Antibodies to granulocytes in patients infected with human immunodeficiency virus

Individuals who are infected with human immunodeficiency virus (HIV) are known to have a high incidence of autoantibodies. In this study, serum samples from 100 individuals with HIV infection were tested for granulocyte antibodies (red cell antibodies, lymphocytotoxic antibodies, circulating immune complexes, and serum immunoglobulin G levels) by granulocyte agglutination (GA) and granulocyte immunofluorescence (GIF) assays. Granulocyte antibodies were detected in 66% of serum samples by GIF and in 21% of serum samples by GA None of the positive sera reacted with granulocyte antigens of known specificity. Antibodies that reacted with red cell antigens other than ABO were detected in only three serum samples, but lymphocytotoxic antibodies were detected in 62% of patients. All serum samples were tested by immunoblotting with granulocyte plasma membranes. Only two samples were found to be positive; one sample reacted with a 58 kd protein and one reacted with a 55 kd protein, but neither serum sample immunoprecipitated any protein from granulocytes that were labeled at the cell surface with iodine 125. Since immune complexes that are bound to granulocyte membranes can be detected by GIF, circulating immune complex levels were measured in all 100 samples. Immune complexes were increased in GIF-reactive serum samples compared with GIF-nonreactive serum samples (23.3 ± 19.5 μg Eq/ml [mean ± SD] vs 9.6 ± 8.1 μg Eq/ml, p < 0.001) but not in GA-reactive serum samples compared with GA-nonreactive sera (24.4 ± 21.3 μg Eq/ml versus 16.9 ± 16.0 μg Eq/ml, p = 0.10). There was no difference in immunoglobulin G levels between GIF-reactive and GIF-nonreactive serum samples (2105 ± 610 mg/dl vs 1915 ± 566 mg/dl, p = 0.20) nor between GA-reactive versus GA-nonreactive sera (2178 ± 533 mg/dl vs 2010 ± 618 mg/dl, p = 0.20). To test whether the positive results in GIF but not GA could have been due to circulating immune complexes rather than granulocyte autoantibodies, GA and GIF assays were performed with heat-aggregated immunoglobulin G in place of serum. Although the GIF assay was strongly positive, the GA assay remained negative. These results suggest that in some cases reactions in GIF were due to immune complexes; however, 33 serum samples that reacted in GIF had normal immune complex levels, which suggests that some positive reactions in GIF were likely due to antibodies to neutrophils. In conclusion, antigranulocyte antibodies were found in only 21% of serum samples, as measured by GA, and antibodies to specific granulocyte membrane antigens could be detected in only two patients. In contrast, the reactivity against granulocytes, which was detected in the GIF assay, was in many cases likely due to the presence of circulating immune complexes in HIV sera. These results should be considered in blood cell antibody testing for acquired immunodeficiency syndrome.