The neoplastic production of the insulin-like growth factor binding protein (IGFBP)-2 often correlates with tumor malignancy and aggressiveness. Since IGFBP-2 contains an RGD motif in its C-terminus, it was hypothesized that this protein may act independently of IGF on tumor cells through integrins. To investigate this, integrin binding, intracellular signaling and the impact of IGFBP-2 on cell adhesion and proliferation were examined in two tumor cell lines. In tracer displacement studies, up to 30% of the added 125 I-hIGFBP-2 specifically bound to the cells. Bound 125 I-hIGFBP-2 was reversibly displaced by IGFBP-2, IGFBP-1 and RGD-(Gly-Arg-Asp)-containing peptides, but not by IGFBP-3, -4, -5, -6 and RGE-(Gly-Arg-Glu)-containing peptides. Blocking with antibodies directed against different integrins and with fibronectin demonstrated that IGFBP-2 cell surface binding is specific for 5 1-integrin. Incubation of IGFBP-2 with equimolar quantities of IGF-I and IGF-II annihilated RGD-specific binding. IGFBP-2 binding at the cell surface led to dephosphorylation of the focal adhesion-kinase (FAK) of up to 37% (P<0·01), and of the p42/44 MAP-kinases of up to 40% (P<0·01). In addition, IGFBP-2 promoted de-adhesion of the cells dose-dependently by up to 30% (P<0·05), and reduced proliferation by 24% (P<0·01). Since one of the cell lines used does not express a functional IGF-I receptor, these data demonstrate that IGFBP-2 can act in an IGF-independent manner, at least in part by an interaction with 5 1-integrin.

The paraneoplastic production of pro-insulin-like growth factor-II (IGF-II) forms causes tumour hypoglycaemias and presumably also has an effect on tumour cell growth. We investigated the molecular weights of IGF-II forms and their ability to form complexes with IGF binding proteins (IGFBPs) in conditioned culture media (CM) from 12 paediatric soft tissue sarcoma (STS) cell lines and from two healthy fibroblast lines. Untreated CM were separated by size exclusion chromatography using biocompatible HPLC. Subsequently, IGF-II, IGFBP-2 and IGFBP-3 were determined in the HPLC fractions by specific RIAs. In the CM, IGF-II concentrations between 0.5 and 8.6 ng/10⁶ cells were measured but no IGF-I was detectable. Parallel to this investigation, a high IGF-II mRNA level averaging 44.4 ± 29.7% was measured by semi-quantitative RT-PCR. The STS cell lines secreted a higher proportion of big-IGF-II forms reaching 10–18 kD (10–33% of the total IGF-II secreted) compared to the healthy fibroblasts (2.5–5%). At the same time, the proportion of IGF-II bound with IGFBP in complexes of 35– 70 kD and 150 kD was reduced by up to 85% in CM from tumour cells. The tumour cell lines apparently secrete a different spectrum of IGF-II forms than healthy fibroblasts. The reduced ability to form complexes with IGFBP and the higher molecular weight of the IGF-II forms produced by the tumour cells indicate that these forms could in fact be the known tumour-associated pro-IGF-II forms. Due to these characteristics, the big-IGF-II forms probably have an altered biological effect on the tumour cells when compared to IGF-II.

No abstract