Expression of AC133, a Novel Hematopoietic Precursor Antigen, on Acute Myeloid Leukemia Cells

To the Editor:

In a recent issue of BLOOD, Yin et al1reported the production of AC133, a monoclonal antibody (MoAb) that binds to a novel cell surface antigen present on a CD34^(bright) subset of human hematopoietic stem cells (HSCs). In the same issue, Miraglia et al2 described that the AC133 antigen is also expressed on subsets of CD34⁺leukemias. Six samples of bone marrow cells from patients with acute myeloid leukemia (AML) were examined for expression of CD34 and AC133 by multiparameter flow cytometry using directly conjugated MoAbs. The cells from five samples were positive for CD34 and three of those coexpressed bright levels of AC133 (above second decade of fluorescence intensity), one expressed dim levels (between first and second decade of fluorescence intensity), and one expressed no AC133 (fluorescence was limited to first decade of fluorescence intensity). The cells of one sample were negative for both AC133 and CD34. Miraglia et al2 reported that only myelomonocytic acute leukemia (French-American-British [FAB] M4 or M5) expressed very high levels of AC133, indicating that AC133 staining may be useful for further subtyping of AML cells. Because some of the CD34⁺ leukemias were AC133⁻, it was suggested that isolation of AC133⁺ HSCs may be a viable purging alternative in these patients, because they would obviously not benefit from a CD34-based selection and autologous transplantation.

We investigated the expression of AC133 antigen on blast cells of acute myeloid leukemias. Thirty cases of AML were examined for expression of AC133, CD34, and other cell surface and cytoplasmatic markers using multicolor flow cytometry in addition to the routinely performed morphological, cytochemical, and cytogenetic analyses. The patient characteristics are listed in Table1. 24/30 (80%) AML samples were positive for AC133. Nine of 24 expressed bright levels and 15 expressed dim levels of AC133. All AC133^bright leukemias expressed CD34, whereas 10 of the AC133^dim leukemias were positive and 3 were negative for CD34. Interestingly, 3 of 6 AML that were negative for AC133 showed expression of CD34. Therefore, in 10% of cases examined, AC133-based cell sorting might provide a purging alternative to a CD34-based selection on the condition that AC133⁺cells in these patients are tumor free.

Our data show that AC133 expression was often but not always associated with CD34 expression. In most cases, the fluorescence signal for CD34 was brighter than for AC133.

In our analysis, AC133 expression was found on all FAB subtypes examined (M1, M2, M4, M5, M6, and secondary leukemia after myelodysplastic syndrome [MDS]). In contrast to the findings of Miraglia et al,2 bright expression levels were not restricted to certain FAB subtypes in our studies.

Chromosomal aberrations were found with the same frequency in both AC133 dim and bright AMLs (64% v 71%) and with a lower frequency in AC133-negative AMLs (50%), but the difference showed no statistical significance. There was no detectable correlation to cytogenetic abnormalities associated with a more favorable prognosis [eg, t(8;21), inv(16), t(16;16)] or unfavorable prognosis [eg, −7, del(7q), del(5q); Mrozek et al³]. However, the number of patients screened in our study might still not be sufficient to give a conclusive answer. Chromosomal aberrations were found with a higher frequency in CD34⁺ than in CD34⁻ AMLs (84%v 13%; P < .001). This result is consistent with the findings of Fruchart et al.4 

Patients with de novo AML received the sequence TAD-HAM as double-induction therapy consisting of cytosine arabinoside (Ara-C) combined with daunorubicin and thioguanine followed by Ara-C combined with mitoxantrone. Patients with relapsed or secondary AML were treated according to the Ida-FLAG protocol consisting of idarubicin combined with fludarabine, Ara-C, and filgastrim until recovery of leukocytes to greater than 1,000/μL.

Neither the number of leukocytes nor lactate dehydrogenase as clinical prognostic markers were significantly different in AC133-positive and -negative AML. Clinical outcome in terms of complete remission (CR), defined as a normocellular bone marrow containing less than 5% blasts and a peripheral blood count with greater than 2,500/μL leukocytes and greater than 100,000/μL platelets, did not significantly differ in AC133-positive and -negative AMLs, although there was a trend towards better outcome in AC133-negative AMLs (61%v 83% reached CR; P > .1).

In conclusion, membrane expression of AC133 in combination with other antigens might facilitate a more precise immunologic characterization of AMLs in the future and could serve as an alternative to CD34 for purging purposes in a selected number of patients. It is found in approximately 80% of AML samples and does not seem to correlate to FAB subtype, cytogenetic aberration, or clinical outcome.