Auer rods and differentiation in acute promyelocytic leukemia

Abstract

Dalal et al (2007) recently reported that the residual detection of Auer rods one month after treatment of an Acute Promyelocytic Leukemia (APL) patient was a sign of maturing granulocytes under differentiation therapy but not of refractory leukemia. Auer rods deriving from the crystallisation of myeloperoxidase (MPO) granules are the hallmark of Acute Myeloid Leukemia (AML), but can also be observed in the Myelodysplastic Syndromes in which they have been associated with a more aggressive disease (Seymour & Estey, 1995). Among AML, two characteristics are specific to APL (AML-M3 subtype in the French-American-British classification): 1) the presence in more than 90% of cases of the oncogenic PML-RARα fusion protein and 2) an exquisite sensitivity to the differentiation induced by All-trans retinoic acid (ATRA) (Cassinat & Chomienne, 2001). We report herein an unexpected correlation between the presence of Auer rods in APL blast cells at diagnosis and ATRA differentiation efficacy. We have been routinely assessing primary APL cells at diagnosis to evaluate their sensitivity to differentiation induction by ATRA. As previously reported (Cassinat et al, 2001), although the percentage (%) of differentiated cells (estimated by the percentage of nitroblue tetrazolium (NBT) positive cells) after 6 d of culture was very high in almost all cases, the levels were more heterogeneous after 3 d of culture (ranging between 0% and 100%) indicating different levels of differentiation sensitivity. 128 consecutive patient samples showed an inverse correlation between the percentage of cells harbouring Auer rods and the percentage of differentiated cells at day 3 (P < 0·015, Pearson test). When patients were studied as three groups according to the proportion of cells with Auer rods (Fig 1A), the median percentages of differentiated cells after 3 d of culture were significantly lower in patients with a higher level of Auer rods (P = 0·014, Kruskal-Wallis non parametric test). (A) Results of ATRA-induced differentiation (measured by NBT test after 3 d of culture) according to the percentage of Auer rod-positive cells at diagnosis. Box 1: 0 to 1% Auer rod-positive cells (n = 48 patients). Box 2: 2 to 15% Auer rod-positive cells (n = 55 patients). Box 3: at least 15% Auer rod-positive cells (n = 25 patients). (B) Distribution of the percentage of Auer rod-positive cells in patients that relapsed (n = 8) or not (n = 30), in the group of patients with a PML-RARα transcript of the Bcr1 subtype. In 62 of these patients, clinical outcome was obtained from the APL93 and APL 2000 trials registries (median follow-up, 9 and 4 years, respectively) (Fenaux et al, 1999; Ades et al, 2006). It appeared that patients with higher levels of Auer rod-positive cells tended to have a higher risk of relapse (Table I). Accordingly, the median percentage of Auer rod-positive cells was higher in the group of patients (n = 12) that relapsed compared to patients that did not relapse (n = 50) (8·5% vs. 4·5%). This difference only trended toward significance when the whole population was considered (P = 0·14; Kruskal-Wallis non parametric test), but was in accordance with our previous observation that in vitro differentiation could be correlated to patient prognosis (Cassinat et al, 2001). Interestingly, when a more homogeneous population, such as patients harbouring a Bcr1 transcript (the most frequent transcript) were analysed (n = 38) (Fig 1B), the percentage of Auer rod-positive cells in the group of patients that relapsed was significantly higher compared to the patients that did not relapse (13·5% vs. 5%, P = 0·045, Kruskal-Wallis non parametric test). These data suggest a link between the presence of Auer rods in APL cells at diagnosis and the sensitivity of these cells to the differentiation effect of ATRA and hence with the clinical outcome. Given that Auer rods are supposed to be a result of MPO crystallisation and their presence is inconstant in AML patients, one hypothesis could be that presence of Auer rods is dependent on a particular MPO genotype. Indeed, several polymorphisms have been reported in the MPO promoter or coding regions and correlation may exist with several types of cancer, such as lung cancer (Gresner et al, 2007) or leukemia (Zhang et al, 2007). We have recently shown that MPO expression during ATRA treatment differs between patients with good or poor differentiation response (Quere et al, 2007). Thus, the present study could be the basis for testing the hypothesis that the presence of Auer rods could reflect a specific genotype correlated to ATRA sensitivity.