Hematopoietic Stem/Progenitor Cell Proliferation and Differentiation Is Differentially Regulated by High-Density and Low-Density Lipoproteins in Mice

Rationale Hematopoietic stem/progenitor cells (HSPC) are responsible for maintaining the blood system as a result of their self-renewal and multilineage differentiation capacity. Recently, studies have suggested that HDL cholesterol may inhibit and impaired cholesterol efflux may increase HSPC proliferation and differentiation. Objectives We hypothesized that LDL may enhance HSPC proliferation and differentiation while HDL might have the opposing effect which might influence the size of the pool of inflammatory cells. Methods and Results HSPC number and function were studied in hypercholesterolemic LDL receptor knockout (LDLr−/−) mice on high fat diet. Hypercholesterolemia was associated with increased frequency of HSPC, monocytes and granulocytes in the peripheral blood (PB). In addition, an increased proportion of BM HSPC was in G2M of the cell cycle, and the percentage of HSPC and granulocyte-macrophage progenitors (GMP) increased in BM of LDLr−/− mice. When BM Lin-Sca-1+cKit+ (i.e. “LSK”) cells were cultured in the presence of LDL in vitro we also found enhanced differentiation towards monocytes and granulocytes. Furthermore, LDL promoted lineage negative (Lin−) cells motility. The modulation by LDL on HSPC differentiation into granulocytes and motility was inhibited by inhibiting ERK phosphorylation. By contrast, when mice were infused with human apoA-I (the major apolipoprotein of HDL) or reconstituted HDL (rHDL), the frequency and proliferation of HSPC was reduced in BM in vivo. HDL also reversed the LDL-induced monocyte and granulocyte differentiation in vitro. Conclusion Our data suggest that LDL and HDL have opposing effects on HSPC proliferation and differentiation. It will be of interest to determine if breakdown of HSPC homeostasis by hypercholesterolemia contributes to inflammation and atherosclerosis progression.


Introduction
Lipoproteins play essential roles in coronary heart disease (CHD). A meta-analysis showed that a 1 mmol/l reduction in the level of LDL cholesterol is associated with 19% reduction in CHD mortality [1]. In contrast, increase of HDL cholesterol by 1 mg/dl decreases the incidence of CHD 2-3% [2]. Apart from CHD, hypercholesterolemia also accelerates other diseases including Alzheimer's amyloid pathology [3].
Atherosclerosis, the underlying pathogenesis of CHD, is characterized by chronic inflammation due to both loss of endothelial integrity and to subendothelial retention of LDL [4]. Once retained in the subendothelial space, both native and modified LDL activate vascular cells, attracting circulating white blood cells to form plaques.
It is well known that granulocytes and monocytes play critical roles in atherosclerosis development. Granulocytes are the first cells recruited to the plaque, where they secrete elastase and metalloproteinase, resulting in extracellular matrix degradation and inflammatory cell adhesion [5]. Monocytes in the plaque take up both native and modified LDL and become foam cells, the hallmark of atherosclerosis [6]. Inflammatory and/or activated vascular cells also secrete inflammatory cytokines such as tumor necrosis factor (TNF)-a, interferon (IFN)-c, granulocyte macrophage colony-stimulating factor (GM-CSF) and monocyte CSF (M-CSF), which further reinforce the inflammation processes [7,8,9,10,11,12,13]. Thus, hypercholesterolemia-associated leukocytosis and transformation of monocytes to macrophages facilitate atherosclerosis progression [14].
In contrast, there is an inverse correlation between the incidence of CHD and plasma levels of HDL and its major apolipoprotein A-I (apoA-I) [15]. Aside from HDL-mediated reverse cholesterol transport and anti-oxidative properties, one of the major functions of HDL is to inhibit vascular cell activation, thereby, prohibiting inflammatory cell infiltration and atherosclerotic progression [16,17,18]. We previously demonstrated that HDL improved endothelial repair, inhibited inflammation and thus attenuated transplant arteriosclerosis and atherosclerosis in vein grafts in the presence of Scavenger Receptor Type B-I (SR-BI) [19,20,21].
Recently, Yvan-Charvet et al., reported the suppressive role of HDL on hematopoietic stem cell (HSC) proliferation via ATPbinding cassette transporters (ABCA1) [22]. In contrast, hypercholesterolemia has been shown to promote BM cell mobilization in man and mice [23,24]. These studies lead to the hypothesis that HDL and LDL may have opposing effects on hematopoietic cell function. We demonstrate here that hypercholesterolemia is associated with enhanced proliferation of HSPC in vivo and with increased myeloid cell differentiation. Both appear to be mediated, at least in part, by extracellular signal regulated kinase (ERK). By contrast, HSPC proliferation was inhibited in BM of C57BL/6J mice infused with purified human apoA-I or reconstituted (r)HDL. We further demonstrated that exposure of HSPC to LDL in vitro induced differentiation to monocytes and granulocytes, whereas HDL decreased myeloid cell differentiation induced by LDL.

White blood cell counts
Leukocytes, lymphocytes, monocytes and granulocytes in PB were quantified with an Ac-Tdiff hematology analyzer (Beckman Coulter; Brea, CA, U.S.A.).

ELISA for human apoA-I
Human apoA-I levels in plasma and fractions obtained by gel filtration were determined by sandwich ELISA (MyBioSource, San Diego, CA, U.S.A.).

PLPC determination
PLPC was purchased from Avanti Polar Lipids. After reconstitution in PBS, PLPC concentration was determined using the EnzyChrom Phospholip Assay Kit (BioAssay Systems, Hayward, CA, U.S.A.).

Separation of plasma lipoproteins by gel filtration
Mouse plasma lipoproteins were fractionated by fast performance liquid chromatography gel filtration of 50 ml plasma on Superose 12 HR 10/30 (Pharmacia, Herts, UK). Cholesterol content in non-HDL fractions and HDL fractions was quantified by Amplex TM Red Cholesterol Kit (Molecular Probe, CA, U.S.A.).

LDL and HDL isolation by density gradient ultracentrifugation
Plasma LDL (1.019 g/ml,d,1.063 g/ml) and HDL (1.063 g/ ml,d,1.21 g/ml) were isolated from healthy volunteers by density gradient ultracentrifugation in a swing-out rotor as described [25]. Subsequently, LDL and HDL were dialyzed against 1 mM EDTA in PBS overnight. Cholesterol concentration was measured by Amplex TM Red Cholesterol Kit.

Identification of lipoprotein fractions by SDS-PAGE
To study the apolipoprotein profiles in the fractions, 20 ml of each fraction was separated on 4-12% NuPage (Invitrogen, Gent, Belgium) and then stained with GelCode Blue Stain (Thermo Fisher Scientific, Rockford, IL, U.S.A.) overnight.

Quantification of murine apoA-I levels
Murine apoA-I expression in plasma was quantified by western blot as described before [26]. A goat anti-mouse apoA-I antibody (sc-23606, Santa Cruz Biotechnology, CA, U.S.A) was used to detect murine apoA-I, without cross-reactivity with human apoA-I.
For in vivo BrdU analysis of HSPCs, mice were injected with 0.2 mg BrdU/g intraperitoneally 12 h before analysis [28]. After staining with Lineage cocktail APC, Sca-1 PE and cKit-APC-Cy7, cells were permeabilized and stained with anti-BrdU FITC using the BrdU Flow Kit according to manufacturer's instruction (Becton Dickinson).
To study adhesion molecules expression, Lin-cells were exposed to 0 or 100 mg/ml LDL for 24 hours. After harvest, cells were stained with Lineage cocktail APC, Sca-1 FITC, cKit APC-Cy7 together with CXCR4 PE, integrin b1 PE, or integrin a5 PE for FACS analysis.
All FACS studies were performed using the appropriate isotype control antibodies. To achieve reliable quantification, at least 100,000 events were acquired.

qRT-PCR
Total RNA from cultured Lin-cells was extracted using RNAeasy microkit (Qiagen, Valencia, CA). mRNA was reverse transcribed to get cDNA using Superscript III reverse Transcriptase (Invitrogen). Primers used in this study are as following: SR-BI: forward 59-GGCTGCTGTTTGCTGCG- 39  HSPC in vitro differentiation assay LSK cells were cultured at a density of 1000 LSK cells per well in SFEM medium supplemented with SCF (20 ng/ml), IL-3 (10 ng/ml) and IL-6 (10 ng/ml) (all from R & D Systems). Immediately after seeding, LDL (100 mg/ml) or LDL (100 mg/ml) plus HDL (600 mg/ml) were added. In parallel, GM-CSF (10 ng/ ml) was used as a positive control. After 14 days, cells were harvested by cytospin and a Giemsa stain was performed. Promonocytes were identified based on an increased nuclear/ cytoplasmic ratio, and granulocyes were identified based on their specific nuclear morphology [29]. Total cells, promonocytes and granulocytes were counted under the microscope to calculate the percentage of differentiated cells. For each condition, at least 5 fields of cells were counted. After 14 days, cells were harvested and stained with antibodies against Ly-6c, CD11b, Ly-6G and F4/80 for FACS. For pERK inhibitor experiments, U0126 (10 mM) (Merck, Darmstadt, Germany) was immediately added to LSK cells upon seeding and maintained till harvest.

In vitro adhesion and migration assay
Adhesion and migration of Lin-cells were tested as described before [19,30]. In brief, Lin-cells isolated from CD45.2 mice were cultured with LDL 0, LDL 0 plus U0126, LDL 100 mg/ml or LDL 100 mg/ml plus U0126 for 24 hours. After numeration, 50,000 cells were allowed to adhere to fibronectin (25 mg/ml)-coated 96well plate for 24 hours. After extensive wash with PBS, adhered cells were fixed with 3.7% paraformaldehyde, stained with crystal violet (5 mg/ml in 2% ethanol, Sigma-Aldrich) and counted under the microscope.
To test motility, 50,000 cells were loaded in the upper chamber of modified Boyden chambers (8 mm pore size, Costar, Avon, France). The lower chamber was filled with IMDM medium. After 4 hours, cells in the upper chamber were removed by swaps. Cells migrating on the lower surface of the transwells were stained with Hoechst and counted under the fluorescent microscope as described before [19]. For the adhesion and migration assays, 5 random fields were counted and the average cell number was calculated.

Statistics
Data are expressed as mean 6 SEM. Unpaired, 2-tailed Student's t test was used if there were 2 experimental groups. For more than 2 experimental groups, one-way ANOVA with Dunnett was applied to test treated groups against control or one-way ANOVA with Bonferroni was used to compare all groups. Statistical analysis was performed using GraphPad Prism (Graph-Pad Software Inc, La Jolla, CA, U.S.A.). A P value less than 0.05 was considered significant.

Hypercholesterolemia leads to increased leukocytosis in blood
We hypothesized that LDL may cause proliferation and mobilization of HSPC from the BM, whereas HDL would have the opposite effects. LDLr deficiency causes impaired LDL clearance, resulting in high plasma levels of LDL-cholesterol. Thus, LDLr 2/2 mice were used to study the effect of LDL on HSPC. LDLr 2/2 mice received either normal or high fat diet from the age of 2 months onwards and this for 2 months. In parallel, WT mice were included as control. The lipoprotein profile of these mice is shown in Table 1 (n = 5 for both WT and high fat diet; n = 11 for normal diet respectively).  Although no differences in white blood cells were observed between WT and LDLr 2/2 mice fed a normal diet, the number of white blood cells, neutrophils and monocytes in the PB was 1.4fold (P,0.05), 1.6-fold (P,0.05), and 2-fold (P,0.05) higher in LDLr 2/2 mice maintained on high fat diet (n = 5 compared to mice on normal diet n = 11, Table 1). Furthermore, the percentage of Ly-6c hi and F4/80 + monocytes and Ly-6G hi granulocytes was increased in LDLr 2/2 mice on normal diet, compared to WT mice. (n = 6-12, P,0.05) (Figure 1, A-C). Moreover, the percentage of Ly-6c hi and F4/80 + monocytes and Ly-6G hi granulocytes was 2.2-, 2.4-, and 1.5-fold higher in LDLr 2/2 mice on high fat diet, respectively, compared to LDLr 2/2 mice on normal diet (n = 5-12, P,0.05) (Figure 1, A-C and Figure S1).
Hypercholesterolemia enhances the frequency of HSPC in the bone marrow and peripheral blood of LDLr 2/2 mice Next, we measured the frequency of HSPC in PB and BM of WT and LDLr 2/2 mice that received either normal or high fat diet for 2 months. Quantification of Lin2 Sca1+ cKit+ (LSK) cells in blood showed that the percentage of HSPC in LDLr 2/2 mice on high fat diet was increased compared to WT and LDLr 2/2 mice on normal diet (Figure 2A). In agreement with other reports [22,23,31,32,33,34,35], the percentage of LSK cells in BM of WT and LDLr 2/2 mice fed on normal diet was 0.09160.0098% and 0.09960.0101%, respectively (n = 9 for WT mice and n = 11 for LDLr 2/2 mice, P = 0.7). However, the percentage of LSK cells in the BM was 1.9-fold higher in LDLr 2/2 mice fed on high fat diet, compared with LDLr 2/2 mice on normal diet (n = 5-11, P,0.01) ( Figure 2B). We found a 1.9-fold increase of the number of CD342/Flk22/LSK cells in BM of LDLr 2/2 mice on high fat diet, compared to LDLr 2/2 mice on normal diet (normal diet: 0.01860.0027%; high fat diet: 0.04060.0058%, n = 5-7, P,0.05). Similarly, the percentage of granulocyte-macrophage progenitors (GMP) was higher in LDLr 2/2 mice on high fat diet, compared to WT and LDLr 2/2 mice on normal diet (LDLr 2/2 on high fat diet: 1.760.04%; LDLr 2/2 on normal diet: 1.160.07%; WT mice: 0.960.14%, n = 5-11, P,0.05 for both) ( Figure 2C). rHDL infusion decreases the frequency of LSK cells in the bone marrow and peripheral blood of mice but does not affect monocyte frequency To assess the effect of HDL on HSPC in vivo, C57BL/6 mice were infused with rHDL at 0, 4, 8, 12, or 16 mg/kg intravenously on days 1, 3 and 5. On day 6, LSK cells were enumerated in PB and BM. In line with other reports [18,36,37,38], we showed that rHDL infusion induced human apoA-I expression without any effect on murine apoA-I expression and lipoprotein profile. The lipoprotein profile in mice infused with saline or rHDL 8 mg/kg is shown in Table 2. Both rHDL and human apoA-I infusion induced human apoA-I expression in plasma ( Figure 3A). Murine apoA-I expression in plasma is shown in Figure 3B. Moreover, the injected human apoA-I facilitated HDL formation as evidenced by its presence in lipoprotein fractions obtained by FPLC ( Figure 3C). Overall, no difference in total white blood cell count was seen in the blood of animals infused with the different rHDL concentrations (table 2). The percentage of LSK cells was decreased in the PB (P,0.05; n = 6) and BM (P,0.05; n = 6-22) of mice infused with 8, 12 and 16 rHDL mg/kg, respectively (Figure 3, D and E). The percentage of GMP cells in BM was not different among the three groups ( Figure 3F). Representative dot plots of quantification of LSK in LDLr 2/2 and rHDL-infused mice are shown in Figure  S2.

SR-BI is expressed on HSPC and apoA-I infusion has similar effect as rHDL
Receptors for HDL include ABCA1, ABCG1 and SR-BI. To determine which of these receptors is expressed on HSPC, we performed in an initial step qRT-PCR to quantify the expression of Abca1, Abcg1 and Sr-BI [39] in Lin-cells of C57BL/6 mice. This demonstrated that Lin-cells express all three receptors (data not shown). We also detected SR-B1 on the cell membrane of 8862% of LSK cells (n = 3, Figure 4A). As apoA-I is the major apolipoprotein of HDL and HDL binds to SR-BI via apoA-I, we tested if, like rHDL, in vivo infusion of apoA-I would affect HSPC frequency. C57BL/6 mice were infused with saline or apoA-I at 8 mg/kg on days 1, 3 and 5, and LSK cells were quantified on day 6 by FACS. As for rHDL infusions, the percentage of LSK cells was decreased by 30% in the BM of mice that received apoA-I infusions, (n = 4-5, P,0.05), compared to control ( Figure 4B).

LDL and HDL have opposing regulatory effect on LSK proliferation in BM
To explore whether the increased LSK cell frequency is due to enhanced proliferation, we studied the proliferative status of HSPC in the BM of LDLr 2/2 mice on normal or high fat diet. We injected Bromodeoxy-Uridine (BrdU) i.p. in LDLr 2/2 mice on high fat vs. normal diet 12 hours prior to sacrifice, and quantified the percentage of proliferating HSPC by FACS. We found that the percentage of BrdU-incorporating LSK cells in the whole LSK cell population was increased from 12.060.64% in control mice to 2062.24% in mice on high fat diet (n = 5-6, P,0.05) ( Figure 5A).
We also tested the effect of HDL on the proliferative status of HSPC in vivo. C57BL/6J mice received 8 mg/kg rHDL on days 1, 3 and 5, and BrdU was injected i.p. 12 hours prior to sacrifice on day 6. The percentage of BrdU-incorporating LSK cells in the entire LSK cell population was significantly lower in rHDLinfused animals compared to controls (12.460.73% vs. 7.860.95%, P,0.05; n = 5) ( Figure 5B). Representative BrdU incorporating LSK cells are shown in Figure 5C.  LDL and HDL modulate HSPC differentiation in an opposite way in vitro As we had noticed an increase in monocytes and myeloid progenitor cells in LDLr 2/2 mice on high fat diet but not in rHDL-infused animals, we next evaluated the effect of LDL and HDL on LSK cell differentiation to monocytes and granulocytes in vitro. LSK cells from WT C57Bl/6 mice were cultured in serum free medium supplemented with cytokines and with either LDL (100 mg/ml), or a combination of LDL plus HDL (600 mg/ml). In parallel, GM-CSF (10 ng/ml) was added as a positive control. After 14 days, cells were harvested and a Giemsa stain was performed. Addition of LDL to LSK cell cultures induced a 93616.1-fold increase in the total cell number, whereas addition of HDL+LDL only induced a 1767.8-fold increase (n = 4, P,0.01). LDL and GM-CSF induced monocytic and granulocytic differentiation of LSK cells, as assessed by immunohistochemistry [29]. However, addition of HDL decreased the LDL-induced HSPC differentiation ( Figure 6A and Figure S3 representing the Giemsa stained progeny from different cultures). Cultured progeny was also stained with antibodies against CD11b, F4/80, Ly-6G and Ly-6c. Consistent with the morphological assessment, the percentage of Ly-6c hi monocytes and F4/80 + monocytes was 8.260.88% and 5.461.73% in cells treated with LDL, but decreased to 2.660.49% and 1.560.40% in cells treated with LDL plus rHDL (P,0.5, for both) (Figure 6, B and C). In addition, the percentage Ly-6G+ granulocytes was 461.27% in LDLtreated cells but decreased to 0.960.25% in the presence of HDL (p,0.05) ( Figure 6D).

LDL-induced HSPC differentiation toward granulocytes is mediated via pERK
As effects of HDL on HSPC are at least in part due to the inhibition of extracellular signal-regulated kinase 1/2 (ERK1/2) [22], we tested if LDL would activate ERK 1/2. After stimulation with 100 mg/ml LDL for 0, 5 or 15 min, total bone marrow cells were fixed and stained with antibodies against lineage markers, Sca-1, c-Kit and pERK. FACS analysis showed that the percentage of pERK positive LSK cells in the LSK population was 9.660.53% at baseline level, which increased 2-fold (P,0.05) within 5 min after LDL addition and returned to baseline levels after 15 min (Figure 7, A and B). We next exposed LSK cells to LDL with or without the pERK inhibitor, U0126, for 14 days. The myeloid lineage differentiation was analyzed as described above. We found that U0126 abrogated LDL-induced granulocyte differentiation ( Figure 7C). Pretreatment of LDL enhances Lin2 cell motility which is partially mediated via ERK phosphorylation As the increase of LSK cells in the PB might be due, as suggested by Gomes, to increased mobilization, we tested whether LDL affects adhesion/migration receptor expression and HSPC motility. Lin2 cells from wt mice were exposed to LDL for 24 hours and adhesion molecule expression was studied by FACS. Median Fluorescence Intersity (MFI) of CXCR4 was higher in Lin2 cells treated with LDL for 24 hours, compared to control (LDL 0 mg/ml: 38576300.6; LDL 100 mg/ml: 47336389, n = 9, P,0.01). In contrast, LDL treatment did not affect integrin b1 and integrin a5 expression on Lin2 cells (integrinb1: 98746869.6 vs. 954961113.5; integrin a5: 36116560.2 vs. 38266496.8, n = 4-5).
Next, we performed adhesion and migration assays to explore the impact of LDL on Lin2 cell mobility. As pERK plays a critical role in the regulation of LDL on HSPC, we also investigated whether modulation of LDL on Lin2 cell function required ERK phosphorylation. Lin2 cells isolated from WT mice were exposed to 0 or 100 mg/ml LDL in the presence or absence of U0126 for 24 hours and then subjected for adhesion and migration assays. LDL and pERK inhibitor did not change Lin2 cell adhesion to fibronectin-coated plates ( Figure 8A, n = 5-7), but Lin2 cells pretreated with LDL showed increased migration to the lower chamber of a modified Boyden chamber, compared to control ( Figure 8B, n = 5-7).

Discussion
Hypercholesterolemia is at least partially associated with monocytosis because of increased monocyte survival and continued cell proliferation [14]. Noteworthy, infusion of rHDL attenuated monocytosis and neutrophilia in apoE 2/2 mice on western diet [32]. We here asked the question whether the effects seen on leukocytosis and monocytosis could at least in part be caused by effects of LDL and HDL on HSPC and progenitors.
We describe that LDL and HDL have opposing effects on HSPC behavior. (1) Hypercholesterolemia was associated with leukocytosis, and in particular with increased levels of Ly-6c hi and F4/80 + monocytes and Ly-G hi granulocyte production in blood.
(2) LDL promoted HSPC differentiation toward atherogenic monocytes and granulocytes in vitro, which was inhibited by HDL.
(3) LDL stimulated ERK phosphorylation in LSK cells and LDLpromoted LSK differentiation toward granulocytes was partially dependent on ERK phosphorylation. (4) LDL also induces an increase in LSK cells in BM and PB, at least in part due to increased proliferation of HSPC. (5) As LDL also increases motility of HSPC, it is possible that part of the increased number of LSK cells in PB is due to mobilization in response to LDL, as suggested by Gomez.
Although the white blood cell count and HSPC percentage in blood and BM of LDLr 2/2 mice on normal diet was similar to that of WT mice, the percentage of Ly-6C hi , F4/80 + monocytes and Ly-6G hi granulocytes was higher in LDLr 2/2 mice on normal diet vs. WT mice. When LDLr 2/2 mice were placed on high fat diet, an increased percentage of HSPC and production of these inflammatory cells was seen, compared to LDLr 2/2 mice on normal diet. As the lipoprotein profile in the blood of WT and LDLr 2/2 mice on normal diet only differs in LDL levels, the difference in the percentage of monocytes and granulocytes can only be explained by the elevated LDL levels. Therefore, (1) LDL appears to be a potent stimulator for HSPC differentiation into Ly-6c hi and F4/80 + monocytes and Ly-G hi myeloid cells; (2) The effect of LDL on HSPC seems to be dose-dependent: mild hypercholesterolemia, such as in LDLr 2/2 mice on normal diet, causes an increase in myelomonocytic cells in the blood, while high cholesterol levels cause a much more pronounced effect both on proliferation, mobilization and differentiation of HSPC. Consistently, when isolated LSK cells were treated with LDL for 14 days in vitro, skewed differentiation of LSK cells toward myeloid cells was seen. Furthermore, LDL-induced HSPC differentiation toward granulocytes was inhibited by addition of the pERK inhibitor, U0126. These in vitro data demonstrate a direct positive regulation of LDL on HSPC differentiation into myeloid cell population. Distinct from LDL, addition of HDL prohibited the After seeding, cells were exposed to LDL 100 mg/ml in the presence or absence of pERK inhibitor, U0126 at 10 mM, for 14 days. Ly-6G hi granulocytes were analyzed by FACS. doi:10.1371/journal.pone.0047286.g007 Figure 8. LDL modulates Lin-cell motility which is partially mediated via pERK. To study the effect of LDL on HSPC function, Lincells of CD45.2 mice were exposed to 0 or 100 mg/ml LDL in the presence or absence of pERK inhibitor U0126 for 24 hours and then subjected for adhesion and migration assays. (A) Cells were seeded onto fibronectin-coated plates for 4 hours. After extensive washing, adhered cells were stained with crystal violet and counted under the microscopy. (B) Cells were loaded into transwells and allowed to migrate for 4 hours. Cells in the upper chamber of the transwells were removed by swaps. Cells migrated to the lower surface of the transwells were stained with Hoechst. Adherent cells or migrated cells were numerated under the microscope, averaging the number of cells in 5 random fields. doi:10.1371/journal.pone.0047286.g008 effect of LDL on HSPC number and differentiation in vitro, suggesting the direct but negative regulation of HDL on HSPC.
The anti-atherogenic properties of HDL include reversed cholesterol transport, anti-inflammation, anti-oxidation, maintenance of endothelial integrity and suppression of HSC proliferation in BM [21,22,40,41]. Consistently, our data demonstrated that rHDL inhibited HSPC proliferation. Our data suggest a direct modulation of HSPC by HDL as ABCA1 and SR-BI, the HDL receptors, are expressed on LSK cells; and infusion of apoA-I, like rHDL, decreased LSK cell frequency and proliferation in vivo.
Gomes et al., reported that hypercholesterolemia promotes HSPC mobilization from BM partially via enhanced SDF-1 production and breakdown of the SDF-1/CXCR4 axis. In vitro, they showed that LDL induced HSPC differentiation into monocytes and this was inhibited by Blt3, a SR-BI inhibitor. Furthermore, LDL-promoted HSPC migration was mediated via Blt3. These data indicate that SR-BI may play a critical role in HSPC biology. During the final preparation of our manuscript, we noticed that Murphy et al., reported the impact of impaired cholesterol efflux on HSPC proliferation and monocyte production in apoE deficient (apoE 2/2 ) mice [32]. Our data in LDLr 2/2 mice are consistent with their findings. Our studies, however, also identify a direct effect of LDL on HSPC differentiation toward Ly-6c hi and F4/80 + monocytes and Ly-6G hi granulocytes which is counteracted by HDL. Moreover, short term treatment of LDL enhanced Lin-cell mobility. Given that stem cell proliferation is in close context with its mobilization activity [30,42,43,44,45,46], our data of HSPC proliferation and increased Lin2 mobility indicate that LDL may promote HSPC mobilization. Consistent with the observation of Yvan-Charvet et al., that pERK is a key signaling molecule in the regulation of lipoproteins in HSPC, we demonstrate that the regulation of LDL on HSPC differentiation into myeloid cells and on Lin2 motility are both mediated at least in part by pERK.
Leukocytosis is a marker of inflammation that is associated with ischemic vascular disease, sickle cell disease, and diabetes [47]. Others and we demonstrate that LDL promotes HSPC proliferation, differentiation and leukocytosis while HDL has the opposite effect [32,48]. However, whether and how LDL-induced HSPC expansion and the associated leukocytosis/monocytosis affects atherosclerosis is still unknown. Moreover, LSK cells contain various subsets of progenitors that may have differential effects on disease development. Lineage tracing studies will be needed to demonstrate the direct contribution of BM LSK cells or other cell populations to atherosclerosis.
Based on our findings and published data, we propose a model defining the influence of lipoproteins on HSPC. LDL promotes HSPC proliferation, leading to increased production of HSPC and GMP, in the blood, as well as myelomonocytic differentiation leading to increased frequency of monocytes and neutrophils in BM and blood. The effects of LDL appear to be counteracted by HDL. Figure S1 Hypercholesterolemia increased the percentage of proatherogenic monocytes and granulocytes in peripheral blood of LDLr -/mice. After red blood cells were lysed, white blood cells were stained with antibodies against CD11b, Ly-6c, F4/80 and Ly-6G for FACS analysis. (A) Ly-6c hi monocytes were indicated in the box. (B) F4/80 + monocytes were shown in the box. (C) Ly-6G hi granulocytes were shown in the box. n = 5-7. The percentage of Ly-6c hi and F4/80 + monocytes and Ly-6G hi granulocytes was shown on the right corner of each plot. (TIF)