Identification of Ex Vivo Myeloma Cells with the PNH Phenotype

Abstract 3994

Myeloma cells can harbor ∼ 35 mutations, whereas according to Loeb's model, if > 2 oncogenic mutations are required for the development of a malignancy, then an elevation in the mutation rate would be required for these mutations to occur in the same cell. Indeed, there is evidence that hypermutability in myeloma may be a result of abnormal homologous recombination or activation induced cytidine deaminase. An alternative model is that successive rounds of clonal selection in an expanding pre-malignant population could result in an accumulation of oncogenic mutations. Using our method for the detection of rare cells that have an acquired somatic mutation of the PIG-A gene, we have previously demonstrated hypermutability in some but not all lymphoma and myeloma cell lines, and we have recently shown that samples of ex vivo blasts from ∼ 50% of patients with ALL demonstrate hypermutability. Here we have hypothesized that genomic instability could be identified in myeloma. PIG-A encodes an enzyme that is necessary for the biosynthesis of glycosylphosphatidylinositol (GPI) and is particularly promising as a sentinel gene for spontaneous mutations because it is X-linked, and thus a single mutation can produce the mutant phenotype. PIG-A is mutated in Paroxysmal Nocturnal Hemoglobinuria (PNH), and it is known from this condition that a broad spectrum of inactivating mutations can produce the PNH phenotype, which is a loss of all GPI-linked surface proteins. PIG-A mutations do not affect transmembrane proteins and are growth-neutral under almost all circumstances. The PNH phenotype can be readily detected by flow cytometry, using antibodies specific for GPI-linked proteins (e.g., CD48, CD55, and CD59), as well the FLAER reagent, which binds directly to GPI. In order to quantitate the frequency of myeloma cells with the PNH phenotype, we analyzed thawed ficolled samples from patients with a heavy burden of myeloma cells in the marrow. Cells were stained sequentially with FLAER-Alexa 488, then with a mixture of murine anti-CD48, anti-CD55, and anti-CD59 antibodies, followed by FITC-conjugated rabbit anti-mouse immunoglobulin, followed by a PE-conjugated antibody specific for a myeloma-specific transmembrane protein–either CD38 or CD138. Live myeloma cells were identified by forward/side scatter and propidium iodide exclusion and expression of CD38 or CD138. Using this approach, in previous studies we have seen that cells with the PNH phenotype have a low FITC/FLAER fluorescence, defined as < 4% of the level of the GPI (+) cell population, after gating for cells with at least 10% of the PE fluorescence of the GPI (+) population. For a negative control, we analyzed 2 non-malignant B-lymphoblastoid cell lines (BLCLs) from normal donors, and for a positive control, we analyzed the mantle cell lymphoma cell line HBL2A (in this case using CD45-PE to identify transmembrane proteins). The normal BLCLs demonstrated a frequency of PNH cells of 6.3 × 10⁻⁶ and 18.4 × 10⁻⁶, which is in the range that we have previously reported for BLCLs and granulocytes from normal individuals. These values are also similar to the frequency of spontaneously arising phenotypic variants in normal donors using other genes. In contrast, as we have previously reported, the mantle cell line demonstrated a markedly higher frequency of cells with the PNH phenotype– 1034 × 10⁻⁶. Of the involved marrow samples analyzed, there were at least 2 distinct groups. One group, representing 14 of the 20 samples (70%), demonstrated a mutant frequency that is comparable to non-malignant cell populations, with a median value of 9.5 × 10⁻⁶ (range 2.4 to 37 × 10⁻⁶). The remaining 6 samples (30%) demonstrated a markedly increased frequency of PNH cells, with a median value of 90 × 10⁻⁶ (range 73 to 11,763 × 10⁻⁶). Most of the samples analyzed were obtained from patients who had received prior therapy, but one of the samples demonstrating a very high frequency of PNH cells (1314 × 10⁻⁶) was derived from a patient who had not had prior therapy and was known to have had an abnormality of p53 based on FISH. This data demonstrates that an increase in inactivating mutations—as determined by this assay– is not essential for the development of myeloma, but it does seem to be a common feature of this condition. This flow-based assay could be applied at the time of diagnosis to facilitate investigations as to whether hypermutability correlates with outcome in patients with myeloma.