Hematopoietic Stem/Progenitor Cells Directly Contribute to Arteriosclerotic Progression via Integrin β2

Recent studies described the association between hematopoietic stem/progenitor cell (HSPC) expansion in the bone marrow (BM), leukocytosis in the peripheral blood, and accelerated atherosclerosis. We hypothesized that circulating HSPC may home to inflamed vessels, where they might contribute to inflammation and neointima formation. We demonstrated that Lin− Sca-1+ cKit+ (LSK cells) in BM and peripheral blood of LDLr−/− mice on high fat diet expressed significantly more integrin β2, which was responsible for LSK cell adhesion and migration toward ICAM-1 in vitro, and homing to injured arteries in vivo, all of which were blocked with an anti-CD18 blocking antibody. When homed LSK cells were isolated from ligated artery and injected to irradiated recipients, they resulted in BM reconstitution. Injection of CD18+/+ LSK cells to immunodeficient Balb/C Rag2− γC−/− recipients resulted in more severe inflammation and reinforced neointima formation in the ligated carotid artery, compared to mice injected with PBS and CD18−/− LSK cells. Hypercholesterolemia stimulated ERK phosphorylation (pERK) in LSK cells of LDLr−/− mice in vivo. Blockade of pERK reduced ARF1 expression, leading to decreased integrin β2 function on HSPC. In addition, integrin β2 function could be regulated via ERK-independent LRP1 pathway. Integrin β2 expression on HSPC is regulated by hypercholesterolemia, specifically LDL, in pERK-dependent and -independent manners, leading to increased homing and localization of HSPC to injured arteries, which is highly correlated with arteriosclerosis. Stem Cells 2015;33:1230–1240


INTRODUCTION
Inflammatory cells in atherosclerotic plaques are exclusively derived from hematopoietic stem/progenitor cells (HSPC). Recently, others and we reported that hypercholesterolemia promotes HSPC proliferation and differentiation toward atherogenic monocytes and granulocytes; of which elevated peripheral white blood cell levels have been described in atherosclerotic patients [1][2][3]. Although the link among HSPC proliferation in bone marrow (BM), leukocytosis in peripheral blood (PB), and accelerated atherosclerosis progression was noted, there is to date no evidence demonstrating the direct involvement of HSPC in plaque development.
It is well-known that HSPC reside in the hypoxic BM niche and give rise to all blood cells in postnatal life [4]. Under steady state conditions, a fraction of HSPC can be found in PB [5]. Circulating HSPC can migrate into peripheral tissues such as spleen, liver, lymph node, and aor-tic adventitia [6][7][8][9]. Although circulating HSPC migrate into lymph nodes and proliferate and differentiate in situ to become resident myeloid cells for host defense [6], the fate of HSPC that traffic into damaged tissues remains largely unknown. Nevertheless, there is evidence that when CD341 hematopoietic progenitor cells migrate into the injured spinal cord, exacerbated neuroinflammation and progression of multiple sclerosis are observed [10]. Likewise presence of HSPC in asthmatic airways is associated with asthma progression [11]. HSPC trafficking in injured liver, by contrast, appears to be associated with improved liver function, even if the mechanism is unknown [8,9]. We here tested the hypothesis that HSPC, under the influence of hypercholesterolemia, can home into the intima of inflamed arteries and contribute to arteriosclerotic plaques formation.
HSPC self-renewal versus differentiation and retention of HSPC within the BM niche are regulated by HSPC intrinsic factors as well as by extrinsic factors via engagement of a cytokine and growth factor receptors and adhesion receptors by their specific ligands. The integrins including a4b1, a5b1, b2, a6 and some chemokines such as C-X-C chemokine receptor type 4 (CXCR4) and CD192 (i.e. CCR2) play important roles in retention and homing of HSPC from blood to the BM niche [12][13][14][15], mobilization of HSPC from the BM into the PB [16][17][18][19], and trafficking into peripheral tissues [8][9][10]20]. In addition, CXCR4 has been identified as a key factor in hypercholesterolemia-induced HSPC mobilization from BM into PB [17,21,22]. However, it remains unknown what other key adhesion molecules are also affected and thus modify HSPC function by hypercholesterolemia. Expression and function of these adhesion receptors are regulated by a variety of signaling cascades, including the mitogen activated protein kinases (MAPK) signaling pathway [23][24][25]. Recently, Yves-Charvet reported extracellular signal regulated kinases 1/ 2(ERK1/2) activation in HSPC from hypercholesterolemic Abca1 2/2 Abcg1 2/2 mice [3]. In line with this study, our data demonstrated that low-density lipoprotein (LDL)-mediated differentiation of HSPC to granulocytes occurs in response to LDL-stimulated ERK1/2 activation [2]. This led us to determine if LDL affects integrin function and hence migration of HSPC into arteriosclerotic plaques via activation of the ERK pathway.
LDL receptor-related protein (LRP) is a member of the LDL-receptor family. It is expressed in a variety of cell types such as hepatocytes and leukocytes. More than 30 ligands have been discovered which explains the multiple functions of LRP [26]. Like other members in this family, LRP1 mediates cholesterol uptake via endocytosis. Aside from its function in cholesterol homeostasis, LRP1 has been found to interact with integrin b2 in leukocytes and therefore modulate integrin clustering on the membrane [27]. LRP1 deficiency abrogated integrin b2-dependent adhesion of leukocytes to endothelial cells [28]. Interestingly, an intimate association between LRP1 expression and ERK phosphorylation has been noticed in different cell types, all of which modulate cell adhesion and migration [29][30][31]. However, it is currently unknown if LRP1 regulates HSPC adhesion, migration or homing.
Here we report that hypercholesterolemia increased the percentage of integrin b 2 1 Lin 2 Sca-1 1 cKit 1 (LSK) cells in LDLr 2/2 mice. Integrin b 2 regulated LSK cell adhesion and migration toward to ICAM and homing to injured artery. Grafted integrin b 2 1/1 LSK cells resulted in enhanced inflammation and neointima formation in the ligated artery, compared to injection of PBS and integrin b 2 2/2 LSK cells. Finally, we demonstrate that LDL effects on integrin b 2 expression and function are mediated by the ERK/ADP-ribosylation factor 1 (ARF1)-dependent and ERK-independent LRP1 pathway.
To determine if b 2 on LSK cells was functional, LSK cells from LDLr 2/2 mice on chow and HFD were isolated by fluorescence-activated cell sorting (FACS), incubated with or without a blocking anti-CD18 Ab or IgG control Ab for 20 minutes prior to the adhesion assay. LSK cells were plated on intercellular adhesion molecular-1 (ICAM-1)-coated 96-well plate for 2 hours. After nonadherent LSK cells were removed, adherent LSK cells were mixed with beads and enumerated by FACS. LSK cells from LDLr 2/2 mice on HFD displayed a 2.6fold greater adhesion capacity to ICAM-1 compared with those from mice on chow diet, which was inhibited by a blocking anti-CD18 Ab (n 5 5-8, Fig. 1D).
To further evaluate whether increased integrin b2 expression was associated with enhanced LSK migration over ICAM-1, Lin2 cells were isolated from LDLr 2/2 mice on chow and HFD, treated with phosphate buffered saline (PBS), anti-CD18 or IgG control Abs for 20 minutes, and then plated in transwells coated with ICAM-1. After 6 hours, cells in the lower chamber were collected, counted, and stained with anti-LSK Abs. The input Lin2 population was also stained with LSK Abs to enumerate the absolute number of LSK cells that had migrated to the lower chamber. LSK cells from HFD mice displayed a 1.6-fold higher migration capacity through ICAM-1coated transwells compared to those from mice on chow diet. Anti-CD18 Abs inhibited LSK migration from mice on chow diet and HFD (n 5 5-11, Fig. 1E).

HSPC Could Home to Inflamed Arteries via Integrin b2
We next assessed the role of integrin b 2 on HSPC homing to injured arteries in vivo. Ligation of carotid artery has been shown to induce ICAM-1 expression on injured endothelial cells [32,33]. Therefore, this model was used in the entire study. Lin2 cells isolated from LDLr 2/2 mice on HFD were labeled with PKH26 and then injected intravenously into mice wherein the carotid artery was ligated 3 days earlier. Prior to injection, cells were incubated with PBS, IgG Ab, or anti-CD18 Ab for 20 minutes. The following day, the injured and Wang, Gao, Schouteden et al.  noninjured carotid arteries were dissected after extensive perfusion to remove blood cells in the arteries, digested, and stained with anti-LSK Abs. FACS data demonstrated that homing of LSK cells to injured artery was blocked by CD18 antibody (% homed LSK cells: PBS group: 0.120% 6 0.0107%; IgG group: 0.118% 6 0.0228%; CD18 group: 0.053 6 0.0059; n 5 5-8) ( Fig. 2A, 2B).
To explore the kinetics of homed LSK cells to injured arteries, carotid artery ligation was performed on CD45.1 mice and then they were received Lin2 cells isolated from LDLr 2/2 mice on HFD. One and ten days after cell injection, CD45.1 mice were sacrificed, and both injured and uninjured carotid arteries were dissected. After digestion, artery cells were numerated and stained with anti-mouse CD45.2 and LSK markers for FACS. The absolute number of homed LSK cells in uninjured and injured arteries was calculated. The number of donor-derived LSK cells was higher in injured arteries compared to uninjured arteries at both day 1 and day 10 after injection (day 1: To prove that LSK cells homed to the injured arteries were HSPC, PKH261 LSK cells were isolated by FACS and injected together with CD45.1 BMCs to irradiated CD45.1 recipients. After 16 weeks of BM transplantation, recipient blood cells were stained with anti-CD45.1 FITC, anti-CD45.2 PerCP-Cy5.5 together with myeloid and lymphoid markers for chimerism analysis. CD45.2-derived granulocytes, monocytes, B cells, and T cells were identified by FACS (n 5 5, Fig. 2C). These data indicate that homed PKH26 1 LSK cells are indeed HSPC.
Injection of CD18 1/1 LSK Cells Resulted in More Neointima Formation than CD18 2/2 LSK Cells To assess the fate of HSPC homing to injured artery, we grafted HSPC to immunodeficient Balb/C Rag2 2 ÇC 2/2 recipients (H-2kd) that underwent a complete ligation of the right carotid artery. Three days after ligation, splenectomy was performed to improve HSPC homing (Fig. 4A). Seven days after artery ligation, 40,000 LSK cells isolated from CD18 1/1 or CD18 2/2 mice were injected to Balb/C Rag22 ÇC 2/2 recipients. Four weeks after ligation surgery, recipients were sacrificed and the histology of the plaques was examined. H&E analysis showed that administration of CD18 1/1 LSK cells resulted in greater neointima formation compared to mice injected with PBS or CD18 2/2 LSK cells (n54-7, Fig. 4B, 4C). Cryosections were stained with antibodies against H-2kb and CD45 and inflammation index was assessed. Quantification of CD451 cells illustrated that mice that were injected with CD18 1/1 LSK cells developed aggravated neointima inflammation compared to mice received PBS and CD18 2/2 LSK cells (Fig. 4D). In addition, the percentage of H-2kb1 CD451 cells among CD451 cells was twofold higher in mice injected with CD18 1/1 LSK cells compared to mice received CD18 2/2 LSK cells (Fig. 4E). Identification of donor-derived CD451 cells, that is, H-2kb 1 CD45 1 cells was illustrated in Figure 4F.
We established in vitro assays to explore the effects of lipoproteins on LSK cells. Lin2 cells were isolated from LDLr 2/2 mice and exposed to equal amount of LDL or HDL (0-100 mg/ ml) for 24 hours. Cells were harvested and stained with antimouse CD18 and LSK markers for FACS as described above.
The percentage of CD181 LSK cells in LSK population was 1.4-fold higher at 25 mg/ml LDL, 1.5-fold higher at 50 mg/ml LDL, and 1.6-fold higher at 100 mg/ml LDL than non-LDL treated cells. Distinct from LDL, HDL treatment did not increase CD18 expressing LSK cells in vitro (Supporting Information Fig. 3).
We next assessed if pERK is involved in the induction of b 2 integrin expression by LDL. Lin2 cells were cultured with 100 mg/ml LDL in the presence or absence of pERK inhibitor, U0126, for 24 hours. Cells were stained with Abs against LSK and integrin b 2 . As was seen in vivo, LDL increased the fraction of LSK cells that expressed CD18 (n 5 5-6, Fig. 5C). Culture of LSK cells with LDL in combination with U0126 attenuated these changes (Fig. 5C). By contrast, incubation with LDL did not alter the percentage of integrin a5 1 LSK cells nor the percentage of integrin b 1 1 LSK cells (%integrin a5 1 LSK cells: 89% 6 3.0% vs. 89% 6 1.8%; %integrin b 1 1 LSK cells: 96.7% 6 1.16% vs. 95.8 6 1.74, n 5 4).
Presentation of adhesion molecules on the cell membrane is a complex process which starts with gene transcription, translation, and transportation. The small G protein ARF1 has been shown to mediate receptor trafficking to the membrane, including the chemokine receptor CXCR4, integrin b1, and CD11a [35][36][37][38]. We thus evaluated whether LDL induced any change in ARF1 expression in Lin2 cells. Western blot data illustrated that exposure of LDL to Lin2 cells increased ARF1 expression in a pERK-dependent manner (Fig. 5D, 5E). To dissect whether ARF1 facilitated CD18 expression or function, Lin2 cells were exposed to PBS or 5 mM Brefeldin A (BFA), an ARF1 inhibitor [36], for 8 hours and then stained with anti-CD18 and LSK Abs. BFA treatment neither altered the percentage of LSK cells in the population (control: 3.8% 6 0.44%; BFA: 3.5% 6 0.59%, n 57) nor led to significant change of CD18 expression on LSK cells. However, blocking ARF1 expression by BFA significantly reduced LSK cell adhesion to ICAM-1 in vitro and LSK cell homing to ligated arteries in vivo (n 5 6-7, Fig. 5F, 5G).

LRP1 is also in Part Responsible for LDL-Mediated
Integrin b2 Induction on HSPC LRP1 belongs to the LDL receptor superfamily. LRP1, together with LDLr, regulate cholesterol homeostasis via endocytosis. Different from LDLr, LRP1 has multiple functions such as regulation of integrin expression and function [27,28,39]. We first determined if LRP1 is expressed on Lin2 cells by Western blot analysis. LRP1 could not be detected in Lin2 cells under control conditions but was induced by exposure of Lin2 cells to LDL. By contrast, exposure to LDL did not significantly affect LDL receptor expression in Lin2 cells (Fig. 6A-6C). Of note, inhibition of ERK by U0126 did not prevent induction of LRP1 expression in Lin2 cells. To further assess the possible involvement of LRP1 in integrinmediated HSPC homing to injured arteries, we isolated LSK cells from LRP1 mutant mice (LRP1 n2/n2 ) [31] and repeated the function assays described above. LSK cells isolated from LRP1 n2/n2 mice showed decreased adhesion to ICAM-1 in vitro and reduced homing to injured arteries in vivo, compared to WT LSK cells (n 5 5-7, Fig. 6D, 6E). These data indicate that LRP1 is at least in part responsible for the regulation of integrin b2 function, while induction of LRP1 by LDL is ERK-independent.

DISCUSSION
Leukocytosis and monocytosis contribute substantially to arteriosclerotic progression [40][41][42]. Others and we have previously described that hypercholesterolemia induces HSPC proliferation and differentiation, resulting in expansion of the inflammatory cell pool in PB. However, it remains unclear if and how HSPC can migrate into lesion where they may locally expand and differentiate into leukocytes that further reinforce plaque formation. Till present, CXCR4 is the main factor identified to be responsible for HSPC mobilization in hypercholesterolemia. It would be very interesting to explore if other key factors that are modulated and in turn regulate HSPC function in the context of hypercholesterolemia. Therefore, we assessed if (a) HSPC directly contribute to arteriosclerosis progression; (b) hypercholesterolemia and LDL modulate Inflammation of blood vessels, which underlies the formation of atherosclerotic plaques, leads to upregulation of ICAM-1, the ligand for integrin b 2 . It is well-known that integrin b2 facilitates neutrophil and monocyte attachment and then transmigration through inflamed endothelium, which is why integrin b2 deficiency attenuates neointima formation in mice [43][44][45]. We demonstrate here that HSPC express significantly higher levels of CD18, when harvested from hypercholesterolemic mice, or following exposure of HSPC to LDL in vitro. We further demonstrate that homing of LSK to injured arteries occurs, which can be inhibited with blocking anti-CD18 antibodies. Grafting CD18 1/1 LSK cells to recipients accelerated inflammation and arteriosclerosis in the injured artery, which is partially due to donor-derived inflammatory cells in the lesion. These data implicate the critical regulation of integrin b 2 in HSPC adhesion, migration, and homing and further identify pro-atherogenic property of HSPC in arteriosclerosis development. Due to technical limitations, it is at present not feasible to trace HSPC proliferation and differentiation following homing to the lesion. Nevertheless, we postulate that both LSK proliferation and differentiation might occur in situ. We previously demonstrated that pERK is responsible for LDL-mediated increased differentiation of HSPC in vitro. In this study, we reported that hypercholesterolemia is associated with pERK activation in LSK cells in vivo. We found that pERK is responsible for upregulation of the G protein ARF1, known to mediate receptor trafficking to the membrane such as CXCR4, integrin b1, and CD11a [35][36][37][38]. We here demonstrate that ARF1 is also responsible for induction of functional b 2 integrin on HSPC. We also identified a second signaling pathway, LRP1, to be involved in mediating integrin b 2 function on HSPC. Aside from LDL uptake, LRP1 also can affect integrin expression and/ or function. In macrophages, LRP1 interacts with multiple sites in a M b 2 , which results in integrin clustering leading to cell adhesion and migration [27,28]. LRP1 also plays a role in inducing maturation of the b1 integrin, causing a substantial increase of this receptor on myeloid cells [39]. In this study, we found that LDL induces expression of LRP1, and that deficiency of LRP1 significantly inhibits HSPC adhesion to ICAM-1 in vitro. LRP1 induction by LDL was, however, independent on pERK. Deficiency of LRP1 did not affect expression levels of b 2 but reduced the adhesion capacity of LRP1-defieicnt LSK cells to ICAM-1. This would be consistent with the fact that LRP1 can not only cause integrin maturation and therefore expression [23] but also integrin function, as was shown in macrophages [27]. How LDL induces LRP1 expression remains unknown. Likewise, whether other forms of modified LDLs, aside from LDL isolated from plasma used in this study, influence LRP1 expression would be of interest. Finally, as LRP1 has multiple functions, we cannot exclude that other functions of LRP1 may play additional role in HSPC biology.
Our proposed model of the molecular mechanisms underlying LDL-mediated effects on HSPC biology is depicted in Figure 7. In vivo hypercholesterolemia and in vitro exposure to LDL causes phosphorylation of ERK1/2 in LSK cells. Phospho-ERK is, as others and we previously showed, responsible for LSK proliferation and differentiation [2,3] and as shown here, adhesion, migration, and homing. Phospho-ERK induces HSPC migration via ARF1, which increases expression of functional b 2 integrin on LSK cells. LDL also increases integrin function, not expression, on HSPC by interacting with LRP1 known to increase integrin function [27]. The proposed model extends our knowledge on the pathogenesis of hypercholesterolemiaassociated arteriosclerosis in the context of HSPC, which may provide novel therapeutic targets for the treatment of arteriosclerosis.

SUMMARY
Hypercholesterolemia-associated leukocytosis accelerates arteriosclerosis progression. The link among HSPC proliferation and differentiation in the BM, leukocytosis, and reinforced arteriosclerosis has been noted in hypercholesterolemic mice. However, it remains undefined whether and how circulating HSPC could directly participate in arteriosclerosis. In addition, how hypercholesterolemia, specifically LDL, affects HSPC trafficking to injured arteries is largely unknown. Using LDLr 2/2 mice model, we identified that integrin b 2 was upregulated on HSPC of LDLr 2/2 on HFD. Increased integrin b 2 expression enhanced HSPC function including their adhesion and migration toward ICAM-1 and homing to injured arteries. Strikingly, grafting HSPC to injured arteries accelerated inflammation and arteriosclerosis progression. Taken together, HSPC bear proatherogenic property and could directly participate in arteriosclerosis. Hypercholesterolemia stimulates arteriosclerosis progression could be partially via its regulation of HSPC as well as mature myeloid cells.

ACKNOWLEDGMENTS
We express our sincere thanks to Evelyn De Schryver and Thomas Vanwelden for their fundamental technical assistance and to Paul Holvoet for his intellectual input. All authors have approved the manuscript submission. Y.F. is a postdoctoral fellow from the Fonds voor Wetenschappelijk Onderzoek-Vlaanderen (FWO). This work was supported by the FWO funding (G1508612N) to Y.M.F., Vanwayenberghe fonds, FWO funding (G085111N), NIH-PO1-CA-65493-06, and Odysseus funding to C.M.V.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest. lated by high-density and low-density lipoproteins in mice. PLoS One 2012;7: e47286.
3 Yvan-Charvet L, Pagler T, Gautier EL et al. ATP-binding cassette transporters and Figure 7. Proposed model of how hypercholesterolemia and LDL modulate HSPC homing to lesion site. Hypercholesterolemia and LDL stimulate ERK phosphorylation in HSPC. ERK activation promotes HSPC proliferation and differentiation into atherogenic myeloid cells. Apart from that, ERK activation enhances ARF1-mediated integrin b 2 function of HSPC. In parallel, LDL increases LRP1 expression on HSPC, which in turn potentiates integrin b2 activity. As results from these changes, HSPC can attach to ICAM-1-expressing endothelial cells in lesion site where they may further contribute to inflammation and arteriosclerosis progression. Abbreviations: ARF1, ADP-ribosylation factor 1; HSPC, hematopoietic stem/progenitor cell; ICAM1, intercellular cell adhesion molecular-1; LDL, low density lipoprotein; LRP1, LDL receptor-related protein 1; pERK, phospho ERK.