Do human myeloma cells directly produce basic FGF?

Basic fibroblast growth factor (bFGF) is a growth factor with proangiogenetic properties. Elevated bone marrow (BM) and peripheral serum bFGF levels have been reported in patients with multiple myeloma (MM)^1-3 ; however, the source of bFGF in patients with MM is not completely elucidated.

Recently, Bisping et al,⁴  in line with others,¹  have reported that human myeloma cell lines (HMCLs) (RPMI-8226, U266, KMS-11, and KMS-18) and sorted CD38^high/CD138⁺ cells obtained from 12 of 15 patients with MM produced bFGF, concluding that myeloma cells are the predominant source of bFGF. In contrast, Gupta et al⁵  have shown that neither human myeloma cells nor Epstein-Barr virus (EBV)–positive B-cell lines secrete bFGF.

In order to better clarify this issue, we wish to present our evidence. Using reverse transcription–polymerase chain reaction (RT-PCR) (bFGF primer pairs: forward, 5′-GGCTTCTTCCTGCGCATCCAT-3; reverse: 5′-GGTAACGGTTAGCACACACTCCTTT-3′) we found that XG-6, RPMI-8226, OPM-2, as well as EBV-positive cell line ARH-77 did not express bFGF mRNA, whereas U266 was positive and XG-1 expressed bFGF at low intensity (Figure 1A). Similarly, we failed to detect bFGF either in HMCL lysates by Western blot analysis (antipolyclonal bFGF antibody [Ab]; R&D Systems, Minneapolis, MN) or in HMCL (10⁶/mL)–conditioned media by enzyme-linked immunosorbent assay (ELISA; R&D Systems; range of sensitivity, 10 to 640 pg/mL), both in the presence and absence of interleukin-6 (IL-6, 20 ng/mL) with the exception of U266 and XG-1 (Figure 1B-C). Consistently, we previously showed that blocking anti-bFGF Ab failed to inhibit HMCL-induced angiogenesis in an in vitro system, suggesting that any bFGF biologic activity was found in HMCLs.⁶ 

Purified CD138⁺ MM cells (purity > 95%) isolated by an immunomagnetic method (magnetic-activated cell sorter [MACS]; Miltenyi Biotec, Bergisch-Gladbach, Germany) were positive for bFGF mRNA expression in 11 of 35 patients with newly diagnosed MM in stages I to III (median age, 64 years [range, 33-88 years]; and median plasmacytosis, 35% [range, 12%-95%]) (Table 1). In contrast, BM stromal cells (BMSCs) obtained from all patients were positive for bFGF mRNA (Figure 1A).

bFGF protein has been found in plasma cell lysates in 8 of 30 patients tested. Consistently, bFGF levels were detected by ELISA assay (R&D Systems) in conditioned media of purified MM cells (10⁶/mL) in 7 of 28 patients (Table 1). Furthermore, a nuclear bFGF immunostaining with low cytoplasmic positivity has been found in bone marrow myeloma cells of 3 of 21 patients (Table 1; Figure 1D).

No correlation has been found between bFGF expression by MM cells and BM plasmacytosis (Pearson Chi-square, P = .35) or the presence of osteolytic lesions. The differences in the microvessel density (MVD) and in the number of microvessels per field, evaluated as previously described,⁶  between bFGF mRNA–positive and –negative patients with MM did not reach a statistical significance (MVD ± SE, 36 ± 4 vs 24 ± 3.2; number of microvessels ± SE, 7.4 ± 5 vs 3.59 ± 0.5; Mann-Whitney test, P = .19 and P = .16, respectively).

In conclusion, our data indicate that bFGF is rarely produced directly by MM cells, suggesting that bFGF is not the major proangiogenetic factor produced by myeloma cells, even if its production could be involved at least in part in the MM-induced angiogenesis.