Loss of YTHDF2-mediated m⁶A-dependent mRNA clearance facilitates hematopoietic stem cell regeneration

Dear Editor,

Hematopoietic stem cells (HSCs) have unique self-renewal capacity to replenish the entire blood system, which is essential for various stress conditions, such as serial transplantation, ionizing radiation, anti-cancer medication, hemorrhage or infection.¹ N⁶-methyl-adenosine (m⁶A), the most prevalent reversible nucleotide modification in cellular mRNA in eukaryotes,² is associated with the maturation, translation, and eventual decay of protein-coding transcripts.^3,4 “Reader” proteins of m⁶A modification mediate many of the properties of m⁶A-methylated transcripts through methyl-specific RNA binding.^5,6 Although the function of m⁶A modification in cell fate decision of embryonic stem cells and hematopoietic stem and progenitor cells (HSPCs) has been recently reported,^7,8 the role of “Reader” proteins in HSC function has not been reported yet, especially under hematological stresses. It is of great importance to understand how HSCs recognize and respond to such modifications. Here, we investigated the “Reader” protein of m⁶A modification in HSC homeostasis and regeneration upon various stresses.

The differentiation and self-renewal of HSCs involve a large amount of mRNA regeneration and degradation. Within the cytoplasm, the m⁶A reader protein YTHDF2 facilitates m⁶A-dependent mRNA decay under normal and stress conditions.^5,9 We hence hypothesize that YTHDF2-dependent post-transcriptional m⁶A-mRNA silencing plays an important role in HSC homeostatic and hematological stress.

We utilized three Cre transgenic mouse lines (Mx1-Cre, Ert-Cre and Vav-Cre) to cross with Ythdf2^fl/fl mice to deplete the expression of YTHDF2 specifically in adult HSCs of mouse bone marrow (BM), respectively (Fig. 1a). The knockout efficiency of YTHDF2 in BM HSCs of Ythdf2^fl/fl Vav-Cre, Ythdf2^fl/fl Mx1-Cre and Ythdf2^fl/fl Ert-Cre was validated by multiple approaches (hereafter, the pIpC-treated Ythdf2^fl/fl Mx1-Cre mice are designated as Ythdf2^Δ/Δ and the pIpC-treated Ythdf2^+/+ Mx1-Cre mice as Ythdf2^+/+) (Supplementary information, Fig. S1a-d). Flow cytometry analysis indicated that the proportion and absolute number of long-term HSC (LT-HSC), short-term HSC (ST-HSC), multipotent progenitor (MPP) and CD150⁺CD48^–Lin^–Sca1⁺c-Kit⁺ (SLAM-LSK) in the Ythdf2^Δ/Δ mice were significantly higher than that in the Ythdf2^+/+ control mice (Fig. 1b; Supplementary information, Fig. S1e, f). The percentage and differentiation capacity of B cells, T cells, myeloid cells, granulocyte/monocyte progenitors (GMP), common myeloid progenitor (CMP), megakaryocyte/erythrocyte progenitor (MEP), common lymphoid progenitor (CLP) and peripheral blood (PB) cells were comparable between the Ythdf2^Δ/Δ mice and control mice (Supplementary information, Fig. S1g-j). These data suggest that YTHDF2 deletion uniquely affects the homeostasis of HSPCs.

Fig. 1

Deletion of Ythdf2 improved HSC regeneration via Wnt signaling and pro-survival genes during hematopoietic stress. a Experimental schematic for generation of mice with HSC-specific deletion of Ythdf2. Ythdf2^flox/flox mice were crossed with the interferon-inducible transgenic Mx1-Cre to generate Ythdf2^+/+ Mx1-Cre mice and Ythdf2^fl/fl Mx1-Cre mice. Animals were then treated with intraperitoneal (i.p.) injections of 300 μg of pIpC every other day for seven times to delete Ythdf2 alleles. Ythdf2^Δ/Δ denotes mice with HSC-specific deletion of Ythdf2, whereas Ythdf2^+/+ denotes mice without Ythdf2 deletion as control. b FACS analysis of LSK (Lin^–c-Kit⁺Sca-1⁺), LT-HSCs (CD34^–FlK2^–LSK), ST-HSCs (CD34⁺FlK2^–LSK), and MPPs (CD34⁺FlK2⁺LSK) in Ythdf2^+/+ and Ythdf2^Δ/Δ BM cells. The absolute cell numbers are shown (n = 5 mice per genotype). c Proliferation analysis of LT-HSC from Ythdf2^+/+ and Ythdf2^Δ/Δ mice. The percentage of BrdU-positive cells is shown. n = 5 mice per group. d Percentage of donor-derived PB cells at the indicated time points in serial competitive transplantation assay (1st, the first competitive transplantation, n = 6 mice per genotype; 2nd, the second competitive transplantation, n = 6 mice per group). e Percentage of donor-derived LT-HSCs 16 weeks after secondary transplantation (n = 6 mice per group). f Absolute number of LT-HSCs from mice treated with 150 mg/kg 5-FU on day 12 (n = 7 per genotype). g Absolute number of LT-HSCs from mice irradiated with 4.5 Gy on day 15 (n = 7 per genotype). h Representative FACS plots (left panel) and frequency (right panel) of apoptotic cells (ANNEXIN V⁺ DAPI^–) in the LSK from mice after IR (n = 7 mice per genotype). i GSEA analysis of the Wnt signaling transcriptional signature of HSCs in Ythdf2^Δ/Δ relative to Ythdf2^+/+ group. NES normalized enrichment score. j qPCR validation of expression of Wnt signaling target genes and survival-related genes in Ythdf2^Δ/Δ and Ythdf2^+/+ LT-HSCs (n = 7 mice per genotype). k m⁶A-RIP-qPCR assay showing the enrichment levels of m⁶A on c-Myc, Axin2, Mcl-1, Notch1 mRNAs in Ythdf2^Δ/Δ and Ythdf2^+/+ LSK cells. l RNA degradation assay shows that c-Myc and Axin2 mRNAs decay slower in Ythdf2^Δ/Δ LSK cells than in Ythdf2^+/+ LSK cells 2 h after actinomycin D treatment. The residual RNAs were normalized to t = 0. Three independent experiments were performed for all of the blots and qPCR assays. m Number of LT-HSCs in Ythdf2^+/+ and Ythdf2^Δ/Δ mice after Wnt inhibitor ICG001 treatment or DMSO in vivo (n = 5 mice per group). n Cell cycle analysis of LT-HSCs in Ythdf2^+/+ and Ythdf2^Δ/Δ mice after Wnt inhibitor ICG001 treatment or DMSO in vivo (n = 5 mice per group). Data represent the mean ± SD from three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant

To acquire further insight into the expansion of HSC pools after Ythdf2 deletion, we conducted cell cycle and cell proliferation analysis on LT-HSCs. Ki67 staining revealed a significantly higher frequency of cycling cells (Ki67⁺) in Ythdf2^Δ/Δ LT-HSCs than in Ythdf2^+/+ cells (~30% in Ythdf2^Δ/Δ vs. ~18% in Ythdf2^+/+; Supplementary information, Fig. S2a). Three-day bromodeoxyuridine (BrdU) labeling further identified a higher frequency of proliferating cells in Ythdf2^Δ/Δ LT-HSCs (BrdU⁺; ~27% in Ythdf2^Δ/Δ vs. ~14% in Ythdf2^+/+) (Fig. 1c). Similarly, Ythdf2^fl/fl Vav-Cre mice also exhibited a larger HSC pool than Ythdf2^+/+ Vav-Cre mice (Supplementary information, Fig. S2b). Together, these data suggest that YTHDF2 plays an important role in maintaining adult HSC quiescence. The loss of YTHDF2 drives proliferation of HSCs.

We hypothesize that although there are a higher number of Ythdf2-null HSCs, their functionality might be impaired upon transplantation or stresses. To this end, we conducted competitive transplantation experiments. Surprisingly, in comparison with Ythdf2^+/+ HSCs, Ythdf2^Δ/Δ HSCs exhibited a 50% increase in their long-term repopulating ability as revealed by PB and LT-HSC chimerism analysis and secondary transplantation (Fig. 1d, e; Supplementary information, Fig. S2c, d). The analysis of recipient mice further revealed no remarkable defects in the homing ability of Ythdf2^Δ/Δ HSC and multi-lineage reconstitution of Ythdf2^Δ/Δ donor-derived cells after the secondary transplantation (Supplementary information, Fig. S2e, f). Limiting-dilution transplantation using four dose concentrations of BM cells revealed a higher frequency of multi-lineage repopulating cells at 12 weeks in Ythdf2^Δ/Δ donor bone marrow (Supplementary information, Fig. S2g). To further strengthen the above results, we used the Ert-Cre line to induce YTHDF2 depletion in adult Ythdf2^fl/fl Ert-Cre mice and tested the function of Ythdf2-null HSCs with competitive repopulation assays. Consistent with the findings in Ythdf2^fl/fl Mx1-Cre HSCs, Ythdf2^fl/fl Ert-Cre HSCs exhibited expansion and a robust long-term repopulation activity (Supplementary information, Fig. S2h-j).

To examine the regenerative capacity of YTHDF2-deficient HSCs after hematopoietic injury, Ythdf2^+/+ and Ythdf2^Δ/Δ mice were treated with 5-FU to ablate the proliferating hematopoietic cells. A dramatic increase in the HSC number and function was observed in Ythdf2^Δ/Δ mice compared with the Ythdf2^+/+ controls (Fig. 1f; Supplementary information, Fig. S3a-c). We also observed a significant delay in 5-FU-induced hematopoietic failure in Ythdf2^Δ/Δ mice (Supplementary information, Fig. S3d). Further, to test the regenerative capacity of YTHDF2-deficient HSCs in vivo after ionizing radiation, we examined the recovery of HSCs in mice following sequential 4.5 Gy exposure on day 0 and day 15. The results showed that Ythdf2^Δ/Δ mice displayed a significant increase in HSCs compared with the Ythdf2^+/+ controls (Fig. 1g and Supplementary information, Fig. S3e). In addition, Annexin V analysis showed reduction of apoptosis rate in Ythdf2^Δ/Δ HSCs compared with the Ythdf2^+/+ controls (Fig. 1h), suggesting that the increased HSC number was related to cell viability. These data indicate that YTHDF2 deficiency resulted in elevated regeneration capacity and protection in HSCs under transplantation and hematopoietic stresses.

To further explore the underlying molecular mechanisms of YTHDF2-mediated HSC homeostasis and regeneration, we performed RNA-seq analysis on freshly isolated Ythdf2^+/+ and Ythdf2^Δ/Δ HSCs. The gene-set enrichment analysis (GSEA) revealed a marked upregulation of genes in Wnt signaling pathway (Fig. 1i and Supplementary information, Fig. S4a). Expression of a list of Wnt signaling downstream targets (including Myc, Ccnd1, Axin2) and pro-survival genes (including Mcl-1, Bcl2) was significantly increased in the Ythdf2^Δ/Δ HSCs as detected by qPCR (Fig. 1j). In addition, m⁶A-RIP-qPCR assay showed that the enrichment of m⁶A on Myc and Axin2 mRNA was notably increased in Ythdf2^Δ/Δ LSK cells (Fig. 1k). Furthermore, YTHDF2-RIP-qPCR assay showed the enrichment of Myc and Axin2 mRNAs in LSK cells (Supplementary information, Fig. S4b). Next, we performed RNA decay assays and found that Myc and Axin2 mRNA levels were both increased in Ythdf2^Δ/Δ HSCs in comparison to Ythdf2^+/+ HSCs 2 h after actinomycin-D treatment (Fig. 1l).

To determine whether the activation of Wnt signaling in Ythdf2^Δ/Δ LT-HSC contributes to HSC proliferation and subsequent functional enhancement, we administered the small-molecule Wnt/β-catenin inhibitor ICG-001 to Ythdf2^+/+ and Ythdf2^Δ/Δ mice and examined the HSC numbers and cell cycle status. The small molecule ICG-001 targets cAMP-responsive element binding (CREB)-binding protein (CBP) to inhibit β-catenin/LEF1-mediated transcription. ICG-001 treatment partially abolished the proliferation and phenotypic expansion of HSCs in Ythdf2^Δ/Δ mice under normal condition and 5-FU stress (Fig. 1m, n; Supplementary information, Fig. S4c, d). Taken together, these data suggest that both the phenotypic expansion and proliferation of YTHDF2-deficient HSCs are partially due to the activation of Wnt signaling.

In summary, we utilized three Cre/LoxP systems to generate the hematopoietic-specific knockout mice of Ythdf2, and identified a unique function of YTHDF2-mediated m⁶A recognition in mouse adult HSC. Mechanistically, the expression of Ythdf2 in HSCs facilitates decay of the m⁶A-modified mRNAs of Wnt target genes, contributing to the repression of Wnt signaling at steady state (Supplementary information, Fig. S4e). Usually, this repression is protective, since the abnormal activation of Wnt signaling results in expansion of HSC number but exhaustion of its functionality upon hematopoietic stresses.¹⁰ Interestingly, upon deletion of YTHDF2, not only the number of HSCs was expanded but also the regenerative capacity of HSCs was increased under stress conditions. Our data suggested that YTHDF2 deficiency simultaneously prevented the degradation of mRNAs of both Wnt target genes and survival-related genes during hematopoietic stresses, and this combined effect synergistically enhanced the regenerative capacity of HSC.

A therapeutic challenge of hematopoietic regeneration is to stimulate stem cell proliferation for immediate tissue repair while avoiding stem cell depletion under stress.¹ Our current findings revealed the functional importance of YTHDF2-mediated reading and processing of RNA m⁶A modification in HSCs, by integrating the proliferation-related genes and the pro-survival genes. YTHDF2 deficiency boosted HSC expansion without losing its viability under stress conditions. These findings established the crucial role of YTHDF2 in regulating HSC protection, self-renewal and regeneration, and thereby provided preliminary results supporting the possibility that these effects could also be translated into effective clinical applications.

Materials and Methods are available in Supplementary Information.