Intravenous immune globulin prevents venular vaso-occlusion in sickle cell mice by inhibiting leukocyte adhesion and the interactions between sickle erythrocytes and adherent leukocytes

Sickle cell disease results from a single missense mutation in the β-globin gene that alters the solubility of the deoxygenated molecule, leading to polymer formation, red blood cell (RBC) sickling, and vaso-occlusion.¹  Previous studies demonstrated that sickle RBCs are more adherent to vascular endothelium and that the endothelium in sickle cell patients and mice is inflamed.^2-5  Clinical studies have suggested a link between elevations in white blood cell (WBC) counts and poor outcome in sickle cell disease,⁶  and in vitro data indicate that sickle cells can bind neutrophils in a static adhesion assay.⁷  Recently, intravital microscopy studies in sickle cell mice revealed that WBC adhesion in venules played a direct role in sickle cell vaso-occlusion by interacting with sickle RBCs.⁸  Thus, sickle cell vaso-occlusion appears to involve multicellular interactions among RBCs, WBCs, and the venular endothelium.⁹  In the course of experiments using monoclonal antibodies to evaluate the role of adhesion receptors in this phenomenon, we noticed that control antibodies inhibited RBC-WBC interactions. Here, we demonstrate that high doses of commercial intravenous human immune globulin (IVIG) dramatically inhibit RBC-WBC interactions and improve blood flow and survival of sickle cell mice subjected to tumor necrosis factor-α (TNF-α)–induced vaso-occlusion.

Sickle cell (SS) mice,¹⁰  a kind gift from Dr Narla Mohandas (New York Blood Center, New York), were bred at Mount Sinai School of Medicine. Genetically identical cohorts of male sickle mice were generated by bone marrow transplantation as described previously.⁸  Mice carrying more than 97% donor chimerism, as assessed using acid-urea gels,¹¹  were used for intravital microscopy studies. All experimental procedures were approved by the Animal Care and Use Committee of Mount Sinai School of Medicine.

Human γ-globulin (IVIG) was a gift from Bayer (Elkhart, IN) and diluted in phosphate-buffered saline (PBS) for injection. Fab and Fc fragments were prepared from IVIG using a protein G column (Pharmacia, Uppsala, Sweden) as described.¹² 

SS mice were injected (200 μL) with either IVIG, Fab or Fc antibody fragments, immunoglobulin G (IgG)–free albumin (Sigma, St Louis, MO), or control vehicle (endotoxin-free PBS) via the lateral tail vein 20 minutes before the start of the surgery. Following anesthesia with a mixture of 2% α-chloralose and 10% urethane in PBS (6 mL/kg), a polyethylene tube was inserted via tracheotomy to facilitate spontaneous respiration, and the cremaster muscle was prepared for intravital microscopy as described.⁸  Fifteen minutes following surgery, venules were videotaped for 75 minutes (intravital microscopy viewing period 1: IVM 1), after which 0.5 μg murine recombinant TNF-α (rTNF-α) (R&D Systems, Minneapolis, MN) was injected intraperitoneally. After exposure to TNF-α for 90 minutes, the same venules were recorded for 90 additional minutes (IVM 2).

The numbers of rolling and adherent leukocytes were quantitated by playback analyses of videotapes as previously described.⁸  RBCs were identified based on their size and shape (discoid and sickle-shaped cells). An RBC-WBC interaction was defined as the arrest of an RBC on an adherent WBC for more than 2 video frames (more than 0.07 seconds). This time interval was selected because it correlates with a readily discernable adhesion event when the videotape is played in real time. RBC-WBC interactions were counted over a 100-μm venular segment for 3 1-minute intervals per venule.

Statistical significance for nonparametric values (RBC-WBC interactions) was assessed using the Kruskal-Wallis test followed by the Mann-Whitney test. Analysis of variance (ANOVA) with Bonferroni/Dunn correction of P was used for parametric values.

Surgical preparation of cremaster muscle for intravital microcopy invariably triggers an inflammatory response leading to rolling and adhesion of leukocytes to venular endothelium.¹³  Thereafter, discoid and sickle-shaped RBCs interact with adherent WBCs.⁸  Administration of IVIG reduced both the number of adherent WBCs attached to the endothelium (Table 1) and the number of interactions between RBCs and WBCs (Figure 1A) in a dose-dependent manner.

Following TNF-α administration, which is known to enhance in vivo leukocyte adhesion, sustained RBC-WBC interactions in the groups injected with PBS, IVIG 1 mg/kg, or IVIG 10 mg/kg led to progressive reductions in venular blood flow, but mice treated with at least 200 mg/kg IVIG were significantly protected from blood flow reductions (Figure 1B-C; wall shear rates in Table 1). Furthermore, most mice pretreated with PBS, IVIG 1 mg/kg, or IVIG 10 mg/kg died before the end of IVM 2 (Table 1), whereas significantly more mice treated with at least 200 mg/kg IVIG were still alive at the end of the experiment (1 of 14 alive for PBS and low-dose IVIG versus 7 of 10 alive for at least 200 mg/kg IVIG; P = .002, Fisher exact test).

To investigate whether the effects of high-dose IVIG were due to an increase in intravascular oncotic pressures by the protein or nonspecific protein coating, we treated sickle mice with 400 mg/kg IgG-free human albumin prior to the intravital microscopy protocol and evaluated its effects on RBC-WBC interactions, blood flow, and survival. We found that the number of interactions between RBCs and WBCs and effects on blood flow were similar to those observed in the PBS-treated groups (Figure 1A-B). Moreover, all 5 albumin-treated mice died during the experimental procedure (Table 1). Thus, the inhibition of RBC-WBC interactions by IgG is not due to protein-mediated improved microcirculatory hemodynamics or to non-specific protein coating of sickle RBCs or WBCs.

Previous studies revealed the presence of bound IgG on the surface of sickle cells^14-17  and that IgG binding was greatest on dense¹⁶  and sickle-shaped¹⁷  RBCs. Hofstra et al found that preincubation of neutrophils with IgG inhibited their binding to sickle RBCs in a static adhesion assay, suggesting that sickle RBCs might interact with adherent neutrophils in vitro through IgG-Fc receptors.⁷  To investigate whether the effect of IVIG was mediated by Fab or Fc fragment of the IgG molecule, antibody fragments were prepared and injected at a dose of 200 mg/kg, which corresponds to about 400 mg/kg whole IgG on a molar basis. Neither fragment significantly affected WBC-endothelial interactions (Table 1). While both the Fab and Fc fragments tended to reduce the numbers of RBC-WBC interactions during IVM 1 (0.25 ± 0.08 and 0.82 ± 0.44 RBC-WBC interactions per minute for Fab and Fc, respectively), the reduction was only significant for the Fab group (P = .02 compared with PBS or albumin-treated). Neither fragment led to improved survival (Table 1). These results indicate that immunoglobulin fragments can influence RBC-WBC interactions, but neither can recapitulate the effects of the intact IVIG.

IVIG has been used to treat select autoimmune disorders, but its mode of action is still incompletely understood. Several mechanisms have been proposed, including blockade of Fc receptors or activation of inhibitory Fc receptors^12,18  and antiinflammatory effects mediated through Fab fragments.¹⁹  Thus, both regions of IVIG may be important, explaining the partial benefit observed using individual fragments. Notably, the number of adherent leukocytes was also significantly reduced in sickle cell mice treated with high-dose (at least 200 mg/kg) IVIG but not with IgG fragments or albumin (Table 1). IVIG may thus protect sickle cell mice from vaso-occlusion by both inhibiting the RBC-WBC interactions and also by reducing the number of adherent leukocytes in venules. The latter may contribute to IVIG's antiinflammatory effects in other disorders and thus may reflect an additional mechanism of IVIG action.

Although the vaso-occlusion observed in our model may have similarities to that in patients with sickle cell disease, the possibility that our model, which uses 2 separate proinflammatory stimuli (surgery and TNF-α), may produce vaso-occlusion by mechanisms different from those causing pain crisis in humans cannot be excluded. Nevertheless, our data suggest a possible beneficial effect of IVIG in sickle cell–associated vaso-occlusion.

We identified 4 reported cases where IVIG was administered in sickle cell patients, in each case for a delayed hemolytic transfusion reaction. While IVIG was beneficial 3 cases,^20,21  in the fourth case there was a temporal relationship between IVIG administration and the occurrence of a sickle cell crisis.²²  Our data raise the possibility that IVIG may be beneficial in preventing or treating sickle cell vaso-occlusion. Caution is necessary, however, because IVIG administration has been associated with both renal toxicity^23,24  and thrombosis,²⁵  including an estimated 0.6% risk of stroke,²⁶  and there is reason to believe that patients with sickle cell disease may be predisposed to such toxicities.