Hemgn Protects Hematopoietic Stem and Progenitor Cells Against Transplantation Stress Through Negatively Regulating IFN‐γ Signaling

Abstract Hematopoietic stem and progenitor cells (HSPCs) possess the remarkable ability to regenerate the whole blood system in response to ablated stress demands. Delineating the mechanisms that maintain HSPCs during regenerative stresses is increasingly important. Here, it is shown that Hemgn is significantly induced by hematopoietic stresses including irradiation and bone marrow transplantation (BMT). Hemgn deficiency does not disturb steady‐state hematopoiesis in young mice. Hemgn −/‐ HSPCs display defective engraftment activity during BMT with reduced homing and survival and increased apoptosis. Transcriptome profiling analysis reveals that upregulated genes in transplanted Hemgn −/− HSPCs are enriched for gene sets related to interferon gamma (IFN‐γ) signaling. Hemgn −/− HSPCs show enhanced responses to IFN‐γ treatment and increased aging over time. Blocking IFN‐γ signaling in irradiated recipients either pharmacologically or genetically rescues Hemgn −/‐ HSPCs engraftment defect. Mechanistical studies reveal that Hemgn deficiency sustain nuclear Stat1 tyrosine phosphorylation via suppressing T‐cell protein tyrosine phosphatase TC45 activity. Spermidine, a selective activator of TC45, rescues exacerbated phenotype of HSPCs in IFN‐γ‐treated Hemgn −/− mice. Collectively, these results identify that Hemgn is a critical regulator for successful engraftment and reconstitution of HSPCs in mice through negatively regulating IFN‐γ signaling. Targeted Hemgn may be used to improve conditioning regimens and engraftment during HSPCs transplantation.


Introduction
Hematopoietic stem and progenitor cells (HSPCs) possess the remarkable ability to replenish the blood system and maintain hematopoietic homeostasis in response to either physiological or ablated stress demands, such as irradiation, chemotherapy, or hematological diseases. [1,2] This ability is exploited routinely in the clinic via HSPCs transplantation (HSPCT). [3] Successful HSPCT depends on the engraftment process, in which donor cells are lodged in the bone marrow (BM) medullary cavity, and subsequent retention and proliferation in the BM space leads to reconstitution of the hematopoietic system. [4] Ionizing radiation is used to condition patients receiving HSPCT, which induces a cytokine release syndrome. [5] In the early stages of engraftment, donor cells operate within a skewed cytokines environment, need to withstand the transplantation stress which includes multiple apoptotic signals and respond to proliferative stimuli, exiting their quiescent phase and undergoing a period of self-renewal and differentiation to restore hematopoietic homeostasis. observed in committed myeloid progenitors (CMPs, Lin − Sca-l − c-Kit + CD34 + CD16/32 low ) and megakaryocyte-erythroid progenitors (MEPs, Lin − Sca-l − c-Kit + CD34 − CD16/32 low ) ( Figure 1A). To assess the functional role of Hemgn in HSPCs, we generated a whole-body Hemgn −/− mouse model on a C57/BL/6J (CD45.2 + ) background. The mutant allele was confirmed by sequencing the Hemgn cDNA ( Figure S1, Supporting Information). Western blot analysis using a specific antibody against mouse HEMGN demonstrated that neither intact nor truncated HEMGN proteins were present in Hemgn −/− HSPCs ( Figure 1B). Hemgn −/− mice appeared normal and showed no internal anatomical abnormalities (data not shown). In 2-month-old animals, Hemgn −/− mice display normal cellularity of peripheral blood (PB) ( Figure 1C) and BM ( Figure 1D), as well as the constitution of leukocytes in PB. The numbers of immunophenotypically-defined enriched HSPCs in the BM of Hemgn −/− mice were comparable to those in WT mice ( Figure 1E). No significant difference of the cell cycle status in LT-HSCs and ST-HSCs was observed between Hemgn −/− and WT mice ( Figure 1F; Figure S2, Supporting Information). Functional colony-forming unit (CFU) assays suggested that the numbers of burst forming unit-erythroid (BFU-E), CFUgranulocyte and macrophage (CFU-GM), and CFU-granulocyte, erythrocyte, macrophage, and megakaryocyte (CFU-GEMM) in BM ( Figure 1G) and spleen (SP) ( Figure S3, Supporting Information) were similar between Hemgn −/− and WT mice. Serial replating assays suggested that Hemgn −/− and WT HSCs displayed similar self-renewal capacity ( Figure S4, Supporting Information), indicating that the number of functional HSCs in Hemgn −/− mice was comparable to that in WT mice. Overall, these findings suggest that Hemgn is dispensable for steady-state hematopoiesis under homeostatic conditions.

Hemgn −/− HSPCs Display Severely Defective Engraftment Activity
HEMGN protein was significantly induced in BM HSPCs 1h following total body irradiation (TBI) exposure (Figure 2A). Hemgn mRNA was increased in donor HSPCs from the recipient BM at 6 h post-BMT ( Figure 2B), indicating that HEMGN may contribute to transplantation stress. Since various pro-inflammatory cytokines could be induced such as tumor necrosis factor (TNF ), interleukin 6 (IL-6), interferon (IFN-), and IFN-under irradiation and BMT conditions, we investigated the expression levels of HEMGN under various cytokines. We found that HEMGN could be induced dramatically by IFN-treatment while other cytokines did not affect HEMGN expression significantly ( Figure S5, Supporting Information). To assess whether Hemgn affects HSPCs regeneration capacity in vivo, a competitive BMT assay was carried out by transplanting WT or Hemgn −/− mice BM (CD45.2 + ) mixed with competitor cells (CD45.1 + CD45.2 + ) at the indicated ratio into lethally irradiated recipient mice (CD45.1 + ) ( Figure 2C). Analysis of the donor chimerism in PB showed that WT BM cells successfully reconstituted recipient's hematopoiesis at 16 weeks post-BMT, however, none of the recipients transplanted with Hemgn −/− BM cells were reconstituted even when transplanted at a 5-fold higher dose. When the ratio increased to 10:1, only 2 of 11 recipients transplanted with Hemgn −/− donor cells were reconstituted but with a significant Figure 1. Hemgn is dispensable for steady-state hematopoiesis in young mice. A) Hemgn mRNA expression levels in purified murine HSPCs (n = 6 mice). The left panel represents the gating strategy for HSPCs and committed progenitors. Data shown are representative of three independent experiments with a total of six mice per group. Each mouse was performed in triplicate. Error bars indicate SEM. B) Validation of Hemgn knockout mice by western blotting analysis of HEMGN protein expression in total BM mononuclear cells, BM Lin − cells and spleen from WT and Hemgn −/− (KO) mice. GAPDH was used as internal control. C) Hemgn −/− mice displayed normal cellularity of PB and D) BM (n = 6 mice per group). Data are the pool of two independent experiments. Error bars indicate SD. E) The total numbers of the HSPCs and committed progenitors in the BM of Hemgn −/− mice were comparable to those in WT mice (n = 8 mice per group). Data are the pool of two independent experiments. Error bars indicate SD. F) Cell cycle analysis of LT-HSCs and ST-HSCs in Hemgn −/− and WT BM using DAPI and Ki-67 staining (n = 5 mice per group). Data are the pool of two independent experiments. Error bars indicate SD. G) CFU assays of Hemgn −/− and WT BM cells (n = 5 mice per group). Data shown are representative of two independent experiments with a total of five mice per group. Each mouse was performed in triplicate. Error bars indicate SEM. For all graphs, *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 2.
Hemgn −/− HSPCs display severely defective engraftment activity. A) Mice were exposed to 4.5Gy TBI for 1h and LSK cells were isolated for measuring HEMGN protein expression. n = 2 mice per group. B) Hemgn mRNA expression levels in LSK cells before and after transplantation was detected by quantitative RT-PCR (n = 3 mice in Before BMT group, n = 6 mice in After BMT group). Each mouse was performed in triplicate. Error bars indicate SEM. C) Whole BM cells from WT and Hemgn −/− mice (CD45.2 + ) were mixed with competitor BM cells (CD45.1 + CD45.2 + , 1 × 10 6 ) at the indicated ratio and injected into lethally irradiated recipients (CD45.1 + ). Donor chimerism in PB was analyzed at 16 weeks after transplantation (n = 10 mice per group). The left panel shows representative flow cytometric plots of mice transplanted with WT or Hemgn −/− BM cells with competitor cells www.advancedsciencenews.com www.advancedscience.com reduction in engraftment (below 5%). Competitive BMT assay using LSK cells showed that Hemgn −/− LSK cells could not engraft and reconstitute hematopoiesis from 4 weeks to 16 weeks post-BMT ( Figure 2D). Noncompetitive BMT assays suggested that WT BM cells (2 × 10 5 ) were able to efficiently engraft 4 weeks post-BMT, but Hemgn −/− BM cells failed to engraft even at cell doses ten times higher than the controls ( Figure 2E). 90% of mice transplanted with 5 × 10 5 WT BM cells survived within 40 days after BMT, while only 10% of mice transplanted with Hemgn −/− BM cells survived ( Figure 2F). These data suggest a significant role of Hemgn in HSPCs in the engraftment process. When donor cells number was increased to 2 × 10 7 , 100% of the recipients receiving Hemgn −/− BM cells were reconstituted with normal lineage distribution at 16 weeks after BMT ( Figure 2E,G), suggesting that Hemgn deficiency may not lead to alteration of HSPCs differentiation ability. Transplantation of WT BM cells into Hemgn −/− recipients demonstrated that the engraftment and reconstitution were normal, indicating that Hemgn acts in the cells autonomously to maintain HSPCs engraftment activity (Figure 2H).

Hemgn Is Required for HSPCs Homing, Survival, and Expansion in Recipient Mice
Defected HSPCs engraftment could result from reduced HSPCs homing activity, impaired self-renewal, and decreased proliferation in recipient. We first investigated whether the homing potential of HSPCs was affected by Hemgn deficiency. Equal numbers of donor cells and competitor cells were transplanted into lethally irradiated recipient. Competitor BM cells served as internal control in this assay. The ratios of the donor Lin − Sca-1 + cells versus competitor cells in both groups were similar before BM transplantation, however, the ratio of the Hemgn −/− Lin − Sca-1 + cells versus competitor cells in recipients BM was significant decreased 18 h after transplantation ( Figure 3A). Congruently, the absolute number of Hemgn −/− donor-derived Lin − Sca-1 + cells in the recipient mice marrow was also significantly lower compared to donor cells derived from WT mice ( Figure 3B), indicating that Hemgn deficiency impaired HSPCs homing ability. Similarly, the ratio of Hemgn −/-HSPCs versus competitor cells in recipients SP was significantly lower at 18 h after transplantation ( Figure 3C). In a congenic model, where cell proliferation dyes-labeled HSPCs were transplanted into lethally irradiated hosts to assess HSPCs homing, a significantly lower number of Hemgn −/− HSPCs homed to BM compared with WT control cells was observed ( Figure 3D). These results indicate that the homing capacity of HSPCs is impaired in the absence of Hemgn. To confirm, we injected WT or Hemgn −/− mice BM mixed with competitor cells at the indicated ratio into the femoral BM cavity of lethally irradiated mice by intrafemoral transplantation (Figure 3E), which is a strategy to avoid inefficient homing associated with intravenous (IV) HSPCT. [20] Analysis of the chimerism in PB for donor cells revealed that the reconstitution ability of donor cells from Hemgn −/− mice was still markedly lower than that of matched WT controls, but intrafemoral transplantation of Hemgn −/− cells significantly improved the engraftment compared with the IV BMT ( Figure 3E; Figure S6, Supporting information). When the ratio increased to 5:1, recipients receiv-ing Hemgn −/− BM cells were equivalently reconstituted with normal lineage distribution at 16 weeks post-BMT ( Figure S6, Supporting information), supporting the conclusion that Hemgn deficiency reduces HSPCs homing to the BM.
We further investigated donor cells survival and expansion in recipient mice. The results revealed that number of Hemgn −/− donor cells that homed to BM was significantly reduced 45 h compared with 18 h and 30 h post-BMT; at 14 days, Hemgn −/− donor cells in recipient's BM were almost undetected ( Figure 3F). In contrast, the number of WT donor cells in recipient's BM was only slightly reduced at 45 h post-BMT and then markedly increased at 14 days. In line with these observations, we detected increased apoptosis in transplanted Hemgn −/− donors in BM (Figure 3G) at early stages after transplantation. Using non-irradiated recipients, we showed that the homing efficiency ( Figure S7A, Supporting information) and apoptosis ( Figure S7B, Supporting information) of Hemgn −/donor cells were comparable to WT donor cells, suggesting that the engraftment defect in Hemgn −/− HSPCs is related to irradiation-induced microenvironments alteration.
During BMT, irradiation increases the production of reactive oxygen species (ROS), thereby promoting DNA damage in donor HSPCs. [5] We found that the ROS levels ( Figure 3H) and DNA damage ( Figure 3I,J) were significantly elevated in Hemgn −/− donor HSPCs compared with those in WT HSPCs, suggesting that Hemgn deficiency caused more serious oxidative stress responses in HSPCs after transplantation.

Hemgn Deficiency Enhances IFN-Signaling Pathway in Transplanted HSPCs and Increases HSPCs Responses to IFN-Administration In Vitro and In Vivo
To identify possible mechanisms of the reduced engraftment capability of Hemgn −/− HSPCs, we performed genome-wide expression analysis using RNA-seq of purified Hemgn −/− and WT HSPCs from recipient mice at 6h post-BMT. This revealed 760 differentially expressed genes (DEGs) (325 upregulated and 435 downregulated, p < 0.05, 1.5-fold cut off) (Table S1, Supporting Information). Reactome pathway analysis of DEGs revealed that upregulated genes in Hemgn −/− HSPCs were significantly associated with genes sets related to IFN-signaling pathway and chemokine receptors bind chemokines, while the downregulated genes were associated with gene sets related to cell cycle and Rho GTPases signaling ( Figure 4A). We validated these genes expression by quantitative PCR analysis ( Figure 4B). GSEA analysis revealed a transcriptional signature associated with the IFN--treated HSPCs (GSE81559) to be the most enriched gene set among upregulated genes in Hemgn −/− HSPCs ( Figure 4C),  indicating that Hemgn deficiency enhanced IFN-signaling in HSPCs after transplantation.
We further examined the response of Hemgn −/− HSPCs to IFN-in vivo. In line with previous report, [23] a single injection of IFN-caused a dramatic increase in the percentage of proliferating HSPCs ( Figure 5H) and absolute number of HSPCs ( Figure 5I) in WT BM, whereas the changes were significantly augmented in Hemgn −/mice. BM cells from IFN--treated Hemgn −/− mice formed less hematopoietic colonies than those from IFN--treated WT mice ( Figure 5J). A similar alternation of HSPCs was also observed in a mouse model of Listeria monocytogenes acute infection ( Figure 5K,L; Figure S10, Supporting information), since IFN-is necessary to influence the behavior of HSPCs in this model. [24] Previous studies suggest that the sustained IFN-signaling promotes HSCs depletion in mice, [25,26] we postulated that chronically dysregulated IFN-signaling in Hemgn −/− mice might lead to diminished self-renewal and, ultimately, depletion of the HSCs compartment. We examined the BM of young (2-month-old) and aged WT and Hemgn −/− mice (28-month-old). BM mononuclear cells (MNCs) in aged Hemgn −/− mice was comparable to those in aged WT mice (Figure 6A), but the number of phenotypic HSCs in aged Hemgn −/− mice was lower compared with those of agematched WT controls ( Figure 6B,C; Figure S11, Supporting information). Aged Hemgn −/− BM cells gave rise to about 40% fewer colonies than WT controls ( Figure 6D). Moreover, a significant increase in accumulated DNA damage in aged HSCs was noted in Hemgn −/− mice ( Figure 6E). Accordingly, the ROS levels in aged Hemgn −/− HSCs were higher than those of age-matched WT con-trols ( Figure 6F). Consistently, the tyrosine-phosphorylated Stat1 (p-Stat1(Y701)) was significantly sustained in aged Hemgn −/− HSPCs with IFN-treatment ( Figure 6G). In line with the increased HSCs aging, 28-month-old Hemgn −/mice showed a significant elevation in neutrophils frequency and decrease in lymphocytes frequency in PB ( Figure 6H) as well as an increase of CD11b + cells frequency in BM ( Figure 6I) compared with the age-matched WT controls. No death was observed in Hemgn −/mice over time. These findings suggest that Hemgn deficiency can lead to depletion of the HSCs compartment and increased HSCs aging over time.

Engraftment Defect of Hemgn −/− HSPCs Is Rescued by Blocking IFN-Signaling Pathway
Based on the results obtained above, we asked if the abnormal activation of IFN-signaling might be the major cause of engraftment defect of Hemgn −/− HSPCs. We   10 6 ). E) The donor chimerism and trilineage differentiation in PB were analyzed at 16 weeks after transplantation. The left panel showed representative flow cytometric plots of trilineage differentiation. F) Noncompetitive BMT assay of BM cells from the indicated mice models at 4 weeks post-BMT (n = 5 mice per group). For all graphs, data are presented as mean ± SD. ***p < 0.001.

Hemgn Regulates IFN-Signaling Pathway through Modulating T-Cell Protein Tyrosine Phosphatase TC45 Activity
IFN-triggers JAK-STAT signaling pathway through binds to its receptors, [27] leading to phosphorylation of JAK1, followed by tyrosine phosphorylation of Stat1. Our result showed that Hemgn deficiency had no effect on IFN-receptors expression ( Figure   S14, Supporting information) and IFN-induced JAK1 phosphorylation in HSPCs (Figure 8A), indicating that Hemgn does not affect the upstream events of IFN-pathway. Given that HEMGN is mainly located in nucleus, and Hemgn deficiency prolonged Stat1 tyrosine phosphorylation in IFN--treated HSPCs, we hypothesized that the molecular target of Hemgn is closely linked to modulate Stat1 phosphorylation level in nucleus. Indeed, we www.advancedsciencenews.com www.advancedscience.com Figure 8. Hemgn regulates IFN-signaling pathway through modulating TC45 activity. A) LSK cells were sorted from WT or Hemgn −/− mice BM and treated with IFN-(100 ng mL −1 ) for the indicated time. Total cell lysates were prepared and the phosphorylation of Jak1 was detected. B) Nuclear and cytoplasmic extracts were isolated from LSK cells treated with IFN-(100 ng mL −1 ) for the indicated time and the tyrosine phosphorylation of Stat1 was detected. C) HepG2 cells were transiently transfected with HEMGN for 36 h and pretreated with PTP inhibitor XIX (2 μm) for 2 h. Then the cells were treated with IFN-(50 ng mL −1 ) for the indicated time and the tyrosine phosphorylation of Stat1 and TC45 expression were detected. D,E) WT and Hemgn −/− mice were treated with spermidine (SPM) for 3 days followed by i.p injection with 10 μg IFN-for 24 h. F,G) The proliferating HSPCs and absolute number of HSPCs were analyzed (n = 3 or 4 mice per group). For all graphs, data are presented as mean ± SD. *p < 0.05, **p < 0.01. H) Model depicting the critical role of Hemgn in donor HSPCs engraftment during BMT. Hemgn is induced by irradiation or BMT stress which in turn regulates the key aspects of the transplanted HSPCs through promoting TC45-medicated Stat1 dephosphorylation, which deactivates IFN-signaling, resulting in protection of HSPCs from transplantation stress. demonstrated that the p-Stat1(Y701) was sustained at higher levels from 30 min to 1 h after IFN-treatment in the nuclear fraction of Hemgn −/− HSPCs compared with that in IFN-treated WT HSPCs, although no obvious difference was observed at 15 min ( Figure 8B). In contrast, the levels of p-Stat1(Y701) in cytoplasmic fraction of HSPCs was not affected by Hemgn deficiency (Figure 8B).
Previous studies demonstrate that dephosphorylation of Stat1 in nucleus is a key mechanism for inactivation of Stat1 activity. [28] TC45 is the major nuclear Stat1 protein tyrosine phosphatase (PTPase), [29] we therefore speculated that HEMGN may be linked to modulate TC45-medicated Stat1 dephosphorylation. We found that overexpression of HEMGN led to an diminished p-Stat1(Y701) levels in HepG2 cells treated with IFN-, how-ever, this effect was rescued by administration of PTPase inhibitor XIX, a selective inhibitor of TC45, [30] suggesting that HEMGN might regulate Stat1 dephosphorylation through TC45 ( Figure 8C). Notably, the levels of TC45 expression were not affected by Hemgn deficiency or overexpression ( Figure 8A,C).
Furthermore, spermidine (SPM), a selective agonist of TC45, [31] was employed to investigate the effect of Hemgn on IFNsensitivity of HSPCs in vivo. Administration of spermidine did not affect the behavior of HSPCs in both genotype mice in the absence of IFN-( Figure 8D-G). Upon IFN-treatment, spermidine administration significantly rescued the exacerbated response to IFN-in Hemgn −/− HSPCs ( Figure 8D-G). These results indicate that Hemgn deletion accelerates HSPCs response to IFN-in Hemgn −/− mice via suppressing TC45 activity.

Experimental Section
Mice: Hemgn-deficient (Hemgn −/− ) mice were successfully constructed using Zinc Finger Nuclease technology (ZFNs) in a C57BL/6 (CD45.2) background. Interferon-gamma receptor 1-deficient (Ifngr1 −/− ) mice in C57BL/6 (CD45.2) background were purchased from Jackson Laboratory (Bar Harbor, ME). Hemgn −/− and Ifngr1 −/− strains were intercrossed to produce mice that were homozygous for disruptions at the Hemgn and Ifngr1 loci (Hemgn −/-Ifngr1 −/− , DKO). Hemgn −/− and GFP transgenic mice (C57BL/6-TgN(act-EGFP)OsbC14-Y01-FM131) were similarly intercrossed to produce mice that were homozygous for Hemgn mutation that expressed the GFP (Hemgn −/− -GFP). In all experiments, male and female mice between 8 and 10 weeks of age were used in the studies. Genetically modified mice were systematically compared to their sex-, age-, and weight-matched wild type (WT) littermates. All mice were housed in individually ventilated cages under specific pathogen-free conditions at the animal facility of the authors' institute with a 12 h light-dark cycle and allowed free access to food and water. All animal experiments were reviewed and approved by the Animal Ethics Committee of the Beijing Institute of Lifeomics.
Bone Marrow Transplantation Experiments: For competitive transplantation, donor cells were mixed with 1 × 10 6 freshly isolated competitor BM cells (CD45.1 + CD45.2 + ) and then transplanted into recipients previously irradiated with a split dose of 9Gy (4.5Gy+4.5Gy) through tail vein injection. For non-competitive transplantation, donor cells were injected into recipients without competitor cells. Engraftment was measured at the time points indicated and analyzed for PB chimerism and multilineage engraftment. For intrafemoral transplantation, recipients were anesthetized with 1% pelltobarbitalum natricum (Sinopharm Chemical Reagent Co., Ltd.), and the joint between the femur and tibia in one leg was surgically exposed. Donor cells mixed with 2 × 10 5 freshly isolated competitor cells were injected into the femur medullary cavity near the medial epicondyle. The insertion hole was immediately sealed with bone wax, and the skin was closed by wound clips.
Flow Cytometry: Antibodies used for flow cytometry analysis or cell sorting were listed in Supporting Information Table S2, Supporting Information. LSRFortessa (BD) were used for flow cytometric analysis and FACSAria III (BD) for cell sorting. HSPCs were identified as viable LSK (Lin − Sca-1 + c-Kit + ) as previously described. [50] Before staining, red blood cells were eliminated by RBC Lysis Buffer (Biolegend). For HSPCs sorting, Lin − cells were enriched using EasySep Mouse Hematopoietic Progenitor Cell Enrichment Kit (Cat.19756, STEMCELL Technologies). Then the enriched Lin − cells were stained with relative surface markers and sorted for further in vivo and in vitro investigations.
Homing Assay: Homing assay was performed as previous described [51] with some modifications. Briefly, BM cells prepared from Hemgn −/− or WT control mice were mixed with competitor BM cells (CD45.1 + CD45.2 + ), and then were injected into lethally irradiated recipient mice. Eighteen hours after injection, the ratio of Lin − Sca-1 + (LS) cells in injected test versus competitor BM cells (percentage of CD45.2 + LS cells to percentage of CD45.1 + CD45.2 + LS cells) was analyzed in BM cells. In addition, homing assays were also performed with Cell Proliferation Dyes (CPD)-labeled donor cells. CPD staining was performed according to the manufactures' instructions. Single cell suspensions of bone marrow cells to be injected were washed once resuspended in PBS (1.0-2.0 × 10 7 cells mL −1 ). To this suspension, the same volume of a 10 μm solution of Cell Proliferation Dye eFluor 670(CPD,65-0840, eBioscience,) was added to give a final concentration of 5 μm CPD. 10 μm CPD solution was prepared by dilute 5 mm stock solution (dissolved in anhydrous DMSO) in PBS (pre-warmed to room temperature). The cells were then incubated at 37°C in the dark for 10 min. Further labeling of the CPD was stopped by addition of 5 volumes of cold complete media (containing 20% serum in DMEM media) and incubate on ice for 5 min. The suspension was washed three times with complete media and finally resuspended in PBS. The CPD-labelled cells and other BM cells were counted respectively by FACS using 123 count beads (1234-42, eBioscience). These cells were mixed in proportion to get the final concentration of 5 × 10 7 cells mL −1 and injected intravenously with the volume of 200 uL.
RNA-Seq and Analysis: LSK cells from primary recipient mice BM (three biological replicates) at 6 h post-BMT were sorted using BD FACSAria III. Total RNA was isolated using the QIAGEN RNeasy Micro Kit (QIAGEN). Smart-Seq2 method was used to produce cDNA. The cDNA production was checked by Qubit 3.0 Flurometer and Agilent 2100 Bioanalyzer to ensure the expected production with length around 1-2 kbp. Then the cDNA was sheared randomly by ultrasonic waves for Illumina library preparation protocol including DNA fragmentation, end repair, 3′ ends A-tailing, adapter ligation, PCR amplification, and library validation. After library preparation, PerkinElmer LabChip GX Touch and Step OnePlus Real-Time PCR System were introduced for library quality inspection. Qualified libraries were then loaded on Illumina Hiseq platform for PE150 sequencing at the ANNOROAD GENOME Co., Ltd. (Beijing, CN) following the vendor's recommended protocol. Only transcripts with FKPM greater than 1 at least one samples were selected for differential testing. Transcripts with p < 0.05 and fold change ±1.5 were considered differentially expressed. The enrichment analysis of pathways was performed using reactome pathway database (https://reactome.org). GSEA analysis with GSE81559 (IFN--dependent genes in HSPCs) was carried out using GSEA 4.0.0 and MsigDB 7.0. [52] The accession number for the RNA-seq data reported in this paper is GEO: GSE146949.
Western Blotting: To obtain whole-cell extracts, cells were washed with PBS and incubated for 20 min in cold lysis buffer containing freshly added protease inhibitors (Beyotime). Protein concentrations were determined using BCA protein assay (Beyotime). Total protein (10 mg) was separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore). Membranes were probed with specific primary antibodies, antibody-protein complex detected by horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence (ECL) exposed (Pierce).
Statistical Analysis: Unless otherwise indicated, data in all figures are expressed as the mean ± SD. Statistical details relevant to each experiment including n and number of times each experiment was repeated are listed in the figure legends. For comparisons between 2 groups, 2tailed Student's-t tests were performed. For multiple group (more than 2 groups) comparisons, One-way ANOVA tests were performed. In the mouse survival experiments, survival rates were plotted in Kaplan-Meyer survival curves and analyzed with the log-rank non-parametric test. Statistical significance is indicated in figures as follows: *p < 0.05; **p < 0.01; ***p < 0.001. Data values were analyzed and graphed using GraphPad Prism software version 7.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.