The t (4; 14) translocation in multiple myeloma (MM) is identified in 15% of patient samples and dysregulates both FGFR3 and multiple myeloma SET domain (MMSET). 1 Activating FGFR3 mutations are observed in cell lines and late-stage MM, but the incidence of mutation in newly diagnosed patients, although not well characterized, 2 is likely pivotal in MM progression. 3 We therefore selected CD138 cells from the bone marrow of 150 newly diagnosed MM patients. cDNA from selected cells was hybridized to HuGeneFL GeneChip microarrays (Affymetrix, Santa Clara, CA). 4 FGFR3 overexpression was noted in 24 patients (16%). Reverse transcriptase–polymerase chain reaction (RT-PCR) and fluorescence in situ hybridization (FISH) subsequently confirmed that all of the FGFR3-positive patients carried the t (4; 14) translocation (Figure 1A). The entire open reading frame of FGFR3 was then amplified and bidirectionally sequenced in triplicate, which identified mutation in 2 patients. One patient harbored a 6-bp insertion 5-AACAGT-3 at nucleotide position 1306, which was unique to the FGFR3 IIIb isoform, implying a splice variant (Figure 1Bi). In a second patient, a point mutation was observed at nucleotide number 761AG resulting in a Tyr241Cys (tyrosine to cysteine change) located in the linker region between immunoglobulin-like domain I and immunoglobulin-like domain II (Figure 1Bii). This genetic change is in close proximity to other previously identified activating mutations. Together these data indicate that mutation of FGFR3 is observed in less than 5% of t (4; 14)-positive and in only 1 of 150 newly diagnosed MM patients overall. Incorporating findings from the literature5-9 with our larger study, 52 (15%) of 348 myeloma patients have at (4; 14) and 3 (6%) of the 52 translocation-positive patients have a potential activating mutation of FGFR3, or 3 (1%) of 348 overall.

To determine the single nucleotide polymorphism (SNP) profile of FGFR3 we next examined the Celera database and identified 21 SNPs pertaining to the coding region of FGFR3. One of these SNPs, 921CT, was identified in 17 of 22 FGFR3 sequences (Figure 1Biii). Using the SNP human database (http://www. ncbi. nlm. nih. gov/SNP/snp_ref. cgi) this 294CT polymorphism has an allele frequency of 0.877. The 77% SNP incidence in the patient samples is therefore not statistically different from that of the general human population.