Absence of CD9 expression in acute myeloid leukemia: possible correlation with t(8;21).


Sir, Recognition of novel immunophenotypic associations with the genetically and prognostically distinct subgroups of acute leukemia is desirable as they can provide valuable addition to the diagnostic algorithms and also help in minimal residual disease monitoring. CD9 is a tetraspanin molecule that regulates integrin-mediated functions such as adhesion and motility [1]. It has been shown earlier that low expression or absence of CD9 on leukemic cells fairly accurately predicts t(12;21) (ETV6RUNX1) in B-cell acute lymphoblastic leukemia (B-ALL) [2]. Similar to ALL with t(12;21), RUNX1-RUNX1T1 (previously AML1-ETO) fusion in acute myeloid leukemia (AML) with t(8;21) results in repression of normal RUNX1 gene [3]. A possibility of RUNX1 having a regulatory role in CD9 expression has been raised earlier [4]. We analyzed our AML cases to ascertain whether differential CD9 expression can discriminate AML with t(8;21), which is associated with a favorable outcome, from other AML subtypes. Sixty consecutive cases of AML over a 6-month period were retrospectively analyzed for immunophenotypic expression of CD9 and correlated with the cytogenetic and molecular analysis data. The patient age ranged from 2 months to 74 years (median 11 years). The median age of 11 years in our study group may indicate referral bias from our pediatric oncology unit. There were 40 males and 20 females (M : F 2 : 1). Routine immunophenotyping was performed for CD9 expression on blasts along with CD13, CD33, CD117, MPO, CD15, CD65, CD11b, CD36, CD19, CD10, CD79a, CD38, cCD3, CD2, CD4, CD7, CD34, CD45, HLA-DR, CD56, CD14, and CD64 using FC-500 (Beckman Coulter) flow cytometer. Additional markers such as CD18, CD41, and CD61 were used in selected cases. Blasts were gated as CD45 dim events with low side scatter (SSC) on CD45/SSC plot [5]. CD9 median fluorescence intensity (MFI) in the blasts was documented. CD9 positivity was defined as expression in at least 20% gated cells, using internal negative controls (lymphocytes gated on CD45/SSC plot). As per the AML classification (WHO, 2008) [3], 17 of 60 cases were AML with recurrent genetic abnormalities. This included 8 cases with t(8;21)(q22;q22), seven cases with t(15;17) (q22;q12), and two cases with inv(16)(p13.1q22). One case was classified as AML with myelodysplasia-related changes. The remaining 42 cases belonged to AML, not otherwise specified category. Cytogenetic and/or molecular analysis regarding common recurrent cytogenetic abnormalities—PML-RARA, RUNX1-RUNX1T1, and CBFBMYH11—was performed as per the methods described earlier [6]. Cases with absence of common recurrent cytogenetic abnormalities mentioned above were grouped along with the patients with normal cytogenetics. So finally, the cases were divided as normal cytogenetics (n = 37), t(8;21)(q22;q22) (n = 8), t(15;17)(q22; q12) (n = 7), inv(16)(p13.1q22) (n = 2), -Y (n = 2), trisomy 8 (n = 1), t(7;11)(p15;p15) (n = 1), hyperdiploidy (n = 1), and complex karyotype (n = 1). Statistical analysis was performed using SPSS, version 20, Fischer’s exact test, and Mann–Whitney U test, and P < 0.05 was considered significant. Among the various cytogenetic subgroups, CD9 was absent in 6/8 (75%) cases (P = 0.011) with t(8;21)(q22; q22) (Figure 1), 9/37 (24.3%) patients with normal cytogenetics, 1/7 (14.3%) with t(15;17)(q22;q12), and none of the two cases with inv(16)(p13.1q22). Two of the other six cases (33.3%) were also negative for CD9 (Table 1). CD9 MFI difference between cases with t(8;21) (q22;q22) and normal cytogenetics was not significant (Table 1). Fusion protein ETV6-RUNX1 in B-ALL with t(12;21) exerts a dominant negative effect on normal RUNX1 function [7]. These cases also have absence of CD9 expression, possibly indicating a direct regulatory role of RUNX1 in CD9 expression. The differential CD9 expression in B-ALL with t(12;21) has also formed the basis of proposed diagnostic algorithm to classify childhood B-ALL into genetically distinct entities in an accurate, cost-effective, and less time-consuming manner [8]. Similar to precursor B-ALL with t(12;21), AML with the t (8;21) has RUNX1 fused with RUNX1T1, resulting in suppression of RUNX1 function. In the study by Racke et al.,

DOI: 10.1111/ijlh.12296

1 Figure or Table

Cite this paper

@article{Gupta2015AbsenceOC, title={Absence of CD9 expression in acute myeloid leukemia: possible correlation with t(8;21).}, author={Sunil Kumar Gupta and Adarsh Chopra and S K Singh and Rakesh Kumar and Sameer Bakhshi and Lalit Kumar and Atul Sharma}, journal={International journal of laboratory hematology}, year={2015}, volume={37 3}, pages={e56-8} }