Human basophils express CD22 without expression of CD19

Background: Even modem automatic cell counters cannot count basophils precisely. Therefore, we need a rapid, accurate, precise, and easy method for counting basophils. Methods: Using flow cytometry, basophils (CD22+/CD19-) and B cells (CD22+/CD19+) were counted. Within a large lymphocyte light scatter gate, % basophils (G%baso) and % B cells (G%B) were determined from the total count. Another method of analysis was to make two regions (R1 for basophils and R2 for B cells) and to determine in those the % basophils (R1%baso) and % B cells (R2%B) without gating. The flow cytometric basophil counts of the blood of 21 normal controls and 43 chronic myelogenous leukemia (CML) patients were compared with manual basophil count (Ma%baso) and basophil count by Coulter electronic cell counter (Hialeah, FL) (Auto%baso). CD22+/CD19- cells were sorted by a FACSCalibur (Becton Dickinson, San Jose, CA). Results: The G%baso of all samples was 4.66 ± 5.35%, and R1%baso was 4.23 ± 4.88%, and they were well-correlated (r = 0.996, P < 0.001). The G%B of all samples was 1.55 ± 1.68%, and R2%B was 1.59 ± 1.67%, and they were also well-correlated (r = 0.993, P < 0.001). Their correlation was better in normal controls than in CML. G%baso was well-correlated to Ma%baso (r = 0.827) and Auto%baso (r = 0.806), and R1%baso was well-correlated to Ma%baso (r = 0.831) but showed poor correlation to Auto%baso (r = 0.734). Auto%baso revealed the poorest correlation to Ma%baso (r = 0.692). The sorted CD22+/CD19- cells were all basophils (99.48 ± 0.30%), and they revealed CD13, CD33, and dim CD45 expression, whereas CD3, CD14, CD16, and HLA-DR were not detected on them. Conclusions: We discovered a specific marker combination to identify basophils (CD22+/CD19-), and we suggest that flow cytometric analysis using these markers is an easy, reliable, and accurate method of basophil counting.