Originally, the immune system was seen as a system that primarily combats infection. But, as discussed in the recent Review article by Grace Chen and Gabriel Nuñez (Sterile inflammation: sensing and reacting to damage. Nature Rev. Immunol. 10, 826–837 (2010))1, most of us accept the idea that the immune system is key to tissue homeostasis and even to homeostatic interactions with the outside antigenic world, including the gut microbiota1,2. However, much remains unknown about the nature of the triggers of pro-inflammatory innate immune responses. In a broad sense, two main types of activators prevail: non-self antigens (known as pathogen-associated molecular patterns (PAMPs)), which are present in or released from infectious invaders; and damage-associated molecular patterns (DAMPs), which are host molecules (such as high-mobility group box 1 protein (HMGB1)) released from damaged cells under necrotic but not apoptotic conditions.

Heat shock proteins (HSPs) are frequently mentioned as prime examples of DAMPs (see, for example, Refs 1,3). There are, however, several qualities inherent to HSPs that disqualify them as DAMPs.

First, DAMPs are intracellular molecules normally hidden from recognition by the immune system, whereas HSPs are freely present in the extracellular fluids4,5 and are also frequently exposed at the outer surface of cells (both eukaryotic and prokaryotic cells)6,7. Second, Toll-like receptor 2 (TLR2) and TLR4 are seen as two of the main receptors involved in the recognition of HSP60 and HSP70. These HSPs are released from cells under necrotic conditions, but it has been shown that TLR2 and TLR4 are not required for the host response to DAMPs that are derived from necrotic cells8. Moreover, other receptors for HSPs, such as the HSP70 scavenger receptor SRA1, have been shown to confer a suppressive rather than an activating signal to host cells9. Third, in vitro-cultured dendritic cells have been shown to adopt a tolerizing phenotype, rather than a mature or activated phenotype, in the presence of HSPs10,11. Fourth, in experimental models of autoimmunity and of tissue or tumour transplants, immunization with HSPs was shown to lead to the induction of regulatory T cells, which suppressed disease or transplant rejection12,13,14,15,16,17,18,19.

Taken together, these phenomena argue against the involvement of HSPs in the induction of the immune response to damage-derived signals. On the contrary, HSPs seem to have a dampening effect on immune activation and have the capacity to promote immune homeostasis12,20.

A possible reason for the proposition that HSPs are DAMPs could be that some of the early studies used recombinant HSP preparations that were contaminated with lipopolysaccharide (LPS), although in most cases contamination levels were not determined10,21. More recent studies (reviewed in Ref. 5) using HSP preparations from which the contaminants had been effectively removed did not provide evidence to support a pro-inflammatory function for HSPs.

Based on the current experimental data, it seems that HSPs are key elements in the type of immune system responsiveness or reactivity that is induced by the following three typical conditions: one, sterile tissue damage, when tissue-derived HSPs in combination with DAMPs activate tissue repair and regulation; two, damage caused by infectious pathogens, when HSPs in the presence of DAMPs and pathogen-derived PAMPs lead to a full pro-inflammatory response with elimination of infection and regulation; and three, homeostatic interactions with commensal symbionts of the gut microbiota, when a combination of PAMPs and HSPs leads to the regulation and maintenance of symbiosis22,23,24. Given the ubiquitous presence and stress-inducible nature of HSPs in both tissue cells and microbial invaders or symbionts, therapeutic targeting of HSPs offers an attractive possibility for the fine-tuning of such immune responses and the dampening of inflammation through the induction or activation of regulatory T cells.