In the absence of information on functional manifestations of carotid body (CB) inflammation, we studied an experimental model in which lipopolysaccharide (LPS) administration to pentobarbitone-anaesthetized cats was performed by topical application upon the CB surface or by intravenous infusion (endotoxaemia). The latter caused: (i) disorganization of CB glomoids, increased connective tissue, and rapid recruitment of polymorphonuclear cells into the vascular bed and parenchyma within 4 h; (ii) increased respiratory frequency and diminished ventilatory chemoreflex responses to brief hypoxia (breathing 100% N(2) for 10 s) and diminished ventilatory chemosensory drive (assessed by 100% O(2) tests) during normoxia and hypoxia; (iii) tachycardia, increased haematocrit and systemic hypotension in response to LPS i.v.; and (iv) increased basal frequency of carotid chemosensory discharges during normoxia, but no change in maximal chemoreceptor responses to brief hypoxic exposures. Lipopolysaccharide-induced tachypnoea was prevented by prior bilateral carotid neurotomy. Apoptosis was not observed in CBs from cats subjected to endotoxaemia. Searching for pro-inflammatory mediators, tumour necrosis factor-alpha (TNF-alpha) was localized by immunohistochemistry in glomus and endothelial cells; reverse transcriptase-polymerase chain reaction revealed that the CB expresses the mRNAs for both type-1 (TNF-R1) and type-2 TNF-alpha receptors (TNF-R2); Western blot confirmed a band of the size expected for TNF-R1; and histochemistry showed the presence of TNF-R1 in glomus cells and of TNF-R2 in endothelial cells. Experiments in vitro showed that the frequency of carotid nerve discharges recorded from CBs perfused and superfused under normoxic conditions was not significantly modified by TNF-alpha, but that the enhanced frequency of chemosensory discharges recorded along responses to hypoxic stimulation was transiently diminished in a dose-dependent manner by TNF-alpha injections. The results suggest that the CB may operate as a sensor for immune signals, that the CB exhibits histological features of acute inflammation induced by LPS, that TNF-alpha may participate in LPS-induced changes in chemosensory activity and that some pathophysiological reactions to high levels of LPS in the bloodstream may originate from changes in CB function.