AC133 antigen as a prognostic factor in acute leukemia

AC133 is a novel 5-transmembrane antigen present on a CD34((bright)) subset of human hematopoietic stem cells (HSCs) and it is also expressed on the s ubset of CD34 positive (CD34(+)) leukemias. But the clinical significance o f AC133 expression on leukemic blasts is not yet known. We investigated the expression of AC133 antigen on blast cells of acute leukemia. Forty-one ca ses of acute leukemia were examined for expression of AC133, CD34, and othe r antigens using multicolor flow-cytometry. Samples were considered positiv e if at least 20% of the cells specifically stained with monoclonal antibod ies (MoAbs) revealed a higher fluorescence intensity compared to cells of c orresponding negative control samples ( = 20% cut-off level). 14/36 (38.9%) acute myelogenous leukemia (AML) samples and 6/20 (30%) acute lymphoblasti c leukemia (ALL) samples were positive for AC133, the difference was not si gnificant, All AC133 positive (AC133(+)) leukemias expressed CD34, whereas 13 of 33 CD34(+) leukemias were negative for AC133, and AC133(+)/CD34(-) le ukemia was not found. Expression rates of CD31, CD62L, CD62E, CD105 and CD1 44 were significantly higher in AC133(+) leukemia compared to those of AC13 3(-) leukemia (P= 0.045, P <0.001, P < 0.00 1, P < 0.00 1, P = 0.003, respe ctively), but bcl-2, CXCR- 1, CXCR4, VLA-4, CD 106 expression rates were no t significantly different between AC133(+) and AC133(-) leukemias. None of the clinical prognostic markers such as age, hemogram, lactate dehydrogenas e, and chromosomal aberration were significantly different between AC133(+) and AC133(-) leukemias. CR rates of AC133(+) AML and AC133(-) AML were not significantly different, although there was a trend toward higher CR rates in AC133(-) AML (18/22[81.8%] AC133(-) AML versus 9/14[64.3%) AC133(+) AML ), but the 1-year relapse rate of AC133- AML was significantly higher than that of AC133(-) AML (8/9 (88.9%) versus 7/19 (36.8%), P=0.016). Median dis ease-free survival (DFS) times of AC133(+) and AC133(-) AML were significan tly different (11 and 18 months, respectively, P = 0.006), although overall survival (OS) times were not significantly different (AC133(+) 15 months v ersus AC133(-) 20 months, respectively, P=0.06). Similar results regarding clinical outcomes were found when AC133(+)/CD34(+) and AC133(-)/CD34(+) wer e analyzed separately, but the difference did not attain statistical signif icance. In ALL, 9/11 (81.8%) AC133(-) and 2/4 (50%) AC133(+) cases achieved CR, but the difference was not significant. Four of 11 AC133(-) ALL (36.4% ) and 2 of 3 AC133(+) ALL (66.7%) relapsed within I year. In survival analy sis, median DFS time and OS time of the AC133(+) group were 7 and 18 months , respectively, and these were not significantly different from those of th e AC133(-) group (median DFS 15, OS 22 months, respectively). Our results d emonstrate that AC133 expression in AML blasts is associated with poor clin ical outcomes in terms of higher early relapse and shorter disease-free sur vival, suggesting that the AC133 antigen might provide the prognostic strat ification of acute leukemia. However, to verify the effect of AC133 express ion on the therapeutic outcomes of adult acute leukemia, further study incl uding more cases is needed. (C) 2001 Elsevier Science Ltd. All rights reser ved.