APPEARANCE OF TUMOR-NECROSIS-FACTOR-ALPHA AND SOLUBLE TNF RECEPTOR-I AND RECEPTOR-II IN PERITONEAL EFFLUENT OF CAPD

Dialysate and serum concentrations of tumor necrosis factor-alpha (TNF-alpha), soluble TNF-receptor I (sTNFRI) and soluble TNF-receptor II (sTNFRII) were measured during stable and infectious CAPD to determine whether these mediators are released intraperitoneally or derived from the circulation. Dialysate/serum ratios were compared to those of various marker proteins for peritoneal transport and to interleukin-6 (lL-6), which is locally produced. Peritoneal immunoreactive TNF-alpha could be detected in 19 of 20 stable CAPD patients after a night dwell, but only occasionally and in lower concentrations during and after a standard four-hour peritoneal permeability test. Both sTNFRs highly exceeded TNF-alpha dialysate concentrations. In case of peritonitis a median 16-fold increase in dialysate TNF-alpha occurred on the first day, which declined towards control values during a longitudinal follow-up of eight consecutive days. sTNFRI and sTNFRII in dialysate increased three- to fourfold. Their peaks, however, appeared on the second peritonitis day. Bioactive TNF-alpha was only detected when immunoreactive levels exceeded 1000 pg/ml. Serum values of all variables were not altered during infection; sTNFRs exceeded TNF-alpha 300- to 400-fold. During stable CAPD indirect evidence was obtained for transperitoneal transport from plasma to dialysate of TNF-alpha (molecular wt 17 kD), sTNFRI (55 kD) and sTNFRII (75 kD). Dialysate/serum (D/S) ratios were higher, the lower the molecular weight; they were related to D/S ratios of those marker proteins with the nearest molecular weight; D/S ratios were unrelated to the intraperitoneally produced IL-6. Furthermore, the observed D/S ratios were as expected theoretically for their molecular weights. Higher than expected D/S ratios were found during peritonitis for TNF-alpha on days 1 and 2, and for sTNFRII on day 2, pointing to local release within the peritoneal cavity only in the acute inflammatory phase. Gel permeation chromatography revealed that TNF-alpha was present in a monomeric 17 kD form, unbound to receptors, whereas in case of peritonitis smaller sTNFRII fragments, transported at a higher rate, could not be excluded. Therefore, these higher than expected D/S ratios indicate local production of TNF-alpha during peritonitis and possibly also of sTNFRII, although transport of smaller receptor fragments might also occur.