A careful reading of the review of fever in “Fever: Friend or Foe?”, reveals the embarrassing deficiency in medical science’s understanding of how fever, much less anorexia, functions in infection. Since fever (as well as anorexia and other components of the acute-phase response) is induced by our own cytokines, it is virtually axiomatic that fever has been more beneficial than harmful on an evolutionary scale (since otherwise the response would have been deleted). As a “fan” of fever, I’ve compiled a list of six potential benefits of fever, each of which is found in the literature and has a reasonable experimental or theoretical basis (and each likely has some degree of correctness). Note that the first four are based on fever being a heat stressor. Have a look and then see my take on it.
Fever may work by:
1) directly harming pathogens,*
2) inducing apoptosis of infected cells (and neoplastic cells),
3) inducing host’s heat shock proteins to protect host cells during infection,
4) inducing heat shock proteins in pathogens—extracellular heat shock proteins activate immune responses as “danger signals”,
5) increasing efficacy of immune responses since they work better at slightly elevated temperatures,*
6) turning down the immune response by causing apoptosis of neutrophils and lymphocytes.
Reasons 1 and 5 (marked by an asterisk) are most commonly given. My take is that reason 2 fits with reason 1, since infected cells and tumor cells are essentially pathogens (or pawns thereof) and need to be killed. As to reason 3, this would imply that any temperature above that needed to induce host heat shock proteins would be harmful, and it’s not clear why the pathogens wouldn’t also gain protection from making their own heat shock proteins. Reason 4 explains that heat shock proteins can stimulate immune responses, but that seems like a wash—the heat shock proteins both protect and harm the pathogens. Reason 6 is unique in suggesting fever acts to turn down immune responses. I find this difficult to accept from an evolutionary/energetic standpoint—why expend so much energy and harm when there are lots of cheaper ways (such as cytokine signaling) to turn the response down? This leaves us with reasons 5 and 1&2, which I believe are both important, but … . While it is true that slightly higher temperatures do enhance many immune functions, I argue that fever didn’t evolve for this reason. Why would a system develop that requires a huge energy input to get optimal performance, rather than have the optimal defense develop at normal temperatures? Given that immune cells also perform best in slightly hypoxic and slightly acidic conditions, it seems most reasonable to me that immune cells evolved to perform optimally in their normal working environment, i.e. slightly warm, slightly hypoxic, and slightly acidic.
That leaves reasons 1 and 2, that fever directly harms pathogens and pathogen-producing cells. The problem never addressed is “How can fever harm pathogens (or infected or neoplastic cells) more than normal host cells or the host as a whole?” I don’t find the following answer satisfying: “Well, some pathogens may be harmed by the heat of fever. Others not. It all depends.”
Joe Alcock and I directly addressed this issue by considering fever as a heat stressor (LeGrand EK, Alcock J. 2012. Turning up the heat: immune brinksmanship in the acute-phase response. Q Rev Biol 87:3-18). Pathogens are typically pathogenic to extent that they grow and replicate rapidly, becoming relatively more vulnerable to stress than host cells and the host as a whole. Additionally, the pathogens are already subjected to the stressors at the infection site. The standard appropriate response to stress of any kind is to devote resources to ameliorating the stress or to become quiescent—it is risky to ignore stress and continue as before. I’ve been working with Dr. Judy Day, a biomathematician at the University of Tennessee, on computer simulations which confirm the concept of immune brinksmanship, that the host can typically benefit from using completely non-specific stress to harm rapidly growing and replicating pathogens relatively more than itself. Yes, it’s energetically costly and dangerous (potentially lethal, as in sepsis) but it is does provide benefit. We’ve also explored the benefits of completely non-specific stressors locally and regionally (think infection-site stressors such as low glucose, glutamine, iron, zinc, oxygen, pH, and increased reactive molecular intermediates). Host-derived local and regional stressors work with even less overall host damage than systemic stressors. However and as expected, host defenses are even more effective if they use specific stressors targeted against pathogens, thereby causing no collateral damage. The view of fever as a host-induced systemic stressor, to which rapidly growing and replicating pathogens are more susceptible, not only explains the “how” of fever, but it suggests times when fever and other systemic stressors may be more harmful than helpful. This would occur when the host finds itself more stressed or at risk than the pathogens, compatible with the reasonable view that it’s better and safer to be in good condition than in poor condition before getting infected.