DNA‐triggered AIM2 inflammasome activation in keratinocytes: Comment on Kopfnagel et al. Exp Dermatol. 2011. 20:1027–9
Y. Dombrowski,Sarah Koglin,J. Schauber
DOI: https://doi.org/10.1111/j.1600-0625.2012.01466.x
2012-06-01
Experimental Dermatology
Abstract:Editor, With great interest, we read the publication ‘Human keratinocytes express AIM2 and respond to dsDNA with IL-1b secretion’ by Kopfnagel et al., which was recently published in Experimental Dermatology (1). AIM2 is a cytosolic sensor for double-stranded DNA (dsDNA) and forms an IL-1b-activating inflammasome upon activation (2,3). DNA from viruses, intracellular bacteria but also endogenous, cytosolic self-DNA is sensed by this receptor suggesting that AIM2 plays a role in infectious, but maybe also in inflammatory skin disease (4,5). As we have extensively studied the expression and function of the cytosolic DNA sensor AIM2 in skin, we would like to comment on the reported findings: While Kopfnagel et al. confirm our observations of increased AIM2 expression in lesional skin in psoriasis published recently (6), they also observed increased AIM2 protein expression in healthy skin. When we analysed AIM2 expression in healthy versus lesional psoriasis vulgaris (PSV) skin and in lesional versus non-lesional skin of patients with PSV, significantly elevated AIM2 mRNA expression in lesional psoriatic skin was detected (6). Furthermore, AIM2 mRNA expression in lesional skin was significantly increased compared with nonlesional skin of the same donor. In contrast, AIM2 mRNA in healthy skin was hardly detectable in a large cohort of non-psoriatic volunteers. In extensive immunohistochemical studies, no AIM2 protein expression in keratinocytes in healthy skin was observed in contrast to lesional psoriatic skin where strong, specific AIM2 staining in the epidermis was detected (6). These results were confirmed by Western blot analyses of healthy and PSV skin in the meantime (Y. Dombrowski, S. Koglin & J. Schauber, unpublished data). The strongest immunohistochemical staining in lesional psoriatic skin was detected in the apical keratinocyte layer, while the dermis remained negative – in contrast to the data presented by Kopfnagel et al. who also show strong staining of dermal collagen fibres (6). How could this discrepancy be explained? We used a different antibody for AIM2 detection in our studies [originally developed by V. Hornung et al. (7)], but initially tested different antibodies against AIM2 including the one used by Kopfnagel et al. All of those antibodies – except for the one we used in our studies – were not useful for AIM2 detection as they produced high background because of unspecific binding. Probably, Kopfnagel et al. had the same difficulties as high unspecific staining of the dermis is visible in Fig. 1a. Hence, as the same antibody was used in Western blot analysis, we believe that AIM2 detection in healthy skin shown in Fig. 1 might reflect unspecific antibody binding. Furthermore, as a member of the HIN200 gene family, AIM2 gene transcription is induced by IFNc, which is high in inflammatory skin diseases such as psoriasis and chronic atopic dermatitis, but not found in healthy skin (8). Our results demonstrated earlier that IFNc production and AIM2 expression closely correlate with inflamed skin in psoriasis (6). As a consequence of the absence of IFNc, AIM2 is scarcely detectable in healthy tissue and unstimulated cultured keratinocytes, respectively. As a functional consequence, only IFNc-prestimulated keratinocytes showed IL-1b secretion after stimulation with cytosolic DNA in our experiments through AIM2 (as confirmed by siRNA knockdown), whereas a functional role for the proposed AIM2 expression in unstimulated keratinocytes is not provided in the recent study (1). Another point which we wish to comment on is the experimental setting used for detection of DNA uptake in keratinocytes. Human keratinocytes are able to take up exogenous DNA by micropinocytosis or through receptor interactions (9). The DNA used by Kopfnagel et al. in their experiments, however, was complexed with the lipophilic tail LyoVec (poly(dA:dT)/LyoVec, InvivoGen, San Diego, CA, USA) – a cationic lipid of the phosphonolipid family, which is used for the transfection of oligonucleotides: LyoVec binds to DNA because of opposite electrostatic charges and enables fusion to cellular membranes (10; InvivoGen). Therefore, it is not surprising that poly(dA:dT)/LyoVec was detected in intracellular compartments [shown in supplementary figure S3, (1)] as it would be with any other transmembrane transfection method. Additionally, DAPI, which was used for DNA detection, also detects nuclear DNA, which makes it difficult to distinguish between cytosolic and nuclear DNA localization in the presented figures. Finally, it is surprising that unstimulated human keratinocytes used in the presented study actively secreted high amounts of IL1b (Fig. 2a). This would suggest constant inflammasome activity and – when translated to the in vivo situation – a chronic background inflammation in the skin as IL-1b is a potent activator of immune responses (11,12). To prevent undesired inflammation, IL-1b secretion is controlled by a two-step process: a first signal increases pro-IL-1b expression through an NFjB-dependent mechanism, while a second (danger-related) signal initiates IL-1b processing into its biologically active form (11,13). It has been shown earlier that human keratinocytes store large amounts of preformed pro-IL-1b (14). Still, in NHEK, the second signal is mandatory for IL-1b processing and activation. Of course, these differences to prior studies could be explained by different isolation and/or culture conditions as suggested by the authors.