PSGL-1 is highly expressed on Ly-6C^hi monocytes and a major determinant for Ly-6C^hi monocyte recruitment to sites of atherosclerosis in mice

Mice

ApoE^-/-/PSGL-1^-/--deficient mice (C57BL/6J background) were generated after crossing PSGL-1^-/- mice¹² with ApoE^-/- mice. Wild-type C57BL/6J (WT) and ApoE^-/- mice were from the Jackson Laboratory (Bar Harbor, ME). Mice were kept in a specific pathogen-free facility. All mouse experiments were approved by the Institutional Animal Care and Use Committees of the Oklahoma Medical Research Foundation and the University of Minnesota.

Flow cytometry

All antibodies were obtained from BD Biosciences (San Diego, CA) unless specified. Mouse peripheral blood was used for monocyte analysis. Leukocytes that were positive for myeloid marker CD11b but negative for the rest of the lineage markers were defined as monocytes.⁴ Anti-Ly-6C mAb (AL-21) was used to classify monocytes into subsets with either high expression of Ly-6C (Ly-6C^hi) or low expression of Ly-6C (Ly-6C^lo).⁴ A rat anti-mouse PSGL-1 mAb (4RA10) was used to measure PSGL-1 expression.

The chemokine receptor CX3CR1 was originally used to classify monocytes.³ In mice, the CX3CR1^loCD11b⁺ monocyte subset is the equivalent of Ly-6C^hi monocytes.³ To examine PSGL-1 expression and P- and E-selectin binding of CX3CR1^loCD11b⁺ monocytes, C57BL/6 CX3CR1^GFP/+ mice were used. In CX3CR1^GFP/+ mice, one CX3CR1 allele was replaced with the gene encoding green fluorescent protein (GFP).¹³

Human monocytes have CD14⁺CD16^- and CD14^-CD16⁺ subsets, which resemble murine Ly-6C^hiCX3CR1^lo and Ly-6C^loCX3CR1^hi subsets, respectively.³ Human leukocytes that were HLA-DR and CD14 double-positive, or HLA-DR and CD16 double-positive were analyzed for PSGL-1 expression using a mAb to human PSGL-1 (KPL1). The protocol was approved by the IRB committee of the University of Minnesota.

Flow cytometry was performed on a FACSCalibur (BD Biosciences). Data were analyzed using Summit Software v4.3 (Dako, Carpinteria, CA).

In vitro flow chamber assay

To obtain sufficient cells, Ly-6C^hi and Ly-6C^lo monocytes from murine spleens were used. The spleen-derived monocytes are surrogates for circulating monocytes.⁴ Ly-6C^hi and Ly-6C^lo monocytes were sorted from the monocyte population using the inFlux V-GS Cytometer Work Bench (Cytopeia, Seattle, WA).

Flow chamber experiments were carried out as described.¹¹ Briefly, murine P-selectin-IgM, E-selectin-IgM, control CD45-IgM, biotinylated 2-glycosulfopeptide-6 (2-GSP-6, gift from Dr. Richard Cummings, Emory University), or human L-selectin IgG chimera was captured on the dishes. 2-GSP-6 is modeled after the NH₂-terminal selectin-binding region of PSGL-1.¹⁴ Human L-selectin is the equivalent of murine L-selectin because they share the same binding activity to murine PSGL-1. Sorted Ly-6C^hi or Ly-6C^lo monocytes (0.5 × 10⁶/ml in HBSS with 0.5% HSA) were perfused over P-selectin, E-selectin, control CD45-IgM, 2-GSP-6, or L-selectin in dishes mounted in a parallel-plate flow chamber. Rolling cells were analyzed using a Silicon Graphics workstation (Silicon Graphics, Sunnyvale, CA).

Ex vivo perfusion of murine carotid arteries

WT or PSGL-1^-/- Ly-6C^hi or Ly-6C^lo monocytes were labeled with calcein AM (Molecular Probes, Eugene, Oregon), and were infused into the ApoE^-/- carotid arteries at a rate of 10 ml/min (1 × 10⁶ cells/ml), resulting in a wall shear stress of 3.0 ± 0.1 dyne/cm².¹⁰ Cell rolling and adhesion were recorded on videotape by stroboscopic epifluorescence illumination with an intravital microscope (Axioskop, 10× water immersion objective, NA 0.5, Carl Zeiss, Thornwood, NY).

Atherosclerosis and carotid artery wire injury models

To induce atherosclerosis, ApoE^-/-/PSGL-1^-/- or ApoE^-/- mice at 5 weeks of age were fed the Western diet for 12 weeks.⁵ Atherosclerotic lesions stained with Oil red O were quantified by en face analysis of the whole aorta and by cross-sectional analysis of the proximal aorta.¹⁵ For the carotid artery wire injury models, 8-week-old male ApoE^-/- mice were fed the Western diet for 2 weeks and carotid artery wire injury was performed.¹⁶ For quantification of neointimas, ten 5-μm aortic sections that were stained with Movat pentachrome (Sigma, St. Louis, MO) were analyzed for each mouse. Images were analyzed using the NIH Image software.¹⁰

Immunostaining

Cryosections of aortas from ApoE^-/- and ApoE^-/-/PSGL-1^-/- mice were stained with anti-F4/80 mAb (MOMA-2) (Accurate Chemical, Westbury, NY).¹⁷ Eight sections every 40 μm from each aorta root were examined. Digital images were used to quantify macrophages with NIH Image J software.

Ly-6C^hi monocytes in cryosections of ApoE^-/- and ApoE^-/-/PSGL-1^-/- carotid arteries were examined using an FITC-conjugated anti-mouse Ly-6C mAb (AL-21, 1:100 dilution, BD Biosciences) as described.⁴ Peripheral Ly-6C^hi or Ly-6C^lo monocytes, which were cytocentrifuged on glass slides, were stained with the same FITC-conjugated mAb and used as controls for the identification of Ly-6C^hi cells. Fluorescence was detected using a fluorescence microscope (4× objective, NA 0.3, Olympus America, Center Valley, PA).

Murine Ly-6C^hiCX3CR1^lo monocytes and human CD14⁺CD16^- monocytes express high levels of PSGL-1

We examined PSGL-1 expression on mouse Ly-6C^hi and Ly-6C^lo monocytes. We defined monocytes as CD11b⁺CD90^loB220^loCD49b^loNK1.1^loLy-6G^lo cells (Figure 1A).⁴ Within this monocyte population, cells were classified into Ly-6C^hi and Ly-6C^lo subsets based on their differential expression of Ly-6C (Figure 1B). The overall monocytes were not significantly different among ApoE^-/- and ApoE^-/-/PSGL-1^-/- mice fed the Western diet for 12 weeks and ApoE^-/- mice fed a standard chow (Figure 1A and data not shown). However, ApoE^-/- mice fed the Western diet had an elevated number of Ly-6C^hi monocytes compared to the number of Ly-6C^hi monocytes from ApoE^-/- mice fed chow (Figure 1B and C), which is consistent with published results.^{4, 5} The percentage of Ly-6C^hi monocytes was similar in ApoE^-/- and in ApoE^-/-/PSGL-1^-/- mice fed the high-fat diet (Figure 1C). Remarkably, PSGL-1 was expressed at a much higher level on Ly-6C^hi than on Ly-6C^lo monocytes from ApoE^-/- mice fed either chow or the Western diet (Figure 1D and E). The increase in PSGL-1 expression was not caused by the ApoE deficiency because Ly-6C^hi monocytes from WT mice also expressed more PSGL-1 (Figure 1E). Real-time PCR demonstrated that PSGL-1 mRNA transcripts in Ly-6C^hi monocytes were also expressed at a higher level than the transcripts in Ly-6C^lo monocytes (Figure 1F).

We confirmed that PSGL-1 was also highly expressed on CX3CR1^loCD11b⁺ monocytes compared to its expression on CX3CR1^hiCD11b⁺ monocytes (Figure 2A and B). In addition, CX3CR1^loCD11b⁺ monocytes exhibited greater binding to P- and E-selectin (Figure 2C and D) compared with CX3CR1^hiCD11b⁺ monocytes.

We also investigated PSGL-1 level on human monocytes. Interestingly, consistent with the expression patterns of PSGL-1 on murine monocytes, PSGL-1 level was also higher on human CD14⁺CD16^- monocytes than on CD14^-CD16⁺ monocytes (Figure 2E, n = 3, P < 0.01).

Ly-6C^hi monocytes interact preferentially with P-, E-, and L-selectin as well as 2-GSP-6 under flow

To evaluate the function of PSGL-1 on Ly-6C^hi monocytes, we first tested fluid-phase binding of P- and E-selectin to WT monocytes using flow cytometry. Compared with Ly-6C^lo monocytes, Ly-6C^hi monocytes had greater binding to P- and E-selectin (Figure 3A-B). The interaction between PSGL-1 and P-selectin was PSGL-1-dependent because 4RA10, a mAb that blocks PSGL-1 function, eliminated this interaction (Figure 3A). 4RA10 substantially reduced but did not abolish the PSGL-1 interaction with E-selectin (Figure 3B), which reflects E-selectin ligand activities other than PSGL-1 on Ly-6C^hi monocytes.¹⁸ Ly-6C^hi monocytes from ApoE^-/- mice fed the Western diet for 12 weeks had similar P- and E-selectin binding profiles (data not shown), suggesting that the 12-week high-fat diet did not significantly alter selectin ligand activities on monocytes. Similar levels of mRNAs of α1,3-fucosyltransferase VII (FTVII) and core 2 β1,6-glucosaminyltransferase-I (C2GlcNAcT-I), the protein products of which are essential for PSGL-1 function, were detected in Ly-6C^hi and Ly-6C^lo monocytes (Figure S1).

We then compared the rolling of Ly-6C^hi and Ly-6C^lo monocytes, respectively, on immobilized P-, E-, or L-selectin under flow conditions. Comparatively more Ly-6C^hi than Ly-6C^lo monocytes rolled on P-, E-, and L-selectin at low shear stress (0.5-1 dyn/cm²) (Figure 3C, D, and E). Ly-6C^hi cells rolled more stably on P- and L-selectin as manifested by their slow rolling velocities (μm/s, 1.37 ± 0.48 for Ly-6C^hi vs. 3.55 ± 1.22 for Ly-6C^lo on P-selectin; 8.9 ± 3.07 for Ly-6C^hi and 117.9 ± 4.38 for Ly-6C^lo on L-selectin, at 0.5 dyn/cm², P < 0.01, n = 15). There was no significant difference in rolling velocities on E-selectin between Ly-6C^hi and Ly-6C^lo cells, which is consistent with previous studies that PSGL-1 is important for tethering to E-selectin and other E-selectin ligand(s) such as CD44 other than PSGL-1 contributes to the slow rolling on E-selectin.^{12, 18} Rolling was PSGL-1- and selectin-dependent because mAbs to PSGL-1, P-, E-, or L-selectin reduced the number of rolling cells to the basal levels observed on the surfaces coated with control reagents (data not shown). Significantly, only Ly-6C^hi monocytes rolled on P- or E-, and L-selectin at relative high shear stress (>2 dyn/cm²) (Figure 3C, D, and E).

Because L-selectin was preferentially expressed on Ly-6C^hi monocytes (Figure 4), we also compared rolling of Ly-6C^hi and Ly-6C^lo monocytes on immobilized 2-GSP-6, which is modeled after the NH₂-terminal L-selectin-binding region of PSGL-1.¹³ At all shear stress tested, Ly-6C^hi cells rolled on 2-GSP-6 (32 ± 4 cells/mm² at 1 dyn/cm²) (Figure S2A). Rolling was PSGL-1-dependent as it was blocked by mAb 4RA10 to PSGL-1 (Figure S2A). In contrast, almost no Ly-6C^lo rolling on 2-GSP-6 was observed (Figure S2A and B). This experiment suggests that L-selectin may contribute to the selective homing of Ly-6C^hi cells to atherosclerotic lesions.

Ly-6C^hi monocytes roll preferentially on early atherosclerotic endothelium in a PSGL-1-dependent manner

To test whether Ly-6C^hi monocytes have increased capacity to interact with early atherosclerotic endothelium, we used the ex vivo perfusion of ApoE^-/- carotid artery model.¹⁰ In this model, rolling of monocytes on early atherosclerotic endothelium of ApoE^-/- mice is primarily mediated by P-selectin and vascular cell adhesion molecule 1 (VCAM-1).^{10, 19} Compared with Ly-6C^lo monocytes, more Ly-6C^hi monocytes rolled on atherosclerotic endothelium at 3 dyn/cm² (Figure 3F). Consequently, the number of adherent Ly-6C^hi monocytes was increased at all time points compared to Ly-6C^lo monocytes (Figure 3G). PSGL-1^-/- Ly-6C^hi monocytes had dramatically reduced rolling and adhesion on atherosclerotic endothelium, indicating primarily PSGL-1-dependent interactions. The observed residual rolling of PSGL-1^-/- Ly-6C^hi monocytes on endothelium may be from interactions between integrin VLA4 and VCAM-1.²⁰

High-fat diet does not change the expression profile of other adhesion molecules or the chemokine receptor CCR2 on Ly-6C^hi monocytes

To exclude the possibility that the increased recruitment of Ly-6C^hi monocytes into atherosclerotic lesions is a result of altered expression of other molecules resulting from consumption of the Western diet, we compared the expression of L-selectin, CD44, CD18 (b2 integrin), CD49d (VLA4), and CCR2 on both Ly-6C^hi and Ly-6C^lo monocytes. Consistent with previous reports,^{3, 5} expression of L-selectin and CCR2 was higher on Ly-6C^hi than on Ly-6C^lo monocytes, whereas CD18 and CD49d were lower on Ly-6C^hi than on Ly-6C^lo monocytes (Figure 4). Both Ly-6C^hi and Ly-6C^lo cells expressed a similar level of CD44 (Figure 4). Ly-6C^hi or Ly-6C^lo monocytes from WT or from ApoE^-/- mice fed either chow or the Western diet for 12 weeks express similar levels of these molecules (Figure 4), suggesting that diet does not affect the expression of these molecules.

ApoE^-/-/PSGL-1^-/- mice have reduced monocyte/macrophage accumulation in atherosclerotic lesions and develop smaller atherosclerotic plaques

Ly-6C^hi monocytes are key contributors to atherosclerosis.⁴ To examine whether PSGL-1 contributes to monocyte recruitment in atherosclerotic lesions and to the development of atherosclerosis, ApoE^-/- and ApoE^-/-/PSGL-1^-/- mice were fed the Western diet for 12 weeks. Compared to ApoE^-/- mice, ApoE^-/-/PSGL-1^-/- mice exhibited less monocyte/macrophage infiltration in the lesions (Figure 5A and B). In addition, the lesions in ApoE^-/-/PSGL-1^-/- aortas were significantly smaller than lesions in ApoE^-/- aortas (Figure 5C, D, E, and F). There was no significant difference between ApoE^-/- and ApoE^-/-/PSGL-1^-/- mice in their plasma lipid levels (Figure S3). These data support the important role of PSGL-1 in mediating Ly-6C^hi monocyte recruitment into atherosclerotic lesions and in the development of early atherosclerotic plaques.

Lack of PSGL-1 results in impaired Ly-6C^hi monocyte recruitment to injured arterial walls and reduces formation of neointimal lesions

Monocytes are critical in arterial neointimal formation.²¹ We examined whether PSGL-1 deficiency reduces monocyte infiltration in neointima and the size of the lesions. Our experiments showed that ApoE^-/-/PSGL-1^-/- mice had less monocyte/macrophage infiltration in the neointima of carotid arteries after wire injury compared to ApoE^-/- mice (Figure 6A and B). To examine Ly-6C^hi monocyte homing to injured arteries, cryosections of ApoE^-/- and ApoE^-/-/PSGL-1^-/- carotid arteries, which were harvested 7 days after wire-induced injury, were stained using an anti-murine Ly-6C mAb as described.⁴ In this model, P-selectin expressed by regenerating endothelial cells and adherent platelets contributes to the monocyte recruitment into injured carotid arterial wall.²² We found that ApoE^-/- carotid arteries had substantial accumulation of Ly-6C^hi monocytes after wire-induced injury (Figure 6C). In contrast, Ly-6C^hi monocytes were rarely observed in wire-injured ApoE^-/-/PSGL-1^-/- carotid arterial sections (Figure 6D). These data suggest the critical contribution of PSGL-1 to the recruitment of Ly-6C^hi monocytes to neointimal lesions.

We also used this model to investigate the role of PSGL-1 in neointima formation. The average size of neointima in the carotid arteries of ApoE^-/-/PSGL-1^-/- mice after wire injury was dramatically reduced compared to that in ApoE^-/- mice (Figure 6E and F), indicating an important role for PSGL-1 in this pathology.