Background: Gemutuzumab ozogamicin (GO) is a molecularly targeted drug which combines a humanized anti-CD33 antibody (p67.6) with calicheamicin. GO has differed therapeutic efficacy in cases of acute myeloid leukemia (AML). Single nucleotide polymorphisms (SNPs) in the CD33 gene have been investigated as one of the reasons for this difference. SNPs (CC, CT, and TT types) in rs12459419 have been reported to be associated with clinical outcomes in pediatric AML patients treated with GO-containing chemotherapy. On the other hand, there are reports of poor association in adult AML patients, and the significance of rs12459419 SNPs in remains to be clarified. CD33 is composed of two extracellular domains, IgV and IgC. p67.6, anti-CD33 antibodies, recognize IgV, while HIM3-4 antibodies recognize IgC. rs12459419 SNP is located in the splicing acceptor region of exon 2 of the CD33 gene, when the SNP is T, transcripts that skip exon 2 which encodes the IgV domain, are more likely to occur. The increasing expression of this exon skipping product may be accompanied by a decreasing amount of wild-type CD33 (CD33FL) and an increasing amount of IgV domain-deficient CD33 (CD33ΔE2). GO only recognizes CD33FL, but not CD33ΔE2. We used CRISPR/Cas9 genome editing system to generate cell lines with specifically modified CD33 SNPs to elucidate the expression of CD33FL and CD33ΔE2 by CD33 SNPs and its effect on sensitivity to GO in a uniform genetic background. Methods and Results: First, we analyzed CD33 SNPs in 23 AML patients treated with GO (CC type: 13; CT type; 9: TT type: 1). The outcome of 17 patients, excluding a history of hematopoietic stem cell transplantation before treated with GO, was evaluated. Three-year overall survival rate for the type of CC vs. CT was 70.7% vs. 83.3% (p value = 0.609), and 3-year disease-free survival rate for the type of CC vs. CT was 80.8% vs. 83.3% (p value = 0.895). There was no significant difference in CC and CT. Next, using p67.6 recognizes only CD33FL and HIM3-4 recognizes both CD33FL and CD33ΔE2, we evaluated the expression ratio of CD33FL and CD33ΔE2 in relation with CD33 SNP. We assumed that the expression levels of MFI(median fluorescence intensity) were roughly similar in HIM3-4, but the MFI evaluated by both antibodies tended to be higher in the CT and CC types than in the TT type (MFI for p67.6 of CC, CT and TT was 2675, 1752 and 245, respectively; MFI for HIM3-4 of CC, CT and TT was 699, 629 and 53, respectively). The patient samples differ in many genetic backgrounds besides CD33 SNPs, making accurate evaluation difficult. Genome editing was performed by co-transfection of guide RNAs targeting CD33 SNPs, Cas9 protein and with single-stranded oligonucleotides to the target leukemia cell line. Single cell cloning was performed by limiting dilution to generate single nucleotide altered cells for the CD33 SNP. OCI-AML3 wild-type (TT type) and the obtained CD33 SNP-modified clones (CC type, 4 clones and CT type, 4 clones) were evaluated by flow cytometry with p67.6 and HIM3-4.The results showed that the MFI of both antibodies was higher in the CT and CC types than in the TT, the wild type (MFI for p67.6 of CC, CT and TT was 10760, 13686 and 4801, respectively; MFI for HIM3-4 of CC, CT and TT was 3115, 4067 and 887, respectively). These results showed similar trends to the clinical samples implying that differences in CD33 SNPs affect the amount of total CD33 expression on the cell surface itself. Conclusion: The CD33 rs12459419 SNP regulated the expression level of total CD33 in clinical samples and gene-edited cell lines. There was no obvious difference in the expression levels of total CD33 and CD33FL at the cell surface between the CC and CT types, but it was higher than that of the TT type. The difference in CD33 expression may influence the therapeutic efficacy of GO.