Atp8a1 deletion increases the proliferative activity of hematopoietic stem cells by impairing PTEN function

Abstract

Purpose

The eukaryotic cell plasma membrane contains several asymmetrically distributed phospholipids, which is maintained by the P4-ATPase flippase complex. Herein, we demonstrated the biological effects and mechanisms of asymmetrical loss in hematopoietic stem cells (HSCs).

Methods

An Atp8a1 knockout mouse model was employed, from which the HSC (long-term HSCs and short-term HSCs) population was analyzed to assess their abundance and function. Additionally, competitive bone marrow transplantation and 5-FU stress assays were performed. RNA sequencing was performed on Hematopoietic Stem and Progenitor Cells, and DNA damage was assayed using immunofluorescence staining and comet electrophoresis. The protein abundance for members of key signaling pathways was confirmed using western blotting.

Results

Atp8a1 deletion resulted in slight hyperleukocytosis, associated with the high proliferation of HSCs and BCR/ABL1 transformed leukemia stem cells (LSCs). Atp8a1 deletion increased the repopulation capability of HSCs with a competitive advantage in reconstitution assay. HSCs without Atp8a1 were more sensitive to 5-FU-induced apoptosis. Moreover, Atp8a1 deletion prevented HSC DNA damage and facilitated DNA repair processes. Genes involved in PI3K-AKT-mTORC1, DNA repair, and AP-1 complex signaling were enriched and elevated in HSCs with Atp8a1 deletion. Furthermore, Atp8a1 deletion caused decreased PTEN protein levels, resulting in the activation of PI3K-AKT-mTORC1 signaling, further increasing the activity of JNK/AP-1 signaling and YAP1 phosphorylation.

Conclusion

We identified the role of Atp8a1 on hematopoiesis and HSCs. Atp8a1 deletion resulted in the loss of phosphatidylserine asymmetry and intracellular signal transduction chaos.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13402-023-00797-7.

Introduction

Eukaryotic cell membranes are constructed from many molecules, including phospholipids, cholesterol, and proteins. The phospholipid bilayer is the basic scaffold of the cell membrane, which is comprised of sphingomyelin (SM), phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol (PI) and phosphatidylserine (PS) [1, 2]. The different phospholipids are unevenly distributed within the outer and inner leaflets. For instance, the phospholipid composition in the outer leaflet of the erythrocyte membrane is rich with SM and PC, whereas the inner leaflet is enriched with PE, PI and PS [3]. The asymmetric nature of the membrane is crucial for proper cellular functioning. Lipid asymmetry within the membrane contributes to the asymmetry of the membrane itself, and most lipids can flip between outer and inner leaflets (except for glycolipids) [4, 5]. The asymmetric distribution of phospholipids within the membrane is required to maintain cellular biological activities. Remarkably, phospholipids can be actively translocated by flippases, floppases and scramblase via ATP hydrolysis to maintain the lipid bilayer asymmetry [6]. The P4-ATPase flippase complex is comprised of a catalytic alpha subunit and an accessory beta subunit, which catalyzes the hydrolysis of ATP to provide energy and facilitates the transport of aminophospholipids from the outer to inner leaflet [7]. PS and PI are negatively charged molecules of the inner leaflet that interact with positively charged transmembrane residues near the cytoplasmic end of single-pass membrane proteins [8].

There are 14 types of P4-ATPases in the human proteome that fulfill multiple important physiological functions, including: the generation and maintenance of membrane phospholipid asymmetry, lipid scavenging, vesicle formation and membrane scaffolding [9]. Previous studies have shown that P4-ATPase deficiency can result in various diseases [9]. For example, 1) ATP8B1 mutations are associated with progressive familial intrahepatic cholestasis (PFIC1) [10]. 2) ATP11C is critical for phosphatidylserine internalization and B lymphocyte differentiation; its deletion/deficiency has been shown to cause B cell–deficiency syndrome in mice [11]. 3) Atp8a1 deficiency or abolishment has been demonstrated to cause physical and functional defects in humans and mice [12–15].

The primary function of ATP8A1 is to catalyze ATP hydrolysis to provide energy for the ATP8A1-TMEM30A flippase complex, which sequesters PS from the outer to inner leaflet to maintain an asymmetric PS distribution in the plasma membrane [9, 16]. PS is a cytoplasmic leaflet protein docking site component that contributes to the activation of several key signaling pathways, including AKT, PKC and RAF-1 signaling, which are known to stimulate neuronal survival, neurite growth and synaptogenesis [17–19]. Indeed, previous studies have demonstrated that Atp8a1 deletion delays murine hippocampus-dependent learning [15, 20].

PS is also involved in many other biological processes. It is placed upon the surface of apoptotic cells, where it can be recognized by protein-S and GAS6 during co-opsonized macrophage phagocytosis [21]. PS participates in the coagulation cascade. Platelets expose PS on the outer surface when the coagulation process is activated. Extracellular PS recruits clotting factors V, VIII, X, and prothrombin, which can attach to the surface of activated platelets [22]. PS also serves as an anchor for some cytoplastic proteins. For example, RAS, RHO-family GTPases and the SRC family kinase interact with PS via the hydrophobic side chain. Some proteins interact with PS via the conventional C2 domains, such as PKC, Annexin V, coagulation factor V, lactadherin and PTEN [23, 24]. PTEN negatively regulates PI3K/AKT signaling pathway activation via dephosphorylating PIP3 to PIP2), inhibiting AKT activation [25]. AKT–mTOR network plays a central role in HSC homeostasis, which responds to multiple intracellular and extracellular signals in regulating several cellular processes, including the cell cycle, mitochondrial metabolism, and protein synthesis [26].

Based upon the aforementioned studies, we hypothesized and demonstrated that Atp8a1 deletion and disordered PS distribution might disrupt HSC function in self-renewal, proliferation, and differentiation.

Materials and methods

Mice

Atp8a1 heterozygous mutant (Atp8a1^±) mice were purchased from The Jackson Laboratory (US). Atp8a1 homozygous mutation (Atp8a1^−/−) and their wild-type littermates (WT) mice were generated by intercrossing Atp8a1^± mice as previously described [20]. All animal studies were performed in accordance with the guidelines approved by the Institutional Animal Care and Use Committee of Sichuan University (CN). The genotype was validated by polymerase chain reaction using tail slips. Additionally, western blot analyses were performed on cultured (in vitro) HSC-like cells to determine whether ATP8a1 protein was absent in the KO mice (Supplementary Data S2a).

Flow cytometry

White blood cells (WBC) from donor mouse peripheral blood (PB), spleen and bone marrow (BM) were incubated in red blood cell lysis buffer (pH 7.4) to lyse erythrocytes. The cellular composition in these organs/tissues was determined by flow cytometry and analyzed using FlowJo (V10) software. Single-cell suspensions were incubated for 30 min at room temperature with the following cell-surface marker antibodies: B220 for B cells; Gr-1 for myeloid cells; CD3 for T cells; Lin^lowc-Kit⁺Sca-1⁺ for HSCs; Lin^lowc-Kit⁺Sca-1⁻CD34⁺CD16/32⁻ for common myeloid progenitors (CMPs); Lin^lowc-Kit⁺Sca-1⁻CD34⁺CD16/32⁺ for granulocyte-monocyte progenitors (GMP); Lin^lowc-Kit⁺Sca-1⁻CD34⁻CD16/32⁻ Megakaryocyte-erythroid progenitors (MEP). Lin^lowc-Kit⁺Sca-1⁺CD150⁺CD48⁻ for long-term HSCs (LT-HSCs). Lin^lowc-Kit⁺Sca-1⁺CD150⁻CD48⁻ for short-term HSCs (ST-HSCs). Lin^lowc-Kit⁺Sca-1⁺CD150⁻CD48⁺ for multiple potential progenitors (MPPs). Pecp-cy5.5-CD150 and APC-cy7-CD48 were purchased from Biolegend Biosciences (US); whereas efluor450-B220, APC-Gr-1, PE-Cy7-CD16/CD32, Alexa fluor700-CD34, PE-cy7-CD45.1, APC-efluor780-CD45.2, PE-IgM, PE-CD3, Pacific blue-Lineage, PE-Sca-1 and APC-c-Kit were purchased from eBioscience (US).

Bone marrow competition assay

B6.SJL-Ptprca-Pepc^b/BoyJ (CD45.1) mice and Atp8a1^mut (CD45.2) donor (8–10 weeks old) BM cells were harvested and mixed at a 1:1 ratio. Total BM cells 1 × 10⁶ were injected into CD45.2 mice, subjected to two rounds of 500 Gy X-ray irradiation. PB, spleen and BM cells were collected post-transplantation 12 weeks and analyzed using FACS to determine the CD45.1 and CD45.2 cell percentage in the recipients.

Hematopoietic stem/progenitor cell isolation

Murine bone marrow cells were extracted by flushing femurs and tibias with PBS. Anti-Phycoerythrin (PE) MicroBeads (Miltenyi Biotec, DE) were bound to PE-conjugated sca-1 antibodies to enrich HSCs among the Sca-1⁺ cells. Sca-1⁺ cells were stained to identify HSC surface markers (Lin^low, c-Kit⁺, and Sca-1⁺) and sorted using an Aria III FACS machine (Becton Dickinson, US).

RNA-sequencing analysis

Total RNA was extracted from the sorted HSPCs (Lin^lowSca1⁺c-Kit⁺, LSK cells) using an RNeasy Mini kit (Qiagen, DE). The RNA samples were reverse transcribed, amplified using Ovation Pico WTA System V2 and biotin-labeled with the Encore Biotin Module (NuGEN Technologies, US). RNA-sequencing (RNA-seq) was performed using an Illumina (US) MiSeq system. The RNA-seq and gene set enrichment analyses were performed as previously described [27]. The raw sequence data reported in this article was deposited at the Genome Sequence Archive [28] within the National Genomics Data Center [29], China (National Center for Bioinformation / Beijing Institute of Genomics, Chinese Academy of Sciences, GSA: CRA009276), and is publicly accessible at https://ngdc.cncb.ac.cn/gsa.

GO/KEGG enrichment analysis

Gene Ontology (GO) enrichment analyses were performed to identify the biological implications of the differentially expressed genes. Fisher's Exact Test was used to identify significantly enriched GO categories, for which the P values were adjusted using the False Discovery Rate (FDR) method [30]. The 20 most significantly enriched GO terms for the different conditions/cell types were displayed using graphics generated with ggplot2.

BM transduction/transplantation CML mouse model

The MSCV-BCR/ABL1-IRES-eGFP retroviral vector was used to generate virus stocks as previously described. Briefly, donor mice were pretreated with 200 mg/kg 5-fluorouracil (5-FU) via a tail vein injection. Four days later, BM cells were harvested from donor mouse femurs and tibia, pre-stimulated with interleukin-3 (Il3), interleukin-6 (Il6), and stem cell factor (SCF), then cosedimented twice with retroviral stock at 1000 g for 90 min. The transduced marrow cells were transplanted by lateral tail vein injection into lethally irradiated syngenetic recipient mice, about 3000 LSC (GFP⁺LinlowSca1⁺c-Kit⁺) cells per mouse. Leukemia cells were identified using FACS as per the description above.

Cell immunofluorescence staining and imagining

Sorted cells were fixed with 4% PFA for 10 min and dispersed on a glass slide. The cells were subsequently permeabilized with 200µL PBS containing 0.2% Triton X-100 at room temperature (RT) for 20 min, blocked with PBST solution containing 2% BSA for one hour at RT and incubated with the primary antibody (1:500) overnight at 4 °C. Following the incubation, the cells were gently washed and incubated with Alexa555-conjugated secondary antibody (1:500) for one hour at RT. After this, the cells were washed with PBST for 15 min and incubated with 5 μl DAPI in the dark for five minutes. The stained cells were imaged by fluorescent or confocal microscopy (Zeiss Axio Imager M2, DE).

DNA damage burden analysis

Sorted HSPCs were used in an alkaline comet electrophoresis analysis with a Trevigen Comet Assay kit according to the manufacturer’s protocols. In brief, purified HSPCs were embedded in Comet low-melting-point agarose, transferred onto a Trevigen CometSlide, and assayed on a single high-throughput slide (to prevent analytical variation). The immobilized cells were lysed overnight and treated with freshly made alkaline unwinding solution, followed by electrophoresis (under alkaline conditions using the Trevigen Comet Assay Electrophoresis System), SYBR Green staining and imaging.

Statistical analysis

The results from the experiments described herein were presented using the mean plus/minus the standard error of the mean (± SEM). The statistical significance was determined using a one-way analysis of variance (ANOVA) or Student’s t-test using GraphPad Prism (version 8). A P value equal to or less than 0.05 was used as the threshold when considering the statistical significance of differences.

Results

Atp8a1 deletion causes a hematopoiesis disorder in mice

To determine whether Atp8a1 deficiency affects hematopoiesis, adult Atp8a1^−/− mice and control littermates were sacrificed for hematological system analysis. The total WBC number was counted in the peripheral blood, spleen, and bone marrow, which revealed that Atp8a1^−/− mice had slightly increased spleen and BM counts (Fig. 1a). B cells, T cells, and neutrophils were analyzed further with FACS to ascertain whether Atp8a1^−/− affected lineage-committed cells. We found that the percentage and number of neutrophils were higher in PB from Atp8a1^−/− mice (Fig. 1b), whereas the B cell percentage and number were lower in Atp8a1^−/− mice (Fig. 1c); there were no apparent differences for T cells (Fig. 1d). In Atp8a1^−/− spleens, the total neutrophil number and percentage were consistently higher (Fig. 1e). Unlike in the PB, the splenic B cell percentage was comparable between the Atp8a1^−/− and WT mice, but the total number was higher (Fig. 1f). In the BM, the percentage and number of neutrophils were increased, B cells were decreased and T cells were not significantly different (Fig. 1g-i). These results show that Atp8a1 loss was associated with B cell deficiency in the BM, increased hematopoiesis, and increased myeloid-biased differentiation.

Fig. 1.

Atp8a1 deletion increases hematopoiesis in mice (a) White blood cell counts in peripheral blood (PB) and bone marrow (BM) and spleen weight were indicated. BM cells were collected and enumerated from femurs and tibias of Atp8a1-/- mice or their WT littermates. (b-d) The percentage and number of neutrophils, B cells, and T cells in PB of WT and Atp8a1^−/− mice. (e) The percentage and total number of neutrophils in the spleen. (f) The percentages and the total number of B cells in the spleen. (g-i) The percentage and the total number of neutrophils, B cells and T cells in BM. (j) The histological sections of BM from Atp8a1^−/− and WT respective mice. Upper: 200 times magnified; Lower: 400 times magnified. (k) Light micrographs of the spleen sections from Atp8a1.^−/− and WT mice. Upper: 200 times magnified; Lower: 400 times magnified. At least three independent experiments were performed for confirmation, representative results are displayed. Data represented using the mean ± SEM. n = 6. * P ≤ 0.05

Histomorphometric analysis showed no significant structural BM differences between Atp8a1^−/−and WT mice (Fig. 1j). The histological Atp8a1^−/− mouse spleen sections revealed that there were fewer lymphoid follicles, and the distinction between the red and white pulp was not clear due to lymphoid architecture distortion, diffuse white pulp, and a greater presence of granular leukocytes in the red pulp in (Fig. 1k). The pathological changes within Atp8a1^−/− spleens were consistent with the results from the cell composition analyses.

Atp8a1 deletion increases the HSC pool and promotes their proliferation

The high hematopoiesis and myeloid-biased differentiation rate might indicate that Atp8a1 loss extends HSC self-renewal capacity. Therefore, BM cells were harvested from adult Atp8a1^−/− and WT group mice, after which hematopoietic stem and progenitor cells (HSPCs) were analyzed via FACS to examine whether Atp8a1 deficiency affected HSC and hematopoietic progenitor cell (HPC) homeostasis (Fig. 2a). We found that Atp8a1 deficiency caused significantly increased HSPC (LSK, lin^lowSca-1⁺c-Kit⁺ cells) percentage and number (Fig. 2b). All the HSPC subpopulations, including long-term HSCs (LT-HSCs, lin^lowSca-1⁺c-Kit⁺CD150⁺CD48⁻), short-term HSCs (ST-HSCs, lin^lowSca-1⁺c-Kit⁺ CD150⁻CD48⁻) and multiple potential progenitors (MPP, lin^lowSca-1⁺c-Kit⁺ CD150⁻CD48⁺) were increased in Atp8a1 deficient BM (Fig. 2c-e). We also detected increased granulocyte-monocyte progenitor (GMP) cell numbers in Atp8a1 deficient mice compared to the WT controls (Fig. 2f). In contrast, common myeloid progenitor (CMP) was not increased in percentage and number (Fig. 2g). CMPs are generated from MPPs, which can differentiate into GMPs and MEPS. However, the number of CMPs was not increased in the BM from Atp8a1 mice. It is possible that their numbers could not rise due to increased CMP to GMP/MEPS transition efficiency.

Fig. 2.

Atp8a1 deletion increases HSC pool (a)The schematic diagram of hematopoietic stem and progenitor cell (HSPC) analysis with FACS. (b) The percentages and numbers of HSPCs (Lin-c⁻Kit⁺Sca1⁺) BM of Atp8a1^−/− and wide-type (WT) mice were indicated. (c-d) The percentage and total numbers of long-term hematopoietic stem cells (LT-HSCs), short-term hematopoietic stem cells (ST-HSCs) and multiple potential progenitors (MPPs) in BM were indicated. (f–h) The percentage and number of granulocyte-monocyte progenitor (GMPs), common myeloid progenitor (CMPs), and megakaryocyte-erythroid progenitor (MEPs) in BM of Atp8a1.^−/− and WT mice were indicated. At least three independent experiments were performed for confirmation. Data represented using the mean ± SEM. n = 6. *, P ≤ 0.05; **, P ≤ 0.01

Atp8a1 deletion improves HSC reconstitution capability

To investigate whether Atp8a1 deficiency affects the capacity of HSCs to repopulate, Atp8a1^−/− (CD45.2) and B6.SJL-Ptprca Pepcb/BoyJ (CD45.1) BM cells were mixed equally and transplanted into lethally irradiated CD45.2 mice (Fig. 3a). Twelve weeks following the bone marrow transplantation, the recipient mice were sacrificed, and the percentage of lineage-committed cells were analyzed in the PB, spleen, and BM. Overall, BM cells from Atp8a1^−/− mice had a stronger capacity for hematopoietic repopulation (Fig. 3b) and were responsible for a significantly greater percentage of the HSPCs (Fig. 3c). These results suggested that Atp8a1 deficiency increased HSC repopulation capability under competitive conditions.

Fig. 3.

Atp8a1 deletion promotes HSPC proliferation (a) Reconstitution assay schematic diagram. Twelve weeks post-BMT transplantation, PB, spleen (SPL), and BM cells were collected and analyzed using flow cytometry. (b) The percentage of committed cells in the PB, spleen, and BM from CD45.2 (Atp8a1^−/−) and CD45.1 (WT) HSCs. (c) The percentage of HSPCs was replenished from CD45.1 (WT) and CD45.2 (Atp8a1^−/−) HSCs. The percentage of CD45.2 and CD45.1 cells was shown in the plots. (d) Neutrophil (NEUT) and Lymphocyte (LYMPH) counts in PB on day four post-5-FU injection. (e) H&E BM staining for 5-FU treated Atp8a1^−/− and WT mice. 200 times magnification. (f) The total cell number collected from 5-FU treated Atp8a1^−/− and WT mice femurs and tibias. (g-h) The number of HSPCs and LT-HSCs in Atp8a1.^−/− and WT mice post 5-FU exposure. Three independent experiments were performed for each assay, representative images are shown. The data was presented using the mean ± SD. n = 3. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001

Cells undergoing proliferation are sensitive to 5-FU-induced cell death. To assess the impact of Atp8a1 during stress-induced hematopoiesis, Atp8a1^−/− and WT mice were administered 5-FU for five days. Atp8a1^−/− mice had symptoms of weakened hematopoiesis relative to the WT mice, including lower WBC counts and reduced neutrophil and lymphocyte numbers in PB (Fig. 3d). A histological BM analysis showed that its cellularity was significantly decreased in Atp8a1-/- mice compared with WT controls five days post-5-FU injection (Fig. 3e). We also found an attenuated total cell number decrease within the BM (Fig. 3f). Moreover, the HSPC and LT-HSC frequency and absolute number were significantly lower in Atp8a1^−/− mice after 5-FU exposure compared to the WT littermates (Fig. 3g and h). These results suggest that Atp8a1 deficiency resulted in greater HSPC sensitivity to 5-FU-induced cytotoxicity.

To validate whether Atp8a1 deletion also promotes HSC proliferation in vitro, LHX2-derived HSC-like cells were employed. Atp8a1^−/− and WT mouse BM cells were transduced with MSCV-LHX2-IRES-GFP retrovirus. The number of cells in the Atp8a1^−/− group was about 1.5 times higher than in the WT group and consistent with the in vivo data (data not shown). The cell proliferation was analyzed and the cell number was significantly higher in Atp8a1^−/− group (Supplementary Data S2b). In the cell cycling assay, we found more cells in the S-M stages in the Atp8a-/- group than in the WT group (Supplementary Data S2c). The results of the BrdU integration assay showed that the percentage of BrdU-positive cells accounted for about 40.8% of total cells in the Atp8a1^−/− group, while they accounted for only 30.1% in the WT group (Supplementary Data S2d). These results indicated that Atp8a1 deletion also promoted HSC proliferation in vitro.

Atp8a1 deletion promotes BCR/ABL1 transformed leukemia stem cell proliferation

Chronic myeloid leukemia (CML) is a canonical stem cell disease in which leukemia cells are replenished by leukemia stem cells (LSCs) derived from HSCs bearing the BCR/ABL1 fusion oncogene [31]. A BCR/ABL1-induced CML mouse model was produced via retroviral transduction/transplantation to study whether Atp8a1 deletion promoted LSC proliferation (Fig. 4a). The number of cells injected into the tail vein during the BMT experiments for all the groups was normalized based upon GFP and LSK cell counts. No significant overall survival (OS) differences were found between the BCR/ABL1-transduced WT recipients and Atp8a1^−/− BM cells (Fig. 4b). However, pulmonary hemorrhage (caused by leukemia cells) was reduced in Atp8a1^−/− group mice (data was not shown). Pulmonary hemorrhage is a major cause of death for CML-bearing mice. Thus, weaker extramedullary hematopoiesis in the Atp8a1^−/− group mice might account for prolonged survival despite the greater leukemia cell burden.

Fig. 4.

Atp8a1 deletion enhances BCR/ABL1 induced CML in mice (a) A schematic diagram for the BCR/ABL1 induced CML mouse model. i.v. Intravenous injection (b)The overall survival (OS) curves for the recipients of the BCR/ABL1 transduced WT or Atp8a1 deletion BM cells. Independent experiments were performed twice. (c) The OS curves for the secondary BMT CML. (d)The percentage of leukemia cells (double positive for Gr-1 and GFP) in PB from Atp8a1^−/− and WT donor recipients were monitored with FACS. (e) Leukemia cells in the afflicted mouse BM and spleens were analyzed on days 15 and 24 post-BMT. (f) The total LSC number in secondary BMT CML recipients. (g-h) H&E staining of lung and spleen sections from recipients receiving BCR/ABL1 transduced WT or Atp8a1.^−/− BM (200 times magnification). The data was expressed using the mean ± SD. n ≥ 3 mice. *, P ≤ 0.05; ****, P ≤ 0.0001

Serial BM transplantation is usually used to evaluate LSC self-renewal and proliferation in BCR/ABL1-induced CML models [32]. By harvesting BM cells from primary CML recipients and transferring them to secondary recipients, we found that Atp8a1^−/− BM significantly promoted CML progression in the secondary recipients. All the recipient mice in the Atp8a1^−/− group died of CML within three to four weeks, whereas the median survival time for the WT group was over five weeks, some of which were leukemia-free at the experimental endpoint (Fig. 4c). Correlatedly, there were more leukemia cells in the Atp8a1^−/− group PB during disease progression (Fig. 4d). Leukemia cells within the BM and spleen were analyzed on days 15 and 24 post-BMT, which was higher in the Atp8a1^−/− donor group mice (Fig. 4e). A quantitative analysis for BM-associated LSCs (GFP⁺Lin^lowSca-1⁺c-Kit⁺) revealed that the LSC percentage and absolute number was higher in the Atp8a1^−/− group mice (Fig. 4f). Consistently, more leukemia cells infiltrated into Atp8a1^−/− group mouse spleen, and more severe pulmonary hemorrhage was observed (Fig. 4g and h). Taken together, these results indicated that Atp8a1 deficiency promoted normal HSC proliferation and LSC proliferation.

Atp8a1 deletion causes AP-1 complex transcriptional activity

To investigate the gene expression profile changes caused by Atp8a1 deletion, Atp8a1^−/− and WT HSPCs were sorted for RNA-sequencing. A total of 1299 genes were determined to be significantly altered (P ≤ 0.05, with a two-fold or greater change in mean expression) between the two groups, among which 536 genes had increased expression and 763 genes had decreased expression in Atp8a1^−/− HSPCs (Supplementary data, S3). In addition, the differentially expressed genes (DEGs) were further analyzed to detect enrichment within the GO biological process database. The GO biological process analysis revealed DEGs that were significantly enriched for terms including: “nucleosome”, “telomerase holoenzyme complex”, “transcription factor AP-1 complex”, “double strand break repair”, and “response to reactive oxygen species” (Fig. 5a). Additionally, whole transcriptome comparisons generated using GSEA showed that gene sets related to “Phosphatidylinositol 3 Kinase Activity”, “PI3K-AKT-mTOR Signaling” and “DNA Repair Complex Assembly” were significantly enriched with increased expression in Atp8a1^−/− HSPCs (Fig. 5b).

Fig. 5.

Atp8a1 deletion leads to JNK-AP-1 signaling pathway activation (a) GO term analysis results comparing the Atp8a1^−/− and WT HSPC RNA-sequencing data. (b) Gene Set Enrichment Analysis (GSEA) results comparing Atp8a1^−/− and WT HSPCs. (c) The STRING analysis shows the predicted molecular action among DEGs regulated by the AP-1 complex in Atp8a1^−/− and WT HSPCs. Colored nodes (orange) show genes that comprise the AP-1 complex. (d) The targets of JUN and FOS were enriched and increased in Atp8a1^−/− HSPCs compared to WT HSPCs. (e) FOS and JUN were increased in Atp8a1^−/− HSC-like cells compared to WT cells. BETA ACTIN served as the sample loading control. (f) JNK phosphorylation level was increased in Atp8a1^−/− HSC-like cells. GAPDH served as the sample loading control

DEGs were further analyzed using the STRING database to view the functional protein association networks. The AP-1 complex subunits, JUN and FOS, were highlighted in the network (Fig. 5c); the function of these genes is consistent with the GO (CC) and GSEA analysis results (Fig. 5a and b). To further verify the AP-1 complex transcriptional activity, JUN and FOS target gene expression levels were compared between Atp8a1^−/− and WT HSPCs. These genes were enriched and transcript levels were significantly increased in Atp8a1^−/− HSPCs (Fig. 5d). Likewise, the JUN and FOS protein levels were significantly increased in Atp8a1^−/− HSC-like cells (Fig. 5e). JNK directly activates the AP-1 complex and promotes its transcriptional activity [33]. Therefore, we also detected JNK activity and found that the JNK phosphorylation level was significantly higher in Atp8a1^−/− HSC-like cells than in WT HSC-like cells (Fig. 5f).

Atp8a1 deletion is associated with reduced DNA damage

The GO (BP) and GSEA analyses showed that genes involved in the DNA repair pathway were significantly enriched with increased expression in Atp8a1^−/− HSPCs (Fig. 5a and b). Therefore, genes involved in DNA repair processes were selected, and those with significant expression changes were compared and clustered (Fig. 6a). These were further analyzed using a STRING database enrichment analysis. These genes were mainly present in the radiation response and the Fanconi anemia pathways (Fig. 6b). To characterize the biological effects of Atp8a1 deletion on DNA stability, HSPCs isolated from 24-week-old Atp8a1^−/− and WT mice were examined with an γH2A immunofluorescence assay, which demonstrated that fewer γH2A foci occurred in Atp8a1^−/− HSPCs and reflected reduced DNA damage (Fig. 6c and d). A further examination of the DNA damage burden was conducted using a comet electrophoresis assay, which also showed that less DNA damage occurred in Atp8a1^−/− HSPCs compared to those from WT mice (Fig. 6e-g).

Fig. 6.

Atp8a1 deletion prevents HSPC DNA damage (a) The heat map shows the expression and categorical aggregation for the major DNA damage repair genes in HSPCs from three Atp8a1^−/− mice versus two WT mice. (b) The STRING analysis shows the predicted molecular action among DEGs involved in DNA damage repair in HSPCs from three Atp8a1^−/− mice versus two WT mice. Colored nodes: genes involved in response to radiation (red) and genes involved in the Fanconi anemia pathway (blue). (c) Immunofluorescence γH2A (red) and DAPI (blue) staining in WT and Atp8a1^−/− HSPCs. (d) The Mean Rhoda Intensity value for each cell was analyzed, with each dot representing the value for an individual cell. At least 30 cellular signals were collected for each group. (e) Comet Assay results for HSPCs isolated from WT and Atp8a1^−/− mouse BM. (f) Quantitative tail DNA content analysis using Comet Assay Analysis Software. (g) Quantitative olive tail moment results for WT and Atp8a1^−/− HSPCs. (h) Immunofluorescent H2A (red) and DAPI (blue) staining for HSPCs isolated from x-ray irradiated WT and Atp8a1^−/− mice. Pictures were taken under the same parameters. (i) Quantitative γH2A fluorescence intensity analysis.(j) Immunofluorescent staining for P53 (red) and DAPI for nuclei (blue) in HSPCs isolated from x-ray irradiated WT and Atp8a1^−/− mice. (k) Quantitative P53 fluorescence intensity analysis. (l) Comet assay for HSPCs isolated from X-ray irradiated WT and Atp8a1^−/− mice. (m) Quantitative comet tail DNA content analysis. (n) Olive tail moment calculation for WT and Atp8a1.^−/− HSPCs. Values are represented using the mean ± SEM, and the significance was evaluated using a t-test. ***, P ≤ 0.001; ****, P ≤ 0.0001

The bioinformatics analysis showed that genes responding to radiation were enriched for Atp8a1 deletion. Hence, Atp8a1^−/− and WT mice were irradiated with low-dose X-rays to determine whether they had a different capacity for acute DNA damage repair. Immunofluorescence γH2A staining showed that the mean intensity in Atp8a1^−/− HSPCs was significantly weaker than in the WT group (Fig. 6h and i), and P53 was found to be significantly increased (Fig. 6j and k). In addition, the DNA damage burden was lower in Atp8a1^−/− HSPCs (Fig. 6l-n), indicating that Atp8a1 deletion prevents murine HSPC DNA damage.

Atp8a1 deletion impairs PTEN function

According to the GSEA analysis results, genes involved in “Phosphatidylinositol 3 Kinase complex” and “PI3K-AKT-mTOR Signaling” were enriched and increased in Atp8a1^−/− HSPCs (Fig. 5b). Therefore, AKT activity was examined, the phosphorylation at T450 and Ser473 was significantly increased in Atp8a1^−/− HSC-like cells when compared to WT cells (Fig. 7a). mTOR protein level also increased (Fig. 7b). To investigate whether PI3K-AKT-mTOR signaling activity reduced PTEN function, the PTEN protein level was examined and shown to significantly decrease in Atp8a1^−/− HSC-like cells (Fig. 7b).

Fig. 7.

Atp8a1 deletion impairs PTEN function and results in PI3K-AKT-mTOR signaling activation (a) Akt phosphorylation levels at sites T450 and Ser473 were detected in Atp8a1^−/− and WT HSC-like cells. (b) The mTOR and PTEN protein levels in WT and Atp8a1^−/− HSC-like cells. (c) The YAP1 protein and p-YAP1 levels in WT and Atp8a1^−/− HSC-like cells. (d-e) Immunofluorescence staining for YAP1 and p-YAP1 in WT and Atp8a1^−/− HSC-like cells. The MFI in the cells was analyzed, with each dot indicating an individual cell. At least 30 cell signals were collected for each sample. Values are presented using the mean ± SEM, and the significance was evaluated with a t-test and represented with **** (P ≤ 0.0001). (f) Cnn1 transcription was measured using RT-PCR and presented using the mean ± SD from three independent experiments. (g) The major signaling pathway changes in Atp8a1^−/− HSPCs. Plasma membrane structures in eukaryotes: PC: phosphatidylcholine; PI: phosphatidylinositol; PS: phosphatidylserine; PIP3: phosphatidylinositol-3, 4, 5 -triphosphate; PIP2; phosphorylating phosphatidylinositol-4, 5-bisphosphate; SM: sphingomyelin; Gly: glycolipid; CHO: cholesterol

Previous studies have reported that the PI3K/AKT pathway regulates YAP1 activation [34] and that YAP1 is involved in HSC homeostasis [35]. YAP1 was lower in Atp8a1^−/− HSC-like cells (Fig. 7c-d), with more phosphorylated YAP1 being detained in the cytoplasm than WT HSC-like cells (Fig. 7c and e). YAP1 transcriptional activity was also detected by examining the Ccn1 mRNA expression, the direct target of YAP1. We found that Ccn1 expression was significantly decreased (Fig. 7f), suggesting that Atp8a1 deficiency causes a transcriptional YAP1 inhibition. In summary, Atp8a1 deletion resulted in PI3K-AKT-mTOR signaling activity, impairing PTEN function. AKT also induces the accumulation of ROS, leading to JNK activation and DNA damage. JNK activity increases the transcriptional activity of the AP-1 complex. AKT activation increases YAP1 phosphorylation and reduces its transcriptional activity (Fig. 7g). All these altered pathways and cell biological processes promote HSPC proliferation and enhance DNA repair ability.

Discussion

The P4-ATPase flippase complex is composed of α-and β-subunits. In humans, the β-subunits include three members: TMEM30A, TMEM30B, and TMEM30C. Most P4-ATPase α-subunits can interact with TMEM30a, such as ATP8A1, ATP8B1, ATP11A, ATP11B, and ATP11C [36]. Our previous studies have demonstrated that Tmem30a deletion results in dramatic hematopoiesis impairments, including murine HSCs [37]. Tmem30a deletion causes HSCs to die rapidly in a short period [37]. But the role of Atp8a1 on hemogenesis and HSCs has not been studied yet. Here we show that Atp8a1 deletion results in HSC in homeostasis disruption. Atp8a1 deletion promoted the HSC pool, which was reflected by increased BM-associated HSC percentage and absolute number (Fig. 2b) and the number of lineage-committed blood cells in PB, spleen, and BM (Fig. 1a-i). Atp8a1 deletion increased HSC capacity for self-renewal and blood-lineage replenishment under a competitive repopulation condition (Fig. 3b and c). The biological effects of Atp8a1 deletion returned BCR/ABL1 transformed LSCs in a CML mouse model to a state similar to normal HSCs (Fig. 4). Based on our findings, Atp8a1 deletion appears to increase HSC numbers due to high HSC self-renewal and division capacity. Cells undergoing proliferation are sensitive to 5-FU-induced cell death. Therefore, a 5-FU administration assay was used to validate whether Atp8a1 deletion promoted more HSPCs within proliferative cell cycle stages. The repeated administration of 5-FU eliminated more cells in Atp8a1^−/− mice than in WT mice (Fig. 3d-h). Overall, the defects caused by Atp8a1 deletion are more moderate than those caused by Tmem30a deletion [37]. This might be due to redundancy in α-subunit function among ATP8A1 members. Thus, the limited number of β-subunits members could prevent Tmem30a loss from being similarly compensated.

Atp8a1 deletion disrupts the asymmetric distribution of PS and results in less PS retention within the inner leaflet. PTEN must anchor the membrane’s inner leaflet via PS to execute its dephosphorylation function [24]. Therefore, Atp8a1 deletion impairs PTEN function to reduce PS asymmetric distribution, resulting in PI3K/AKT aberrant activation. Aberrant PI3K/AKT drives many pathways to increase ROS levels through directly modulating mitochondrial bioenergetics and activating NADPH oxidases (NOXs) or indirectly producing ROS as a metabolic by-product [38, 39]. In addition, ROS generation promotes the activation of JNK [40]. Aberrant PI3K/AKT activity to modulate mitochondrial bioenergetics is reflected in the gene expression profiling experiment. In the GSEA analysis, genes involved in the mitochondrial proton transporting ATP synthase complex are enriched, and their expression was increased in HSPCs with Atp8a1 deletion (Fig. 5b). This also may indicate that ROS generation is aberrantly activated following Atp8a1 deletion in HSPCs. ROS promote DNA damage response, which could induce a positive feedback loop with P53 [40]. Moreover, the GO (CC) analysis results also indicate that genes involved in telomerase holoenzyme complex are enriched, and their expression is increased in HSPCs with Atp8a1 deletion (Fig. 5a). Telomerases play essential roles in genomic stability and cellular replicative potential [41]. This might explain the reason Atp8a1 deletion prevents HSPC DNA damage.

Previous studies showed that knocking down Atp8a1 with siRNA inhibited YAP1 transcriptional activity in COS-1 cells [42]. YAP1 activity restrains HSC proliferation, and that Yap1/Taz deletion resulted in the loss of HSC quiescence and increased HSPC proliferation. However, Yap1 deletion alone only moderates effects on HSC proliferation [35]. Additionally, the PI3K/AKT pathway is known to regulate YAP1 activation [34]. This indicates that the increase of HSC proliferation following Atp8a1 deletion might partially contributed to YAP1 transcriptional inhibition. Moreover, the biological functions of YAP1 is complex and varying. Atp8a1 deletion represses its transcriptional activity but increases its phosphorylation level in the cytoplasm (Fig. 7c and e). YAP1 interacts with more than three hundred proteins (as observed in the IntAct Database), including LATS1, CTNNB1, JUN, HIF1A, JUN, and GSK3B, all of which play essential roles in HSC homeostasis [43–45]. A high level of phosphorylated YAP1 in the cytoplasm inevitably alters its interactor function. Therefore, the final outcome from YAP1 modulation following Atp8a1 deletion could be more complex than Yap1 deletion alone.

Due to the HSC scarcity, longer processes and complex multi-step operations are acquired for sorting HSPCs. During the procedure, unexpected factors can be potentially introduced, leading to HSPC intracellular signaling modifications. Thus, the accuracy and reproducibility of the results are seriously challenged. When studying signaling, we chose HSC-like cells transformed with LHX2, which previous studies have demonstrated are HSC-like and have many of the most important HSC properties. These include, self-renewal and the capacity to differentiate into all lineage mature functional blood cells [46, 47]. Furthermore, HSC-like cells can be produced in bulk under controllable culture conditions. The system is remarkably convenient for some assays, especially for the assay acquiring a large amount of protein. Notably, the results of intracellular signal transduction obtained from HSC-like cells are consistent with the results of transcriptomic data analysis. Therefore, HSC-like cells are reliable resources for HSC signaling studies.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

L.Z, C.P and Y.H conceived research ideas, designed experiments, analyzed data, and wrote the manuscript. L.Z., W.T. and C.L. performed experiments; Y.F and B.W performed the bioinformatics; H. Q, Q.Q, N.L; W.H. Y.S and Z. Y helped complete the experiments. L.Z, X.Z and K.S reviewed and edited the manuscript, discussed the results and commented on the manuscript. All authors have read and approved the article.

Funding

This work was supported by grants from the National Natural Science Foundation of China (82170114 to Y. Hu) and National Clinical Research Center for Geriatrics, West China Hospital, Sichuan University (Z20201008 to Y. Hu), Guizhou Provincial Science & Technology Support Program (NO [2020]4Y061 to Y. Hu). National Natural Science Foundation of China (81802468 to LZh). Chengdu Science and Technology innovation project (2021-YF05-00800-SN to LZh).

Data availability

The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request. In addition, the raw sequence data reported in this paper have been deposited in the Genome Sequence Archive[28] in National Genomics Data Center[29], China National Center for Bioinformation / Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA009276), which is publicly accessible at https://ngdc.cncb.ac.cn/gsa.

This paper has not been previously published and is not under consideration by another journal. All authors have approved and agreed to submit the manuscript to this journal.

Declarations

Ethical approval

All mouse studies were approved by the Institutional Animal Care and Use Committee (IACUC) at the Sichuan University. All animals were monitored for abnormal behavior to minimize pain and suffering. Animals were euthanized when excessive deterioration of health was noted.

Competing interests

The authors declare that the research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.