a) UMAP representation and clustering of a new analysis of publicly available scRNAseq data from CD4⁺ T cells from young and old mice. Clusters are colored as naïve (blue), rTregs (yellow), aTregs (orange), kTregs (red), T_EM (light green), cytotoxic (brown) and exhausted (dark green). (b) Quantification of the percentage of kTreg cells in young and old mice (n = 4 mice per group). (c) Comparative expression of genes related to Treg function in aTreg and kTreg (n = 4 mice per group). (d) Violin plots showing the expression of several pro-inflammatory markers in each cluster of CD4⁺ T cells (n = 4 mice per group). Each dot represents an individual mouse. Box-and-whisker plots show the median, the maximum, the minimum and 25th and 75th percentiles. Statistical analysis was performed using two-tailed unpaired Welch’s t test (b). Exact P values and additional statistical parameters can be found in the source data.

Source data

Extended Data Fig. 4 Analysis of the expression of different transcription factors in the Treg subsets.

Representative histograms and quantifications of the expression of different transcription factors by flow cytometry in splenic rTreg (yellow), aTreg (orange) and kTregs (red) from young (2-month-old) mice (n = 6 mice per group) measured as gMFI. gMFI values are relative to the MFI in rTregs (n = 6 per group). gMFI: Geometric Mean Fluorescence Intensity. Each dot represents an individual mouse. Data are presented as mean values ± SEM. Statistical analysis was performed using Friedman test with post hoc Dunn’s correction. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Exact P values and additional statistical parameters can be found in the source data.

Source data

Extended Data Fig. 5 KLRG1⁺ Tregs show a pro-inflammatory phenotype and induce inflammation.

(a-b) In vitro assessment of the suppressive activity of KLRG1⁺ Tregs. KLRG1^- or KLRG1⁺ Tregs were sorted from old (22-month-old) and co-cultured with CD45.1 naïve CD4⁺ T responder cells labelled with CellTrace Violet at 2:1 ratio (2 naïve:1 Treg) for 72 h in the presence of α-CD3/α-CD28 (n = 3 mice). (a) Representative histogram of the CellTrace Violet signal measured by flow cytometry in CD45.1 CD4⁺ T conventional responder cells after 72 h of culture. (b) Quantification of the capacity of KLRG1^- and KLRG1⁺ Treg cells to suppress the proliferation of CD45.1 CD4⁺ responder T cells. The percentage of suppression was determined as: 100 − (% of proliferating cells with Tregs)/(% of proliferating cells without Tregs). (c) Representative histograms and quantifications of the expression of different cytokines by flow cytometry in splenic rTreg (yellow), aTreg (orange) and kTregs (red) from old (21-month-old) mice measured as gMFI. gMFI values are relative to the MFI in rTregs (n = 6 mice per group). (d-e) Analysis of the secretion of different cytokines by the Treg subsets in activating conditions using multiplex. Treg subsets were sorted from young mice injected with IL-2 + IL-33 and incubated during 24 h in the presence of α-CD3/α-CD28 before the assay (n = 3 mice). (d) Heatmap and statistical comparisons of the secretion of different cytokines by rTreg, aTreg and kTregs. Heatmap values are relative to the maximum of each cytokine. (e) Quantification of IL-10, IL-6, GM-CSF and CXCL2 secreted by activated rTreg, aTreg and kTregs. (f-h) In vivo assessment of the pro-inflammatory activity of KLRG1⁺ Tregs: CD3ε^-/- T cell deficient mice were injected with splenic KLRG1^- or KLRG1⁺ Tregs isolated from young mice injected with IL-2 + IL-33. The mice were analyzed four months after the adoptive transfer (n = 3 mice per group). (f) Schematic diagram depicting the adoptive transfer of kTregs. (g) Spleen weight from CD3ε^-/- mice non-injected or injected with KLRG1^- or KLRG1⁺ T cells. (h) Quantification of the percentage of circulating neutrophils or basophils in CD3ε^-/- mice non-injected or injected with KLRG1^- or KLRG1⁺ T cells. gMFI: Geometric Mean Fluorescence Intensity. Each dot represents an individual mouse. Data are presented as mean values ± SEM. Statistical analysis was performed using two-tailed unpaired Student’s t test (b), Friedman test with post hoc Dunn’s correction [(c), (e): IL-6], RM one-way ANOVA with post hoc Tukey’s correction [(d); (e): IL10, GM-CSF, CXCL2] or ANOVA with post hoc Tukey’s correction [(g), (h)]. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Exact P values and additional statistical parameters can be found in the source data.

Source data

Extended Data Fig. 6 Gating strategy to identify KLRG1⁺ Tregs in human PBMCs.

Gating strategy followed to identify KLRG1⁺ Tregs (CD45⁺CD3⁺CD4⁺CD8^-CD25⁺IL7R^-KLRG1⁺) in human PBMCs.

Supplementary information

Reporting Summary

Supplementary Table 1

Differentially expressed genes in the kT_reg cluster.

Source data

Source Data Fig. 1

Statistical source data.

Source Data Fig. 2

Statistical source data.

Source Data Fig. 3

Statistical source data.

Source Data Fig. 4

Statistical source data.

Source Data Fig. 5

Statistical source data.

Source Data Fig. 6

Statistical source data.

Source Data Fig. 7

Statistical source data.

Source Data Fig. 8

Statistical source data.

Source Data Extended Data Fig. 1

Statistical source data.

Source Data Extended Data Fig. 2

Statistical source data.

Source Data Extended Data Fig. 3

Statistical source data.

Source Data Extended Data Fig. 4

Statistical source data.

Source Data Extended Data Fig. 5

Statistical source data.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Soto-Heredero, G., Gabandé-Rodríguez, E., Carrasco, E. et al. KLRG1 identifies regulatory T cells with mitochondrial alterations that accumulate with aging. Nat Aging 5, 799–815 (2025). https://doi.org/10.1038/s43587-025-00855-9

Download citation

Received: 19 June 2023
Accepted: 27 March 2025
Published: 30 April 2025
Issue date: May 2025
DOI: https://doi.org/10.1038/s43587-025-00855-9

This article is cited by

Systemic inflammation as a central player in the initiation and development of Alzheimer’s disease
- Irem Bayraktaroglu
- Natalia Ortí-Casañ
- Ulrich L. M. Eisel
Immunity & Ageing (2025)
Immunosenescence and cancer: molecular hallmarks, tumor microenvironment remodeling, and age-specific immunotherapy challenges
- Qianwen Liu
- Jingfeng Li
- Yihao Liu
Journal of Hematology & Oncology (2025)
Aging reshapes the adaptive immune system from healer to saboteur
- Sandra Delgado-Pulido
- Matthew J. Yousefzadeh
- Maria Mittelbrunn
Nature Aging (2025)

Subjects

Abstract

Similar content being viewed by others

Main

Results

T cells with mitochondrial alterations accumulate with aging

KLRG1 identifies age-associated regulatory T cells

KLRG1+ Treg cells show mitochondrial alteration and senescence features

KLRG1+ Treg cells express ST2 and are induced by IL-33 in vivo

KLRG1+ Treg cells exhibit a pro-inflammatory phenotype

KLRG1+ Treg cells increase in human blood samples during aging

Discussion

Methods

Animal procedures

IL-2 + IL-33 treatment

Adoptive transfer of Treg cells

Suppression assays

Human samples

Tissue processing for flow cytometry

Spleen and lymph nodes

Blood

Colonic lamina propria

White adipose tissue

Peyer’s patches

Liver

Bone marrow

Human samples

Flow cytometry

Analysis of mitochondrial fitness

Dimensional reduction and clustering analysis of flow cytometry data

RNA-seq

Analysis of cytokine secretion by Treg subsets using Multiplex

Electron microscopy

Reanalysis of scRNA-seq data

Statistics and reproducibility analysis

Reporting summary

Data availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Extended data

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

This article is cited by

Search

Quick links

KLRG1⁺ T_reg cells show mitochondrial alteration and senescence features

KLRG1⁺ T_reg cells express ST2 and are induced by IL-33 in vivo

KLRG1⁺ T_reg cells exhibit a pro-inflammatory phenotype

KLRG1⁺ T_reg cells increase in human blood samples during aging

Adoptive transfer of T_reg cells

Analysis of cytokine secretion by T_reg subsets using Multiplex