Aging-declined RNA exportation impairs hematopoietic stem cells by inducing R-loop

Abstract

Aging-related accumulation of DNA damage adversely affects hematopoietic stem cell (HSC). However, the mechanisms underlying this accumulation and strategies for its elimination to rejuvenate aged HSC remain largely obscure. This study uncovers a notable surge in R-Loop presence within aged HSC, notably co-localized with γH2AX and replication protein A (RPA), and correlated with RNA residency in the nucleus. Targeted induction of R-Loop impairs the function of HSC. Mechanistically, RNA exportation is compromised in aged HSC due to a decline in Alyref, the primary constituent of the transcription- export complex (TREX). Specifically, Alyref dysfunction results in RNA retention within the nucleus, mimicking the functional characteristics of aged HSC. The nuclear accumulation of RNA leads to the formation of RNA:DNA hybrids, known as R-Loop structures, consequently inducing replication stress and DNA damage. Introducing a quantitative boost of Alyref in aged HSC notably reinstates RNA transportation, diminishes R-Loop formation and replication stress, and ultimately enhances the performance of aged HSC. Taken together, our research demonstrates the initial revelation that aging-triggered replication stress stems from abnormal RNA transportation-propelled R-Loop configurations, hinting at the potential of quantitatively modulating RNA transportation to mitigate the physiological drawbacks of aging on HSC.

Introduction

The diminished regenerative capability of tissue-specific stem cells due to aging contributes significantly to disruptions in tissue equilibrium during the aging process.^1-3 Hematopoietic stem cells (HSC) are responsible for life-span cell production,⁴ and experience a decline in reconstitution capacity with aging.⁵ Moreover, aged HSC display a bias towards myeloid lineage differentiation,^6-8 posing a risk factor for various hematopoietic malignancies.⁹ Therefore, elucidating the molecular mechanisms underlying HSC aging and devising strategies to anti-aging intervention for aged HSC function will unquestionably benefit the elderly and individuals afflicted by hematopoietic system degenerative changes.

We and others have previously shown that DNA damage accumulation, characterized by γH2AX, has been considered one of the driving forces of HSC aging.^10-15 Faithful DNA replication is pivotal for the maintenance of genome stability. Any molecular event interfering with DNA replication, collectively named replication stress, makes DNA molecules more vulnerable to damage, mutations, or genomic instability.^16-18 A previous study has shown that replication stress is a potent driver of functional decline in aged HSC.¹⁰ While, the critical cellular signaling leading to replication stress in aged HSC has not been clearly elucidated. In this study, we observed that R-Loop, a three-stranded nucleic acid structure consisting of a DNA:RNA hybrid and a displaced strand of DNA, which is generated during DNA transcription,^19,20 accumulated in aged HSC, which impairs HSC by inducing replication stress and DNA damage. Mechanistically, the retention of RNA in nucleus of aged HSC, induced by the aging-declined RNA exportation, results in the increase of R-Loop.

Methods

Mice

Alyref^lox/flox mice were generated in Cyagen Biosciences Inc. (Guangzhou, China). Alyref^flox/flox mice (C57BL/6N) were generated by inserting loxP sites spanning the third exon of Alyref via homologous recombination. To achieve hematopoietic-specifically knockout mice, Alyref^flox/flox were crossed to Mx1-Cre mice. To induce Cre expression in Mx1-Cre⁺; Alyref^flox/flox mice were intraperitoneally injected with Poly I:C (25 mg/kg) every other day for 14 days. All genotyping primers are listed in Online Supplementary Table S1.

The recipients used in the competitive transplantation assays were CD45.1/2 that were the first generation of C57BL/6 (CD45.2) and B6.SJL (CD45.1) mice. All mice were housed in specific-pathogen-free facilities, and all procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Tsinghua University.

RNA fluorescent in situ hybridization

RNA fluorescent in situ hybridization (FISH) was conducted using a Ribo fluorescence in situ hybridization kit (RiboBio, C10910) in accordance with the manufacturer’s directions. For the R-Loop and mRNA co-localization assay, cells were performed for RNA FISH after cells incubated by anti-DNA-RNA hybrid, clone S9.6 and secondary antibodies.

Cytoplasmic and nuclear RNA fractionation and sequencing

The cytoplasmic and nuclear fractions derived from CD34-LSK cells were used to generate cDNA libraries, which were sequenced with 150-bp paired-end reads on an Illumina instrument by Novogene.

R-Loop peak calling

Strand-specific peaks were called by MACS3 (v3 3.0.0b1) with R-Loop and input samples. Peaks with q values less than 0.01 and enrichment score more than 10, and located outside the blacklisted region were defined as R-Loop peaks. We annotate peaks with genomic regions by R package ChIP seeker.

Single-strand DRIP-sequencing and single-strand DRIP-quantitative polymerase chain reaction

Single-strand DRIP-sequencing (ssDRIP-seq) library construction was performed according to published procedures.²¹ Freshly isolated CD34⁻ LSK cells were lysed in DNA lysis buffer and incubated at 37°C. Genomic DNA was fragmented using restriction enzymes and subjected to DRIP with the S9.6 antibody. DRIP DNA was validated by quantitative polymerase chain reaction (qPCR) or sonicated for ssDRIP-seq library construction, followed by sequencing on an Illumina NovaSeq system.

dCas9 coupled RNaseH1 D209N-mediated site-specific R-Loop

dCas9 coupled RNaseH1 D209N-mediated site-specific R-Loop was performed as previously published.²² In brief, dCas9 infused with RNaseH1 D209N was co-expressed with gRNA targeting the selected R-Loop sites in NIH-3T3 cells. mCherry⁺ cells were purified for ssDRIP-qPCR and immunofluorescence staining. The gRNA and qPCR primers are listed in Online Supplementary Table S6.

RNA sequencing

In brief, 50 HSC were sorted directly into lysis buffer. Then the RNA sequencing (RNA-seq) library was prepared using the method of Smart-seq2. After the library construction, the insertion size was assessed by Agilent Bioanalyzer 4200 system (Agilent Technologies), and the accurate insertion size was quantified by Taqman fluorescence probe of AB Step One Plus Real-Time PCR system. Clustering of the index-coded samples was performed using HiSeq PE Cluster Kit v4-cBot-HS (Illumina) following the manufacturer’s instructions. The libraries were sequenced by an Illumina Hiseq platform with 150-bp paired-end. The transcriptome sequencing was performed by ANNOROAD Gene Technology Company.

Statistical analysis

Data are shown as mean ± standard deviation (SD). Student’s t test (two-tailed unpaired) were used for comparisons (GraphPad Prism v.7.0). NS: not significant; *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. All experiments were repeated two or three times.

Results

R-Loop is increased in aged hematopoietic stem cells

We conducted immunofluoresence for R-Loop using the S9.6 antibody,²³ and observed accumulated R-Loop foci in aged HSC (Figure 1A-D), with RNase H, but not RNase III, serving as a positive control to dissolve R-Loops.^24,25 We then profiled genome-wide R-Loop at single-nucleotide resolution in young and aged HSC to systemically examine the correlation of R-Loops with aging by conducting ssDRIP-seq according to a recently developed method²¹ (Online Supplementary Figure S1A). As expected, R-Loops correlate with high G/C skewing in both young and aged HSC (Figure 1E), indicating accurate mapping of R-Loops by ssDRIP-seq data. We then examined the global distribution of age-associated changes in R-Loops across the mouse genome. The results show that most R-Loops are located in the intergenic region (~50%) and gene body region (~40%), which includes the 5’UTR, 3’UTR, exons, and introns, with no significant differences observed between young and aged HSC (Figure 1F). However, 6.6% of R-Loop peaks are located in the promoter region of young HSC, and this significantly increases to 10% with aging (Fisher’s exact test, odd ratio =1.5; P=1.97e-10) (Figure 1F). To further characterize the increase of R-Loops with aging, we performed a genome browser inspection and observed a significant increase in R-Loop signals from 2 kb upstream of the TSS to downstream of the TES in aged HSC (Online Supplementary Figure S1B). Additionally, we noted increased R-Loop coverage in aged HSC (Online Supplementary Figure S1C), which aligns with previous reports.²⁶ We quantified the peak width in aged and young samples. Consistent with these observations, the results show that the width of R-Loop peaks exhibit significant increase in aged HSC (Online Supplementary Figure S1D). To delve deeper into the molecular characteristics influenced by aging-associated increases in R-Loops, we conducted gene set enrichment analysis (GSEA) of genes ranked based on their differential R-Loop signals in the gene body region between aged and young HSC. The results indicate that the augmented presence of R-Loops is enriched with genes associated with cell aging and myeloid functions, which is a classical feature of aged HSC (Online Supplementary Figure S1E; Online Supplementary Table S7).

Figure 1.R-Loop is increased in aged hematopoietic stem cells. (A) Representative images depicting immunofluorescence staining of S9.6 in both young and aged hematopoietic stem cells (HSC). Freshly isolated HSC (CD34^- Flt3^- LSK) from young (2 months) and aged (24-30 months) mice were subjected to immunofluorescence staining. The cells were either left untreated (-RNase H) or were instead treated with RNase H (+RNase H). The panels are shown at the same exposure. The cell nuclei were stained with DAPI. (B) Representative images depicting immunofluorescence staining of S9.6 in both young and aged HSC. Freshly isolated HSC from young (2 months) and aged (24-30 months) mice were subjected to immunofluorescence staining. The cells were either left untreated (-RNase III) or were instead treated with RNase III (+RNase III). The panels are shown at the same exposure. The cell nuclei were stained with DAPI. (C-D) S9.6 foci distribution in both young and aged HSC treated with RNase H (C) or RNase III (D). (E) Metaplot showing G/C skews, A/T skews, G/C contents and base compositions in 500 bp upstream and downstream regions of R-Loop peak summits in young and aged HSC. (F) Barplot showing the genomic distribution of R-Loop strand-specific peaks (forward and reverse strand separately) in both young and aged HSC.

Accumulated R-Loop induces DNA damage and impairs the function of hematopoietic stem cells

Previous studies have demonstrated that aged HSC accumulate DNA damage, as evidenced by increased γH2AX foci.¹¹ To investigate this further, we conducted whole-genome sequencing of young and aged HSPC to identify such damage between them. Surprisingly, we observed no significant increase in insertions and deletions (InDels) in aged HSPC, suggesting that physiological DNA breaks are not elevated with aging (Figure 2A). Subsequently, we performed immunofluorescence staining for γH2AX and R-Loops in aged HSC. The results revealed co-localization between them (Figure 2B). Since γH2AX is also indicative of replication stress,¹⁰ we further investigated immunofluorescence staining for RPA, a marker of replication stress.^10,17,27,28 Remarkably, R-Loops co-localized with RPA foci in aged HSC (Online Supplementary Figure S2A), indicating an association between R-Loops and replication stress during aging.

To further explore the causal relationship between R-Loops and DNA damage, we focused on Nfe2l2, which modulates the migration and retention of HSC²⁹ and is enriched with R-Loop in aged HSC (Figure 2C). Upon introducing an R-Loop in the intron of Nfe2l2 (Online Supplementary Figure S2B; Figure 2D, refer to the Methods section for details), we observed the accumulation of RPA and γH2AX foci (Figure 2E, F), suggesting that the elevation of R-Loops indeed triggers DNA damage in HSC.

To further investigate the causation between R-Loop and the functional decline of HSC, we generated two efficient small hairpin (sh)RNA against RNaseH1 and RNaseH2A (Online Supplementary Figure S2C), the key enzymes responsible for dissolving R-Loops.^30-32 By conducting a competitive transplantation assay, we observed that knockdown of these enzymes impairs the reconstitution capacity of HSC and promotes a myeloid differentiation bias (Online Supplementary Figure S2D, E), which are key features of aged HSC. Bone marrow (BM) analysis exhibited compromised reconstitution capacity and a reduction of HSC upon knockdown of these enzymes (Online Supplementary Figure S2F, G). Moreover, we purified GFP⁺ hematopoietic stem and progenitor cells from the recipients at the end of the third month and examined the level of R-Loop. The results show that the level of R-Loop is significantly increased in cells carrying shRNaseH1 and shRNaseH2A (Online Supplementary Figure S2H, I). It is notable that these cells exhibit significant increased γH2AX foci (Online Supplementary Figure S2J, K).

In brief, the above data suggest that the accumulation of R-Loops co-localizes with DNA damage signal and results in the functional decline of HSC.

Aging-increased R-Loop correlates with the lingering RNA in the nucleus of hematopoietic stem cells

Given that the R-Loop is a microstructure of chromosomes composed of DNA:RNA strands, we then hypothesize whether the increase in R-Loops correlates with the retention of RNA in the nucleus of aged HSC. To explore this hypothesis, we categorized the genes into four groups based on the fold change of R-Loop signals between aged and young HSC: category-1 having the greatest loss of R-Loops and category-4 with the greatest gain of R-Loops. The results reveal that genes showing increased R-Loop formation display a higher ratio of nuclear to cytoplasmic RNA with aging (Figure 3A). This suggests that the accumulation of RNA in the nucleus of aged HSC might contribute to the formation of R-Loops.

To corroborate this observation, we focused on two genes: Nfe2l2 and Adgrg1, which play crucial role in regulating HSC.^29,33 Our observations indicate that the mRNA of Nfe2l2 and Adgrg1 accumulate in the nucleus of aged HSC (Figure 3B, C). Moreover, the R-Loop signal across the Nfe2l2 and Adgrg1 loci increases with aging (Figures 2C and 3D). RNA FISH and immunofluorescence staining reveal the co-localization of their RNA and the accumulated R-Loops within them (Figure 3E).

In summary, these findings suggest that R-Loops are heightened in aged HSC and are associated with the accumulation of RNA in nucleus during aging.

RNA exportation is impaired in aged hematopoietic stem cells

We then conducted an RNA FISH assay to assess the bulk poly(A)⁺RNA distribution between the nucleus and cytoplasm of both young and aged HSC, aiming to scrutinize RNA distribution patterns associated with aging. Our findings indicate a notable accumulation of mRNA within the nucleus of aged HSC compared to their younger counterparts (Figure 3F; Online Supplementary Figure S3A). To corroborate this observation, we analyzed the profiles of nuclear and cytoplasmic RNA from both young and aged HSC. The results unequivocally demonstrate a significant increase in mRNA accumulation within the nucleus of aged HSC (Online Supplementary Figure S3B).

Figure 2.Accumulated R-Loop induces DNA damage and impairs the function of hematopoietic stem cells. (A) The absolute number of insertions and deletions (InDels) in young (2 months) and aged (24-30 months) hematopoietic stem and progenitor cells (HSPC). Freshly isolated LSK cells from young and aged mice were subjected to whole-genome sequencing. (B) Representative fluorescence images depicting the co-localization of the R-Loop with γH2AX in aged hematopoietic stem cells (HSC). (C) Representative tracks showing aged HSC-specific R-Loop peaks (red, aged HSC; blue, young HSC) in the region comprising Nfe2l2. (D) single strand DRIP-quantitative polymerase chain reaction (ssDRIP-qPCR) monitoring of the R-Loop levels at specific regions. Data are presented as mean ± standard deviation. (E, F) Representative fluorescence images depicting the co-localization of the R-Loop with RPA (E) and γH2AX (F) in NIH-3T3 cells.

Figure 3.Aging-increased R-Loop correlates with the lingering RNA in the nucleus of hematopoietic stem cells. (A) Boxplot of the relative nucleus/cytoplasm ratio in four groups of genes with different R-Loop signal changes that occurred during aging. DESeq2 was performed to measure the difference in single strand (ss)DRIP reads in the gene body and its regulatory region in both young and aged hematopoietic stem cells (HSC). The genes were divided into 4 groups according to the quantiles of the DESeq2 test statistic. Genes belonging to level 1 have the highest R-Loop loss, while genes in level 4 have the highest R-Loop gain. The Y-axis indicates the nucleus/cytoplasm ratio of aged samples divided by the same ratio in young samples (one sided Wilcoxon signed rank test). (B) Histograms depicting the nuclear and cytoplasmic mRNA ratio of Nfe2l2 and Adgrg1 in both young and aged HSC. The data were obtained from nuclear and cytoplasmic RNA sequencing (RNA-seq). (C) Representative images depicting the Nfe2l2 and Adgrg1 mRNA in aged HSC. Freshly isolated HSC from young (2-month-old) and aged (24-27-month-old) mice were subjected to the assay. The panels are shown at the same exposure. The cell nuclei were stained with DAPI. (D) Representative tracks showing aged HSC-specific R-Loop peaks (red, aged HSC; blue, young HSC) in the region comprising Adgrg1. (E) Representative images depicting the co-localization of the R-Loop with Nfe2l2 and Adgrg1 mRNA in aged HSC. Freshly isolated HSC from young (2-month-old) and aged (24-27-month-old) mice were subjected to the co-localization assay. The panels are shown at the same exposure. The cell nuclei were stained with DAPI. (F) Representative images showing the distribution of Poly(A)⁺RNA in young and aged HSC. These images were obtained by subjecting freshly isolated HSC (CD34^- Flt3^- LSK) from young (2 months) and aged (24-30 months) mice to RNA fluorescence in situ hybridization (FISH) using a Cy3-labeled oligo(dT)₅₀ probe to visualize poly(A)⁺RNA. Nuclei were stained with DAPI. The panels are shown at the same exposure for accurate comparison.

To unravel the molecular mechanisms underlying this phenomenon, we initiated an integrative analysis of five previously published datasets (RNA-seq data of young and aged HSC).^34-38 Our examination revealed a significant enrichment of Gene Ontologies related to RNA localization with aging (Online Supplementary Figure S3C). Delving deeper, we found a notable impairment in the RNA transportation pathway within aged HSC (Online Supplementary Figure S3D), with Alyref emerging as the most significantly downregulated gene (Online Supplementary Figure S3E). Subsequently, we embarked on validating the alteration of Alyref expression between young and aged HSC, confirming a marked decrease in Alyref at both the mRNA and protein levels with aging (Online Supplementary Figure S3F, G).

Targeted dysfunction of Alyref induces R-Loop and disrupts hematopoietic homeostasis

To probe the functional implications of Alyref in HSC aging, we generated Alyref^flox/flox mice (Online Supplementary Figure S4A), and bred them with Mx1-Cre mice to produce Mx1-Cre; Alyref^flox/flox and Mx1-Cre; Alyref^flox/+ offspring. Upon poly I:C treatment for 14 days, Alyref is uniformly and heterozygously deleted in hematopoietic cells yielding Alyref^-/- and Alyref^+/- mice, respectively (Online Supplementary Figure S4B, C). We observed increased and co-localized signals among R-Loops, γH2AX and RPA in Alyref^-/- and Alyref^+/- HSC (Figure 4A, B; Online Supplementary Figure S4D, E). Additionally, replication fork assays exhibited increased replication stress in Alyref^+/-(Figure 4C), Alyref^-/- HSC (Figure 4D) and aged HSC (Figure 4E). Collectively, these data indicate that dysfunction of Alyref lead to an increase in R-Loops, DNA damage and replication stress in HSC.

We then sought out to examine the hematopoietic phenotype of Alyref^+/- and Alyref^-/- mice. Complete blood count assay revealed a notable decrease in white blood cell (WBC), neutrophil (NEUT), lymphocyte (LYM), basophilic granulocyte (BAS) and platelet (PLT) counts in Alyref^-/- mice, contrasting with Alyref^+/- counterparts (Figure 4F). However, both strains exhibited a drop in red blood cell (RBC), hemoglobin (HGB) and hematocrit (HCT) levels (Figure 4F; Online Supplementary Figure S4F). The BM cellularity of Alyref^-/- mice exhibited a significant decrease compared to their littermate controls (Online Supplementary Figure S4G). Subsequently, we delved into investigating the lineage composition in the peripheral blood (PB) and BM of both Alyref^+/- and Alyref^-/- mice, focusing on T cells, B cells, and myeloid cells. Our analysis unveiled a significant increase in T-lineage cells and a decrease in B and myeloid lineages in both PB and BM of Alyref^-/- mice, contrasting with Alyref^+/- mice (Online Supplementary Figure S4H, I). Moving forward, we scrutinized the hematopoietic stem and progenitor cells of these mice, encompassing common myeloid progenitor (CMP), granulocyte-macrophage progenitor (GMP), megakaryocyte-erythroid progenitor (MEP), common lymphoid progenitor (CLP), multipotent progenitor cell (MPP), shortterm hematopoietic stem cell (ST-HSC) and long-term hematopoietic stem cell (LT-HSC). Strikingly, our findings demonstrated a dramatic decrease in all populations within Alyref^-/- mice, while no such effect was observed in Alyref^+/- mice (Online Supplementary Figure S4J-L). These findings suggest that short-term heterozygous deletion of Alyref does not significantly impact hematopoietic homeostasis. To further explore the long-term effects, we retained Alyref^+/- mice for 10 months and observed an increase in RNA accumulation within the nucleus of HSC (Online Supplementary Figure S4M, N). Notably, complete blood count assay revealed evident decreases in WBC, LYM, HGB and HCT levels (Online Supplementary Figure S4O). Moreover, the BM cellularity of Alyref^+/- mice displayed a mild decrease compared to their littermate controls (Online Supplementary Figure S4P). Further analysis of lineage composition unveiled a significant decrease in B lineage cells and a mild increase in myeloid lineage cells in the peripheral blood, but not bone marrow, of Alyref^+/-mice (Online Supplementary Figure S4Q, R). Additionally, hematopoietic stem and progenitor cells (LSK), ST-HSC, GMP and CLP exhibited a significant decrease in Alyref^+/-mice (Online Supplementary Figure S4S, T).

In summary, these data collectively indicate that Alyref plays a crucial and dosage-dependent role in maintaining hematopoietic homeostasis.

Figure 4.Targeted dysfunction of Alyref induces R-Loop and disrupts hematopoietic homeostasis. (A, B) Representative fluorescence images depicting the co-localization of the R-Loop with RPA (A) and γH2AX (B) in Alyref^-/- hematopoietic stem cells (HSC). (C, left representative image) Representative images depicting the CldU/IdU-labeled DNA replication tracks in wild-type (WT) and Alyref^+/-HSC (12 months). (C, right dot plot) Dot plots depicting the DNA replication fork symmetry in WT and Alyref^+/- HSC. The dot plot of ldU to CIdU track length ratios for individual replication forks. Data are represented as mean ± standard deviation (SD). A minimum of 493 molecules were analyzed in each group. (D, left representative image) Representative images depicting the CldU/IdU-labeled DNA replication tracks in WT and Alyref^-/- HSC (2-4 months). (D, right dot plot) Dot plots depicting the DNA replication fork symmetry in WT and Alyref^-/- HSC. The dot plot of ldU to CIdU track length ratios for individual replication forks. Data are represented as mean ± SD. A minimum of 358 molecules were analyzed in each group. (E, left representative image) Representative images depicting the CldU/IdU-labeled DNA replication tracks in young (3 months) and aged HSC (24 months). (E, right dot plot) Dot plots depicting the DNA replication fork symmetry in young and aged HSC. The dot plot of ldU to CIdU track length ratios for individual replication forks. Data are represented as mean ± SD. A minimum of 423 molecules were analyzed in each group. (F) The dot plot displays the counts of white blood cells (WBC), neutrophils (NEUT), lymphocytes (LYM), red blood cells (RBC), basophilic granulocytes (BAS) and platelets (PLT) in WT, Alyref^+/- and Alyref^-/- mice. The data are based on at least 7 mice per group and are presented as mean ± SD.

Targeted dysfunction of Alyref impairs hematopoietic stem cells

Subsequently, we conducted a competitive transplantation assay to assess the reconstitution capacity of Alyref^+/- and Alyref^-/- HSC (Online Supplementary Figure S4U). The findings revealed compromised reconstitution capacity in both Alyref^+/- and Alyref^-/- HSC (Online Supplementary Figure S4V, W). Notably, the reconstitution capacity of Alyref^-/- HSC was significantly lower than that of Alyref^+/- HSC, suggesting a dosage-dependent effect of Alyref on HSC function. To further evaluate the self-renewal capacity of Alyref^+/- and Alyref^-/- HSC, we conducted a secondary transplantation assay using BM cells from primary recipients (Online Supplementary Figure S4U). The results demonstrated a significant impairment in the self-renewal capacity of both Alyref^+/- and Alyref^-/- HSC (Online Supplementary Figure S4V). Additionally, both groups exhibited a myeloid-biased potential (Online Supplementary Figure S4X), characteristic of aged HSC.

Further exploration of the molecular distinctions among Alyref^+/-, Alyref^-/- and wild-type (WT) HSC, involved RNA-seq analysis of these cells, revealing distinct groupings by principal component analysis (PCA) analysis (Online Supplementary Figure S4Y). GSEA unveiled the loss of HSC fingerprint genes and activation of cell aging-related genes in both Alyref^-/- and Alyref^+/- HSC (Online Supplementary Figure S4Z; Online Supplementary Table S7).

In summary, these data underscore that targeted dysfunction of Alyref recapitulates the characteristics of aged HSC, and demonstrates a dosage-dependent effect.

Dysfunction of RNA transportation impairs hematopoietic stem cells

The aforementioned results strongly suggest a potential role for Alyref-mediated RNA transportation in HSC aging. To further investigate this hypothesis, we sought to determine whether the dysfunction of other RNA transporters affects the reconstitution capacity of HSC. To this end, we developed two efficient shRNA targeting Thoc1 and Thoc5 (Figure 5A), essential components of the RNA transportation complexes TREX and AREX,³⁹ respectively. RNA FISH analysis revealed that knockdown of these genes resulted in the accumulation of poly(A)⁺RNA in the nucleus of hematopoietic stem and progenitor cells (Figure 5B). Subsequently, we conducted competitive transplantation assays to assess the functional impact of their knockdown on HSC. The results demonstrated that knockdown of Thoc1 and Thoc5 severely impaired the reconstitution capacity of HSC (Figure 5C, D).

Bulk mRNA nuclear export relies on the RNA export receptor Nxf1, which interacts with FG-repeat containing nucleoporins to facilitate RNA translocation.^40,41 We subsequently designed and tested two highly effective shRNA targeting Nxf1 (Online Supplementary Figure S5A). Upon performing the same assay, we observed that Nxf1 knockdown leads to the accumulation of poly(A)⁺RNA in the nucleus (Figure 5E) and significantly diminishes the reconstitution capacity of HSC (Figure 5F).

Taken together, these findings demonstrate that dysfunction of RNA transportation genes causes RNA buildup in the nucleus of HSC, thereby compromising their function.

Quantitative refill of Alyref rescues the function of aged hematopoietic stem cells by dampening R-Loop and replication stress

We then explored whether overexpression of Alyref could rescue the function of aged HSC. To test this hypothesis, we cloned the cDNA of mouse Alyref into a lentiviral vector,⁴² which resulted in efficient overexpression of ALYREF (Online Supplementary Figure S6A). Subsequently, we conducted competitive transplantation assays to assess the functional impact of enforced Alyref on HSC. Surprisingly, it was found that overexpression of Alyref impaired the reconstitution capacity of aged HSC (Figure 6A). A previous study has indicated that overexpression of Alyref leads to the accumulation of RNA in nucleus.⁴³ To investigate this further, we performed RNA FISH assays to examine the RNA distribution in Alyref-overexpressed HSC. The results confirmed that RNA is indeed accumulated in the nucleus upon Alyref overexpression (Figure 6B).

The earlier results have demonstrated that Alyref modulates the function of HSC in a dosage-dependent manner (Online Supplementary Figure S4V). Subsequently, we quantified the protein level of ALYREF in young and aged HSC, revealing a 50% decrease in ALYREF levels in aged HSC (Online Supplementary Figure S3G, Figure 6C; left histogram). However, lentiviral-mediated ALYREF overexpression resulted in a 45-fold increase (Figure 6C; right histogram), explaining the cytotoxic effect of Alyref overexpression on aged HSC. To precisely replenish ALYREF levels in aged HSC, we developed a quantitative gene expression system (QGES) to compensate for the drop of ALYREF. In this system, QGES-2^Alyref exhibits a 2-fold increase (Figure 6D). We conducted competitive transplantation assays to assess the functional role of quantitative refill of ALYREF on aged HSC using QGES-2^Alyref system. LSK cells isolated from 18 months old WT mice were infected with QGES-2^Alyref and empty control lentivirus. After 72 hours, 4000 CD48^- Sca1⁺ GFP⁺ cells were FACS-purified and transplanted into lethally irradiated recipients along with 3×10⁵ competitor cells, which were isolated from 2- to 3-month-old CD45.1 mice (Online Supplementary Figure S6B). Donor-derived chimerism was examined every 4 weeks until the 16^th week. The results showed a mild, though not significant (P=0.18), increase in the QGES-2^Alyref group (Figure 6E). Subsequently, we performed secondary transplantation from primary recipients, and the results revealed that aged HSC with quantitative expression of Alyref significantly outcompeted the control group (Figure 6E), indicating that precise compensation of ALYREF in aged HSC significantly rescues their self-renewal capacity.

Figure 5.Dysfunction of RNA transportation impairs hematopoietic stem cells. (A) Real-time polymerase chain reaction (RT-PCR) analysis of the knockdown efficiency of small hairpin (sh)RNA against Thoc1 and Thoc5 using freshly isolated LSK cells. (B, left representative image) Representative images depicting the distribution of poly(A)⁺RNA in shNT, shThoc1 and shThoc5 hemtopoietic stem and progenitor cells (HSPC). Freshly isolated LSK cells from young mice (2 months) were infected with lentivirus carrying Thoc1 shRNA, Thoc5 shRNA or control shRNA. After 3 days, GFP⁺ cells were purified for Poly(A)⁺RNA fluorescence in situ hybridization (FISH). The panels are shown at the same exposure. Poly(A)⁺RNA was detected using a Cy3-labeled oligo(dT)₅₀ probe. The cell nuclei were stained with DAPI. (B, right dot plot) The dot plots show the ratio of nuclear to cytoplasmic poly(A)⁺RNA signals ratio in shNT, shThoc1 and shThoc5 HSPC. Data are presented as mean ± standard deviation (SD). N and C indicate nuclear and cytoplasmic FISH signals, respectively. N/C ratios were determined for at least 33 cells in each experiment. (C) Experimental design of the competitive transplantation assay. Freshly isolated LSK cells from young mice (2 months old) were infected by lentivirus carrying Thoc1 shRNA, Thoc5 shRNA or control shRNA, labeled by GFP fluorescence. Three days later, 2000 GFP⁺S-ca1⁺CD48^- cells were purified and transplanted into lethally irradiated recipients (CD45.2) along with 3×10⁵ competitor cells (CD45.1). Chimerism in peripheral blood was evaluated every 4 weeks until the 12^th week. (D) These line plots depict the percentage of donor-derived cells (overall, B cells, T cells, myeloid cells) in the peripheral blood of recipients. N=4-7 recipients per group, data are presented as mean ± SD. (E, left representative image) Representative images depicting the distribution of Poly(A)⁺RNA in both shNT and shNxf1 HSPC. Freshly isolated LSK cells from young mice (2 months old) were infected with lentivirus carrying Nxf1 shRNA or control shRNA. After 3 days, GFP⁺ cells were purified for Poly(A)⁺RNA FISH. The panels are shown at the same exposure. Poly(A)⁺RNA was detected using a Cy3-labeled oligo(dT)₅₀ probe. The cell nuclei were stained with DAPI. (E, right dot plot) The dot plot shows the ratio of nuclear to cytoplasmic poly(A)⁺RNA signals ratio in shNT and shNxf1 HSPC. Data are presented as mean ± SD. N and C indicate nuclear and cytoplasmic FISH signals, respectively. N/C ratios were determined for at least 78 cells in each experiment. (F) These line plots depict the percentage of donor-derived cells (overall, B cells, T cells, myeloid cells) in the peripheral blood of recipients. N=6-7 recipients per group, data are shown as mean ± SD.

By conducting RNA FISH, we observed that the aberrant RNA accumulation in the nucleus of aged HSC is rescued by precise refill of ALYREF (Figure 6F). Moreover, we observed dramatic reduction in R-Loop (Online Supplementary Figure S6C, D), γH2AX (Online Supplementary Figure S6E, F) and RPA (Online Supplementary Figure S6G, H) foci in HSC carrying the QGES-2^Alyref- construct. The replication fork assay showed a decrease in replication stress in the QGES-2^Alyref group (Online Supplementary Figure S6I, J).

Collectively, these data indicate that precise refill of ALYREF in aged HSC rescues the self-renewal capacity of aged HSC by diminishing DNA damage and replication stress.

Discussion

Our study demonstrates that the imbalance of RNA distribution between the nucleus and cytoplasm, mediated by compromised RNA transportation, contributes to HSC aging by inducing R-Loops and subsequent replication stress (Online Supplementary Figure S6K). Furthermore, quantitative refill of ALYREF ameliorates the function of aged HSC. This study delves into the functional role of R-Loop in HSC aging and unveils a novel mechanism driving HSC aging.

RNA imbalance promotes hematopoietic stem cell aging

The homeostasis of RNA distribution between nucleus and cytoplasm is vital for cellular function, wherein RNA transportation playing a central role. In this study, we observed that the RNA distribution between the nucleus and cytoplasm of aged HSC became unbalanced, stemming from the decline of Alyref. Targeted dysfunction of Alyref recapitulated the phenotype of aged HSC, while precise refill of it rejuvenated the function of aged HSC by restoring RNA shuttling and dampening R-Loops. Since stable expression of Alyref is crucial for hematopoietic homeostasis and HSC function, investigating the molecular mechanisms compromising Alyref during aging is worthy of further study. A previous study has reported that Alyref specifically interacts with m⁵C RNA.⁴⁴ It is intriguing to explore the specificity of RNA transported by Alyref, whether it is mediated by m⁵C or other RNA modifications.

DNA damage versus aging and aging-related diseases

Aging is the highest risk factor for many diseases. It has been reported that aged blood induces cell and tissue senescence and vice versa, young blood revitalizes aged organism.^45-49 Aged HSC compromise the function of blood system. Therefore, rejuvenating aged HSC will benefit the function of aged organism.

On the basis of R-Loop findings of this study, we open a new aspect in the aging study. Firstly, we demonstrated that aging-accumulated R-loop is the resource of replication stress. Accumulated R-Loop results in more displaced ssDNA, which is vulnerable to damage.²⁰ In addition, accumulated R-Loop interfere with proper DNA repair machinery, especially the homologous recombination pathway. Secondly, R-Loop is a multi-faced player in modulating transcription by influencing promoter activity or chromatin structure, altering the local chromatin environment, promoting epigenetic modifications like DNA methylation or histone modifications.²⁰ Hence, in addition to R-Loop-induced replication stress, whether other aforementioned molecular events, such as those involving long non-coding RNA and DNA damage response RNA,⁵⁰ participate in regulating hematopoietic stem cell aging is worthy of further exploration.

Figure 6.Quantitative refill of Alyref rescues the function of aged hematopoietic stem cells by dampening R-Loop and replication stress. (A) Line plots depicting the proportion of donor-derived cells (overall, T cells, B cells, myeloid cells) in the peripheral blood of recipients. N=5-7 recipients per group. Data are represented as mean ± standard deviation (SD). (B, left representative image) Representative images depicting the distribution of Poly(A)⁺RNA in empty vector (EV) and Alyref overexpressed (OE) hematopoietic stem cells (HSC). The panels are shown at the same exposure. Poly(A)⁺RNA was detected using a Cy3-labeled oligo(dT)₅₀ probe. The cell nuclei were stained with DAPI. (B, right dot plot) Dot plots depicting the nuclear and cytoplasmic (N/C) poly(A)⁺R-NA signals ratio in both EV and Alyref OE HSC. Data are represented as mean ± SD. N and C indicate nuclear and cytoplasmic fluorescence in situ hybridization (FISH) signals, respectively. N/C ratios were determined for a minimum of 73 cells in each experiment. (C) Histograms illustrate the quantitative analysis of ALYREF protein levels in young/aged HSC and EV/Alyref OE HSPC. Data were analyzed using image J. (D, left representative western blot) Representative western blot showing the expression of ALYREF. NIH-3T3 cells were infected with QGES-^1Alyref, QGES-2^Alyref, KC^Alyref and control lentivirus (KC). After 3 days, GFP⁺ cells were purified for western blot. (D, right histogram) Histograms depicting the quantitative ALYREF protein in QGES-1^Alyref, QGES-2^Alyref, KC^Alyref and KC (Figure 6D, left representative western blot). Data were analyzed using image J. (E) Line plots depicting the proportion of donor-derived cells (overall, B cells, T cells, myeloid cells) in the peripheral blood of recipients. N=6-7 recipients per group. Data are represented as mean ± SD. (F, left representative image) Representative images depicting the distribution of Poly(A)⁺RNA in QGES-2^Alyref and KC HSPC using freshly isolated GFP⁺ LSK cells from the recipients of (Online Supplementary Figure S6B) at the end of the 16^th week after transplantation. The panels are shown at the same exposure. Poly(A)⁺RNA was detected using a Cy3-labeled oligo(dT)₅₀ probe. The cell nuclei were stained with DAPI. (F, right dot plot) Dot plots depicting the nuclear and cytoplasmic poly(A)⁺RNA signals ratio in KC and QGES-2^Alyref HSPC. Data are represented as mean ± SD. N and C indicate nuclear and cytoplasmic FISH signals, respectively. N/C ratios were determined for a minimum of 108 cells in each experiment.

Footnotes

• Received March 6, 2025
• Accepted June 26, 2025

Correspondence

*RC, QD and LZ contributed equally as first authors

T. Cheng chengtao@ihcams.ac.cn

J. Wang wangjianwei@ihcams.ac.cn

Disclosures

No conflicts of interest to disclose.

Contributions

Conceptualization by JW. Methodology by RC, QD and LZ. Investigation by RC, QD, LZ, HH, YD, QS, TS, TC and JW. Formal analysis by JW. Resources by JW. Writing by TC and JW. Funding acquisition by QD and JW. Supervision by TC and JW.

Funding

This work was supported by grant numbers 2024-I2M-ZD-010, W2441024, 82250002 and 92249305 (to JW) from CAMS Innovation Fund for Medical Sciences (CIFMS) and the National Natural Science Foundation of China; and by 62103397 (to QD) from National Natural Scientific Foundation of China.

Acknowledgments

We thank Dr. Hong Cheng (Institute of Biochemistry and Cell Biology, University of Chinese Academy of Sciences) for providing constructive suggestions for this project.