The physiological and pathophysiological functions of C-reactive protein (CRP), the classical acute-phase protein, are not well established, despite many reports of biological effects of CRP in vitro and in model systems in vivo. Limited, small scale experiments have suggested that rabbit and human CRP may both protect mice against lethal toxicity of Gram-negative bacterial LPS. However, in substantial well-controlled studies in C57BL/6 mice challenged with Escherichia coli O111:B4 LPS, we show in this work that significant protection against lethality was conferred neither by an autologous acute-phase response to sterile inflammatory stimuli given to wild-type mice 24 h before LPS challenge, nor by human CRP, whether passively administered or expressed transgenically. Male mice transgenic for human CRP, which mount a major acute-phase response of human CRP after LPS injection, were also not protected against the lethality of LPS from either E. coli O55:B5 or Salmonella typhimurium. Even when the acute-phase human CRP response was actively stimulated in transgenic mice before LPS challenge, no protection against LPS toxicity was observed. Indeed, male mice transgenic for human CRP that were pretreated with casein to stimulate an acute-phase response 24 h before LPS challenge suffered significantly greater mortality than unstimulated human CRP transgenic controls. Rather than being protective in this situation, human CRP may thus have pathogenic proinflammatory effects in vivo.