Cyclin A1, an alternative A-type cyclin, is frequently overexpressed in acute myeloid leukemia (AML). The roles of cyclin A1 in AML cells are controversial. To explore role(s) of cyclin A1 in AML cells, we generated two myeloid leukemia cell lines, K562.A1 and U937.A1, which constitutively express cyclin A1 by transfecting pCDNA3.1carryng full length cyclin A1 cDNA. K562.A1 and U937.A1 significantly overexpressed both mRNA (k562.A1=130-fold and U937.A1 =11-fold,real-time PCR) and protein(k562.A1=−4.1-fold and U937.A1=2.5-fold, immunoblot) of cyclin A1 compared with K562.C and U937.C, both of which were transfected with pcDNA3.1Hisc as a control,respectively.

U937.A1 showed increased in vitro cell growth compared with U937.C (U937. A1 =16.8 hrs and U937.C=18hrs, doubling time [DT]) and also showed 3.3-fold increased colony formation in soft agar. Incubation with 1μM cytosine arabinoside (CA) for five days significantly inhibited cell growth in U937.C compared with U937.A1(U937.C=12.2±2% and U937.A1=46.9±2%,p<0.01,cell No. with CA/cell No. without CA) and induced apoptosis as judged by flow cytometry (FCM) after staining Annexin V-FITC/PI(U937.C=53.5±1% and U937.A1=15.1±0.4%, P<0.01).

In contrast, K562.A1 showed decreased cell growth compared with K562.C (K562.A1- =43.2hrs and K562.C=33.6hrs, DT) and decreased colony formation (25.6±6% of those of K562.C). FCM analysis of cells after staining with PI showed that decreased cell growth of K562.A1 was accompanied with inhibition cell cycle progression and G1 accumulation. Small but substantial increase of subG1 DNA content portion were observed in K562.A1 compared with K562.C(K562.A1 =2.4±0.1% and K562.C=0.7±0.2%, P=0.01), suggesting the presence of low proliferating or apoptosis-prone G1 cells in K562.A1. cDNA microarray and real-time PCR analyses revealed marked alteration of expression of genes involved in growth arrest and/or apoptosis in k562.A1 compared with in K562.C,including increase of cell division cycle 20 homolog (CDC20) ( more than 40,000 fold), decrease of both GADD45A and Nijmegen Breakage Syndrome (NBS)1 ( more than 1,000 fold). No significant difference were observed in expression of other cell cycle regulators, such as cyclin A(A2),B1,B2, D1,D2, E1 and PCNA between two cell lines. Cell surface expression of glycophorin A was decreased in K562.A1 compared in K562.C. K562.A1, compared with K562.C showed significantly impaired CA-induced erythroid differentiation (K562.C=51.5 ±4.5% and K562.A1=6.7 ±1.7%, p<0.01 ,the % of benzidine –positive cells at 4 days incubated with 1μM CA )and also decreased induction of apoptosis (K562.C=50.9 ±1.5% and K562. A1=29.6±1.2%, p<0.01 , Annexin V-FITC/PI stained cells). The percentage of Hemin-induced benzidine –positive cells was also decreased in K562.A1 compared in K562.C (K562. C=51.3 ±2.5% and K562.A1=7.3 ±1.7%, p<0.01 ,at 6 days incubated with 0.1mM Hemin). Quantitative enzyme immunoassay using DNA binding NF-κB consensus sequence showed significant increase of activated NF-κB in the nuclear extract in K562. A1 compared in K562.C(K562.C=1.0±0.3pg/μg nuclear protein and 7.9±1.3pg/ μg, P=0.01). Furthermore, 72 hrs incubation with 1μM CA progressively induced activated KF-κB in K562.A1, while the induction of NF-κB in K562.C was negligible (K562.C=2.1±0.7 pg/μg and K562.A1=23.6±0.9pg/μg, P=0.01). It suggests that impaired CA-induced erythroid differentiation and apoptosis in K562. A1 was, at least in part, associated with aberrant induction of NF-κB in these cells. The present results demonstrate that cyclin A1 shows anti-apoptic effect against CA when it is overexpressed in AML cells, while it’s effects on cell growth are variable. Although the details of link between overexpressed cyclin A1 and the cell cycle regulation in myeloid leukemia cells are still remained to be elucidated, the present study suggest cyclin A1-induced NF-κB pathway(s) are candidate for therapeutic target in chemo-resistant myeloid leukemia.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author