Summary The production of insulin-like growth factor I (IGF-I) and IGF-I binding proteins (BPs) by human lung tumour cell lines in vitro has been examined and the levels of these substances in the serum of lung cancer patients investigated. While small cell lung cancer (SCLC) cell lines secreted both IGF-I and BPs, non-small cell lung cancer (NSCLC) cell lines secreted BPs only. No evidence of increased serum IGF-I levels was obtained in a cohort of 52 lung cancer patients having SCLC and NSCLC histologies. In contrast, serum levels of low molecular weight BPs were markedly elevated in the majority of lung cancer patients.The detection of elevated immunoreactive insulin-like growth factor-I (IGF-I) in human lung tumours (Minuto et al., 1986;Macaulay et al., 1988a) Healthy adult male and female non-smokers (n = 32), and male and female normal smokers (n = 31) were included in the study as controls. The age range for controls was 23-81 years and for lung cancer patients 39-79 years.Pre-treatment serum samples were prepared immediately after collection and stored at -70'C before assay.
IGF-I determinationsA radioimmunoassay (RIA) kit (Amersham International, Aylesbury, UK) and a somatomedin C (SM-C) immunoradiometric (IRMA) kit (Immunodiagnostics Ltd, Tyne and Wear, UK) were used for the quantitative measurement of IGF-I in conditioned media and serum samples. Recombinant human IGF-I was used as standards in both assays. The rabbit antiserum used in the competitive RIA was raised against a recombinant analogue of IGF-I, shows 100% cross-reactivity with human IGF-I and 0.5% crossreactivity with human IGF-II. Fifty per cent displacement of tracer occurs with insulin at 2,000 ftunits ml-' (normal range of insulin in fasting individuals is 4-30 gunits ml-' in serum). Phase separation was achieved using Amerlex-M donkey anti-rabbit reagent (Amersham UK). Assay sensitivity is 100pg per tube.The non-competitive SM-C IRMA employed a two site immunoradiometric assay. Briefly standards/samples were incubated simultaneously with a mouse monoclonal anti-SM-C lgG immobilised on the inside walls of test tubes and a '25I-labelled mouse monoclonal antibody directed against a second IGF-I epitope. Unbound materials were then removed by decanting and washing the tubes. The antisera show 3% cross-reactivity with IGF-II and did not cross-react at all with insulin or pro-insulin. The lowest detectable level of SM-C that could be distinguished from the zero standard was 8 mU ml' at the 95% confidence limit.To separate binding protein from IGF-I, serum samples and cell conditioned media were extracted by incubation in 50 of acid extraction solution (supplied by Immunodiagnostics Ltd) for 10min at room temperature. Neutralising solution (500pl) (supplied by Immunodiagnostics Ltd) was then added to each sample. This extraction procedure yields aproximately 100% recovery of SM-C in patient samples. The neutralised, extracted conditioned media and serum samples, and unextracted serum samples were then used in the Amersham RIA and the IRMA.