Targeting BIRC6 rejuvenates hematopoietic stem cell aging and immunosenescence

Abstract

Immunosenescence represents a central hallmark of organismal aging, characterized by a progressive decline in immune function, which compromises host defense and accelerates systemic aging. Hematopoietic stem cell (HSC) aging is a key contributor to this process, characterized by aberrant expansion, myeloid-biased differentiation, and impaired self-renewal, culminating in hematopoietic-immune imbalance. Although the expansion and survival advantages of aged HSCs have been well-demonstrated, the underlying mechanisms remain elusive. Here, we reveal that regulatory T cells (Tregs) within the bone marrow (BM) microenvironment actively safeguard the survival of aged HSCs via a previously uncharacterized signaling pathway. We identify a novel aged HSC subpopulation characterized by high expression of Baculoviral IAP Repeat Containing 6 (BIRC6), an apoptosis inhibitor. This BIRC6-high subpopulation is markedly expanded in aged mice and recapitulates the hallmarks of HSC aging. Mechanistically, cAMP derived from BM Tregs activates the PKA–CREB pathway in HSCs, activating Birc6 transcription, which reduces apoptotic priming in aged HSCs, thereby promoting hematopoietic-immune imbalance. Strikingly, targeted BIRC6 inhibition in HSCs using antibody-conjugated lipid nanoparticle-encapsulated antisense oligonucleotides (LNP-ASOs) significantly reverses hematopoietic-immune aging phenotypes and ameliorates age-associated immune dysfunction in middle-aged mice. LNP-ASO treatment dramatically rebalances immune cell production, reduces immunosenescence markers, and enhances vaccine responses in middle-aged mice. More importantly, this strategy was also effective in HSCs from middle-aged human donors, highlighting its potential for clinical translation. These findings elucidate a key microenvironmental pathway (Treg–cAMP–PKA–CREB–BIRC6) driving HSC aging and offer a novel strategy to ameliorate the aged hematopoietic system and combat age-related immune decline.

Introduction

Aging represents a systemic decline in physiological function that occurs over time. The aging process disrupts organismal homeostasis across molecular, cellular, and organ levels, with the progressive decline of the immune system, known as immunosenescence.¹ Immunosenescence not only compromises the ability to resist pathogen invasion but also contributes to the development of cardiovascular diseases and degenerative disorders in solid organs, ultimately accelerating the systemic aging process.² Therefore, elucidating the mechanisms underlying immunosenescence is crucial for counteracting age-related immune decline and mitigating overall aging progression in the elderly. It should be noted that although closely interrelated with organismal aging, immunosenescence is not entirely a consequence of the latter. The stress-response mechanism (independent of aging) can occur at any stage of the human lifespan.³ At the cellular level, immunosenescence is characterized in part by the accumulation of lymphocyte populations that have undergone replicative senescence.^4,5 This study focuses on elucidating the functional effects and underlying mechanisms of aged immunity during physiological aging.

Hematopoietic stem cells (HSCs), the common progenitors of all immune cells, directly influence the function of the immune system. HSC aging is a major driver of immunodeficiency and diminishes infection resistance in both physiological and pathological aging states.⁶ The principal mechanisms underlying this age-related immune decline involve the loss of regenerative capacity in HSCs and their progenies, accompanied by the acquisition of exhausted, pro-inflammatory, and self-reactive phenotypes in aged immune cells.⁷ However, the mechanisms driving HSC aging remain incompletely understood. Under normal conditions, bone marrow (BM) HSCs maintain their pool size and functional integrity through precisely regulating homeostasis between survival, proliferation, and differentiation, thereby continuously meeting the demand for immune cell production.⁸ Nevertheless, during aging, HSCs undergo characteristic changes, including aberrant expansion, enhanced myeloid-biased differentiation, accumulated genetic mutations, increased reactive oxygen species (ROS) production, impaired mitochondrial fitness, and weakened self-renewal capacity.⁹ Collectively, these age-related phenotypic and functional alterations of HSCs drive immune dysfunction during aging. Notably, a small fraction of aged HSCs adapt to the aging environment and exhibit a survival advantage, ultimately dominating the hematopoietic compartment and contributing to hematopoietic-immune homeostatic imbalance,¹⁰ while the precise identification of these aged HSC subpopulations and their dynamic evolution remains incompletely elucidated. Deciphering the mechanisms of HSC aging is therefore critical for understanding the origins of immunosenescence and developing targeted interventions.

Both intrinsic and extrinsic mechanisms synergistically modulate stem cell aging and survival fate. It is well-established that the aged microenvironment accelerates stem cell aging by transmitting pro-aging signals,¹⁰ particularly the senescence-associated secretory phenotype (SASP). In our previous study, we characterized a subpopulation of aged HSCs featuring elevated major histocompatibility complex class II (MHCII) expression and accumulated mutational load, concurrent with the expansion of niche-resident regulatory T cells (Tregs) in mice with both physiological and pathological aging. Thereupon, the strengthened MHCII–TCR interaction between aged HSCs and Tregs facilitates functional gap‑junction (GJ) formation, ultimately allowing Treg‑derived cyclic AMP (cAMP) transfer to promote aged HSC survival,¹¹ although the precise mechanisms remain elusive.

In the current study, we demonstrated that the diminished apoptotic priming in MHCII-high (MHCII^hi) HSCs from aged mice is dependent on upregulation of the apoptosis inhibitor Baculoviral IAP Repeat Containing 6 (BIRC6, also known as Apollon or BRUCE). Aged HSCs with high BIRC6 expression exhibit markedly elevated aging scores and an aging phenotype closely resembling MHCII^hi aged HSCs. Further investigation revealed that upregulated BIRC6 in aged HSCs is primarily triggered by niche Tregs, which undergo expansion during aging. Mechanistically, Treg-derived cAMP activated the PKA–CREB pathway and thus activated the Birc6 expression in aged HSCs, ultimately conferring a survival advantage onto aged HSCs. Intriguingly, targeting BIRC6 using lipid nanoparticle-encapsulated antisense oligonucleotides (LNP-ASOs) significantly reversed HSC aging phenotypes and ameliorated age-associated immune dysfunction in middle-aged mice. LNP-ASO treatment increased common lymphocyte progenitors, naive T cells, and mature B cells, while reducing markers of immunosenescence and improving adaptive responses to vaccination in middle-aged mice. We further evaluated the therapeutic potential of BIRC6 inhibition in aged human HSCs. Collectively, these findings indicate a novel role of BIRC6 signaling in modulating the survival fate of aged HSCs and offer a conceptual foundation for future interventions targeting HSC aging and immunosenescence.

Results

Upregulated BIRC6 signaling diminishes apoptotic priming in aged HSCs

We previously identified a subpopulation of aged HSCs with high major histocompatibility complex class II expression (MHCII^hi) in both physiological and pathological aging mouse models, which exhibits prominent aging features in both phenotype and function.¹¹ Notably, MHCII^hi HSCs from middle-aged (12-month-old, equivalent to human middle age) mice have already increased in the HSC pool (Fig. 1a, b). MHCII^hi HSCs from middle-aged mice showed dramatically distinct transcriptomic and metabolic features (Fig. 1c, Supplementary Fig. 1a), increased aging score (murine HSC aging signature) (Fig. 1d) and compromised reconstitution potential (Fig. 1e). Nevertheless, the mechanisms by which MHCII^hi HSCs adapt to the aging environment and exhibit a survival advantage remain unclear.

Fig. 1.

Upregulated BIRC6 signaling diminishes apoptotic priming in aged HSCs. a Flow cytometric quantification of BM HSC frequency of 2- and 12-month-old (mo) mice (n = 6). b Flow cytometric gating strategy for MHCII^hi and MHCII^lo HSC subsets, and the relative proportion of these populations in 2 and 12 mo mice (n = 6). c Principal component analysis (PCA) of transcriptomes of MHCII^hi and MHCII^lo BM HSCs of 2 and 12 mo mice (n = 3). d Aging signature scoring of MHCII^hi and MHCII^lo BM HSCs of 12 mo mice. e Peripheral blood (PB) chimerism and donor-derived myeloid, B, and T cells, at 16 weeks post-competitive bone marrow transplantation (n = 6). f KEGG pathway enrichment analysis of upregulated and downregulated pathways in MHCII^hi and MHCII^lo BM HSCs of 12 mo mice. g Flow cytometric quantification of apoptotic priming in the BM HSCs from 2 mo and 12 mo mice following BIM and BID peptide challenge (n = 6). h Heatmap of gene expression of anti-apoptotic IAP and BCL-2 family members in MHCII^hi and MHCII^lo BM HSCs of 2 mo and 12 mo mice (n = 3). i Flow cytometric quantification of BIRC6 protein expression of MHCII^hi and MHCII^lo BM HSCs of 12 mo mice (n = 6). j Aging score and expression of the top 5 age-associated genes of Birc6^hi and Birc6^lo HSCs from young and aged mice (n = 248, 63, 240, and 121 left to right). k Flow cytometric quantification of apoptotic priming in BM HSCs from 12 mo Birc6^ΔHSC or Birc6^f/f mice following BIM and BID peptide challenge (n = 6). l Flow cytometric quantification of frequency of BM HSCs of 12 mo Birc6^ΔHSC or Birc6^f/f mice (n = 6). m PB chimerism and donor-derived myeloid, B, and T cells, at 16 weeks post-competitive bone marrow transplantation (n = 6). All quantitative data are presented as the mean ± SD. *P < 0.05, **P < 0.01. Unless otherwise specified, comparisons were performed using two-tailed unpaired Student’s t-test. P values for panel j were calculated via one-way ANOVA

Pathway enrichment analysis revealed pronounced alterations not only in metabolism and signal transduction but also in cell death and survival pathways in MHCII^hi HSCs from middle-aged mice. Among them, we noticed that the apoptosis pathway was predominantly enriched (Fig. 1f). In line with these observations, although the proportion of apoptotic cells was equivalent between MHCII^hi and MHCII^lo HSCs from middle-aged mice under homeostatic conditions (Supplementary Fig. 1b), BH3 profiling assays uncovered a pronounced reduction in apoptotic priming within MHCII^hi HSCs following challenge with BIM and BID (Fig. 1g). This aligns with an established physiological aging phenotype wherein aged HSCs acquire survival advantages by suppressing apoptotic priming.¹² Given the reduced apoptotic priming in aged HSCs, we further analyzed the expression of key anti-apoptotic genes, particularly B-cell lymphoma-2 (BCL-2) and inhibitor of apoptosis (IAP) family members. Transcriptomic profiling revealed that Birc6 exhibited the most pronounced upregulation among anti-apoptotic genes in aged HSCs (Fig. 1h), confirmed by flow cytometry (Fig. 1i). Although the structural basis and effects of BIRC6-mediated apoptosis regulation have recently been elucidated,¹³ the role of BIRC6 upregulation in stem cell aging remains unclear. Therefore, we divided young and aged mouse-derived HSCs into Birc6 high-expression (Birc6^hi) and low-expression (Birc6^lo) subpopulations based on single-cell transcriptome sequencing data.¹⁴ Then, we assessed the HSC aging signature with established methods.¹⁵ Intriguingly, Birc6^hi HSCs from aged mice showed significantly higher aging score and MHCII score as well as activated myeloid differentiation-associated networks (Fig. 1j, Supplementary Fig. 1c–e). Flow cytometry confirmed a significant expansion of BIRC6-high HSCs from middle-aged mice and higher expression of aging markers (MHCII and CD150)^11,16 in these cells (Supplementary Fig. 1f–h). These findings demonstrate that aged HSCs undergo profound alterations in apoptotic regulators, which suggests a potential role for BIRC6 upregulation in HSC aging.

To elucidate the underlying pathophysiological significance of upregulated BIRC6 in HSC aging, we conditionally deleted Birc6 from HSCs (Birc6^ΔHSC) by using Cre-loxP recombination with Mx1-Cre mice and Birc6^f/f mice. HSCs from young Birc6^ΔHSC mice displayed no alterations in frequency or function compared to littermate controls (Supplementary Fig. 1i, j). However, HSCs from middle-aged Birc6^ΔHSC mice exhibited markedly enhanced apoptotic priming and an ameliorated aging phenotype compared to middle-aged littermate controls (Fig. 1k–m). These data indicate that upregulated BIRC6 signaling reduces apoptotic priming and promotes the aging phenotype in aged HSCs.

Activated BIRC6 signaling in aged HSCs is directed by niche Tregs

HSC aging is influenced by the crosstalk between intrinsic and extrinsic factors.¹⁷ We therefore investigated the mechanisms underlying both intrinsic and extrinsic BIRC6 signaling activation in aged HSCs. Consistent with our previous findings, the cAMP–PKA–CREB signaling pathway is pronouncedly activated in aged HSCs (Fig. 2a, b). Given its key role in modulating cell survival,¹⁸ we proposed that this pathway could be a central mediator of the survival advantage in aged HSCs. Since CREB is a crucial transcriptional regulator, we further investigated whether it directly regulates Birc6 expression. Bioinformatic analysis of the Birc6 promoter using the JASPAR and AnimalTFDB 3.0 databases identified three putative CREB binding motifs (P1–P3) (Fig. 2c). Chromatin immunoprecipitation (ChIP)-qPCR indicated that CREB binds preferentially to the P1 and P2 sites, instead of P3, in the promoter region (Fig. 2c, Supplementary Fig. 2a), with stronger binding in middle-aged HSCs (Fig. 2d). To assess functional relevance, we constructed luciferase reporters containing wild-type (WT) or mutant (MUT) versions of the P1 and P2 sites (Supplementary Fig. 2b). Luciferase assays showed that CREB over expression enhanced activity only in WT reporters (Fig. 2e), demonstrating that CREB transcriptionally activates Birc6. Consistently, ex vivo induction of cAMP–PKA–CREB signaling via forskolin (a cAMP activator) elevated BIRC6 expression and reduced apoptotic priming in HSCs (Supplementary Fig. 2c, d). Conversely, PKA inhibition using its inhibitor H89 suppressed BIRC6 expression and abolished the survival advantage in middle-aged HSCs (Supplementary Fig. 2c, d). In addition to PKA, cAMP can transduce signals via exchange proteins directly activated by cAMP (EPAC), a pathway distinct from the classical PKA cascade.¹⁹ To delineate the downstream pathways, we utilized the PKA-selective agonist 6-MB-cAMP (6-MB) and the EPAC1-specific activator 8-pCPT-2’-O-Me-cAMP (007). 6-MB treatment significantly upregulated BIRC6 expression in HSCs, whereas 007 did not significantly affect BIRC6 expression (Supplementary Fig. 2e). Together, these findings indicate that activated BIRC6 signaling and reduced apoptotic priming in aged HSCs are dependent on the cAMP–PKA–CREB pathway.

Fig. 2.

Activated BIRC6 signaling in aged HSCs is directed by niche Tregs. a KEGG enrichment analysis of the top upregulated signaling pathways in the BM HSCs of 2 and 12 mo mice. b ELISA quantification of intracellular cAMP levels and PKA kinase activity, plus flow cytometric analysis of p-CREB protein expression, in BM HSCs from 2 and 12 mo mice (n = 6). c Schematic and ChIP analysis of the potential CREB binding sites in the Birc6 promoter (n = 6). d ChIP analysis of CREB occupancy at the Birc6 promoter in the BM Lineage^– cells of 2 and 12 mo mice (n = 6). e Luciferase reporter assays of 293T cells overexpressing CREB and transfected with reporter plasmids containing WT and MUT Birc6 promoters (n = 6; NC negative control). f cAMP contents in the BM cells of 2 and 12 mo mice (n = 6; LSK, Lineage^–Sca-1⁺c-Kit⁺ cells; Tconv conventional T cells, megakaryo megakaryocytes, mac macrophages, mono monocytes, DC dendritic cells, MSC mesenchymal stromal cells). g, h Intracellular cAMP levels, BIRC6 protein expression, apoptotic priming status, and BM HSC frequency in 12 mo Foxp3^GFP-DTR mice with or without DT treatment and Treg adoptive transfer (n = 6). i PB chimerism and donor-derived myeloid, B, and T cells, at 16 weeks post-competitive bone marrow transplantation (n = 6). j Schematic of the co-culture assay, plus flow cytometric analysis of 488-cAMP transfer from autologous BM Tregs to HSCs isolated from 2 and 12 mo mice (n = 6). k 488-cAMP transfer from autologous BM Tregs to HSCs from 12 mo mice, assessed in a transwell system, standard co-culture, or co-culture with αMHCII blocking antibody or Cx43-knockout BM Tregs (n = 6). l BIRC6 protein expression and apoptotic priming status in HSCs from 12 mo mice, assessed in a transwell system, standard co-culture, or co-culture with αMHCII blocking antibody or Cx43-knockout BM Tregs (n = 6). m Intracellular cAMP levels, BIRC6 protein expression, and apoptotic priming status in HSCs from 12 mo Foxp3^GFP-DTR mice with DT treatment, with or without Treg adoptive transfer post-DT administration (n = 6). n Scheme for αCD25 or Gap27 treatment. o cAMP contents, BIRC6 expression, apoptotic priming and frequencies of the BM HSCs of 12 mo mice with αCD25 or Gap27 treatment (n = 6). p PB chimerism and donor-derived myeloid, B, and T cells, at 16 weeks post-competitive bone marrow transplantation (n = 6). All quantitative data are presented as the mean ± SD. *P < 0.05, **P < 0.01. Unless otherwise specified, comparisons were performed using one-way ANOVA. P values for panels b, d, f, and j were calculated via a two-tailed unpaired Student’s t-test

The elevated cAMP levels were observed in middle-aged HSCs, although downregulated cAMP synthetases and upregulated cAMP-degrading enzymes (Supplementary Fig. 2f), suggesting a potential influence from the aged BM niche. Among niche cells, Tregs exhibited selective enrichment and a marked increase in cAMP in the BM of middle-aged mice (Fig. 2f). Accordingly, in vitro coculture experiments revealed that activation of the cAMP–PKA–CREB pathway, BIRC6 upregulation, and reduced apoptotic priming in aged HSCs were all dependent on Tregs, but not other niche cells (Supplementary Fig. 2g, h). In line with our previous study,¹¹ BM Tregs expand during aging and augment interaction with HSCs, impacting HSC aging (Supplementary Fig. 2i–k). Consistently, in vivo Treg depletion in middle-aged Foxp3^GFP-DTR mice significantly abrogated activation of the cAMP–PKA–CREB–BIRC6 pathway and ameliorated HSC aging, whereas adoptive transfer of BM Tregs reversed the rescue effects (Fig. 2g–i, Supplementary Fig. 2l, m). These results indicate that Tregs within the aged BM niche contribute to aberrant BIRC6 signaling in aged HSCs.

During physiological aging, the closer spatial proximity between HSCs and Tregs, together with enhanced bidirectional interaction, facilitates the formation of functional GJs.¹¹ We assessed whether Tregs trigger HSC Birc6 signaling through MHCII–TCR interactions and GJ connections, thereby conferring a survival advantage onto HSCs. Building upon a previously established method for evaluating cAMP transfer via GJ, with an Alexa Fluor 488-labeled cAMP (488-cAMP) tracer, we observed substantial cAMP transfer from Tregs to aged HSCs, whereas transfer to young HSCs was minimal (Fig. 2j). This cAMP transfer from Tregs to aged HSCs was dependent on cell contact, MHCII, and connexin 43 (Cx43)-formed GJ function (Fig. 2k). Similarly, activated BIRC6 signaling and diminished apoptotic priming in aged HSCs were also dependent on cell contact, MHCII and functional GJs, both in vitro and in vivo (Fig. 2l, m, Supplementary Fig. 2n). In addition, either Treg ablation via administration of an anti-mouse CD25 antibody (αCD25) or pharmacological blockade of GJ using the specific inhibitor Gap27 markedly attenuated activation of the cAMP–PKA–CREB–BIRC6 axis and aging phenotype of HSCs in middle-aged mice (Fig. 2n–p, Supplementary Fig. 2o, p). These findings suggest that the activation of BIRC6 signaling and the maintenance of the aging phenotypes of aged HSCs are at least partially dependent on Treg-mediated cAMP transfer via GJs.

Targeting BIRC6 mitigates HSC aging

To further clarify the intervention effects of directly targeting BIRC6 on HSC aging, we designed a panel of antisense oligonucleotides (ASOs) specifically targeting BIRC6 and profiled their effects on BIRC6 expression and the HSC phenotype. Given the higher BIRC6 expression in aged HSCs, these cells were selected for initial in vitro ASO screening (Fig. 3a). Among these ASOs, ASO-3 treatment profoundly decreased BIRC6 expression and increased apoptotic priming in aged HSCs (Fig. 3b, c, Supplementary Fig. 3a, b). Nevertheless, young HSCs displayed a moderate downregulation of BIRC6 expression following ASO-3 treatment (Supplementary Fig. 3c, d), with unaltered apoptotic priming (Supplementary Fig. 3d), suggesting reduced sensitivity to BIRC6-ASOs in young HSCs. Dose-response assessment confirmed dose-dependent BIRC6 reduction and apoptotic priming elevation in aged HSCs following ASO-3 treatment in vitro (Fig. 3d, e, Supplementary Fig. 3e). To evaluate the in vivo therapeutic efficacy of ASO-3, middle-aged mice were treated with various doses of ASO-3 via intra-bone marrow injection (Fig. 3f). ASO-3 treatment significantly reduced BIRC6 expression in aged HSCs and prompted their apoptotic priming in a dose-dependent manner (Fig. 3g, h, Supplementary Fig. 3f). In contrast, HSCs from young mice treated with ASO-3 exhibited reduced BIRC6 expression, with no significant difference in apoptotic priming (Supplementary Fig. 3g). To further enhance the in vivo targeting ability of ASOs, a widely applied HSC-targeting lipid nanoparticle (LNP) approach was utilized in ASO encapsulation in subsequent experiments (Fig. 3i).^20,21 This strategy employs LNPs surface-modified with anti-CD117 antibody for targeted HSC delivery, significantly enhancing specificity (Supplementary Fig. 3h). Middle-aged mice treated with LNP-ASOs showed markedly reduced BIRC6 expression and elevated apoptotic priming in HSCs, along with effective reversal of HSC aging phenotypes and functional impairments (Fig. 3j–l, Supplementary Fig. 3i). In contrast, HSCs from young mice treated with LNP-ASOs exhibited moderately reduced BIRC6 targeting and unchanged apoptotic priming compared to controls (Supplementary Fig. 3j). In addition, we conducted a preliminary safety assessment of LNP‑ASO treatment. Over 28 days, the body weights of LNP‑ASO-treated mice remained comparable to those of controls (Supplementary Fig. 3k), without pathological alterations detected in major organs (Supplementary Fig. 3l), indicating good tolerance of in vivo LNP‑ASO at the therapeutic dose. Together, these results indicate that targeted LNP-ASO treatment effectively inhibits BIRC6 signaling and mitigates HSC aging in middle-aged mice.

Fig. 3.

Targeting BIRC6 mitigates HSC aging. a Schematic of the ex vivo assessment of various ASO designs targeting BIRC6 in HSCs. b, c Flow cytometric quantification of BIRC6 protein expression and apoptotic priming status in the ex vivo cultured HSCs sorted from 12 mo mice. Sorted HSCs were ex vivo treated with 1 μM non-targeting, scrambled control ASO (ASO-Ctrl) or the same concentration of target ASO (ASO-1–ASO-6) (n = 6). d, e Relative Birc6 expression in the BM HSCs, flow cytometric quantification of BIRC6 protein expression and apoptotic priming status in the ex vivo cultured HSCs sorted from 12 mo mice. Sorted HSCs were ex vivo treated with ASO-Ctrl or various concentrations of ASO-3 (n = 6). f Schematic of the in vivo treatment of ASO by intra-BM injection. g, h Flow cytometric quantification of BIRC6 protein expression and apoptotic priming status in the BM HSCs from 12 mo mice treated with ASO-3 at doses of 25, 50, or 100 mg/kg (mpk) for 4 weeks (n = 6). i Schematic of the targeted LNP-ASO and in vivo treatment of the targeted LNP-ASO by intravenous injection. j, k Flow cytometric quantification of BIRC6 protein expression, apoptotic priming status and frequency of the BM HSCs from 12 mo mice treated with 10 µg LNP-ASO per mouse for 4 weeks (n = 6). l PB chimerism and donor-derived myeloid, B, and T cells, at 16 weeks post-competitive bone marrow transplantation (n = 6). All quantitative data are presented as the mean ± SD. *P < 0.05, **P < 0.01. Unless otherwise specified, comparisons were performed using One-way ANOVA. P values for panels j–l were calculated via a two-tailed unpaired Student’s t-test

Targeting BIRC6 restores youthful immune features in middle-aged mice

To determine whether reversing HSC aging could restore immunosenescence markers, we compared young mice to middle-aged mice receiving weekly LNP-ASO treatment for 4 weeks. Analyses were conducted at 8 weeks post-treatment, a duration sufficient for the clearance of non-self-renewing cells, revealed that untreated middle-aged mice exhibited an increased proportion of myeloid-biased HSCs (my-HSCs) and a reduction in common lymphoid progenitors (CLPs) compared to young controls (Fig. 4a–d). Conversely, middle-aged mice treated with targeted LNP-ASO showed a depletion of my-HSCs and a restoration of CLP proportions (Fig. 4a–d), demonstrating the rejuvenation of the HSC and progenitor compartments in middle-aged mice.

Fig. 4.

Targeting BIRC6 restores youthful immune features in middle-aged mice. a–d Representative flow cytometric quantification and the frequency of my-HSCs and CLPs in the BM of 2 and 12 mo mice with or without targeted LNP-ASO treatment (n = 6). e Flow cytometric quantification of the percentage of naive T cells (CD44⁻CD62L⁺) within the total CD4⁺ and CD8⁺ T cell populations, as well as the percentage of mature B cells (IgM⁺ IgD⁺) within the total B cell population (CD19⁺ B220⁺) in the PB (n = 6). f, g Representative flow cytometric analysis of the ratio of T cells with an exhausted phenotype relative to those with a non-exhausted phenotype (percentage of PD1⁺CD62L⁻ cells)/(percentage of PD1⁻CD62L⁺ cells), as well as the percentage of aged B cells (CD21/CD35⁻ CD23⁻) of total mature B cells (CD19⁺IgM⁺CD93⁻CD43⁻) in the PB (n = 6). h Heatmap showing the relative cytokine (IL-1α, IL-6, CXCL5, TNF-α, IFN-γ and MCP-1) levels in the plasma (n = 6). All quantitative data are presented as the mean ± SD. *P < 0.05, **P < 0.01. Comparisons were performed using One-way ANOVA

Given that aging is characterized by a critical impairment in the ability of the immune system to generate naive T and B lymphocytes capable of recognizing novel antigens, we investigated whether targeting BIRC6 could restore lymphopoiesis in middle-aged mice. Immune reconstitution was analyzed following 4 weeks post-treatment, a timeframe corresponding to lymphoid output from HSCs.²² Administration of LNP-ASO resulted in a significant elevation in the frequency of circulating naive T cells (CD44^–CD62L⁺) and mature B cells (CD19⁺B220⁺IgM⁺IgD⁺) in middle-aged mice, relative to age-matched controls (Fig. 4e, Supplementary Fig. 4a, b). Notably, LNP-ASO delivery did not trigger significant alterations in thymic weight or the composition of progenitor-cell subsets within middle-aged mice (Supplementary Fig. 4c, d). Collectively, these findings indicate that targeted inhibition of BIRC6 effectively restores the pools of naive T cells and mature B cells in the setting of aging. Beyond numerical alterations, lymphocytes from aged hosts display profound functional impairment, which is characterized by a well-documented exhausted state and pro-inflammatory phenotypic profiles.¹⁶ In middle-aged mice, we observed a marked expansion of T cells bearing an exhausted phenotype (PD1⁺CD62L^–), alongside a reduction in non-exhausted T cells (PD1^–CD62L⁺) (Fig. 4f). Targeted intervention with LNP-ASO reversed this phenotypic imbalance, as evidenced by a decreased ratio of PD1⁺CD62L^– T cells relative to PD1^–CD62L⁺ T cells (Fig. 4f). Middle-aged mice exhibit progressive accumulation of age-associated B cells (ABCs) (Fig. 4g), a cell population linked to humoral immunodeficiency.²³ Targeted LNP-ASO conditioning significantly reduced the proportion of these cells compared to untreated controls (Fig. 4g). These findings collectively demonstrate that targeting BIRC6 not only augments naive lymphocyte production but also partly mitigates aging phenotypes of lymphocytes. Otherwise, after in vivo LNP-ASO treatment, BIRC6 expression in CLPs and naive lymphocytes showed no significant changes (Supplementary Fig. 4e), suggesting a minimal direct effect of LNP-ASO on BIRC6 expression in these cells.

Aging is also associated with systemic inflammation driven by pro-inflammatory cytokines, which have been linked to HSC dysfunction and myeloid skewing. Assessment of plasma cytokine levels revealed elevated IL-1α, IL-1β, and CXCL5 in middle-aged mice, and these levels were reduced following LNP-ASO treatment (Fig. 4h). Additionally, senescence affects the phenotype and function of immune cells.^2,3,24 Compared to young mice, BM cells from middle-aged mice exhibited increased levels of senescence and SASP markers (Supplementary Fig. 4f), which were reversed following LNP-ASO treatment (Supplementary Fig. 4f), suggesting that targeting BIRC6 partially alleviates cellular senescence of BM cells in middle-aged mice. These data indicate that targeting BIRC6 in aged HSCs restores both lymphoid compartment homeostasis and systemic inflammatory balance during aging.

Targeting BIRC6 enhances functional immunity in middle-aged mice

A defining feature of immunosenescence is impaired anti-infection immunity and diminished vaccine-induced protective responses, a phenomenon that was highlighted throughout the COVID-19 pandemic.²⁵ To investigate whether targeting BIRC6 could augment functional protective immunity against pathogenic infection, we assessed the immune responses of mice immunized with a SARS-CoV-2 Spike protein vaccine formulated with aluminum hydroxide (alum) adjuvant (Fig. 5a). Given that effective immune protection induced by protein vaccines relies on the indispensable and non-redundant functions of both B and T cells, we first examined antibody responses. As anticipated, middle-aged mice exhibited significantly lower vaccine-elicited antibody titers compared to young mice (Fig. 5b). However, targeted LNP-ASO treatment in middle-aged mice led to a robust humoral immune response, demonstrating a 22-fold increase in titer compared to the untreated group at 28 days post-vaccination (Fig. 5b). Moreover, LNP-ASO treatment ameliorated the age-related impairment observed in middle-aged mice regarding the expansion of germinal center (GC) B cells, spike-specific GC B cells, and memory B cells (Fig. 5c–e, Supplementary Fig. 5a). Similarly, age-associated alterations in splenic follicular T helper (Tfh) cells, plasmablasts (PBs) and plasma cells (PCs) levels were reversed in aged animals following LNP-ASO administration (Fig. 5f, g), while the proportion of follicular regulatory T (Tfr) cells within GC responses remained nearly unaltered during aging and upon LNP-ASO therapy (Supplementary Fig. 5b). Nevertheless, there was no significant difference in antigen-specific antibody titers nor germinal center responses between the LNP-ASO-treated group and the control group of young mice (Supplementary Fig. 5c–e). Collectively, these results suggest that targeting BIRC6 in aged HSCs at least partially enhances functional immunity in middle-aged mice.

Fig. 5.

Targeting BIRC6 enhances functional immunity in middle-aged mice. a Schematic of mouse immunization. b Spike-specific IgG titers of 2 and 12 mo mice with or without targeted LNP-ASO treatment post immunization (n = 6). c–e Representative flow cytometric analysis and the frequency of germinal center (GC) B cells, spike-specific GC B cells, and memory B cells (n = 6). f, g Representative flow cytometric analysis and the frequency of splenic follicular T helper (Tfh) cells, plasmablasts (PBs) and plasma cells (PCs) (n = 6). All quantitative data are presented as the mean ± SD. *P < 0.05, **P < 0.01. Comparisons were performed using one-way ANOVA

Modulating BIRC6 in aged human HSCs

HSC functional deterioration driven by physiological aging is a well-conserved trait observed in both murine models and humans.⁶ Consistent with this cross-species conservation, augmented MHCII-mediated antigen presentation, increased GJ formation, hyperactivation of the cAMP–PKA signaling cascade, and attenuated apoptotic priming have been identified in aged human HSCs (hHSCs).¹¹ Consistent with our murine data, reanalysis of published single-cell RNA-seq (scRNA-seq) datasets from aged hHSCs demonstrated that BIRC6 exhibits the most pronounced upregulation among all anti-apoptotic members of the IAP and BCL-2 gene families (Fig. 6a, Supplementary Fig. 6a). Using these scRNA-seq data, we stratified hHSCs from both young and aged donors into Birc6 high (BIRC6^hi) and low expression (BIRC6^lo) subpopulations. Transcriptomic profiling showed that BIRC6^hi hHSCs from older donors markedly differed from BIRC6^hi and BIRC6^lo hHSCs from young donors, whereas BIRC6^lo hHSCs from older donors were only moderately distinct (Fig. 6b). Then, we assessed the hHSC aging signature with established methods.²⁶ Intriguingly, BIRC6^hi hHSCs from older donors were significantly enriched for aging signatures (Fig. 6c). We isolated highly pure populations of circulating hHSCs (Lin^–CD34⁺CD38^–CD90⁺CD45RA^–)¹⁶ from peripheral blood mononuclear cells (PBMCs) of young and middle-aged donors (Fig. 6d). RNA-seq profiling demonstrated that, compared to young hHSCs, hHSCs from middle-aged donors displayed marked enrichment of the aging signature, exclusive BIRC6 upregulation and potential HSC–Treg interaction (Fig. 6e, f, Supplementary Fig. 6b), consistent with observations in aged BM hHSCs. Flow cytometry further confirmed profoundly elevated BIRC6 levels, declined apoptotic priming, and aging phenotypes in middle-aged donor hHSCs (Fig. 6g, h). Intriguingly, treatment with a human BIRC6-targeting ASO significantly downregulated BIRC6 expression, enhanced apoptotic priming, and attenuated aging phenotypes in middle-aged donor-derived hHSCs (Fig. 6i–k, Supplementary Fig. 6c–g), without affecting young hHSCs (Supplementary Fig. 6h).

Fig. 6.

Modulating BIRC6 in aged hHSCs. a Expression of the anti-apoptotic IAP family members in BM hHSCs of young and old donors (n = 186 and 121, left to right). b Heatmap showing differential gene expression of BIRC6^hi and BIRC6^lo BM hHSCs of young and old donors. c Bubble plots showing enrichment of aging signatures of BIRC6^hi and BIRC6^lo BM hHSCs from young and old donors. d Representative flow cytometry analysis of circulating hHSCs (Lin^–CD34⁺CD38^–CD90⁺CD45RA^–). e GSEA of the aging signature of circulating hHSCs from young and middle-aged donors. f Heatmap showing the expression of genes of the anti-apoptotic IAP and BCL-2 family members in circulating hHSCs from young and middle-aged donors (n = 5). g, h Flow cytometric quantification of BIRC6 protein expression and apoptotic priming status in circulating hHSCs from young and middle-aged donors (n = 5). i GSEA of the aging signature of hHSCs treated with human BIRC6-tagarting ASO ex vivo. j, k Flow cytometric quantification of BIRC6 protein expression and apoptotic priming status in hHSCs treated with the human BIRC6-tagarting ASO ex vivo (n = 5). l Schematic illustration of how targeting BIRC6 ameliorates HSC aging and immunosenescence. All quantitative data are presented as the mean ± SD. *P < 0.05, **P < 0.01. Comparisons were performed using a two-tailed unpaired Student’s t-test

Discussion

HSC aging is one of the central processes of age-related immune decline, chronic inflammation, and hematologic malignancies.⁷ However, the signaling pathways that promote the survival and functional decline of aged HSCs remain poorly defined, hindering the development of rejuvenation therapies. We previously identified a subpopulation of aged HSCs with upregulated MHCII expression and accumulated mutant load, accompanied by expansion of microenvironmental Tregs in aged mice. The expanded Tregs confer a survival advantage onto aged HSCs by transferring cAMP via GJs,¹¹ although the precise mechanisms remain elusive. In this study, we identified a novel aged HSC subpopulation characterized by high expression of BIRC6 and elucidated a novel regulatory mechanism mediated by BM Tregs via the cAMP–PKA–CREB–BIRC6 axis. Notably, we developed an efficient HSC-targeted LNP-ASO delivery system and demonstrated that targeting BIRC6 signaling effectively reverses HSC aging and immunosenescence, thus providing a new therapeutic target and a highly promising strategy for intervening in age-related immune decline (Fig. 6l).

Despite the fact that the structural basis and effects of BIRC6-mediated apoptosis regulation have recently been elucidated,¹³ its direct targeting strategies are lacking. Here, we implemented BIRC6-targeted intervention using a published HSC-targeting LNP approach,²⁰ a critical advantage over conventional delivery methods, which has been widely applied in HSC‑targeted delivery.^27–29 Targeted LNP-ASO treatment effectively modulated BIRC6 signaling in aged HSCs, not only reversing age-related hematopoietic-immune imbalance and enhancing vaccine responses but also minimizing potential off-target effects. Otherwise, targeting BIRC6 modestly altered other apoptotic regulators, suggesting that BIRC6 inhibition may perturb the broader apoptotic network, affecting HSC survival homeostasis. Further work would benefit from more detailed preformulation and stability assessments of LNP‑ASOs for targeting specificity and safety. Indeed, the survival fate of aged HSCs is likely regulated by integrated intrinsic and extrinsic signals.^6,12 While BIRC6 targeting partly accounts for the increased apoptotic priming we observed, the precise molecular mechanisms of BIRC6 and its downstream network remain unclear. For instance, classical Wnt signaling maintains young HSC homeostasis, while aged HSCs upregulate Wnt5a, suppressing classical Wnt signaling and activating non-canonical Wnt pathways, leading to loss of polarity, reduced regenerative capacity, and myeloid-biased differentiation.^30,31 While the cAMP–PKA–CREB–BIRC6 and non-canonical Wnt5a-Cdc42 pathways function independently in HSC aging, they may engage in crosstalk and share regulatory nodes. Future studies are needed to perform serial transplantation and in vivo tracing of BIRC6-high HSCs and to investigate if targeting BIRC6 indirectly modulates alternative signalings, such as the Wnt5a–Cdc42 axis, thereby comprehensively elucidating the HSC aging signaling network.

HSCs are essential for lifelong immune cell production. While young individuals possess predominantly multipotent (balanced) HSCs with some lineage-biased subpopulations, aging drives a significant shift towards myeloid-biased differentiation. Myeloid skewing in the elderly correlates with impaired adaptive immunity, chronic inflammation, and heightened risk of myeloid malignancies.³² Consequently, interventions that restore ‘youthful’ HSC function hold promise for reversing aged immunity, reducing chronic inflammatory disease burden, and preventing hematological disorders.¹⁶ Notably, cross-sectional studies showed that aging-associated hematopoietic phenotypes manifest in mice during midlife (9–12 months), which corresponds to humans aged 37–47 years.³³ While middle-aged HSCs share certain molecular features with old HSCs, transplantation studies showed that their function can be restored by a young BM microenvironment,³⁴ suggesting a potential intervention window. In this study, we demonstrated that targeted inhibition of BIRC6 signaling in middle-aged HSCs rebalances hematopoiesis, enhances adaptive immunity, and reduces systemic inflammation, with minimal off-target effects. These findings establish HSC-targeting as a viable strategy to ameliorate aged immunity in older adults, with the potential to reduce infection-related morbidity and mortality. In infectious disease contexts, this approach could serve as an immune adjuvant,³⁵ improving vaccine responsiveness and durability in older adults against pathogens such as influenza and SARS-CoV-2. Beyond rebalancing hematopoiesis and revitalizing adaptive immunity, BIRC6 inhibition in middle-aged mice also lowered circulating levels of key pro-inflammatory cytokines and senescence markers, underscoring its potential to counter immunosenescence and reduce age-related inflammatory conditions. In summary, our findings not only delineate a new signaling axis driving HSC aging but also establish a versatile therapeutic avenue for systematically countering aging-related immune decline and its wide-ranging clinical sequelae.

Thymic alterations during LNP-ASO treatment may also influence immunosenescence by affecting CLP or naive T cell compartments.³⁶ Nevertheless, our current analyses revealed that LNP-ASO administration did not induce significant changes in thymic weight or thymic progenitor‑cell subsets in middle‑aged mice. These observations align with previous studies,^16,37 indicating that age-related immune phenotypic and functional changes are primarily driven by HSC functional alterations and largely thymus-independent. The phenotypic outcomes and underlying mechanisms likely differ between our aged HSC-targeted strategy for immunosenescence reversal and thymus-targeted approaches for the same indication.³⁸ Moreover, our analysis does not fully delineate the effects of BIRC6-targeted interventions in hHSCs. In the future, lineage tracing and transplantation assays performed in NSG or other immunodeficient mouse models will be required to comprehensively characterize the aging-related functional properties of hHSCs. While previous studies have confirmed that most hHSC subsets are detectable in peripheral blood and that circulating hHSCs exhibit high transcriptomic similarity to BM hHSCs,^39,40 the direct relationship between BM hHSCs and the circulating compartment in humans remains to be elucidated. Although mouse models have defined conserved aging transcriptional signatures,¹⁵ the direct translation of these findings to humans is complicated by significant biological differences. The lack of a unified, robust hHSC aging transcriptomic signature hinders its utility as a standardized metric in research. To address these limitations, emerging multidimensional immunosenescence assessment frameworks⁴¹ and refined multi-marker algorithms for senescence and SASP markers^5,42 provide standardized methodologies for cross-study comparisons. Collectively, further studies should investigate the systemic and long-term effects of immunosenescence interventions on different immune organs and cell populations across maturation stages under physiological aging or stress responses⁴³ and validate the utility of aging signature tools (spanning both stem cell aging and cellular senescence metrics) to guide precise intervention.

Tregs play a crucial role in the maintenance of tissue homeostasis and stem cell fitness.⁴⁴ In the BM of young mice, Tregs support HSC homeostasis and immune privilege via paracrine signals such as adenosine and IL-10.^45,46 Under lethal irradiation, Tregs enhance HSC survival and mitigate radiation-induced BM failure.⁴⁷ These findings suggest that the non-immunomodulatory roles of niche Tregs may promote HSC survival. Our study reveals a fundamental, age-dependent shift in Treg–HSC crosstalk from paracrine-mediated homeostasis in youth to direct cell contact-mediated survival support during aging. This paradigm may provide insights into rejuvenating aged HSCs by targeting the aged Treg niche. During aging, we propose that Treg-mediated support for HSCs, while initially beneficial for maintaining the pool, ultimately constitutes a maladaptive symbiosis that preferentially sustains dysfunctional aged HSCs, disrupts balanced hematopoiesis, and contributes to aged immunity.⁷ Although this aberrant cooperation transiently maintains HSC numbers, it ultimately drives a vicious cycle of immune aging at the cost of immune diversity and function.⁴⁸ Therefore, targeting this pathological Treg–HSC interaction,¹¹ rather than depleting Tregs, represents a more precise and potentially safer therapeutic strategy. Future work should systematically elucidate the cAMP–PKA–CREB–BIRC6 axis in regulating aged HSC survival and lineage commitment at single-cell resolution in vitro and in vivo,⁴⁹ along with the development of more specific and efficient targeting strategies, to advance HSC aging and immunosenescence intervention.

Materials and methods

Animals

Wild-type (WT) C57BL/6 mice, including young (2-month-old) and middle-aged (12-month-old) experimental cohorts, were obtained from HFK Bioscience Co., Ltd. (Beijing, China) and SPF Biotechnology Co., Ltd. (Beijing, China). Multiple transgenic mouse lines on a congenic C57BL/6 background were acquired from validated commercial sources as follows: B6.Cg-Tg(Mx1-cre)1Cgn/J (Mx1-Cre), B6.Cg-Foxp3^tm2Tch/J (Foxp3^EGFP) and B6.129(Cg)-Foxp3^{tm3(DTR/GFP)Ayr}/J (Foxp3^GFP-DTR) lines were sourced from The Jackson Laboratory (Bar Harbor, ME, USA); B6-Foxp3^{tm1(YFP/icre)/}Cya (Foxp3^YFP-DTR) mice were procured from Cyagen Biosciences (Guangzhou, China); B6/JGpt-Ptprc^{em1Cin(p.K302E)}/Gpt (CD45.1) mice were obtained from GemPharmatech (Nanjing, China); B6.Cg-Birc6^em1(flox)St (Birc6^f/f) mice were generated by Gene Sci-Tech Targeting Biotechnology Co., Ltd (Guangzhou, China). To generate HSC-specific Birc6 knockout (KO) mice, Birc6^f/f mice were interbred with Mx1-Cre transgenic mice. For all in vivo assays, only male transgenic mice and age-matched WT littermate controls were utilized. All mice were background-matched, randomly allocated to experimental groups, and housed in strict specific pathogen-free (SPF) barrier facilities with ad libitum access to autoclaved rodent chow and sterile drinking water. All animal experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of the Laboratory Animal Center (approval No. IACUC-DWZX-2025-025), and all procedures were performed in strict compliance with institutional ethical guidelines for laboratory animal care and use.

Human HSC enrichment and multiplication

Peripheral blood mononuclear cells (PBMCs) from healthy donors were obtained from Shanghai Liquan Hospital by Milestone Biotechnologies (China), and the collection of PBMCs was approved by the Ethics Committee of Shanghai Liquan Hospital (Z-ZJMS-21-08-001), and the study was carried out in accordance with the Declaration of Helsinki. Frozen PBMCs were thawed in a 37 °C-water bath and washed with RPMI 1640 medium. Then, CD34⁺ cells were magnetically enriched per the manufacturer’s instructions (Miltenyi Biotec). These CD34⁺ cells were stained with antibodies (CD34, CD90, CD38, CD45RA, lineage markers) and circulating hHSCs were subsequently sorted using a BD FACSAria cell sorter.

Antisense oligonucleotide (ASO) preparation and in vitro treatment

All ASOs were designed and synthesized by Synbio Technologies Co., Ltd. (Suzhou, China). A non-targeting, scrambled control ASO (ASO-Ctrl) was used at the same concentration as the target ASOs (ASO-1–ASO-6). To analyze the cellular uptake of ASOs, FAM-conjugated ASOs (Synbio Technologies) were utilized for the experiments. For all in vitro experiments, cells were treated with ASOs targeting BIRC6 at concentrations ranging from 0.25 to 4 µM. ASOs were kept at −20 °C and used within 12 h of preparation. ASO sequences and chemical modification were detailed in Supplementary Table 1.

Targeted lipid nanoparticle (LNP)-ASOs preparation

ASOs were encapsulated into LNPs using a self-assembly process adapted from established methods.²⁰ Briefly, an ethanolic lipid mixture containing an ionizable cationic lipid, phosphatidylcholine, cholesterol, and a PEG-lipid was rapidly mixed with an acidic aqueous solution containing the ASO. For subsequent experiments requiring cell targeting, LNPs were functionalized with a specific mouse CD117 monoclonal antibody (mAb). Purified rat anti-mouse CD117 mAb (clone 2B8, BioLegend) was conjugated to LNPs using N-succinimidyl S-acetylthioacetate (SATA)-maleimide chemistry, as previously described.⁵⁰ Briefly, LNP-ASOs were first modified with maleimide groups via the post-insertion technique using DSPE-PEG-mal, while the purified rat anti-mouse CD117 mAb was thiolated using SATA (Thermo Fisher Scientific). Following SATA deprotection with 0.5 M hydroxylamine hydrochloride and removal of unreacted components using Zeba spin desalting columns (Thermo Fisher Scientific), the reactive thiol groups on the antibody were conjugated to the maleimide groups on the LNPs via thioether bond formation. The resulting conjugated LNP-ASO-mAb complexes were then purified by size-exclusion chromatography using Sepharose CL-4B gel filtration columns (Millipore Sigma), and the ASO content within the purified targeted LNPs was quantified using a modified Quant-iT RiboGreen RNA Assay Kit (Thermo Fisher Scientific). Targeted LNP-ASOs were kept at 4 °C and used within 24 h of preparation. The role of the major material in LNP-ASO preparation was detailed in Supplementary Table 2.

Pharmacological treatment

For in vivo ASO intervention, mice were administered once-weekly intra-bone marrow (intra-BM) injections of either control ASO (ASO-Ctrl) or ASO-3 at escalating dosages of 25, 50, or 100 mg/kg body weight, for a continuous 4-week treatment. For interventions using LNP-encapsulated ASO, mice received once-weekly intravenous (i.v.) injections of LNP-formulated ASO at a fixed dose of 10 µg per animal, over 4 weeks. In vivo Treg depletion via αCD25 was performed in accordance with a previously validated protocol that enables efficient depletion of BM Tregs.¹¹ For this intervention, 11-month-old mice were given once-weekly intraperitoneal (i.p.) injections of 0.5 mg per mouse of InVivoPlus anti-mouse CD25 antibody (BioXCell, Lebanon, NH, USA) across 4 weeks. For in vivo gap junction inhibition with Gap27, 11-month-old mice received i.p. injections of 0.5 µg/kg GAP27 peptide (MedChem Express, Monmouth Junction, NJ, USA) on alternate days for 4 weeks.

Flow cytometry and cell sorting

Prior to antibody staining, cells were Fc-blocked by incubation with 1 mg/mL rat IgG (Abcam) for 30 min on ice. Antibody staining was performed in Flow Cytometry Staining Buffer (eBioscience, San Diego, CA, USA). Incubations were carried out on ice for 30 min. For immunophenotypic analysis of HSCs gated as Lineage^–Sca-1⁺c-Kit⁺CD150⁺Flt3^–CD48^–, BM cells were stained with the panel of fluorochrome-conjugated antibodies listed below, each of which was pre-validated for binding specificity and optimized to its ideal working titer: Sca-1 (D7), c-Kit (2B8), CD150 (TC15-12F12.2), Flt3 (A2F10), CD48 (HM48-1), and a mature lineage marker mix [CD3e (145-2C11), B220 (RA3-6B2), Gr-1 (RB6.8C5), Mac-1 (M1/70), and Ter-119 (Ter119)] antibodies (all eBioscience). To analyze the cellular uptake of ASOs, FAM fluorescence was used as a fluorescent label for ASOs. For testing of my-HSCs and CLPs, cells were stained with the antibodies mentioned above, including CD127 antibody (A7R34, eBioscience). My-HSCs and CLPs were defined as follows: my-HSCs (Lineage^–Sca-1⁺c-Kit⁺ Flt3⁻CD34⁻CD150^high), CLP (Lineage^–CD127⁺ Sca-1⁺c-Kit⁺). For hHSC (Lineage^–CD34⁺CD38^–CD90⁺CD45RA^–) phenotypic analysis, PBMCs were stained with fluorochrome-conjugated antibodies according to a previous study.⁵¹

For T cell phenotyping, cells were stained with CD3 (17A2), CD8α (53-6.7), CD4 (GK1.5), CD44 (IM7), CD62L (MEL-14), and PD-1 (RMP1-30) antibodies (all eBioscience). For B cell analysis, cells were surface-stained with the following antibodies: CD43 (S7), CD21/CD35 (7E9), IgM (eB121-15F9), CD19 (1D3), IgD (11-26 c.2a), CD93 (AA4.1), CD23 (B3B4), and CD45R/B220 (RA3-6B2) antibodies (all eBioscience).

For immune protection phenotyping, cells were stained with IgD (11-26c.2a, Biolegend), CD38 (90, Thermo Fisher), FAS (Jo2, BD), GL7 (GL7, Thermo Fisher), CD45R/B220 (RA3-6B2, Biolegend), CXCR5 (L138D7, Biolegend), CD138 (281-2, Biolegend), PD-1 (RMP1-30, Biolegend), IgM (RMM-1, Biolegend), and streptavidin-Brilliant Violet 421^TM (Biolegend) antibodies. SPIKE-binding MBCs and GCs were detected using biotinylated SPIKE (prepared in-house by the laboratory, and the preparation method is detailed in the Supplementary Materials). For testing follicular T helper (Tfh) and follicular regulatory T (Tfr) cells, after cell surface staining, fixation and permeabilization were carried out using the Foxp3/transcription factor staining buffer set (eBioscience) in accordance with the manufacturer’s instructions. Cells were then stained with a Foxp3 antibody (clone FJK-16s, BioLegend) for 30 min at room temperature, followed by flow cytometric analysis.

For intracellular ROS assessment, following surface staining, the DCFDA/H2DCFDA kit (Abcam) was used according to the manufacturer’s protocol. Cells were stained with 20 μM DCFDA in 1× buffer for 30 min at 37 °C, washed once with 1× buffer, and analyzed by flow cytometry.

Cells were sorted using a FACSAriaⅡ or analyzed using an LSRFortessa (all BD Biosciences, San Jose, CA, USA) flow cytometer, as well as Sony ID7000^TM spectral analyzer (Sony Biotechnology, San Jose, CA, USA), with optimization for antigen accessibility and fluorochrome compatibility. All flow cytometry data were analyzed using FlowJo V10 software (Treestar Inc., San Carlos, CA, USA). The antibodies used in Flow cytometry were detailed in Supplementary Table 3.

Apoptotic priming assay

Apoptotic priming of HSCs was measured using a well-established protocol.¹² Briefly, Sca-1⁺ cells were isolated via immunomagnetic enrichment using the EasySep^TM Mouse SCA1 Positive Selection Kit (StemCell Technologies) and stained with validated HSC markers. Cells were then resuspended at 5 ×10⁵/mL in DTEB buffer (0.1% BSA, 5 mM succinate, 10 mM HEPES–KOH, 50 mM KCl, 135 mM Trehalose, 20 μM EDTA and 20 μM EGTA, final pH 7.4) and incubated with 8 μM BIM or 3 μM BID peptide (both MedChem Express) for 30 min. Tetramethylrhodamine (TMRE; Thermo Fisher Scientific; 20 nM final concentration) was added for the last 15 min of the incubation period. Mitochondrial depolarization was analyzed by flow cytometry, with DMSO (Sigma-Aldrich) and FCCP (MedChem Express) included as negative and positive controls, respectively. For ex vivo cultured HSCs, 5 ×10⁵ Sca-1⁺ cells were either cocultured with 1 ×10⁵ autologous BM Tregs or treated with Forskolin (10 μM) or H89 (10 μM, all MedChem Express). For hHSCs, cells were incubated with 1 μM human BIM (EIAab) according to a previous study.⁵² The percentage of apoptotic priming (mitochondrial depolarization) was calculated as

$$\begin{array}{c}\mathrm{\%}\mathrm{D}\mathrm{e}\mathrm{p}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r}\mathrm{i}\mathrm{z}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\left(\mathrm{a}\mathrm{p}\mathrm{o}\mathrm{p}\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{p}\mathrm{r}\mathrm{i}\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{g}\right)\hfill \\ =100(1-([{\mathrm{M}\mathrm{F}\mathrm{I}}_{\mathrm{B}\mathrm{H}3\mathrm{p}\mathrm{e}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{d}\mathrm{e}}-{\mathrm{M}\mathrm{F}\mathrm{I}}_{\mathrm{F}\mathrm{C}\mathrm{C}\mathrm{P}}]/[{\mathrm{M}\mathrm{F}\mathrm{I}}_{\mathrm{D}\mathrm{M}\mathrm{S}\mathrm{O}}-{\mathrm{M}\mathrm{F}\mathrm{I}}_{\mathrm{F}\mathrm{C}\mathrm{C}\mathrm{P}}]\left)\right)\hfill \end{array}$$

HSC culture and HSC–Treg interaction assay

Ex vivo HSC culture was carried out with an established HSC functional maintenance protocol as previously reported.¹¹ Briefly, purified HSCs were seeded into fibronectin-coated 24-well culture plates, and maintained in F12 basal medium supplemented with the following components: 1% insulin–transferrin–selenium–ethanolamine, 1% penicillin/streptomycin/L-glutamine mix, 10 mM HEPES (all reagents from Gibco); recombinant murine SCF (10 ng/mL), recombinant murine TPO (100 ng/mL; both from PeproTech, Rocky Hill, NJ, USA); and polyvinyl alcohol (PVA; 1 mg/mL, Sigma-Aldrich). All cultures were incubated in a humidified incubator at 37 °C with 5% CO₂. To model the direct cellular crosstalk between HSCs and BM Tregs ex vivo, 1 × 10³ purified HSCs were co-cultured with 1 × 10⁴ autologous BM-isolated Tregs at a ratio of 1:10. For mechanistic intervention assays, co-cultures were either set up using a transwell insert system to physically separate HSCs and Tregs, or supplemented with anti-MHCII (αMHCII; 10 µg/mL, clone M5/114.15.2; BioXCell), GAP27 (300 µM, MedChem Express), Forskolin (10 μM, MedChem Express), H89 (10 µM, MedChem Express), 6-MB-cAMP (50 μM, MedChem Express), or 007 (50 µM, MedChem Express).

Human cells in vitro culture and assays

hHSCs (Lin^–CD34⁺CD38^–CD90⁺CD45RA^–) purified by FACS were cultured in StemSpan^TM SFEM II medium (StemCell Technologies) formulated with SCF (10 ng/mL, PeproTech), TPO (10 ng/mL, PeproTech), IL-6 (50 ng/mL, PeproTech), and FLT3L (50 ng/mL, PeproTech) according to an established protocol.⁵³ For the human HSC differentiation assay, 1 × 10⁴ MS-5 cells (Immocell Biotechnology, China) were first seeded into each well of a 24-well plate containing 1 mL α-MEM medium supplemented with ribonucleosides, deoxyribonucleosides, 10% fetal bovine serum (FBS), 2 mM L-glutamine, and 2 mM sodium pyruvate. After 24 h, the medium was replaced with differentiation medium consisting of StemPro-34 SFM with Nutrient Supplement (Gibco), supplemented with SCF (20 ng/mL, PeproTech), TPO (20 ng/mL, PeproTech), EPO (10 ng/mL, PeproTech), Flt3-L (20 ng/mL, PeproTech), IL-2 (10 ng/mL, PeproTech), IL-3 (10 ng/mL, PeproTech), IL-6 (50 ng/mL, PeproTech), IL-7 (20 ng/mL, PeproTech), IL-11 (50 ng/mL, PeproTech), GM-CSF (20 ng/mL, PeproTech), human LDL (50 ng/mL, Stem Cell Technologies), 1% penicillin–streptomycin (Gibco), and 1% L-glutamine (Gibco). hHSCs treated with hASO were added to the MS-5 co-culture system; half-medium changes were performed every 4 days.

RNA-seq

Cells were sorted and total RNA was extracted using the RNeasy Micro Kit (QIAGEN, Hilden, Germany), with RNA concentration and purity quantified via a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific). Libraries were prepared and sequenced on an Illumina NovaSeq^TM 6000. Adaptors, undetermined bases, and low-quality bases were trimmed using Fastp (https://github.com/OpenGene/fastp). Reads were aligned, and gene expression was quantified using RSEM (v1.2.12). Differentially expressed genes (DEGs) were identified with DESeq2 (fold change > 2.0, adjusted P < 0.05). Gene set enrichment analysis (GSEA, Broad Institute) was performed using Molecular Signatures Database v6.0. Signaling pathways were analyzed by Ingenuity Pathway Analysis (QIAGEN) and Kyoto Encyclopedia of Genes and Genomes (KEGG).

Aging Signature, CellRadar and MHCII score

The aging signature of murine HSCs was assessed using an established Aging Signature Web tool (https://eriba.umcg.nl/agingsignature).¹⁵ The lineage potential of HSCs was analyzed through the CellRadar Web tool (https://karlssong.github.io/cellradar), utilizing public data (GSE156807).⁵⁴ The MHCII score was calculated using the score_genes function in Scanpy. The MHCII genes, including H2-Aa, H2-Ab1, and H2-Eb1 were defined in a previous study.⁵⁵ Prior to score calculation, the data were log-normalized and scaled. The aging signature of hHSCs was assessed using an established aging signature according to a previous study.²⁶

Treg depletion and adoptive transfer

In vivo Treg depletion was achieved via diphtheria toxin (DT) administration, using a previously validated protocol.¹¹ 11-month-old Foxp3^GFP-DTR mice were administered i.p. injections of DT at a dose of 10 ng per gram of body weight (Sigma-Aldrich, St. Louis, MO, USA) twice weekly, with injections spaced 1 day apart, for 4 weeks. For BM Treg adoptive transfer assays, 1 × 10⁵ BM-isolated Tregs (either wild-type or Cx43-knockout) were purified from 2-month-old Foxp3^YFP-DTR donor mice and delivered via i.v. infusion into 11-month-old Foxp3^GFP-DTR recipient mice after completion of the first week of the DT depletion regimen. All recipient mice were humanely euthanized for downstream experimental analysis at 4 weeks following the initiation of DT treatment.

Statistical analysis

All statistical analyses were carried out using GraphPad Prism 9.0 software (GraphPad Software, San Diego, CA, USA). Quantitative data throughout the study are presented as the mean ± standard deviation (SD), with the number of independent biological replicates (n) for each experiment explicitly stated in the corresponding figure legends. Comparisons between two groups utilized unpaired two-tailed Student’s t-tests, while comparisons across three or more groups employed one-way analysis of variance (ANOVA) with Tukey–Kramer post hoc testing, with statistical significance defined as P < 0.05. The vast majority of ex vivo functional experiments were repeated across multiple independent biological replicates.

Author contributions

Conceptualization: W.L., J.Z., J.W., and J.X. Methodology: W.L., X.Z., Y.Y., Y.Y., Y.W., J.Z., J.W., and J.X. Investigation: W.L., F.S., S.W., X.C., Y.W., Y.Z., Z.Z., and X.Z. Visualization: W.L. and Y.W. Funding acquisition: W.L., J.Z., J.W., and J.X. Project administration: W.L., J.Z., J.W., and J.X. Supervision: J.Z., J.W., and J.X. Writing—original draft: W.L., J.Z., J.W., and J.X. Writing—review and editing: W.L., XZ., Y.W., J.Z., J.W., and J.X. All authors have read and approved the article.

Data availability

The transcriptomic dataset generated in this study is deposited in the Genome Sequence Archive (GSA) database BioProject under accession number PRJCA042290. Additionally, published datasets (GSE156807, GSE59114, and GSE175604) were reanalyzed herein with explicit permission from the respective authors. All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41392-026-02787-1.