A High Frequency of Blasts with the PNH Phenotype in Patients with ALL.

Abstract 1575

Poster Board I-601

It has been proposed that if more than 2 mutations are required for leukemia to arise, then an elevation in the mutation rate would be necessary for these mutations to coincide in the same cell, based on the rarity of spontaneous mutations (Loeb 1991). Hypermutability could occur transiently, for example, due to mutagens, or permanently, due to inactivation of repair pathways. Genome-wide analysis has indicated that in acute lymphoblastic leukemia (ALL), more than 2 somatic mutations are often found (Mullighan et al 2007), suggesting that hypermutability is likely to occur at least in some cases. To investigate this hypothesis, we have used our method for the detection of rare cells that have acquired a spontaneous inactivating mutation in the PIG-A gene. Expanded populations of stem cells with PIG-A mutations in the setting of bone marrow failure are the hallmark of paroxysmal nocturnal hemoglobinuria (PNH), but under most other circumstances, PIG-A mutations are thought to be growth neutral in vitro and in vivo. PIG-A is on the X-chromosome, and therefore, a single inactivating mutation is sufficient to produce the PNH phenotype, characterized by a lack of glycosylphosphatidylinositol (GPI)-linked proteins and an inability to bind to the FLAER reagent. Furthermore, it is known from patients with PNH and from cell lines that a very broad spectrum of mutations can inactivate PIG-A. This combination of features makes PIG-A uniquely suited as a sentinel gene for spontaneous mutagenesis in humans, which is otherwise very difficult to quantitate. Here we have analyzed a panel of blast cells derived from ficolled marrow samples from patients with B-lineage ALL, obtained from the Children's Oncology Group ALL Cell Bank. To determine the frequency of spontaneous PIG-A mutations, samples were stained sequentially with an Alexa-488 conjugate of the FLAER reagent, a mixture of murine anti-CD55 and anti-CD59 antibodies, FITC-conjugated rabbit anti-mouse immunoglobulin secondary antibody, and PE-conjugated anti-CD45 antibody. The blast population was identified based on forward and side scatter, dim expression of CD45, and exclusion of propidium iodide. Control GPI (+) and GPI (-) cells from a patient with PNH served as a control. Among 19 cases of B lineage ALL studied, the frequency of blast cells with the PNH phenotype varied over a very wide range, and the distribution of mutant frequency values was seen to be bimodal. One group, representing about 40% of the cases of ALL, had a median mutant frequency of 6.8 × 10^-6, which is very similar to the proportion of granulocytes that acquire the PNH phenotype due to spontaneous PIG-A mutations in normal individuals (Araten et al, 1999). These values are also close to the frequency of mutants reported for the HPRT, GPA, and XK genes in normal individuals. The other 60% of the samples had a median mutant frequency of 465 × 10^-6, which is far above the range of this parameter in normal individuals. Among 6 established cell lines derived from T-ALL, the mutant frequency ranged from 4.1 ×10^-6 to 265 × 10^-6. Here as well, the distribution suggested that in this subtype of ALL, there may be two distinct phenotypes with respect to this parameter. Based on this data, we believe that an increase in inactivating mutations is not essential for the development of ALL, but that it is a common feature of this condition. We cannot rule out the possibility that samples with a low mutant frequency would demonstrate some form of genomic instability that would not be detected by our assay, such as an increase in gene duplications or translocations. Further investigations are needed to determine whether hypermutability as detected by this assay is associated with distinct cytogenetic abnormalities, clinical features at presentation, and outcome.