Diagnostic evaluation of t(4;14) in multiple myeloma and evidence for clonal evolution

The presence of a t(4;14) in multiple myeloma (MM) is associated with significantly inferior outcomes following both conventional chemotherapy and high-dose melphalan-based treatment regimens.1, 2 Detection of patients bearing the t(4;14) is therefore important in an MM molecular workup. Furthermore, since targeted inhibitors of the associated fibroblast growth factor receptor 3 (FGFR3) tyrosine kinase are now in clinical trials,3 detection of MM patients who express the FGFR3 protein is also of potential therapeutic relevance. With this in mind, we have employed the infrastructure of the Multiple Myeloma Research Consortium (MMRC) tissue bank to evaluate four different methodologies for the molecular diagnostic detection of this translocation. The methods employed are published previously4, 5, 6 or are provided as Supplementary Information. Fresh bone marrow samples were collected in a uniform fashion from 84 patients with active MM. Immunocytochemistry (ICC)4 and flow cytometry for FGFR3 were performed on fresh samples at the retrieval site, while split samples were then shipped by overnight courier to the MMRC central laboratory, where they were processed according to standard operating procedures, under good laboratory practice (GLP) conditions. Processed samples were then analyzed for detection of the t(4;14) by both IgH-MMSET polymerase chain reaction (PCR)5 – on blood and bone marrow mononuclear cells – and dual fusion cytoplasmic immunoglobulin fluorescence in situ hybridization (cIg-FISH) of bone marrow using a commercially available probe for the t(4;14) (Vysis Inc., DesPlaines, IL, USA).6 Only 60 of the samples contained sufficient plasma cells for cIg-FISH analysis owing to either marrow dilution at the time of collection or low plasma cell numbers reflecting treatment response. It is important to note that the low percentage of successfully studied cases in this series reflects the research nature of the sample provided as the last pull from a relatively large volume and thus hemodilute bone marrow aspiration. In the clinical laboratory setting, adequate cells are available for study in the majority of cases (Dr R Ketterling, Mayo Clinic, personal communication). Of the samples analyzed, eight were positive by cIg-FISH [13%] consistent with the expected frequency. For all other tests, the analysis was conducted in a blinded fashion. When FISH was used as the gold standard for detection of the translocation in bone marrow, the sensitivity and specificity of the other diagnostic tests were as shown in Table 1. Using only data in which FISH and the experimental diagnostic test were successfully performed on the same sample, IgH-MMSET PCR is both the most sensitive (7 out of 7 FISH positive detected) and the most specific test with all negatives correctly identified and no false positives. To evaluate PCR of peripheral blood as a potentially more accessible test, we then conducted a blinded examination of 32 peripheral blood samples which included four patients with known t(4;14) by PCR. PCR detected three of the four positives (sensitivity 75%) and, in this small study, all of the true negatives (100% specificity). Thus as might be expected in the absence of nesting, PCR of peripheral blood will likely miss some true positives owing to very low circulating tumor cell numbers. By gene expression profiling, up to 30% t(4;14) patients have lost expression of FGFR3.7 If patients are to be targeted for therapy based on protein expression, accurate detection of FGFR3 is then essential. In this regard, flow cytometry appeared more sensitive than ICC in the detection of the expressed protein. Notably, the correlation between FC and ICC for protein detection was only 0.46. In an expanded and blinded analysis, 19 of 21 (90%) t(4;14)-positive samples were also flow cytometry positive. Of interest, in the only other reported study of flow cytometry, 20 of 24 t(4;14) cases or 83% were also FGFR3 protein positive,8 suggesting that the frequency of loss of FGFR3 expression may be less common (13% loss by flow cytometry in the 45 patients from the combined studies) than previously reported using gene expression profiling as a read out, where loss in 25–30% of patients has been reported.7 An alternative and perhaps more likely explanation for the higher than expected frequency of flow cytometry positive FGFR3 expression in our series may be nonspecificity of antibody binding. Consequently, we examined a second antibody (R&D Systems, Minneapolis, MN, USA; anti-FGFR3 monoclonal mouse IgG1 clone 136334) along with cIg-FISH in 23 consecutive patients submitted to the tissue bank – flow cytometry was positive in three of the four t(4;14) patients (the more expected frequency), and in the FGFR3-negative patient, a single-fusion signal by cIg-FISH was detected, most likely the result of an unbalanced t(4;14) with loss of der(14). Although the numbers are small, our impression is that this antibody is more reliable and better reflects the true incidence of FGFR3 expression in the t(4;14). We were also interested in better understanding the stability of the t(4;14) over time. To better determine clonal selection/heterogeneity of the translocations, we therefore ascertained the percentage of plasma cells in each patient with an unbalanced t(4;14) translocation (loss of one chromosome bearing the translocation). In 42 patients with the t(4:14), 13 (31%) had a pattern consistent with a balanced translocation (75% of cells with a double fusion signal), 14 (33%) had a pattern consistent with a predominantly unbalanced translocation (>75% cells with only one signal) and 15 (36%) had a chimeric picture. This heterogenous pattern is highly suggestive of an evolving picture towards an unbalanced translocation with presumed loss of the derivative 14 and thus of FGFR3 protein expression as a consequence of disease progression. In summary, our results confirm that 15% of MM patients harbor the t(4;14), but indicate that the translocation is likely to evolve over time, first to a chimeric (dual population) and ultimately to an unbalanced translocation with associated loss of FGFR3 expression consequent to disease progression. We conclude that for the detection of the t(4;14) translocation, cIg-FISH remains a reliable diagnostic methodology; however, this technology is limited in some patients by low plasma cell numbers. The most specific and sensitive test in our study was IgH-MMSET PCR on unsorted bone marrow for detection of the translocation, with flow cytometry most reliably detecting the FGR3 protein. Peripheral blood PCR is specific but lacks sensitivity. This work was supported by the Multiple Myeloma Research Consortium and Multiple Myeloma Research Foundation. Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)