Sickle red cells induce adhesion of lymphocytes and monocytes to endothelium

Abstract

Infusion of epinephrine-activated human sickle erythrocytes (SS RBCs) into nude mice promotes both SS RBC and murine leukocyte adhesion to vascular endothelium in vivo. We hypothesized that interaction of epinephrine-stimulated SS RBCs with leukocytes leads to activation of leukocytes, which then adhere to endothelial cells (ECs). In exploring the underlying molecular mechanisms, we have found that coincubation in vitro of epinephrine-treated SS RBCs with human peripheral blood mononuclear cells (PBMCs) results in robust adhesion of PBMCs to ECs. Sham-treated SS RBCs had a similar but less pronounced effect, whereas neither sham- nor epinephrine-treated normal RBCs activated PBMC adhesion. PBMC activation was induced via at least 2 RBC adhesion receptors, LW and CD44. In response to SS RBCs, leukocyte CD44 and beta2 integrins mediated PBMC adhesion to ECs, a process that involved endothelial E-selectin and fibronectin. SS RBCs activated adhesion of both PBMC populations, lymphocytes and monocytes. Thus, our findings reveal a novel mechanism that may contribute to the pathogenesis of vaso-occlusion in sickle cell disease, in which SS RBCs act via LW and CD44 to stimulate leukocyte adhesion to endothelium, and suggest that RBC LW and CD44 may serve as potential targets for antiadhesive therapy designed to prevent vaso-occlusion.

Figures

Figure 1

SS RBCs, but not normal RBCs, induced increased PBMC adhesion to HUVECs. (A,B) Sham-treated (A) or epinephrine (epi)–treated (B) SS RBCs (green) were coincubated with PBMCs (red). PBMC-RBC mixtures were then tested for adhesion to HUVECs. Photomicrographs using ×20 magnification show that sham-treated SS RBCs (A), which adhered to some degree to nontreated HUVECs, induced adhesion of PBMCs to HUVECs. Epinephrine-treated SS RBCs (B), which adhered strongly to nonactivated HUVECs, also induced PBMC adhesion to HUVECs. Photomicrograph of adhesion of PBMCs (not coincubated with SS RBCs) is not shown because such PBMCs did not visibly adhere to HUVECs. (C) PBMCs were tested for adhesion to HUVECs in the presence of sham- or epi-treated SS RBCs, or after lysis of sham- or epi-treated SS RBCs. Results are presented as percentage of adherent PBMCs at a shear stress of 1 dyne/cm2. Error bars show SEM of 3 different experiments. *P < .001 compared with unstimulated PBMCs. (D) PBMCs separated from blood obtained from 2 different donors (donor 1, lanes 1 and 2; donor 2, lanes 3 and 4) were analyzed alone (lanes 1 and 3) or after coincubation with ABO-matched SS RBCs (lanes 2 and 4). For PBMC-RBC mixtures, after 30 minutes of incubation, cells were treated with RBC lysis buffer, then washed free of lysed RBCs. Proteins (50 μg protein per lane) obtained from both types of PBMC preparations, as well as RBCs only (lane 5), were then analyzed for the presence of RBC proteins by Western blot using antiglycophorin C and anti-CD44 (a positive control for both RBCs and leukocytes) antibodies, and P3 myeloma protein as a negative control (data not shown). Western blot analysis showed that, after RBC lysis, leukocyte preparations were free of detectable RBC proteins. (E) PBMCs not coincubated with SS RBCs did not significantly adhere to HUVECs previously coincubated with SS RBCs for 15 minutes. Similarly, the supernatant potentially containing free heme and reactive oxygen species obtained from lysed SS RBCs did not induce PBMC adhesion to ECs. Error bars show SEM of 3 different experiments. *P < .001 compared with unstimulated PBMCs. (F) Separation of SS reticulocytes (retic) and mature SS RBCs was accomplished using anti–transferrin receptor mAb 5E9 and goat anti–mouse IgG-coated magnetic microbeads. PBMCs were coincubated with sham-treated mature SS RBCs, epi-treated mature SS RBCs, sham-treated SS retic, epi-treated SS retic, or epi-treated unseparated SS RBCs. Adhesion of PBMCs to HUVECs was then tested. Results are presented as percentage of adherent PBMCs at a shear stress of 1 dyne/cm2. Error bars show SEM of 3 different experiments. *P < .05 compared with unstimulated PBMCs; **P < .001 compared with PBMCs coincubated with sham-treated mature SS RBCs; ***P < .001 compared with PBMCs coincubated with sham-treated SS retic. (G) Adhesion of PBMCs to HUVECs after coincubation with sham- or epi-treated normal RBCs versus sham- or epi-treated SS RBCs, respectively. One representative experiment is presented (n = 3).

Figure 2

Activation of PBMC adhesion to endothelium is induced by SS RBC LW and CD44. Inhibition of PBMC interaction with epi-treated SS RBCs was attempted by preincubation of SS RBCs with antibodies against the RBC receptors CD47, LW, and CD44; P3 was used as a nonreactive control antibody. Adhesion of PBMCs coincubated with such epi-treated SS RBCs was then assayed. Results are presented as percentage of adherent PBMCs at a shear stress of 1 dyne/cm2. Error bars show SEM of 3 different experiments. *P < .001 compared with unstimulated PBMCs; **P < .001 compared with PBMCs coincubated with SS RBCs preincubated with P3, then treated with epi.

Figure 3

Leukocyte β2 integrins and CD44 are involved in interaction with SS RBCs. (A,B) Inhibition of PBMC interaction with epi-treated SS RBCs was performed as described in “Inhibition assays.” Results are presented as percentage of adherent PBMCs at a shear stress of 1 dyne/cm2. (A) PBMC adhesion was measured after preincubation of PBMCs with anti-β2 integrin antibody followed by exposure to epi-treated SS RBCs. Error bars show SEM of 3 different experiments. *P < .001 compared with unstimulated PBMCs; **P < .001 compared with PBMCs preincubated with P3. (B) PBMC adhesion was measured after exposure to epi-treated SS RBCs, or epi-treated SS RBCs preincubated with sCD44 or sLW. Error bars show SEM of 4 different experiments. *P < .001 compared with unstimulated PBMCs; **P < .001 compared with PBMCs coincubated with SS RBCs preincubated with sLW, then treated with epi. (C) Phosphorylation of leukocyte CD44. Inorganic 32P radiolabeled intact PBMCs were coincubated or not with sham-treated SS RBCs for 15 minutes or 30 minutes, or with epi-treated SS RBCs for 15 minutes or 30 minutes. Leukocyte CD44 was immunoprecipitated (IP) with anti-CD44 antibody or P3 as a control, as indicated. RBC CD44 was IP from sham-treated SS RBCs with anti-CD44 antibody. The cpm shown are quantitative data of the radioactive band at 85 kDa. One representative experiment is shown (n = 3). (D) Total immunoprecipitates (lanes 1-8) obtained using the same conditions as described in panel C but immunostained with A3D8 mAb against CD44.

Figure 4

Activated leukocyte β2 integrins and CD44 are involved in PBMC adhesion to ECs. Inhibition of PBMC adhesion with antibody (A,B) and recombinant protein (C) was performed as described in “Inhibition assays.” For all experiments, PBMC adhesion was activated by epi-treated SS RBCs as indicated. Results are presented as percentage of adherent PBMCs at a shear stress of 1 dyne/cm2. Error bars show SEM of 3 different experiments for panels A and B and SEM of 4 different experiments for panel C. (A) Inhibition of PBMC adhesion was performed by incubation of activated PBMCs washed free of lysed epi-treated SS RBCs with anti-β2 integrin antibody. P3 protein was used as a nonreactive control for both panels A and B. *P < .001 compared with unstimulated PBMCs; **P < .001 compared with activated PBMCs preincubated with P3. (B) Inhibition of PBMC adhesion was performed by incubation of activated PBMCs washed free of lysed epi-treated SS RBCs with anti-CD47, anti-LW, or anti-CD44 antibody. *P < .001 compared with unstimulated PBMCs; **P < .001 compared with activated PBMCs preincubated with P3. (C) Confluent cultures of HUVECs were incubated without recombinant protein or with sLW or sCD44 protein, washed, and then tested for their ability to support adhesion of activated PBMCs washed free of lysed epi-treated SS RBCs. *P < .001 compared with unstimulated PBMCs; **P < .001 compared with PBMCs coincubated with epi-treated SS RBCs.

Figure 5

Endothelial cell E-selectin and fibronectin are involved in adhesion of activated PBMCs. (A,B) HUVECs were preincubated with or without antibody against CD44, αvβ3 integrin, thrombospondin, fibronectin, E-selectin, or P-selectin. P3 protein was used as a nonreactive control antibody in panel B. HUVECs were washed and then tested for their ability to support adhesion of PBMCs activated with epi-treated SS RBCs. (A) Error bars show SEM of 4 different experiments measuring adhesion at a shear stress of 1 dyne/cm2. *P < .001 compared with unstimulated PBMCs; **P < .01 compared with PBMCs preincubated with epi-treated SS RBCs. (B) Error bars show SEM of 3 different experiments measuring adhesion at a shear stress of 1 dyne/cm2. *P < .001 compared with unstimulated PBMCs; **P < .01 compared with PBMCs preincubated with epi-treated SS RBCs.

Figure 6

SS RBCs induce increased adhesion of both lymphocytes and monocytes to HUVECs. Lymphocytes (lymphs) and monocytes (monos) were isolated as described in “Adhesion of lymphocytes and monocytes.” Results are presented as percentage of adherent lymphocytes (A) or monocytes (B) at a shear stress of 1 dyne/cm2. Error bars show SEM of 3 different experiments for both panels A and B. (A) Sham-treated SS RBCs or epi-treated SS RBCs were coincubated with isolated lymphocytes. Lymphocytes were then tested for adhesion to HUVECs. *P < .05 compared with unstimulated lymphs. (B) Sham-treated SS RBCs or epi-treated SS RBCs were coincubated with isolated monocytes. After incubation, monocytes were then tested for adhesion to HUVECs. *P < .01 compared with unstimulated monos.