Mutant ASXL1 induces age-related expansion of phenotypic hematopoietic stem cells through activation of Akt/mTOR pathway

Somatic mutations of ASXL1 are frequently detected in age-related clonal hematopoiesis (CH). However, how ASXL1 mutations drive CH remains elusive. Using knockin (KI) mice expressing a C-terminally truncated form of ASXL1-mutant (ASXL1-MT), we examined the influence of ASXL1-MT on physiological aging in hematopoietic stem cells (HSCs). HSCs expressing ASXL1-MT display competitive disadvantage after transplantation. Nevertheless, in genetic mosaic mouse model, they acquire clonal advantage during aging, recapitulating CH in humans. Mechanistically, ASXL1-MT cooperates with BAP1 to deubiquitinate and activate AKT. Overactive Akt/mTOR signaling induced by ASXL1-MT results in aberrant proliferation and dysfunction of HSCs associated with age-related accumulation of DNA damage. Treatment with an mTOR inhibitor rapamycin ameliorates aberrant expansion of the HSC compartment as well as dysregulated hematopoiesis in aged ASXL1-MT KI mice. Our findings suggest that ASXL1-MT provokes dysfunction of HSCs, whereas it confers clonal advantage on HSCs over time, leading to the development of CH.


B
illions of blood cells are produced daily to maintain homeostasis in a human body. Hematopoietic stem cells (HSCs) are a rare population, but they are capable of selfrenewal and multi-lineage differentiation, which play a pivotal role in maintaining life-long hematopoiesis. In the aging process, HSCs are subject to both cell-intrinsic and cell-extrinsic stresses, which may result in cellular senescence and increased incidence of malignancy [1][2][3] . It has been reported that aging is associated with a significant expansion in surface marker-defined immunophenotypic HSCs (pHSCs), while these cells reduce repopulation ability [4][5][6][7] . Progenitors with limited differentiation potential and impaired repopulation ability accumulate within the pHSC compartment, which may serve as the mechanistic basis for this phenomenon [5][6][7][8][9] . Such pHSCs may also contribute to the development of clonal hematopoiesis (CH) or clonal hematopoietic disease, but this concept has yet to be validated 1,9 . ASXL1, a gene frequently mutated in CH, is one of three human homologs of the Drosophila Asx gene and is involved in epigenetic regulation [10][11][12][13] . In addition to CH, somatic mutations of ASXL1 gene are detected in myeloid neoplasms including myelodysplastic syndromes (MDS), chronic myelomonocytic leukemia (CMML), and acute myeloid leukemia (AML) [14][15][16][17][18] . Most ASXL1 mutations detected in CH and hematological malignancies are frameshift or nonsense mutations in the last exon, generating a C-terminally truncated form of ASXL1 14,15,[17][18][19][20][21] . Such truncated ASXL1 proteins are indeed expressed in leukemic cells and likely confer change-of-function 22 . Recently, we and others have shown that the mutant ASXL1 interacts with a deubiquitinase BAP1 to form the Polycomb-repressive deubiquitinase (PR-DUB) complex, which efficiently deubiquitinates H2AK119Ub 23,24 . The hyperactive PR-DUB complex of mutant ASXL1 and BAP1 upregulates several target genes, including HOX genes, by deubiquitinating H2AK119Ub to promote myeloid leukemogenesis 23 . In hematopoietic lineage-specific conditional knockin (KI) mice expressing a C-terminally truncated form of mutant ASXL1 (ASXL1-MT; 1900−1922del;E635RfsX15) 25 , global reduction was observed in H3K4me3 and H2AK119Ub. In particular, levels of H3K4me3 at the loci of erythroid differentiation-related genes, such as Id3 and Sox6, were markedly decreased, suggesting that ASXL1-MT impairs hematopoiesis through dysregulated epigenetic modifications.
It is now generally recognized that cells that are well-adapted to selective pressures acquire a fitness advantage, leading to clonal expansion during aging [26][27][28][29] . Clonal expansion of such blood cells harboring somatic mutations is termed CH. CH is commonly observed in elderly individuals in the absence of apparent hematological abnormalities [19][20][21] . Despite normal hematological parameters, individuals with CH are at increased risk of hematological malignancies, indicating that CH can be a preleukemic condition. Mutations most commonly detected in CH include the epigenetic regulators DNMT3A, TET2, and ASXL1. From the results of transplantation experiments using Tet2-deficient or Dnmt3a-deficient mice, the increased self-renewal of HSCs could explain why loss-of-function mutations of DNMT3A and TET2 are frequently detected in CH [30][31][32][33] . On the other hand, several recent studies have shown reduced numbers and functions of HSCs in mutant ASXL1 KI mice 25,34,35 . Thus, identifying how ASXL1 mutations promote the development of CH has remained an unsolved issue.
In the present study, we examined the influence of ASXL1-MT on physiological aging using ASXL1-MT KI mice. We found that ASXL1-MT confers a competitive disadvantage on HSCs after transplantation. On the other hand, in genetic mosaic mouse model, we observed stronger growth advantage in LT-HSCs expressing ASXL1-MT specifically in native hematopoiesis, recapitulating CH in humans. As for the molecular mechanisms by which ASXL1-MT induces CH, we show that ASXL1-MT binds and activates Akt by stabilizing phosphorylated Akt in concert with Bap1. Activation of Akt/mTOR pathway induced by ASXL1-MT causes aberrant cell cycle progression and proliferation in the HSC compartment. At the same time, it also provokes dysfunction of HSCs associated with mitochondrial activation, elevated ROS levels, and increased DNA damage. These molecular changes can cause CH with increased risk of leukemogenesis, which can be ameliorated by rapamycin treatment.

Results
ASXL1-MT reduces the number and function of HSPCs in young mice. To study the effects of ASXL1-MT on hematopoiesis, we crossbred ASXL1-MT KI mice with Vav-Cre transgenic mice. Consistent with our previous reports 25 , young Vav-Cre ASXL1-MT KI mice did not show significant changes in hematological parameters (hereinafter, all experiments were performed with 6-12-week-old for young mice) (Fig. 1a). We also observed no significant changes in the frequency of myeloid cells (CD11b + ), B cells (B220 + ), and T cells (CD3 + ) between control and Vav-Cre ASXL1-MT KI mice (Fig. 1b). Young Vav-Cre ASXL1-MT KI mice showed normal bone marrow cellularity (Fig. 1c), but exhibited a significant decrease in the frequency of hematopoietic stem and progenitor cells (HSPCs), including Lin − Sca1 + c-kit + (LSK) cells, multipotent progenitors (MPPs; CD48 + CD150 − LSK) and long-term HSCs (LT-HSCs; CD48 − CD150 + LSK) ( Fig. 1d and Supplementary Fig. 1). Competitive transplantation assays using whole bone marrow cells or defined populations (MPPs and LT-HSCs) revealed impaired repopulation ability and skewed differentiation potential toward myeloid lineage in ASXL1-MT KI HSPCs (Fig. 1e, f, Supplementary Fig. 2a-d). Thus, consistent with the previous reports, ASXL1-MT reduced the number and function of HSPCs in young mice.
To characterize the defective HSPC function of young Vav-Cre ASXL1-MT KI mice, we evaluated apoptosis and cell cycle status. The frequency of Annexin V-positive cells was significantly increased in HSPCs from Vav-Cre ASXL1-MT KI mice (Fig. 1g), indicating that ASXL1-MT KI HSPCs are more prone to apoptosis than are normal HSPCs. Ki-67/DAPI staining revealed a significantly lower frequency of LT-HSCs in G0 phase, and a higher frequency of both MPPs and LT-HSCs in S/G2/M phase in Vav-Cre ASXL1-MT KI mice (Fig. 1h). These data suggest that increased apoptosis and loss of quiescence contribute to the defective repopulation ability in HSPCs of young Vav-Cre ASXL1-MT KI mice.
ASXL1-MT confers a clonal advantage on LT-HSCs specifically in native hematopoiesis. Next, we assessed age-associated changes in hematopoiesis driven by ASXL1-MT. Analyses of peripheral blood cells showed mild leukocytopenia and mild anemia, which are prominent when compared with age-matched control mice, as well as moderate thrombocytosis in aged Vav-Cre ASXL1-MT KI mice (hereinafter, all experiments were performed with 20-24-month-old for aged mice) (Fig. 2a). Aged Vav-Cre ASXL1-MT KI mice exhibited a myeloid-biased hematopoiesis and hypocellular bone marrow, suggesting dysfunction of hematopoiesis (Figs. 2b, 9d, and e). Aged Vav-Cre ASXL1-MT KI mice do not develop apparent hematological diseases and their lifetime is equal to wild-type mice. Interestingly, the frequencies of LT-HSCs were markedly increased in aged Vav-Cre ASXL1-MT KI mice (Fig. 2c). We then assessed the functions of LT-HSCs in aged Vav-Cre ASXL1-MT KI mice. Single-cell liquid cultures revealed a decrease in the frequency of colony-forming LT-HSC from aged Vav-Cre ASXL1-MT KI mice compared with age-matched control mice (Supplementary Fig. 3a and b).
Furthermore, competitive transplantation assays using LT-HSCs from aged Vav-Cre ASXL1-MT KI mice and age-matched control mice showed that aged LT-HSCs expressing ASXL1-MT exhibited markedly compromised production of peripheral blood and tended to exhibit impaired regeneration of bone marrow cells including LT-HSCs (Fig. 2d-f). We also performed homing assays using LSK cells from both young and aged Vav-Cre ASXL1-MT KI mice. These analyses revealed that LSK cells from aged Vav-Cre ASXL1-MT KI mice exhibited a significant decrease in homing capability compared to age-matched control mice in addition to intrinsic defects detected by in vitro colony forming assays ( Supplementary Fig. 3a-d). Thus, the increased LT-HSCs in aged Vav-Cre ASXL1-MT KI mice are not functional HSCs with long-term repopulation potential (hereinafter, referred to as "immunophenotypic LT-HSCs (pLT-HSCs)").
It appears that ASXL1-MT promotes age-related changes in the hematopoietic system of Vav-Cre ASXL1-MT KI mice (e.g. anemia, myeloid-skewed differentiation, and hypocellular bone marrow). Therefore, we next evaluated the frequency of CD41positive pLT-HSCs, which has been reported to skew differentiation toward myeloid and megakaryocytic lineages, and to expand during aging within the pLT-HSC compartment 8,9,36,37 . This analysis revealed that most pLT-HSCs were positive for CD41 in aged Vav-Cre ASXL1-MT KI mice, while 50-70% were positive in age-matched control mice ( Supplementary Fig. 3e). These data imply that ASXL1-MT accelerates the physiological aging of LT-HSCs.
Next, to assess the competitive growth of ASXL1-MT KI cells in native hematopoiesis without transplantation, we bred ASXL1-MT KI mice with Mx1-Cre transgenic mice and partially induced expression of ASXL1-MT by polyinosine-polycytidine (pIpC) injections (Fig. 3a). As these mice carry the floxed allele of ASXL1-MT-IRES-GFP, cells expressing ASXL1-MT also express GFP ( Supplementary Fig. 1). In these genetic mosaic mice, frequency of ASXL1-MT expressing cells (GFP positive) gradually increased in peripheral blood (Fig. 3b). Similar to the phenotypes of aged Vav-Cre ASXL1-MT KI mice, expression of ASXL1-MT in this model also caused anemia, thrombocytosis, and myeloid-biased differentiation with age ( Fig. 3c and Supplementary Fig. 4a-d). The frequency of LT-HSCs expressing ASXL1-MT was increased in aged Mx1-Cre ASXL1-MT KI mice along with progression of cell cycle ( Fig. 3d and e). Eventually, LT-HSCs expressing ASXL1-MT occupied bone marrow to outcompete their normal counterpart 2 years after pIpC injections, recapitulating CH in humans (Fig. 3f). Given that levels of pro-inflammatory cytokines including TNF-α and IFN-γ were similar in aged control and aged Mx1-Cre ASXL1-MT KI mice (Supplementary Fig. 4e and f), these phenotypic changes were caused by ASXL1-MT itself, not by the inflammation-induced Mx1-Cre expression.
Thus, ASXL1-MT causes a competitive disadvantage of LT-HSCs during transplantation; nevertheless, it confers a clonal advantage on LT-HSCs in native hematopoiesis. To investigate cell intrinsic and extrinsic mechanisms that mediate the effect of ASXL1-MT on HSC function, we performed transplantation assays with three different settings ( Supplementary Fig. 5a, d, and f). First, we transplanted bone marrow cells from young Mx1-Cre ASXL1-MT KI mice into wild-type recipient mice and induced ASXL1-MT expression by pIpC injections one month after transplantation. Induction of ASXL1-MT after engraftment did not lower chimerism of peripheral blood but rather tended to increase the frequency of donor-derived bone marrow cells, suggesting that ASXL1-MT exerts competitive disadvantage on LT-HSCs, specifically during recovery from transplantation ( Supplementary Fig. 5a-c). Second, we transplanted bone marrow cells from young Vav-Cre ASXL1-MT KI mice into young or aged wild-type recipient mice. The frequencies of ASXL1-MTexpressing cells were similar regardless of the age of recipient mice ( Supplementary Fig. 5d and e). Finally, we transplanted bone marrow cells from young Vav-Cre ASXL1-MT KI mice into Mx1-Cre ASXL1-MT KI mice that had been induced expression of ASXL1-MT by pIpC injections one month before transplantation. Expression of ASXL1-MT in non-hematopoietic cells of the recipient bone marrow did not affect the chimerism of ASXL1-MT-expressing cells in peripheral blood (Supplementary Fig. 5f and g). Taken together, we concluded that expression of ASXL1-MT in LT-HSCs promotes the age-related clonal expansion independent of either aging of microenvironment or ASXL-MTexpressing stromal cells in native hematopoiesis.
ASXL1-MT induces aberrant expansion of the LT-HSC compartment through activation of Akt/mTOR pathway. To gain insight into the mechanism underlying the growth advantage of pLT-HSCs expressing ASXL1-MT, we performed RNA-seq analysis using HSPCs from young Vav-Cre ASXL1-MT KI mice and littermate control mice. Geneset enrichment analysis (GSEA) suggested that Akt/mTOR pathway was activated in ASXL1-MT KI HSPCs (Fig. 4a). Consistent with the results of GSEA, intracellular flow cytometry analyses revealed that phosphorylation of Akt and S6, which indicates activation of Akt and mTORC1, respectively, were elevated in young ASXL1-MT KI HSPCs (Fig. 4b, c, Supplementary Fig. 6a, b). Next, we evaluated the age-related change of Akt/mTOR activity in Vav-Cre ASXL1-MT KI mice. LT-HSCs of both young and aged mice exhibited activation of Akt/mTOR signaling to the same degree in Vav-Cre ASXL1-MT KI mice compared with age-matched control mice ( Fig. 4d and e).
As Akt/mTOR signaling positively controls cell proliferation, we inferred that aberrant proliferation of ASXL1-MT-expressing pLT-HSCs was caused by enhanced Akt/mTOR signaling. Therefore, we treated aged Vav-Cre ASXL1-MT KI mice and age-matched control mice with an mTOR inhibitor rapamycin for 8 weeks (Fig. 5a). Intriguingly, treatment with rapamycin reduced the frequency of pLT-HSCs in aged Vav-Cre ASXL1-MT KI mice ( Fig. 5b and c). Similar to young mice, pLT-HSCs in G0 phase were significantly decreased in aged Vav-Cre ASXL1-MT KI mice, which was normalized by rapamycin treatment (Fig. 5d). In contrast, ASXL1-MT had no effect on the frequency of apoptotic LT-HSCs in aged mice (Fig. 5e). Treatment with an Akt inhibitor perifosine also normalized cell cycle status in aged Vav-Cre ASXL1-MT KI mice ( Fig. 5f and g). These data suggest that abnormal cell cycle progression contributes to gradual expansion of the LT-HSC compartment during aging in Vav-Cre ASXL1-MT KI mice. Collectively, the activated Akt/mTOR pathway plays a pivotal role to confer growth advantage on LT-HSCs expressing ASXL1-MT.
ASXL1 interacts with phosphorylated AKT 43,44 . Therefore, we first examined whether ASXL1 and its partner protein BAP1 interact with AKT. We found that wild-type/mutant ASXL1 and BAP1 bound to AKT in both cytoplasm and nucleus in 293T cells, suggesting that they form a complex (Fig. 6a, b, and Supplementary Fig. 7a). Because mutant ASXL1 is known to stabilize BAP1 and enhance its DUB activity 23,24 , we next examined whether ubiquitination of AKT is removed by the ASXL1-MT/BAP1 complex. Consistent with our previous reports 23 , ASXL1-MT and BAP1 were mutually stabilized ( Fig. 6c). Notably, expression of BAP1 together with ASXL1-MT, but not with wild-type ASXL1, effectively deubiquitinated AKT mainly through K48-linked ubiquitin (Fig. 6c, d, and Supplementary Fig. 7b). Thus, these data suggest that AKT is a nonhistone targets of the ASXL1-MT/BAP1 DUB complex.
To determine the role of endogenous Bap1 on Akt signaling, we then assessed the effect of Bap1 deletion in murine bone marrow cells transformed by combined expression of SETBP1-D868N and ASXL1-MT (cSAM cells: cells with combined expressing of mutant SETBP1 and ASXL1-MT) in which Akt is activated 45 . As described in our previous report 23 , Bap1 depletion destabilized and reduced ASXL1-MT protein levels ( Fig. 6e). In addition, Bap1 deletion in cSAM cells decreased expression of phosphorylated Akt (Fig. 6e). A time course experiment with IL-3 stimulation revealed attenuated and shortened phosphorylation of Akt in Bap1-depleted cSAM cells ( Fig. 6f and g). These data suggest that ASXL1-MT cooperates with Bap1 to deubiquitinate and stabilize phosphorylated Akt, leading to activation of Akt/mTOR pathway.  Table 1). As recent studies have shown that mitochondrial dynamics are closely linked to stem cell function 46-48 , we therefore examined the possible influence of ASXL1-MT on mitochondrial activity. MitoTracker staining showed increased mitochondrial membrane potential in HSPCs of Vav-Cre ASXL1-MT KI mice (Fig. 7b). In addition, extracellular flux analyses revealed increased oxygen consumption rates (OCR) in c-kit + cells of Vav-Cre ASXL1-MT KI mice (Fig. 7c). We then conducted ion chromatography-mass spectrometry (IC-MS)-based metabolomics to analyze intracellular metabolites using HSPCs. As shown in Supplementary Fig. 8b and c, the pool of TCA cycle-intermediates as well as levels of ATP were increased in HSPCs of Vav-Cre ASXL1-MT KI mice. These data suggest enhanced mitochondrial respiration in HSPCs of Vav-Cre ASXL1-MT KI mice. Interestingly, immunofluorescence staining of the mitochondrial protein Tom20 revealed that mitochondria in HSPCs of Vav-Cre ASXL1-MT KI mice tended to form a few large aggregates, while those in HSPCs of control mice showed a more punctuated and dispersed morphology ( Fig. 7d and e). These morphological changes suggest altered mitochondrial fission and fusion in HSPCs of Vav-Cre ASXL1-MT KI mice 47,49,50 . Collectively, these results indicate that ASXL1-MT enhances mitochondrial activity and alters mitochondrial dynamics in HSPCs.
ROS-mediated DNA damage causes dysfunction of HSPCs in ASXL1-MT KI mice. Mitochondria are the major source of ROS, which can cause DNA damage 51,52 . We therefore assessed the levels of ROS in Vav-Cre ASXL1-MT KI mice. As expected, ASXL1-MT KI HSPCs showed increased ROS levels (Fig. 7f). Moreover, alkaline comet assay using pLT-HSCs revealed increased DNA strand breaks in young Vav-Cre ASXL1-MT KI mice compared to young control mice (Fig. 7g). We also observed increased nuclear foci of γ-H2AX in ASXL1-MT KI HSPCs after transplantation ( Supplementary Fig. 8d). Notably, aged Vav-Cre ASXL1-MT KI mice harbor more DNA damage than do young Vav-Cre ASXL1-MT KI mice and age-matched control mice ( Fig. 7g), implicating stronger age-related accumulation of DNA damage in pLT-HSCs of Vav-Cre ASXL1-MT KI mice.
Next, we investigated whether treatment with the antioxidant scavenger N-acetylcysteine (NAC) mitigates DNA damage and restores the impaired repopulation potential of HSPCs in Vav-Cre ASXL1-MT KI mice. We transplanted bone marrow cells from Vav-Cre ASXL1-MT KI mice or littermate control mice, treated with NAC for 8 weeks, into recipient mice (Fig. 8a). NAC treatment effectively normalized ROS and γ-H2AX levels of ASXL1-MT KI HSPCs ( Fig. 8b and c), and partially restored their repopulation ability (Fig. 8d). Combined overexpression of catalase and a ROS-detoxifying enzyme, manganese superoxide dismutase (Cat/SOD2), also rescued the defective repopulation potential of HSPCs of Vav-Cre ASXL1-MT mice ( Fig. 8f and g). Thus, reduction of ROS efficiently reverts the dysfunction of HSPCs in Vav-Cre ASXL1-MT KI mice. p53 serves to maintain genomic stability through DNA damage response and regulation of ROS levels, but it can also cause dysfunction of HSCs [53][54][55] . To assess the influence of p53 on ASXL1-MT KI HSPCs, we crossed Vav-Cre ASXL1-MT KI mice with p53 −/− mice. ASXL1-MT KI/p53 −/− mice showed significantly higher frequency of LT-HSCs than ASXL1-MT KI mice (Supplementary Fig. 9a and b). Competitive transplantation assays revealed that loss of p53 restored the repopulation potential of ASXL1-MT KI HSPCs ( Supplementary Fig. 9c). Meanwhile, ASXL1-MT KI/p53 −/− mice are prone to early onset of thymic lymphomas compared with p53 −/− mice (Supplementary Fig. 9d-f). Transduction of dominant-negative form of p53 (p53DD) increased the repopulation potential of ASXL1-MT KI HSPCs associated with increased ROS levels and DNA damage (Supplementary Fig. 9g-j). These findings indicate that p53 plays a critical role in suppressing ROS levels and preventing tumor formation at the expense of HSPC functions in Vav-Cre ASXL1-MT KI mice, presumably by eradicating cells with critical DNA damage.
perifosine as well as rapamycin effectively decreased the mitochondrial membrane potential and intracellular ROS levels in HSPCs of young Vav-Cre ASXL1-MT KI mice ( Supplementary  Fig. 10a-f). We next assessed whether the enhanced Akt/mTOR signaling reduced engraftment of ASXL1-MT KI HSPCs. Bone marrow cells from young Vav-Cre ASXL1-MT KI mice or littermate control mice treated with rapamycin were transplanted into recipient mice ( Supplementary Fig. 10d). Rapamycin treatment effectively improved engraftment of ASXL1-MT-expressing bone marrow cells (Supplementary Fig. 10g). These data indicate that enhanced Akt/mTOR-signaling induced by ASXL1-MT provokes mitochondrial activation, ROS overproduction, and dysfunction of HSPCs. It should be noted, however, that neither antioxidant nor rapamycin inhibits the myeloid-skewed hematopoiesis in young Vav-Cre ASXL1-MT KI mice ( Fig. 8e and Supplementary Fig. 10h). Therefore, it is likely that ASXL1-MT promotes myeloid skewing through Akt/mTOR-independent functions.
To examine the relationship between the enhanced Akt/mTOR signaling and dysregulated hematopoiesis in aged ASXL1-MT KI mice, we treated aged Vav-Cre ASXL1-MT KI mice and agematched control mice with rapamycin for 8 weeks (Fig. 9a). As was the case with young mice, mitochondrial membrane potential and intracellular ROS levels were increased in LT-HSCs of aged Vav-Cre ASXL1-MT KI mice compared with age-matched control mice (Fig. 9b and c). ASXL1-MT-induced elevation of mitochondrial membrane potential and ROS levels was further validated in LT-HSCs of aged Mx1-Cre ASXL1-MT KI mice ( Supplementary Fig. 11a-c). Rapamycin treatment tended to reduce the mitochondrial membrane potential and normalized ROS levels in aged LT-HSCs expressing ASXL1-MT ( Fig. 9b and c). Treatment with perifosine also abrogated mitochondrial activation in these cells ( Supplementary Fig. 11d and e). Importantly, treatment with rapamycin recovered the number of differentiated white blood cells as well as bone marrow cellularity in aged Vav-Cre ASXL1-MT KI mice ( Fig. 7d and e). Furthermore, mitochondrial activation, increased DNA damage, and differentiation defects in aged ASXL1-MT KI HSPCs were ameliorated by rapamycin treatment (Fig. 9f-j). Taken together, these data suggest that overactive Akt/mTOR pathway causes dysregulated hematopoiesis, whereas it confers a growth advantage on pLT-HSCs in aged ASXL1-MT KI mice.
ASXL1-MT enforces age-associated patterns of gene expression in LT-HSCs. To further characterize the impact of ASXL1-MT on physiological aging of LT-HSCs, we performed RNA-seq analysis using LT-HSCs from aged Vav-Cre ASXL1-MT KI mice and agematched control mice ( Fig. 10a and b). As was the case with young mice, GSEA suggests activation of Akt/mTOR pathway in aged Vav-Cre ASXL1-MT KI mice, in concordance with flow cytometry analyses (Figs. 4d, e, and 10c). In aged ASXL1-MTexpressing LT-HSCs, up-regulated genes significantly overlapped with genesets related to Akt/mTOR signaling, platelet signature, and senescence; conversely, down-regulated genes significantly overlapped with genesets related to myeloid differentiation, erythroid differentiation, and HSC signature (Fig. 10d). Since ASXL1-MT appears to accelerate aging of LT-HSCs (e.g. anemia, myeloid-skewed differentiation, and hypocellular bone marrow), we compared alteration of gene expression profiles associated with physiological aging of LT-HSCs (GSE48893) with this RNAseq analysis data. This analysis revealed that ASXL1-MT promotes the expression of age-associated patterns, suggesting that ASXL1-MT accelerates aging of LT-HSCs ( Fig. 10e and f).
AKT/mTOR pathway seems to be activated in MDS patients harboring ASXL1 mutations compared with those harboring DNMT3A or TET2 mutations. Finally, we examined whether the development of MDS, a clonal HSC disease frequently arising from CH, harboring ASXL1 mutations is associated with enhanced Akt/mTOR signaling. To this end, we analyzed a public transcriptome data of CD34-positive bone marrow cells from MDS patients with DNMT3A, TET2, or ASXL1 mutations, and healthy control subjects (Gene Expression Omnibus accession number: GSE58831). GSEA suggested that AKT pathway is upregulated in MDS patients harboring ASXL1 mutations, but not those harboring DNMT3A or TET2 mutations, compared to healthy control subjects (Supplementary Table 2). We also found downregulation of negatively regulated genes by mTOR signaling in MDS patients with ASXL1 mutations (Supplementary Table 2). These data suggest that AKT/mTOR pathway is activated in MDS patients harboring ASXL1 mutations.

Discussion
Somatic mutations in epigenetic regulators, ASXL1, TET2, and DNMT3A, are recurrently detected in CH [19][20][21] . TET2 or DNMT3A deficiency in mice enhances self-renewal of LT-HSCs [30][31][32][33] , suggesting that mutations in TET2 and DNMT3A induce clonal expansion of hematopoietic cells, leading to CH in humans. In contrast, ASXL1 mutations in mice reduce the number and function of HSCs 25,34,35 . In the present study, we confirmed that ASXL1-MT causes a competitive disadvantage of LT-HSCs after transplantation. Nevertheless, LT-HSCs expressing ASXL1-MT acquire a growth advantage and eventually occupy the HSC compartment during aging in genetic mosaic mouse model. This observation is clearly distinct from that of DNMT3A or TET2 mutations, which exhibits an increase in the long-term regenerative potential of HSCs. Our results provide an insight into the pathogenesis of CH, in that CH can be induced by clonal expansion of not only "true" HSCs but also by long-lived pLT-HSCs lacking repopulation potential 1,9 . The enhanced Akt/ mTOR pathway is expected to provoke replication stress along with increased ROS-mediated DNA damage. Indeed, we observed increased DNA strand breaks in aged ASXL1-MT KI pLT-HSCs, indicating that ASXL1-MT expands the pLT-HSC compartment with accumulation of DNA damage. This could be one mechanistic explanation for why individuals with CH are at increased risk for subsequent hematological malignancies, associated with secondary mutations.
The present results suggest that the activated Akt/mTOR pathway, which promotes cell cycle progression, is responsible for aberrant expansion of the LT-HSC compartment in aged ASXL1-MT KI mice. At the same time, we also found that overactive Akt/ mTOR signaling induced by ASXL1-MT provokes dysfunction of HSCs. It has been shown that the enhanced Akt/mTOR activity by depletion of Pten results in transient expansion and subsequent depletion of HSCs 58,59 . It appears that ASXL1-MT has a much milder impact on Akt/mTOR pathway than disruption of a core repressive component comprising Akt/mTOR pathway. The modestly activated AKT/mTOR signaling induced by ASXL1-MT presumably has negative effects on HSC function, whereas it causes abnormal expansion of pLT-HSCs in steady-state hematopoiesis over time. These molecular changes should underlie the expansion of pLT-HSCs along with impaired repopulation potential in ASXL1-MT KI mice. Related to these findings, we observed enhancement of age-associated phenotypes in hematopoiesis including anemia, myeloid-biased differentiation, hypocellular bone marrow, and expansion of surface marker-defined LT-HSCs in aged ASXL1-MT KI mice. These phenotypes were partly rescued by inhibition of Akt/mTOR pathway (Figs. 5c and   9d), suggesting that ASXL1-MT-induced activation of Akt/mTOR pathway is involved in enhanced aging of the hematopoietic system. In addition, RNA-seq analyses revealed that ASXL1-MT promotes the gene expression of age-associated patterns. These results indicate that ASXL1-MT could promote HSC aging via activation of Akt/mTOR pathway.
One of the most intriguing findings of this work is the identification of AKT as a non-histone target of BAP1. BAP1, which is stabilized and given an enhanced DUB activity by ASXL1-MT, is a known DUB for histone H2A, and was also shown to deubiquitinate several non-histone proteins [60][61][62][63][64] . It was previously demonstrated that multiple E3 ligases induce ubiquitination and  degradation of phosphorylated AKT, thereby quenching activated AKT signaling 38,[40][41][42] . Conversely, ASXL1-MT/BAP1 complex deubiquitinates and stabilizes AKT, which leads to weak but possibly continuous activation of Akt/mTOR pathway. Thus, these findings revealed a function of ASXL1-MT as a signaling modulator beyond its known roles as an epigenetic regulator. On the other hand, one recent study has shown that Asxl1 loss activates Akt/ mTOR pathway due to an epigenetic dysregulation at the Pten locus 65 . Thus, we cannot completely exclude the possibility that ASXL1-MT activates Akt/mTOR pathway through an epigeneticdependent manner as well.
Among individuals with CH, the increased risk for developing hematological malignancies and cardiovascular diseases is associated with higher variant allele frequency 21,66 . Thus, it is possible to avoid the onset of these diseases by inhibiting the expansion of mutated CH clone. In the present study, we demonstrate that the aberrant cell cycle progression and the increased frequency of pLT-HSCs are suppressed by treatment with rapamycin in aged NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-22053-y ARTICLE ASXL1-MT KI mice. Hence, a pharmacological inhibition of the Akt/mTOR pathway to individuals with CH harboring ASXL1 mutations is expected to reduce a clonal advantage of the CH clone.
Besides CH, recent reports have revealed that normal tissues also accumulate somatic mutations with aging, which may confer a selective advantage 67 . However, mutational landscapes are clearly distinct; while mutations in epigenetic factors and splicing factors are frequently detected in CH, most mutations detected in oncogenes are signal transduction molecules in other tissues. At present, the reason for this discrepancy remains totally elusive. Clarification of the mechanisms by which CH progress to hematological malignancies may provide us a clue to understand the difference.
In conclusion, we show that ASXL1-MT collaborates with BAP1 to activate Akt/mTOR pathway in an epigeneticsindependent manner. The activated Akt/mTOR pathway causes aberrant expansion of pLT-HSCs to occupy the HSC compartment during aging. Overactive Akt/mTOR signaling also causes mitochondrial activation, overproduction of ROS, increased DNA damage, and subsequent dysfunction of HSCs. The expansion of pLT-HSCs along with increased DNA damage caused by ASXL1 mutations can result in the development of CH, leading to hematopoietic malignancy with secondary mutations. A pharmacological inhibition of Akt/mTOR pathway may pave the way for a preventive intervention to individuals with CH harboring ASXL1 mutations.

Methods
Mice. Wild-type C57BL/6J mice were bred in-house. p53 −/− mice were provided by K. Matsuda (Tokyo University). Conditional ASXL1-MT KI mice have been generated in our laboratory and crossed to Vav-Cre or Mx1-Cre transgenic mice. Experiments were performed with 6-12-week-old for young mice or 20-24-monthold for aged mice. All mice were housed with a 12 h dark/light cycle at a temperature between 20-25°C and a humidity between 40% and 60%. The experiments were approved by the Committee on the Ethics of Animal Experiments and all these mice were maintained according to the guidelines of the Institute of Laboratory Animal Science (PA13-19 and PA16-31).
CRISPR/Cas9-mediated gene knockout. To obtain short guide RNA (sgRNA) constructs targeting for Bap1, annealed oligos were cloned into lentiGuide-puro vector (Addgene, #52963) as previously described 23 . 293T cells were transfected with Cas9-expressing vector (lentiCas9-Blast; Addgene, #52962) or sgRNAexpressing vector together with lentiviral packaging vectors (pMD2.G; Addgene, #12259 and psPAX2; Addgene, #12260) using polyethylenimine (PEI), and then lentiviral supernatant was collected. cSAM cells were transduced with the viruses for 24 h and were subsequently subjected to drug selection with 10 μg/mL blasticidin for stable expression of Cas9 and 1.5 μg/mL puromycin for stable expression of sgRNA.  expressing vectors. Total cell lysates were subjected to immunoprecipitation using anti-FLAG antibody followed by immunoblotting. b 293T cells were transfected with AKT1 and BAP1 expressing vectors. Total cell lysates were subjected to immunoprecipitation using anti-AKT antibody followed by immunoblotting. c 293T cells were transfected with AKT1, FLAG-ASXL1-MT, BAP1, and Myc-Ubiquitin expressing vectors. Total cell lysates were subjected to immunoprecipitation using anti-AKT antibody followed by immunoblotting. Relative levels of Myc-Ubiquitin were quantified by densitometry and normalized to total Myc-Ubiquitin levels. d 293T cells were transfected with AKT1, FLAG-ASXL1-WT, FLAG-ASXL1-MT, BAP1, and Myc-Ubiquitin expressing vectors. Total cell lysates were subjected to immunoprecipitation using anti-AKT antibody followed by immunoblotting. Relative levels of Myc-Ubiquitin were quantified by densitometry and normalized to total Myc-Ubiquitin levels. e-g Murine bone marrow cells transformed by combined expression of SETBP1-D868N and ASXL1-MT (cSAM cells) were transduced with Cas9 and sgRNA targeting Bap1 (e). Bap1-depleted cSAM cells were then starved of IL-3 for 3 h followed by stimulation with IL-3 (1 ng/mL) for the indicated times (f). Relative levels of phosphorylated Akt were quantified by densitometry and normalized to total Akt levels (g). Similar result was obtained from Bap1-depleted cSAM cells using another sgRNA-targeting BAP1 (sgBAP1 #2). A representative experiment from at least n = 2 independent experiments is shown.
with BD FACSuite software or FACSAria with FACSDiva software and were analyzed with FlowJo software. Intracellular staining. Bone marrow cells stained for surface markers were fixed in 4% paraformaldehyde for 10 min at room temperature, permeabilized with 0.2% Triton X-100 for 15 min at room temperature and blocked with 5% goat serum for Cell cycle analysis. Bone marrow cells stained for surface markers were fixed in 4% paraformaldehyde for 10 min at room temperature, permeabilized with 0.2% Triton X-100 for 15 min at room temperature and blocked with 5% goat serum for 15 min at room temperature. Cells were then stained with the Ki-67-eFluor660 (eBiosciense, 50-5698-82) and DAPI for 30 min at 4°C.
Apoptosis analysis. Bone marrow cells were labeled with surface marker antibodies, as described previously, and were then stained with Annexin V-APC (BioLegend, 640920) antibody in binding buffer (10 mM Hepes (pH 7.4), 140 mM NaCl, 2.5 mM CaCl 2 ) for 30 min at room temperature. Serial transplantation assay. 1.0 × 10 6 bone marrow MNCs harvested from ASXL1-MT KI mouse were mixed with 1.0 × 10 6 bone marrow MNCs harvested from littermate control mouse and injected intravenously into lethally irradiated C57BL/6 recipient mice. 6 months after transplantation, 6 × 10 6 million bone marrow MNCs were isolated from three recipient mice and serially transplanted into lethally irradiated recipients. Donor-derived cells were evaluated by GFP positivity.
Homing assay. 5 × 10 4 LSK cells harvested from Vav-Cre ASXL1-MT KI mice were stained with 1 μM CellTrace Far Red (Thermo Fisher Scientific, C34564) for 20 min at room temperature. Cells were then incubated with IMDM, washed, and injected into lethally irradiated mice. 16 h after transplantation, bone marrow cells isolated from femurs and tibias of recipient mice were analyzed to count CellTracepositive cells.
Quantification of cytokines. The levels of TNF-α and IFN-γ in blood serum and bone marrow fluids were measured with TNF-alpha Quantikine ELISA Kit (R&D Systems, MTA00B) and IFN-gamma Quantikine ELISA Kit (R&D Systems, MIF00), respectively, according to the manufacturer's protocol.
Analysis of ROS levels. For intracellular ROS analysis, bone marrow cells were incubated for 30 min at 37°C with 5 μM CellROX Deep Red (Thermo Fisher Scientific). Cells were then washed and analyzed by flow cytometry.
Quantification of mitochondrial membrane potential. Bone marrow cells were incubated with 100 μM MitoTracker Red (Thermo Fisher Scientific) for 15 min at 37°C. Cells were then washed and analyzed by flow cytometry.
Immunofluorescence. MPPs or LT-HSCs were sorted into the slide coated with poly-L-lysine (SIGMA). Cells were placed on ice for 30 min to permit cells to settle onto the slide and fixed with 4% paraformaldehyde for 20 min at room temperature. Cells were then permeabilized with 0.2% TritonX-100 for 15 min at room temperature and blocked with 5% goat serum for 1 h at room temperature. Cells were incubated with the anti-p-H2A.X (Cell Signaling Technology, #9718, 1:200) or anti-Tom20 (Santa Cruz Biotechnology, sc-11415, 1:500) antibody overnight at 4°C, washed and incubated in the Alexa Fluor 568 goat anti-rabbit IgG (Invitrogen, A11011, 1:1000) for 1 h at room temperature. Cell nuclei were stained with DAPI and mounted using fluorescence mounting medium (DAKO). Images were acquired with a Confocal Microscope A1R or super resolution microscope N-SIM (Nikon), and were processed with NIS-Elements software (Nikon). For quantitative analysis of mitochondrial morphology, images were first thresholded and then converted to binary images by using ImageJ software 69,70 . Individual particle (mitochondria) were analyzed for mitochondrial area and number of fragments.
Metabolome analysis. Sort-purified 5 × 10 4 LSK cells from three mice were pooled, washed and then frozen. Metabolite extraction from sorted cells for metabolome analysis was performed as described previously 71 . Briefly, frozen sorted cell fractions together with internal standard (IS) compounds (2-morpholinoethanesulfonic acid (MES) and 1,3,5-benzenetricarboxylic acid (trimesate)) was suspended in ice-cold methanol (500 μL) followed by the addition of an equal volume of chloroform and 0.4 times the volume of ultrapure water (LC/MS grade, Wako). The suspension was then centrifuged at 15,000 × g for 15 min at 4°C. After centrifugation, the aqueous phase was filtered using an ultrafiltration tube (Ultrafree MC-PLHCC, Human Metabolome Technologies). The filtrate was concentrated with a vacuum concentrator (SpeedVac, Thermo Fisher Scientific). The concentrated filtrate was dissolved in 25 μL of ultrapure water and used for ion chromatography (IC)-MS analyses as described below. For metabolome analysis focused on TCA-cycle intermediates, anion metabolites were measured using an orbitrap-type MS (Q-Exactive focus, Thermo Fisher Scientific), connected to a high performance IC system (ICS-5000+, Thermo Fisher Scientific) that enables highly selective and sensitive metabolite quantification owing to the IC-separation and the Fourier Transfer MS principle 72 . The IC was equipped with an anion electrolytic suppressor (Thermo Scientific Dionex AERS 500, Thermo Fisher Scientific) to convert the potassium hydroxide gradient into pure water before the sample enters the mass spectrometer. The separation was performed using a Thermo Scientific Dionex IonPac AS11-HC, 4-μm particle size column. IC flow rate was 0.25 mL/min supplemented post-column with 0.18 mL/min makeup flow of MeOH. The potassium hydroxide gradient conditions for IC separation are as follows: from 1 to 100 mM (0-40 min), 100 mM (40-50 min), and 1 mM (50.1-60 min), at a column temperature of 30°C. The Q Exactive focus mass spectrometer was operated under an ESI-negative mode for all detections. Full mass scan (m/z 70−900) was used at a resolution of 70,000. The automatic gain control (AGC) target was set at 3 × 10 6 ions, and maximum ion injection time (IT) was 100 ms. Source ionization parameters were optimized with the spray voltage at 3 kV and other parameters were as follows: transfer temperature at 320°C, S-Lens level at 50, heater temperature at 300°C, Sheath gas at 36, and Aux gas at 10.
RNA-seq analysis. Total RNA was isolated from sorted LSK cells using RNeasy Micro-kit (QIAGEN) and Agilent TapeStation high sensitivity RNA ScreenTape or Bioanalyzer RNA 6000 Pico kit (Agilent) was used to check for RNA quality and quantity. RNA libraries for young mice were prepared using 15 ng total RNA with the QIAseq FX Single cell RNA library kit (QIAGEN) according to the manufacturer's protocol. In this protocol polyA-selected mRNA was converted in cDNA, and then enzymatic fragmentation (incubation time for "fragment size = 300 bp" was used) and library preparation were performed using 1 μg cDNA. RNA libraries for aged mice were prepared using 1 ng total RNA with the NEBNext Single Cell/ Low Input RNA Library Prep kit (New England Biolabs). Briefly, cDNA synthesis and amplification were performed by template-switching PCR, and libraries were prepared using the amplified cDNA fragments. The quality and quantity of these libraries were checked using Agilent TapeStation D1000 or Agilent high sensitivity DNA kit (Agilent) and KAPA Library Quantification Kits (KAPA BioSystems)/ real-time PCR Systems Step One Plus (Applied Biosystems). These libraries were  Gene set enrichment analysis (GSEA) was performed using the GSEA tool from the Broad Institute (http://software.broadinstitute.org/gsea/). Heat maps were generated using heatmap.2 function of gplots package (3.0.4), from the Bioconductor. The microarray data for HSCs of young and aged wild-type mice are publically available at Gene Expression Omnibus (GEO) with the reference series tag "GSE 48893". Fig. 9 The dysregulated hematopoiesis in aged ASXL1-MT KI mice is reverted by treatment with rapamycin. a Aged Vav-Cre ASXL1-MT KI mice and age-matched control mice were treated with rapamycin intraperitoneally every other day for 8 weeks. At the end of administrations, end-point analyses of bone marrow cells were conducted. b, c Mitochondrial membrane potential (b) and intracellular ROS levels (c) in LT-HSCs were analyzed (n = 4 (Control-Vehicle), 5 (Control-Rapamycin), 4 (ASXL1-MT-Vehicle), and 5 (ASXL1-MT-Rapamycin)). Results are expressed as fold change relative to vehicle control group. d Absolute numbers of bone marrow cells per leg (n = 6 (Control-Vehicle), 7 (Control-Rapamycin), 5 (ASXL1-MT-Vehicle), and 7 (ASXL1-MT-Rapamycin)). e Absolute numbers of white blood cells in bone marrow per leg (n = 5 (Control-Vehicle), 6 (Control-Rapamycin), 5 (ASXL1-MT-Vehicle), and 6 (ASXL1-MT-Rapamycin)). Data are mean ± s.d. f The experimental design for transplantation using whole bone marrow cells treated with rapamycin. Whole bone marrow cells isolated from aged Vav-Cre ASXL1-MT KI mice and age-matched control mice that had been treated with rapamycin for 4 weeks were transplanted into lethally irradiated recipient mice with competitor cells. Recipient mice were continuously treated with rapamycin after transplantation. g, h 4 weeks after transplantation, effects of rapamycin treatment on mitochondrial membrane potential (g) and γ-H2AX levels (h) in LSK cells were determined (n = 4). i, j 4 weeks after transplantation, the frequency of donor-derived cells in peripheral blood and LSK cells were analyzed (n = 8) (i). Differentiation index, which is obtained by dividing the frequency of donor-derived cells in peripheral blood by LSK cells, was calculated (j). Statistical significances are assessed by one-way ANOVA with Tukey-Kramer's post-hoc test (b-e) or two-tailed Student's t-test (g-j). *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Comet assay. Alkaline comet assays were performed with CometAssay Kit (Trevigen). Sort-purified LT-HSCs were embedded in LM Agarose and transferred onto Comet Slides. Cells were then lysed with lysis solution and soaked in alkaline solution (200 mM NaOH, 1 mM EDTA). Comet Slides were subjected to electrophoresis in alkaline solution followed by staining with SYBR Green. Analysis was performed using the ImageJ software.
Statistics and reproducibility. Data are expressed as mean ± standard deviation (s.d.) or standard error of mean (s.e.m.). P values were calculated using unpaired, two-tailed Student's t-test, Mann-Whitney rank sum test, one-way ANOVA with Tukey-Kramer's post-hoc test, or two-sided Fisher's exact test as indicated in figure legend. The significance between the survival of cohorts was assessed by Kaplan-Meier survival analysis and log-ranked tests. P values < 0.05 were considered significant. All calculations were performed by using GraphPad Prism software. Sample size 'n' indicates biological replicates. No randomization or blinding was used and no animals were excluded from the analysis.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
RNA-seq data are available in sequence read archive (SRA) database (accession ID: PRJNA673672). Metabolome analysis data are available in MassIVE (accession ID: MSV000086816). The datasets generated or analyzed during the current study are available from the corresponding author (kitamura@ims.u-tokyo.ac.jp) on reasonable request. Source data are provided with this paper.
Received: 5 December 2019; Accepted: 23 February 2021; Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/.