In rats,
fever can be induced by injection of bacterial
lipopolysaccharide (LPS) into a subcutaneous air pouch. This febrile response is in part dependent on the local action of
interleukin-1beta within the pouch. In the present study, we tried to find out if this model of
fever induction can be applied in guinea pigs and if the local LPS-induced formation of
tumor necrosis factor-alpha (
TNF-alpha) participates in the development of the febrile response. A dose of 100 microg/kg LPS was injected into a subcutaneous air pouch along with
solvent (
0.9% saline), or 100 mg/kg
pentoxifylline, or 1 mg/animal soluble 55-kD
TNF receptor (referred to as TNF-
binding protein, TNF bp). The mean LPS-induced concentration of TNF in the lavage of the air pouch (measured 60 min after injection) was 17,765 pg/ml in animals injected with LPS and
solvent. This value was reduced to 7,631 pg/ml if
pentoxifylline was injected along with LPS into the pouch. If LPS was injected along with TNF bp, no bioactive TNF was detected in the lavage of the air pouch. Simultaneously, TNF was measured in blood plasma. Circulating concentrations of TNF were about 5-8% of the values detected in the lavage of the air pouch (mean values: 1,366 pg/ml in response to LPS plus
solvent; 377 pg/ml in response to LPS plus
pentoxifylline; no circulating TNF at all in response to LPS plus TNF bp). These data indicated a small spill-over of TNF from the air pouch into the circulation. Neither the reduction of local TNF formation in the air pouch by
pentoxifylline nor the complete neutralization of TNF within the pouch by TNF bp resulted in a significant attenuation of the febrile response induced by injection of LPS into the subcutaneous air pouch. If there were an activation of cutaneous nerves to induce
fever by a local formation of
cytokines within the air pouch, TNF would not represent a likely candidate to be responsible for such a neural stimulation.