Rheb1-Deficient Neutrophils Promote Hematopoietic Stem/Progenitor Cell Proliferation via Mesenchymal Stem Cells

Myeloid cells have been identified as hematopoietic stem cell (HSC)-regulating cells. However, the mechanisms by which myeloid cells regulate the function of HSCs are not fully defined. Our previous study indicated that the HSCs are over-expanded in Vav1-Cre;Rheb1fl/fl mice. Here, using in vivo and in vitro models, we found that Rheb1-deficient neutrophils remodeled the bone marrow environment and induced expansion of HSCs in vivo. Further studies showed that loss of Rheb1 impaired neutrophils’ ability to secrete IL-6, led mesenchymal stem cells (MSCs) to produce more SCF, and promote HSC proliferation. We further found that IL-6 suppressed SCF mRNA expression in human MSCs. Interesting, the high level of IL-6 was also related with poor survival of chronic myeloid leukemia (CML) patients, and higher expression of IL-6 in CML cells is associated with the lower expression of SCF in MSCs in patients. Our studies suggested that blocking IL-6 signaling pathway might stimulate MSCs to secrete more SCF, and to support hematopoietic stem/progenitor cells proliferation.


INTRODUCTION
Hematopoietic stem cells (HSCs) play an important role in the regulation of hematopoietic homeostasis. They can self-renew and differentiate into all cell types of the hematopoietic system according to proliferative or stress cues throughout life (Feng et al., 2008;Rathinam and Flavell, 2008). This process requires fine regulation by intra-and extracellular signaling in the bone marrow microenvironment. HSCs primarily reside in the bone marrow cavity (or niche), where they interact with a variety of cell types, including perivascular cells, osteoblast cells (OBCs) and mesenchymal stem cells (MSCs). HSCs directly adhere to mesenchymal cells and possibly to osteoblasts in the bone marrow stroma that express important regulatory molecules including stem cell factor (SCF) and C-X-C motif chemokine 12 (CXCL-12) (Frenette et al., 2013;Greenbaum et al., 2013). In addition, differentiated hematopoietic cells have been proposed to regulate HSCs. For example, macrophages have been shown to indirectly promote the retention of HSCs by regulating mesenchymal cells and osteoblasts (Chow et al., 2011). Megakaryocytes (MK) are physically associated with HSCs in the bone marrow. MK ablation leads to the activation of quiescent HSCs and increases proliferation of HSCs (Meng et al., 2014). Neutrophils can produce ROS to stimulate the proliferation of myeloid progenitors (Kwak et al., 2015). These studies suggested that the hematopoietic cells and niche cells interact with HSCs, regulate HSCs division and proliferation, and maintain hematopoietic system balance in vivo.
It has been known that the change of normal hematopoietic niche to a hostile HSC growth environment is the tipping point for the development of myelodysplastic syndromes (MDS) and leukemia (Raaijmakers, 2012). For example, study has shown that leukemia cells could secrete more proinflammatory cytokines and establish a feedback loop that drive over-expansion of immature myeloid cells and chronic myeloid leukemia (CML) development (Reynaud et al., 2011). Additionally, leukemia cells could also stimulate MSCs to differentiate into OBCs to support LSC proliferation (Schepers et al., 2013), while MSCs may accelerate abnormal HSCs over-proliferation through secreting more SDF-1 in myeloproliferative neoplasm (MPN) patients (Arranz et al., 2014). mTOR is a serine/threonine protein kinase that responds to multiple signals and maintains homeostasis. Increased or decreased mTORC1 activity can alter HSC function and cause hematological disorders (Wang et al., 2016a). Rheb1 acts as a key activator of mTOR to play vital roles in maintaining proper hematopoiesis and myeloid differentiation (Aspuria and Tamanoi, 2004). Previously, we reported that Rheb1-deficient mice showed increased phenotypic HSCs, immature neutrophils in bone marrow, and splenomegaly, which are reminiscent of the hematopoiesis seen in MPNs (Wang et al., 2018). Meanwhile, Rheb1 deficiency inhibits the development of macrophages and neutrophils, thus impairing their phagocytic ability (Wang et al., 2016b). Rheb1 cooperated with MLL-AF9 to promote acute myeloid leukemia progression, and deletion of Rheb1 in the MLL-AF9 acute myeloid leukemia mouse model prolonged the survival of mice by inhibiting the mTORC1 signaling pathway (Gao et al., 2016).
Interestingly, our study found that the absolute number of HSCs were increased in Vav1-Cre;Rheb1 f l/fl mice, while HSCs were not over expanded under in vitro culture conditions (Wang et al., 2018). This suggested that the expansion of these HSPCs in the bone marrow of Vav1-Cre;Rheb1 f l/fl mice might be caused by additional extrinsic factors rather than intrinsic factors. Here in this study, we found that loss of Rheb1 impaired neutrophils ability to secrete IL-6, and this in turn stimulated MSCs to produce more SCF, leading to HSPCs over proliferation.

Mice and Genotyping
Vav1-Cre mice (JAX stock #008610, background CD45.2) were purchased from Jackson Lab. Rheb1 f l/fl mice (background CD45.2) were kindly provided by Dr. Bo Xiao (Zou et al., 2011). The mice were crossed with B6.SJL mice (background CD45.1) and were backcrossed, respectively, to generate CD45.1 congenital Vav1-Cre mice and Rheb1 f l/fl mice (Supplementary Figure 1A). Then the Vav1-Cre mice and Rheb1 f l/fl mice were crossed to generated mice in which Rheb1 was specifically deleted in the hematopoietic system (Vav1-Cre;Rheb1 f l/fl or Rheb1 / , CD45.1). C57BL/6 mice (6-8 weeks old, CD45.2) were the recipient mice in transplantation experiments. Mice were maintained at the specific pathogenfree (SPF) animal facility of the State Key Laboratory of Experimental Hematology (SKLEH). All animal surgeries were approved by the Institutional Animal Care and Use Committee (IACUC), Institute of Hematology and Blood Diseases Hospital, CAMS/PUMC. All efforts were made to minimize mouse suffering.

Isolation of Neutrophils From Bone Marrow
Percoll-based (GE Healthcare, 17144003, Little Chalfont, United Kingdom) density gradient centrifugation was used for the purification of neutrophils from bone marrow cells (Swamydas et al., 2015). A "100% Percoll" solution is generated by adding 5 ml of 10X HBSS (Gibco, 14065056, United States) to 45 ml of Percoll. Percoll dilutions of 52, 62, and 76% were generated from the "100% Percoll" solution using 1X HBSS-EDTA (Thermo Fisher Scientific, 14025092, United States). The 76, 62, and 52% Percoll separation solutions were successively added to a 15 ml centrifuge tube (avoiding mixing of the three concentrations of Percoll separation solutions). The bone marrow cell suspension was overlaid on the Percoll separation layer, followed by centrifugation for 30 min at 2,800 rpm (1,420 g) at room temperature without braking. Cells were harvested from the 76 and 62% Percoll interface and washed twice with 1X HBSS buffer. 3 ml of Histopaque-1119 (Sigma-Aldrich, 11191, United States) was added to a 15-ml conical centrifuge tube, which was then overlaid with the cell suspension, followed by centrifugation for 30 min at 2,000 rpm (724 g) at room temperature without braking. The neutrophils were then collected at the Histopaque-1119 interface. The surface markers of neutrophils (Ly-6G + CD11b + ) were analyzed by flow cytometry.

Isolation of MSCs From Bone and MSCs Culture
Mesenchymal stem cells (MSCs) from the compact bones of mice were obtained as previously described (Zhu et al., 2010). To deplete hematopoietic cells from the tibiae and femurs, the bone cavities were washed thoroughly at three times using a syringe until the bones become pale. Hold the humeri, tibiae and femurs with forceps and excise the compact bones carefully into chips of approximately 1-3 mm 3 with scissors. The bone chips were transferred into a 25-cm 2 plastic culture flask with forceps, then suspend the chips in 3 ml of α-MEM (Hyclone, SH30265.01, United States) containing 10% (vol/vol) FBS (Gibco, 16000-044, United States) in the presence of 1 mg/ml (wt/vol) of collagenase II (Gibco, 17101015, United States). The chips were digested for 1-2 h in a shaking incubator at 37 • C with a shaking speed of 200 rpm. The collagenase digestion was stopped when the bone chips become loosely attached to each other. The digestion medium and released cells were aspirated and discarded. Enzyme-treated bone chips were placed in a 10-cm 2 dish containing 6 ml of α-MEM supplemented with 10% FBS. Each replanting was considered a passage. Passage 3 MSCs were used for all experiments. The surfaces marker of MSCs (Lin − CD45 − CD31 − CD51 + Sca-1 + ) were analyzed by Flow cytometry.

LKS + Isolation and Culture
Rheb1 f l/fl or Rheb / BM cells were isolated from the tibias, femurs and ilia of 8-week-old mice. LKS + cells were stained with the antibodies indicated above and sorted with a BD FACS Aria III flow cytometer (BD Bioscience, United States). Lin − cells and Lin − c-kit + (LK + ) cells were sorted with a c-Kit (CD117) Microbead Kit (MACS, 130-091-224, Germany) and a Lineage Cell Depletion Kit (MACS, 130-090-858, Germany) according to the manufacturer's protocol. 2 × 10 5 wt Lin − cells were cocultured with 2 × 10 7 Rheb1 f l/fl or Rheb1 / BMCs. The total cells were analyzed for the percentage of CD45.2 + /CD45.1 + cells by flow cytometry.

MSCs Coculture With Neutrophils
For the neutrophils and MSCs coculture assay, 1 × 10 5 MSCs were cultured in 24-well plate in a volume of 500 µl α-MEM with 15% FBS. After 24h of culture, the MSCs were cultured with 1 × 10 6 Rheb1 f l/fl or Rheb1 / neutrophils using cell culture inserts (FALCON, 353095, United States). After 24 h of coculture, the MSCs were harvested, and the relative expression of stem cell factor (SCF) was measured. For the IL-6 neutralization experiment, IL-6 antibody (R&D, MAB406-SP, United States) was added to the coculture system at 10 ng/ml. After 24 h of coculture, MSCs were harvested, and the relative expression of SCF was measured. All cells were incubated at 37 • C in a 5% CO 2 incubator.

Isolation of MSCs and MNCs From CML Patients
CML patients' BM cells were obtained from the Blood Bank of the State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, CAMS/PUMC. Specimen acquisition was approved by the Ethics Committee of Blood Diseases Hospital, Chinese Academy of Medical Sciences. All donors signed informed consent forms. One-fifth the volume of hydroxyethyl starch was added to the patient samples, and let stand for 1 h at room temperature to allow the red blood cells to fully sediment. Gently suck the supernatant and divide it into two according to the volume of 3:1. A large volume of supernatant fluid was selected to culture MSC in containing 6 ml of α-MEM supplemented with 10% (vol/vol) FBS. The other was separated using a Ficoll (GE, United States) gradient to generated mononuclear cells (MNCs).

Quantitative Real-Time PCR
RNA was extracted using the RNeasy Mini Kit (QIAGEN, 74106, Germany) according to the manufacturer's protocol. cDNA synthesis was performed using a cDNA reverse transcription kit (Takara, RR047A, Japan) according to the manufacturer's protocol. Quantitative PCR assays were performed in 96-well Micro Amp Fast Optical Reaction Plates (Applied Biosystems, 4344904, United States) using SYBR Green Mix (Roche, 04913914001, Switzerland). The signal was detected using the Step-One Plus Real-Time PCR System (QuantStudio5). GAPDH was used as an endogenous control for gene expression assays.
ELISA 500 µl PB was obtained from the retro-orbital bleeding of mice and clotted for 1 h at room temperature, then centrifuged at 5,915 rpm (3,000 g) for 10 min, and the serum was collected for determination. Then ELISA was performed using the Mouse SCF ELISA Kit (Quantikine, MCK00, United States) and the Mouse IL-6 ELISA Kit (NRC, TAE-385, Canada) according to the manufacturer's protocols. A total of 5 × 10 6 Rheb1 f l/fl or Rheb1 / neutrophils in 200 µl of PBS were frozen and thawed three times and centrifuged at 5,915 rpm (3,000 g) for 10 min, and the liquid supernatants were collected for IL-6 determination. The cell culture medium was concentrated with an ultragentrification device (Merck, UFC900308, Germany). The ELISA tests were read on a Synergy H 4 Hybrid Reader at 450 nm.

Statistical Analysis
GraphPad Prism 6.0 was used for statistical analyses. Every experiment was compared as two groups. The data are presented as the mean ± standard deviation (SD). The unpaired two-tailed Student's t-test was used to compute the P-values. P < 0.05 was considered significant. Significant differences are indicated with asterisks ( * P < 0.05; * * P < 0.01; * * * P < 0.001).

HSPCs Are Expanded in the Rheb1 / Blood Cell-Remodeled Bone Marrow Environment
We have shown that Vav1-Cre;Rheb1 f l/fl mice present HSCs and immature myeloid cells expansion in BM, which resemble the phenotype of MPN patients (Wang et al., 2018). To investigate whether Rheb1-deficient cells regulate the proliferation of HSPCs, we established a chimeric mouse model with wt HSPCs and a Rheb1 / blood cell-remodeled bone marrow environment. We isolated MSCs from bone of wt mice (wt MSCs, Supplementary   Figure 1B), and transplanted 5 × 10 5 wt BMCs (CD45.2 + ) together with 7.5 × 10 5 wt MSCs and 1 × 10 6 Rheb1 / or Rheb1 f l/fl BMCs (CD45.1 + ) into lethally irradiated recipient wt mice (CD45.2) ( Figure 1A). The percentage of Rheb1 / and Rheb1 f l/fl BMC-derived cells (CD45.1 + ) in PB was approximately 80%, 4 months after the transplantation, and the percentage of wt BMC-derived cells (CD45.2 + ) in PB was close to 20% in mice after the transplantation (Figure 1B and Supplementary  Figure 1C). The absolute number of Rheb1 / -derived LKS + and LKS − cells (CD45.1 + ) were significantly increased when compared with those of Rheb1 f l/fl -derived LKS + and LKS − cells (CD45.1 + ), respectively (Figures 1C,D). Rheb1 / BMCderived myeloid cells were also increased in the PB of mice (Supplementary Figure 1D). Moreover, the percentage of Rheb1 / Ly-6G low CD11b + immature neutrophils (CD45.1 + ) was increased in mice when compared with that of Rheb1 f l/fl Ly-6G low CD11b + immature neutrophils (Figures 1E,F). These phenotypes were similar with Rheb1 / mouse (Wang et al., 2018). Then we analyzed the absolute number of wt HSPCs (CD45.2 + ) in the Rheb1 / blood cell-remodeled bone marrow environment in the chimeric mouse model, and found that the absolute number of wt LKS + cells (CD45.2 + ) transplanted with Rheb1 / BMCs were higher than those of the cells transplanted with Rheb1 f l/fl BMCs (Figures 1G-I). The percentage of wt BMC-derived myeloid cells (CD45.2 + ) was increased in PB, while the ratio of wt BMC-derived Ly-6G high CD11b + and Ly-6G low CD11b + neutrophils (CD45.2 + ) was not changed in the BM of mice co-transplanted with Rheb1 / BMCs (Supplementary Figures 1E,F). It indicated that Rheb1 / blood cell remodeled bone marrow environment and promoted HSPC proliferation. To explore whether Rheb1 / BM cells directly promoted phenotypic HSPCs over proliferation, we cocultured wt Lin − cells (CD45.2 + ) with BMCs from Rheb1 f l/fl and Rheb1 / mice (CD45.1 + ) (Supplementary Figure 2A). Interestingly, the number of wt Lin − cells cocultured with BMCs from Rheb1 / mice was similar with that in the control group at 24 h (Supplementary Figure 2B). These data suggested that Rheb1 / BM cells indirectly promoted HSPCs overproliferation.

The Proliferation of wt HSPCs Was Normal in Rheb1 / Recipient Mice After Transplantation
To investigate whether the niche cells affect the expansion of HSPCs in Rheb1 / mice, we transplanted wt whole bone marrow cells (CD45.2 + ) into lethally irradiated Rheb1 / or Rheb1 f l/fl recipient mice (CD45.1 + ) and analyzed the donorderived HSPCs (CD45.2 + ) at 4 months after transplantation (Figure 2A). The rate of donor chimerism was approximately 80% in both Rheb1 / and control mice ( Figure 2B). The absolute number of donor-derived LKS + cells and LKS − cells (CD45.2 + ) in Rheb1 / mice were similar when compared with those in the control mice (Figures 2C,D). Since Rheb1 deletion caused increased number and immaturity of neutrophils in steady condition (Wang et al., 2018), we also analyzed neutrophils by flow cytometry (FACS) with CD11b and Ly-6G antibodies, that have been used as neutrophil subpopulation markers for the identification of myelocytes or promyelocytes, as well as immature or mature neutrophils. The CD11b + Ly-6G + subpopulation of donor-derived cells (wt) in Rheb1 / mice was similar to that in the control mice (Figure 2E,F). These data indicated that niche cells in Rheb1 / mice contributed little to HSCs proliferation in vivo, while Rheb1 / BM cells cooperated with niche cells to promote HSCs proliferation.

Rheb1-Deficient Neutrophils Stimulate MSCs to Secrete More SCF
To investigate whether Rheb1-deficient BM cells interacted with MSCs and stimulated HSCs expansion, we evaluated the expression of Rheb1 and SCF in MSCs from Rheb1 / mice. We found that the mRNA expression of SCF was higher in primary MSCs from Rheb1 / mice than that from Rheb1 f l/fl mice, while Rheb1 expression was equal in these two cell types ( Figure 3A). The SCF level in serum was also increased in Rheb1 / mice when compared with that in Rheb1 f l/fl mice (Figure 3B), while the expression of EPO, G-CSF, TPO IL-3 and IL-6 showed no difference between two groups (Supplementary Figure 3A). Furthermore, we cultured MSCs derived from Rheb1 f l/fl and Rheb1 / mice in vitro and analyzed the expression of SCF in MSCs after serial passages. We found that in comparison with the MSCs from Rheb1 f l/fl mice, the expression of SCF remained at a higher level in MSCs from Rheb1 / mice at passage 3 and then decreased to about the same level as in MSCs from that of Rheb1 f l/fl mice at passage 5 ( Figure 3C). Interestingly, when wt BMCs were transplanted into lethally irradiated Rheb1 f l/fl and Rheb1 / mice (Figure 2A), SCF mRNA expression in MSCs and SCF serum levels were restored to normal levels (Figures 3D-E). These results demonstrated that the higher expression of SCF in MSCs from Rheb1 / mice was caused by extrinsic factors from the bone marrow cells of Rheb1 / mice and was reversible under native conditions.
To evaluate the specific role(s) of Rheb1 / BMCs in influencing MSCs, we cocultured wt MSCs with Rheb1 / or Rheb1 f l/f T cells, B cells and myeloid cells for 12 h and then analyzed the expression of SCF in MSCs under each condition. Interestingly, the expression of SCF was higher in MSCs in the presence of Rheb1 / myeloid cells than in the presence of Rheb1 / T cells or B cells (Supplementary Figure 3B). In our previous study, we showed that the differentiation of neutrophils was abnormal in the BM of Rheb1 / mice. Hence, we isolated neutrophils (Supplementary Figure 3C) from bone marrow cells and cocultured wt MSCs with Rheb1 f l/fl and Rheb1 / neutrophils for 24 h. We found that the expression of SCF was significantly increased in MSCs in the presence of Rheb1 / neutrophils after cultured for 24 h when compared with the control (Figure 3F). Taken together, our results suggested that Rheb1 / neutrophils stimulated MSCs to produce more SCF.

Neutrophil-Trained-MSCs Promote wt HSPCs Expansion
We then cocultured 1 × 10 6 LK + cells with 6 × 10 4 MSCs from Rheb1 / or Rheb1 f l/fl mice for 24 h and found that the . The data are presented as the mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. number of LK + cells cocultured with MSCs from Rheb1 / mice was more than the number of LK + cells cocultured with MSCs from Rheb1 f l/fl mice (Figures 4A,B). Next, we blocked SCF/c-kit signaling by adding a SCF inhibitor in the culture medium. We found that the number of LK + cells was equally decreased in coculturing with MSCs from Rheb1 / or Rheb1 f l/fl mice in medium with SCF inhibitor (Supplementary  Figures 3D,E). These results indicated that Rheb1-deficient neutrophil-trained MSCs could promote wt HSPCs expansion through SCF signaling pathway.
We further cocultured 200 wt LKS + (CD45.1 + ) with 2 × 10 4 MSCs from Rheb1 / or Rheb1 f l/fl mice for 24 h, then transplanted the cultured LKS + (CD45.1 + ) cells with 5 × 10 5 newly isolated wt WBMCs (CD45.2 + ) into lethally irradiated recipient mice by intravenously injection respectively ( Figure 4C). We found that the percentage of chimerism was higher in mice transplanted with LKS + cells cocultured with MSCs from Rheb1 / mice when compared with those from Rheb1 f l/fl mice ( Figure 4D). Accordingly, the absolute number of LKS + (CD45.1 + ) cells treated with MSCs from Rheb1 / mice was significantly increased in comparison with the controls at 4 months after transplantation (Figures 4E,F). However, the percentage of CD11b + (CD45.1 + ) myeloid cells was similar in the PB of the two groups ( Figure 4G). The ratio of Ly-6G high CD11b + and Ly-6G low CD11b + neutrophils was also not changed between the two groups (Figures 4H,I). These data indicated that HSPCs cocultured with MSCs from Rheb1 / mice exhibited a higher expansion capacity while their ability to differentiate into myeloid cells was not changed. MSCs educated by Rheb1deficient neutrophils induced LKS + over proliferation but not myeloid differentiation.

Rheb1-Deficient Neutrophils Stimulate MSCs to Produce More SCF by Decreasing IL-6 Expression
To investigate the underlying mechanisms by which Rheb1deficient neutrophils stimulate MSCs to increase SCF production, we measured the mRNA expression of several potential interleukin and chemokine candidates secreted by neutrophils (Supplementary Figure 3F). We found that the mRNA and protein expression levels of IL-6 were decreased in Rheb1 / neutrophils when compared with those from the control (Figures 5A,B, 0 h). We confirmed that IL-6 is mainly expressed in myeloid cells (Supplementary Figure 3G), and found that IL-6 mRNA and protein expression levels in Rheb1 / neutrophils remained lower after cultured with wt MSCs for 24 h in vitro (Figures 5A,B, 24 h). In addition, the level of IL-6 in the Rheb1 / neutrophil coculture medium was also lower than that in the control ( Figure 5C). These data indicated that Rheb1 / neutrophils expressed and secreted less IL-6 than Rheb1 f l/fl neutrophils. Since MSCs are immunomodulatory cells and secrete a variety of cytokines, including IL-6, we measured the mRNA expression of IL-6 in MSCs and found that its expression in MSCs from Rheb1 / mice was similar to that in MSCs isolated from Rheb1 f l/fl mice (Supplementary Figure 3H).
To determine the role of IL-6 in MSCs, we cultured wt MSCs with different concentrations of IL-6 and analyzed the relative (E) The relative mRNA expression of SCF in wt MSCs cocultured with Rheb1 f l/fl or Rheb1 / neutrophils after being exposed to an IL-6 antibody at 10 ng/ml for 24 h (n = 3). *P < 0.05; **P < 0.01; ***P < 0.001. SCF mRNA expression in MSCs at 24 h. We found that higher IL-6 treatment suppressed SCF mRNA expression in MSCs ( Figure 5D). We then blocked IL-6 signaling by adding an IL-6 neutralizing antibody in the media in which Rheb1 / or Rheb1 f l/fl neutrophils were cocultured with wt MSCs, and measured the expression of SCF mRNA in MSCs (Figure 5E, left panel). We found that the mRNA expression of SCF was significantly increased in MSCs after treatment with the IL-6-neutralizing antibody in coculture with either Rheb1 f l/fl or Rheb1 / neutrophils ( Figure 5E, right panel). These data suggested that the lower expression of IL-6 in Rheb1 / neutrophils stimulated the expression of SCF in MSCs.

IL-6 Regulates SCF Expression in Human MSCs
It has been reported that IL-6 was higher in CML patients (Panteli et al., 2005) and higher level of IL-6 is associated with poor prognosis in CML patients (Nievergall et al., 2016). To evaluate the relationship of IL-6 and SCF expression level in human cells, we first investigated whether IL-6 regulates SCF expression in human MSCs. We cultured MSCs from human umbilical cord blood with different concentrations of IL-6 and analyzed the relative SCF mRNA expression in MSCs at 24 h. The result showed that higher IL-6 treatment suppressed SCF mRNA expression in human MSCs ( Figure 6A). Next we isolated MNCs and MSCs from the BM of 39 CML patients, and measured the expression of IL-6 in mononuclear cells and the expression of SCF in MSCs in patients ( Figure 6B, left panel). We found a negative correlation of SCF and IL-6 by analyzing the relationship of IL-6 and SCF expression in CML patients (n = 39) ( Figure 6B, right panel). The data indicated that higher expression of IL-6 in myeloid cells is associated with the lower expression of SCF in MSCs in CML patients.

DISCUSSION
We previously found that the HSPCs were over-expanded in Vav1-Cre;Rheb1 f l/fl mice (Wang et al., 2018). Here, using in vivo and in vitro models, we found that loss of Rheb1 impaired neutrophils' ability to secrete IL-6, which stimulated MSCs to secrete more SCF and in turn stimulated HSPCs to proliferate ( Figure 6C). FIGURE 6 | IL-6 regulates SCF expression in human MSCs. (A) The relative mRNA expression of SCF in H-MSCs after exposure to different concentrations IL-6 for 24 h (n = 3). The data are presented as the mean ± SD. (B) Correlation analysis for SCF and IL-6 mRNA expression in CML patients (n = 39) (p = 0.0094). R: Pearson correlation coefficients; R 2 : indicates "the goodness of fit." Statistical significance was calculated by Pearson correlation coefficients. (C) Model depicting hematopoietic regulation in the absence of Rheb1. *P < 0.05; **P < 0.01; ***P < 0.001.
The microenvironment in BM comprises multipotent stromal cells (MSCs), osteoblasts, endothelial cells, mature blood cells and the cytokines that they produced . We previously showed that the differentiation and maturation of neutrophils in BM were abnormal in Rheb1 / HSCs, leading aged Rheb1 / mice to show MPN-like symptoms (Wang et al., 2018). Neutrophils and their progenitors have been shown to secrete a variety of cytokines to influence the development and proliferation of HSCs either directly or indirectly. For example, apoptotic neutrophils can stimulate macrophages to secrete G-CSF and regulate HSCs indirectly in zebrafish (Hall et al., 2016). Gr-1 + neutrophils in mouse bone marrow produce ROS during acute infection, which contributes to the proliferation of HSPCs via a paracrine mechanism (Kwak et al., 2015). Senescent neutrophils are markedly increased in aged mice and promote an increase in platelet-biased HSCs via IL-1β (Frisch et al., 2019). Our results demonstrated that Rheb1 / neutrophils remodeled the BM niche, and induced HSPCs to undergo overexpansion in BM (Figure 3) probably via a feedback in MSC through the proinflammatory factor IL-6 (Figures 4, 5). This partially supported that the abnormal myeloid cells could regulate the proliferation of HSCs in MPN patients (Schepers et al., 2013). IL-6 was higher in CML patients and was elevated during the transformation phase of CML (Panteli et al., 2005), while blocking