Abstract
This study examined the adhesive interactions of peripheral blood NK cells with P- and E-selectin and analyzed the effect of IL-12 on the binding of NK cells to these selectins. P-selectin glycoprotein ligand-1 (PSGL-1) is expressed on most resting and IL-12-activated NK cells. However, the percentage of resting NK cells bound to P-selectin-IgG was 15%, and that of activated NK cells bound to P-selectin-IgG was 65%. Furthermore, the number of IL-12-activated NK cells bound to P-selectin-transfected Chinese hamster ovary cells was significantly higher than that of resting NK cells under flow conditions. These interactions were abolished by the incubation of these NK cells with anti-PSGL-1 (PL-1) mAb. Thus, PSGL-1/P-selectin interaction is important in the binding of resting and activated NK cells to P-selectin. NK cells express sialyl-Lewisx (sLex) structure recognized by anti-sLex mAb (KM-93), and IL-12 activation of NK cells increased the mean fluorescence intensity of KM-93-reactive NK cells. Adhesion of IL-12-activated NK cells to E-selectin-transfected Chinese hamster ovary cells was stronger than that of resting NK cells under flow conditions. These interactions were reduced markedly by incubation with anti-sLex mAb. Thus, sLex is the major ligand of resting and activated NK cells for E-selectin. These findings indicate that IL-12 stimulation of NK cells promotes their adhesion activity to endothelial selectins.
The NK cell is a mononuclear lymphoid cell that acts as the first line of defense against foreign cells, microbes, and tumors via its spontaneous or Ab-dependent cytotoxic activity (1, 2). Leukocyte recruitment and migration into inflamed tissue and certain pathologic sites are subject to a multistep adhesion cascade (3, 4). The interaction of leukocytes with endothelium requires at least three steps: initial tethering; arrest and adhesion; and transendothelial migration. Interactions of the selectin family with carbohydrate ligands play an important role in the initial step of leukocyte rolling and adhesion to endothelial cells. E-selectin is the adhesion molecule induced on human vascular endothelium by inflammatory cytokines and tumor interstitial fluid (5). On breast cancer endothelium, the expression of E-selectin, P-selectin, and ICAM-3 increased in contrast to expression in normal tissues (6). Granulocytes, monocytes, and various myeloid leukemic cell lines adhere to E-selectin via the carbohydrate ligands sialyl-Lewisx (sLex)3 or sialyl-Lewisa and a sulfated Lewis ligand structure. Furthermore, L-selectin (CD62L), expressed on neutrophils and monocytes, plays a role in their adhesion to the E-selectin molecule (7, 8). In the adhesion of NK cells to endothelial E-selectin, the important ligand on NK cells for E-selectin is the sLex ligand (recognized by KM-93 and FH-6 mAb) (9). Furthermore, lymphocytes can bind to endothelial P-selectin via P-selectin glycoprotein ligand-1 (PSGL-1) (10, 11). It has been reported that NK cells can bind to P-selectin (10). However, the mechanisms involved in the regulation of NK cell adhesion to endothelial selectins under flow conditions have not been analyzed.
The biologic functions of NK cells are modulated by certain cytokines. IL-2 increases NK cell adhesion on endothelial cells (12). IFN-γ and IL-8 stimulate locomotive ability of NK cells (13, 14). In this study, we analyzed the effect of IL-12 on NK cell adhesion to endothelial selectins. IL-12 is a 70-kDa disulfide-linked heterodimeric glycoprotein composed of 35- and 40-kDa subunits. Some biologic properties of human IL-12 have been evaluated in vitro. IL-12 is able to enhance NK/LAK cell cytolytic activity, to augment cytolytic T cell responses, and to induce secretion of cytokines, particularly IFN-γ, from NK cells as well as T cells. Since NK cells have been implicated as antitumor effector cells and IFN-γ has been shown to have antitumor activity in animals, IL-12 has the potential to be used as an immunomodulatory cytokine in the therapy of malignancies (15, 16). Furthermore, IL-12 is a chemotactic factor and increases binding activity of NK cells to endothelium (17). However, the effect of IL-12 on the adhesion of NK cells to selectins has not been analyzed.
To better understand the effect of IL-12 on NK cell adhesion to vascular endothelium and on accumulation of NK cells to inflamed tissue and tumors, we studied the expression of sLex (recognized by KM-93 mAb) and PSGL-1 on IL-12-activated NK cells and analyzed the adhesion activity of activated NK cells to endothelial selectins under flow conditions.
Materials and Methods
Antibodies
PL-1 and PL-2 (anti-PSGL-1, mouse IgG1k) were gifts of Dr. R. P. McEver and Dr. K. L. Moore (University of Oklahoma Health Sciences Center, Oklahoma City, OK). KM-93 (mouse IgM) against sLex was purchased from Seikagaku-Kogyo (Osaka, Japan). H18/7 (anti-E-selectin, IgG2a) mAb and GA6 (anti-P-selectin, IgG1k) mAb were purchased from Becton Dickinson (San Jose, CA). UCHT-1 (anti-CD3, IgG1) mAb, M5E2 (anti-CD14, IgG2a), HIB19 (anti-CD19, IgG1) mAb, DREG-56 (anti-CD62L, IgG1κ) mAb, C8.6 (anti-human IL-12, IgG1) mAb, phycoerythrin-conjugated anti-CD16 (3G8, IgG1) mAb, isotype-matched control mouse IgG1κ, IgG2a, and IgM were purchased from PharMingen (San Diego, CA). FITC-labeled rabbit anti-mouse Ig and FITC-labeled anti-human IgG were purchased from Dako (Glostrup, Denmark). Control human IgG was purchased from Sigma Japan (Tokyo, Japan). Rat anti-mouse IgG1 MACS beads were purchased from Miltenyi Biotec (Auburn, CA).
Chemicals, reagents, and cytokines
Human IL-2 was a gift from Takeda Pharmaceutical (Osaka, Japan). Recombinant human IL-12 was purchased from R&D Systems (Minneapolis, MN).
Fractionation of cells
Mononuclear cells were isolated from heparinized peripheral blood collected by venipuncture from healthy donors. The blood was diluted with an equal volume of 0.9% NaCl and layered over Ficoll/Hypaque-Lymphoprep (Nycomed, Oslo, Norway). Mononuclear cells were enriched through depletion of cells that adhered to plastic by incubation at 37°C for 2 h, and plastic-nonadherent cells were harvested. To isolate NK cells, this population was incubated with anti-human CD3 and CD19 mAbs for 30 min at 4°C. The cells were washed twice in PBS with 1% BSA and further incubated with goat anti-mouse IgG-conjugated magnetic beads for 30 min at 4°C. The cells were then washed twice and applied to the column in a magnetic cell separator to deplete T cells and B cells. The negative fraction was harvested. Flow cytometric analysis revealed the resulting NK cell population contained CD16+ > 95%, 2% > CD3+, 2% > CD19+, and 2% > CD14+. Neutrophils were isolated from heparinized blood using Monopoly resolving medium (Dainippon Pharmaceutical, Osaka, Japan). Neutrophils were 85% pure as assessed by Wright-Giemsa staining.
In cell culture studies, NK cells (106 cells/ml) were cultured in AIM V medium (Life Technologies, Gaithersburg, MD). NK cells were activated with 0.5 ng/ml IL-12 in the culture medium for 96 h.
Cell lines
Chinese hamster ovary (CHO) cells expressing a stably transfected human P-selectin cDNA (CHO P-selectin) were obtained from Dr. G. R. Larsen (Genetics Institute, Cambridge, MA). E-selectin molecules expressing CHO cells were constructed as described previously (18). Transfected CHO cells were cultured in MEMα (Life Technologies) containing 10% FCS, antibiotics, l-glutamine, and essential amino acids.
FACS analysis of NK
For indirect immunofluorescence, resting NK cells and IL-12-activated NK cells were incubated with KM-93 or PL-1 for 30 min at 4°C. After being washed with PBS, cells were stained with FITC-conjugated anti-mouse Ig. Furthermore, the resting NK cells and IL-12-activated NK cells were preincubated on ice for 30 min with 10% human serum in PBS to block Fc receptor and washed three times with 10% human serum in PBS. These cells (106 cells) were incubated with P-selectin-IgG chimera (50 μl, 20 μg/ml) and 2 mM Ca2+, after preincubation with anti-human IgG-FITC (final concentration, 5.6 μg/ml) on ice for 1 h for conjugation of P-selectin-IgG with FITC. A P-selectin-IgG chimera was produced as described previously (19). In some experiments, NK cells were pretreated with anti-PSGL-1 mAb (10 μg/ml, for 30 min on ice) before incubation with P-selectin-IgG. In others, the cells were incubated with P-selectin-IgG in presence of anti-P-selectin mAb. For negative control staining, we used human IgG preincubated with anti-human IgG-FITC. Expression was analyzed on a Becton Dickinson FACScan with CellQuest analysis software.
Attachment of NK cells to E-selectin and P-selectin under flow conditions
Transfected CHO cells were cultured in capillary tubes. The model of flow conditions was described previously (19). A tube was attached to the end of a glass capillary tube, and the glass capillary tube was connected to a syringe pump (TERUMO, Tokyo, Japan) fitted with a 50-ml syringe to establish laminar flow. The wall shear stress was calculated by Poiseuille’s law of Newtonian fluids, with viscosity of 0.01 P (at room temperature). The wall shear stress in dynes/cm2 = [(mean flow velocity × 8)/(tube diameter × viscosity)]. The tube was mounted on an Olympus (Tokyo, Japan) inverted microscope. Interaction of NK cells with selectin-transfected CHO cells was observed and videotaped for 10 min.
In the flow system, 1 × 106 NK cells/ml of binding medium with 2 mM Ca2+ were caused to flow across transfected CHO cells in the capillary tubes using a syringe pump. The adhesion of NK cells to selectins expressed on transfected CHO cells was determined at variable rates of shear stress. The numbers of adherent cells was examined by counting four to six fields (×200) under an inverted microscope during the entire 10-min experiment.
Rolling velocities of at least 10 cells under each microscope field were measured. At a given shear stress, we observed four to six microscopic fields. Photographs were taken as real time images to mark the location of NK cells. The displacement of the center of individual cells was measured 10 s later.
For Ab inhibition assays, E-selectin-transfected CHO cells in the capillary tubes were first treated with anti-E-selectin mAb or isotype control mouse IgG2a (20 μg/ml for 30 min). P-selectin-transfected CHO cells were treated with anti-P-selectin mAb or isotype-matched control. Furthermore, NK cells were treated with anti-sLex mAb (KM-93, 50 μg/ml for 30 min), anti-CD62L (DREG-56, 20 μg/ml for 30 min), or anti-PSGL-1 mAb (PL-1 and PL-2, 10 μg/ml for 30 min) in some experiments.
The data shown in this study are the results of a representative experiment among more than three independent experiments. The error bar represents the SD. The results are expressed as means ± SD/field.
Statistical analysis
Data were compared through analysis of variance using the unpaired Student’s t test, and p < 0.05 was considered to represent a significant difference between group means.
Results
Flow cytometric analysis of PSGL-1 expression on resting and IL-12-activated NK cells and of binding of P-selectin-IgG to resting and IL-12-activated NK cells
We analyzed the expression of PSGL-1 on resting and IL-12-activated NK cells and the binding of soluble P-selectin-IgG to resting and IL-12-activated NK cells. IL-12-activated NK cells contained CD16+ > 95%, 3% > CD3+. Most resting and IL-12-activated NK cells expressed PSGL-1 recognized by PL-1 mAb (Fig. 1, A and B). The mean channel fluorescence value for PSGL-1 on resting NK cells was 143 ± 26, and that on activated NK cells was 161 ± 30. The percentage of resting NK cells bound to soluble P-selectin-IgG was ∼15% (Fig. 1,C). Binding of soluble P-selectin-IgG to NK cells was completely abolished in the presence of anti-P-selectin mAb (Fig. 1,E). The percentage of IL-12-activated NK cells bound to soluble P-selectin-IgG was about 65% (Fig. 1,D). Addition of anti-P-selectin mAb decreased the binding of soluble P-selectin-IgG to activated NK cells (Fig. 1,F). When NK cells were cultured with IL-12 in the presence of anti-IL-12 mAb (C8.6, 10 μg/ml), binding of soluble P-selectin-IgG was similar to that of resting NK cells (data not shown). Neutrophils, as a positive control, bound efficiently to soluble P-selectin-IgG (Fig. 1 G). Binding of soluble P-selectin-IgG to neutrophils was inhibited in the presence of anti-P-selectin mAb.
Expression of sLex on resting NK cells and IL-12-activated NK cells
We analyzed the expression of sLex on resting and IL-12-activated NK cells. For detection of sLex on NK cells, we used KM-93 mAb. KM-93 was reactive with resting and activated NK cells (Fig. 2). The mean channel fluorescence value for sLex on resting NK cells was 110 ± 28. On the other hand, the mean fluorescence value for sLex on IL-12-activated NK cells was 520 ± 130. The expression of sLex on NK cells incubated with IL-12 in the presence of anti-IL-12 mAb (C8.6 10 μg/ml) was similar to that of resting NK cells (data not shown). Thus, the expression of sLex on NK cells increased significantly by IL-12 activation.
Binding of resting and IL-12-activated NK cells to P-selectin-transfected CHO cells under flow conditions
We examined the adhesion activity of resting and IL-12-activated NK cells to P-selectin transfectant under flow conditions. Resting and activated NK cells tethered and rolled on P-selectin transfectant under shear stress between 0.8 and 2.5 dynes/cm2. At 3.0 dynes/cm2, the binding of NK cells to P-selectin transfectant was not significant (Fig. 3 A). The number of activated NK cells bound to P-selectin transfectant was higher than that of resting NK cells under shear stress below 1.5 dynes/cm2. Above 2.0 dynes/cm2, the difference between the number of adherent resting NK cells and adherent activated NK cells decreased. The interactions of resting and activated NK cells with P-selectin transfectant were completely abolished by pretreatment of P-selectin transfectant with anti-P-selectin mAb.
Next, we analyzed the rolling velocities of resting and activated NK cells under shear stress at 1.0, 1.5, and 2.0 dynes/cm2. The differences between the velocities of resting and activated NK cells were not significant under any shear stress (Fig. 3 B).
Effect of treatment with anti-L-selectin, anti-sLex, or anti-PSGL-1 (PL-1 or PL-2) mAbs on NK cell adhesion to P-selectin transfectant under flow conditions
We examined the P-selectin ligands of NK cells. Resting (Fig. 4,A) and IL-12-activated NK cells (Fig. 4 B) were preincubated with anti-L-selectin, anti-sLex, anti-PSGL-1 (PL-1 or PL-2), or control mAbs. The treated resting and activated NK cells were perfused across the P-selectin transfectant monolayers under shear stress at 1.5 dynes/cm2. Treatment of resting and activated NK cells with anti-PSGL-1 (PL-1) mAb reduced the level of NK cell adhesion to P-selectin transfectant. Treatment of resting and activated NK cells with anti-L-selectin, anti-sLex, or anti-PSGL-1 (PL-2) mAbs did not reduce the level of NK cell adhesion. Treatment with control isotype-matched mAb (IgG1κ) had no inhibitory effect (data not shown). The PL-1 epitope is within or distal to a serine/threonine-rich region of the PSGL-1 polypeptide that is cleaved by O-sialoglycoprotein endopeptidase, whereas the PL-2 epitope is more proximal to the membrane (11). PL-1, but not PL-2 inhibited the binding of NK cells to P-selectin. These findings indicated that PSGL-1 was the major ligand of NK cells for P-selectin, and the O-linked glycans on PSGL-1 on NK cells played an important role in P-selectin/PSGL-1 interaction.
Binding of resting and IL-12-activated NK cell to E-selectin-transfected CHO cells under flow conditions
We examined the adhesion activity of resting and IL-12-activated NK cells to E-selectin under flow conditions. Resting and activated NK cells tethered and rolled on E-selectin transfectant under shear stress between 0.8 and 2.5 dynes/cm2. At 3.0 dynes/cm2, the binding of NK cells to E-selectin transfectant was not significant (Fig. 5 A). The number of activated NK cells bound to E-selectin transfectant was higher than that of resting NK cells under shear stress below 1.5 dynes/cm2. Above 2.0 dynes/cm2, the differences between the number of adherent resting NK cells and adherent activated NK cells decreased. The interactions of resting and activated NK cells with E-selectin transfectant were completely abolished by pretreatment of E-selectin transfectant with anti-E-selectin mAb.
Next, we analyzed the rolling velocities of resting and activated NK cells under shear stress at 1.0, 1.5, and 2.0 dynes/cm2. The differences between the velocities of resting and activated NK cells were not significant under any shear stress (Fig. 5 B).
Effect of treatment with anti-L-selectin, anti-PSGL-1, or anti-sLex mAbs on NK cell adhesion to E-selectin transfectant under flow conditions
We examined the E-selectin ligands of NK cells. Resting (Fig. 6,A) and IL-12-activated (Fig. 6 B) NK cells were preincubated with anti-L-selectin, anti-PSGL-1, anti-sLex, or control mAbs. The treated resting and activated NK cells were perfused across the E-selectin transfectant monolayers under shear stress at 1.5 dynes/cm2. Treatment of resting and activated NK cells with anti-sLex mAb reduced the level of NK cell adhesion to E-selectin transfectant. Treatment of resting NK cells with anti-L-selectin mAb slightly reduced the level of adhesion, but the reduction was not significant. In contrast, treatment of resting NK cells with anti-PSGL-1 mAb and of activated NK cells with anti-L-selectin or anti-PSGL-1 mAb did not block adhesion. Treatment with control isotype-matched mAb (IgG1κ and IgM) was not inhibitory (data not shown). These findings indicated that sLex was the major ligand of NK cells for E-selectin.
Discussion
This study provides the first evidence that peripheral blood NK cells tether and roll on P- and E-selectin under flow conditions. Also, we analyzed the effect of IL-12 on NK cell adhesion to these selectins. PSGL-1 is the major ligand of leukocyte for P-selectin. PSGL-1 is expressed on hemopoietic cells (20). Most NK cells expressed PSGL-1, but only a few NK cells expressed the form of PSGL-1 that could bind to P-selectin (Figs. 1,C and 4A). PSGL-1 is expressed on most lymphocytic cells, while soluble P-selectin binding is evident in only some lymphocytic populations (21). During in vitro stimulation, the percentage of peripheral blood T lymphocytes bound to P-selectin increased. In our study, IL-12 activation increased the percentage of NK cells bound to soluble P-selectin-IgG from 15% to 65% (Fig. 1, C and D). Therefore, NK cells increased the expression of the P-selectin ligands by IL-12 activation. During activation, the difference between the mean fluorescence value for PSGL-1 on resting and activated NK cells was not high (Fig. 1, A and B). Other populations of lymphocytes are similar to NK cells. Most CD4+ T cells express PSGL-1. Although memory T cells or Th1 can bind to P-selectin, naive or Th2 cannot bind to P-selectin (22-24). In contrast, soluble-phase P-selectin can bind to PSGL-1 on almost all neutrophils and monocytes (11). Neutrophils and monocytes were reported to express the high affinity form of PSGL-1 for P-selectin (11). In our studies, the treatment of resting and activated NK cells with anti-sLex mAb did not reduce the level of adhesion to P-selectin (Fig. 4). In contrast, sLex structure of PSGL-1 molecule plays a role in neutrophil and HL-60 adhesion to P-selectin (20, 25). Fuhlbrigge et al. (26) have reported recently that cutaneous lymphocyte Ag (CLA)-positive and CLA-negative T cell populations tethered and rolled on P-selectin with comparable efficiency and that neither CLA nor sLex is required for P-selectin ligand activity on T cells. Alone et al. (27) reported that HECA-452 reactivity on different T clones did not correlate with the level of P-selectin binding activity. Thus, PSGL-1 expressed on NK cells may contain carbohydrate ligands other than the sLex structure for binding to P-selectin, and NK cells may require alteration of PSGL-1 glycosylation, induced by cytokine activation, for up-regulation of its adhesive function.
NK cells can roll on E-selectin via the sLex recognized by KM-93 mAb (Fig. 5,A). Pinola et al. (9) reported that NK cells adhered to E-selectin and expressed the sialylated Lex structure. This sLex epitope on NK cell was weakly reactive with the CSLEX-1 mAb (9, 28) but was identified strongly by KM-93 and FH-6 mAbs (Fig. 2,A) (8). Pinola et al. reported that treatment with KM-93 mAb partially inhibited NK cell adhesion to E-selectin but that HL-60 adhesion to E-selectin was completely inhibited by treatment with KM-93 mAb. In our study, under flow conditions, treatment of resting NK cells with KM-93 mAb resulted in 70% reduction of NK cell binding to E-selectin (Fig. 6 A). However, the omission of the number of adherent resting NK cells to E-selectin by treatment with anti-L-selectin mAb was not clear.
IL-12 activation of NK cells increased the expression of sLex and the level of adhesion to E-selectin under flow conditions (Figs. 2 and 5 A). Leung et al. (29) reported that IL-12 could increase expression of CLA on T cells activated with Ag or superantigen and thereby increase their efficiency of recirculation to the skin, including areas with vascular E-selectin. Furthermore, CD4+ T cells incubated in the presence of IL-12 and IFN-γ for the generation of Th1 could bind to E-selectin more effectively than those stimulated by IL-2 and IL-4 for the generation of Th2 (23, 24). Thus, IL-12 could increase the E-selectin ligands on T and NK cells and thereby promote their binding to E-selectin. Furthermore, Sasaki et al. (30) reported that fucosyltransferase VII transfected Namalwa KJM cells could bind to E-selectin and that this binding was blocked by anti-sLex mAb (KM-93). Wagers et al. have reported that transfection of fucosyltransferase VII cDNA could confer E-selectin binding ability with K-562 (31). IL-12 stimulation may increase the level of fucosyltransferase VII of NK cells.
NK cells are cytotoxic without prior sensitization for a wide range of target cells, including some microorganisms and virus-infected and transformed cells. and have a central role in the regulation of immune response and hemopoiesis. The cytotoxic activity and cytokine production of NK cells is regulated by IL-12. In recruitment, extravasation, and accumulation of leukocytes at certain pathologic sites, carbohydrate ligand/endothelial selectin interactions play an important role. Our study demonstrates that IL-12 increased the expression of sLex, E-selectin ligand, augmented the binding of PSGL-1 for P-selectin, and thereby increased the adhesion activity of NK cells to endothelial selectins under flow conditions. The IL-12 activation of NK cell adhesion is suggested to be important for migration and accumulation of NK cells into infection sites or tumor sites. At local infection sites, macrophages, Langerhans cells, or other cell types produce IL-1 and TNF-α. Local production of IL-1 and TNF-α can induce the expression of E-selectin on endothelium, and TNF-α can induce the expression of P-selectin on endothelium (32, 33). Furthermore, Melder et al. (5) reported that tumor interstitial fluid increased the expression of E-selectin and VCAM-1 on endothelium. Fox et al. (6) reported that the expression of E-, P-selectin, and ICAM-3 increased on breast cancer endothelium in contrast to normal breast endothelium. Under these pathologic conditions, it is suggested that IL-12 effectively enhances the infiltration of NK cells at target sites.
Acknowledgements
We thank Dr. Rodger P. McEver (University of Oklahoma Health Sciences Center) for supplying the PL-1 and PL-2 mAbs and Dr. Glenn R. Larsen (The Genetics Institute Inc.) for supplying P-selectin-transfected CHO cells.
Footnotes
This work was supported in part by grants from the Kanagawa Nanbyo Fundation and the Japan Health Sciences Foundation and by grants-in-aid from the Japanese Ministry of Education, Science, Sports and Culture.
Abbreviations used in this paper: sLex, sialyl-Lewisx; PSGL-1, P-selectin glycoprotein ligand-1; CHO, Chinese hamster ovary; CLA, cutaneous lymphocyte antigen.