Inflammation in Acne vulgaris: Failure of Skin Micro-Organisms to Modulate Keratinocyte Interleukin 1α Production in vitro

Despite decades of research, the pathogenesis of inflammatory acne vulgaris is still not fully understood. Histological studies have demonstrated that inflammatoray foci develop around acne lesions before overt spongiosis or rupture of the pilosebaceous follicle wall [1]. The initial cellular infiltrate in early inflammatory lesions of less than 6 h duration is mononuclear [1, 2] and has been shown to comprise predominantly CD4-positive T cells. The primary inflammatory events have been characterized as an increased expression of ICAM-1 and E-selectin on dermal microvascular endothelial cells, CD4-positive T-cell foci perivascularly and periductally associated with CD1-positive Langerhans cells and extensive MHC class II plus ICAM-1 expression on the cells within the infiltrate, endothelial cells and some basal keratinocytes [2]. The immunocytological data are consistent with a classical cutaneous type IV hypersensitivity reaction which may be initiated non-specifically by comedonal interleukin (IL) 1α and perpetuated by a specific CD4-positive T-cell response to persistent lesional antigens [3].The normal skin commensal bacterium Propionibacterium acnes has been implicated as an aetiological agent of inflammation in acne because treatments which reduce the population density of P. acnes are beneficial [4] and the failure of antibiotic therapy has been associated with the emergence of antibiotic-resistant P. acnes [5]. Numerous investigations, however, have failed to demonstrate an asssociation between the type or number of P. acnes colonizing lesions and the severity of acne. The normal cutaneous microflora also includes Staphylococcus and Malassezia species which are readily isolated from the surface and pilosebaceous follicular ducts of normal and acne-affected skin [6, 7]. In an extensive study of the pro-inflammatory cytokine content of acne comedones, it was shown that bioactive IL-1α-like material was present in 76% of open comedones and in 58% the levels exceeded 100 pg/ml – sufficient to initiate non-specific inflammation if released into the dermis [3]. An interesting association between the total number of micro-organisms present and the levels of IL-1α was demonstrated [3] which was independent of microbial species.It was therefore important to determine whether the cutaneous microflora was able to stimulate IL-1α production by keratinocytes and thus contribute indirectly to the initiation of inflammation in acne vulgaris.Isolates of P. acnes, P. granulosum, Staphylococcus epidermidis, S. capitis, S. hominis and Malassezia furfur serovar B were obtained by swabbing the facial skin of normal volunteers. Species were typed and cultured in 1-litre volumes in their appropriate culture media as described previously [8, 9, 10]. Cultures were harvested during the stationary phase, and intact cells were washed and either treated with formaldehyde overnight [1% (v/v) formaldehyde (Merck) in phosphate-buffered saline (PBS; Oxoid)] at 4 °C, rewashed in PBS and resuspended in RPMI-1640 (Gibco) or, after freezing at –70°C, were sheared by vortexing with glass beads in ice-cold PBS. Cell wall debris was sedimented by centrifugation at 12,000 g (Sorval; Dupont) for 10 min. The supernatant containing the soluble cellular fraction was freeze dried. The microbial culture supernatants were adjusted to pH 7.4, filtered (0.2 µm) and freeze dried. The freeze-dried fractions were reconstituted in tissue culture medium at 100 µg/ml protein (Bio-Rad Microassay).Normal human keratinocytes (NHK) were isolated from neonatal foreskins and cultured for 3 passages in keratinocyte serum-free medium (Life Technologies) supplemented with 5 ng/ml epidermal growth factor (Life Technologies) at 37°C in 5% (v/v) CO2 in air using standard procedures. Virus-transformed keratinocytes of the SKV14 cell line were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Life Technologies) plus suppplements.Keratinocytes were seeded in 96-flat-well plates (Falcon) at 1.5×104/well in keratinocyte serum-free medium (200 µl; NHKs) or DMEM (200 µl; SVK14s). Following overnight incubation at 37°C in 5% (v/v) CO2 in air to allow attachment, the culture medium was aspirated and replaced with 200 µl of medium plus microbial fraction. Final concentrations of microbial cells were 6×105/well and final protein concentrations of culture supernatants and cellular fractionates were 1 µg/ml. Appropriate microbial growth medium controls were included and all tests were performed in triplicate. Keratinocytes were cultured at 37°C for 72 h in 5% (v/v) CO2 in air in a humidified modular incubation. Co-incubation supernatants were harvested at 0, 24 and 72 h and centrifuged at 8,000 g (Microcentaur; MSE) for 3 min. Aliquots were then stored at –20°C for IL-1α determinations.The viability of the keratinocytes was determined at 0, 24 and 72 h by the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide) cleavage assay [11]. The absorbances at 570 nm were determined (MR7000 plate reader; Dynatech).Keratinocyte supernatants were assayed for IL-1α by ELISA (Biotrak IL-1α; Amersham International). Levels of IL-1α were expressed in units of specific activity (pg IL-1α per optical density at 570 nm). Levels of biologically active IL-1α were determined using a murine thymocyte proliferation assay as described previously [3]. Specificity was determined with anti-human-IL-1α-neutralizing antibodies (R & D Systems).Data were analysed by one-way analysis of variance and minimum significant differences between means calculated by the T method.The viability of the NHK cells and SVK14 cells as determined by the MTT cleavage assay was not significantly affected by any of the microbial fractions throughout the 72-hour co-incubation period.The levels of IL-1α released by keratinocytes determined by ELISA ranged from 0 to 136 pg/ml and were generally maximal at 72 h. Levels of IL-1α specific activity for the negative controls at 72 h were as follows: NHK strain 1, 0.23±0.09; NHK strain 2, 0.12±0.10; SVK14, 0.5±0.01.The results obtained when NHK strains 1 and 2 and SVK14 cells were cultured with whole microbial cells after 72 h are shown in figure 1. When co-incubated with microbial fractions there was a trend towards increased IL-1α production with NHK strain 1 producing more IL-1α than NHK strain 2. The values obtained using SVK14 cells were extremely low irrespective of microbial stimulus (≤0.1 units of specific activity). As shown in figure 1, NHK strain 1 incubated with S. epidermidis whole cells and SVK14 cells incubated with S. capitis whole cells released significantly more IL-1α than the negative controls (p<0.01). None of the other microbial cells tested significantly increased IL-1α production by the keratinocytes at either 24 or 72 h. NHK strain 2 co-incubated with the cellular fraction (1 µg/ml protein) of M. furfur and the cellular fraction of P. granulosum produced significantly more IL-1α than the negative control (p<0.05; data not shown). The culture supernatant of P. granulosum also stimulated increased IL-1α production at 72 h by NHK strain 2 (p<0.05; data not shown). For the remaining cell fractions and culture supernatants tested there were no significant effects of IL-1α production by the keratinocytes.Biologically active IL-1α was only detected in 26% of the samples tested (limit of detection 1.5 pg/ml) and therefore it was not possible to analyse these data by ANOVA. Maximal bioactivity was detected in the 72-hour cultures which was consistent with the IL-1α ELISA data. There was no correlation between the levels of bioactive IL-1α in the positive samples and the equivalent values for immunochemical IL-1α (r=0.02; Pearson’s product moment correlation coefficient).Low levels of IL-1α were release constitutively by cultured human keratinocytes and there was a trend towards slightly higher levels being produced in response to the microbial stimulants. Significant increases in specific IL-1α production were, however, only observed in isolated tests, e.g. one strain of NHK with whole cells of S. epidermidis and S.capitis. Overall, the results did not support the hypothesis that cutaneous micro-organisms or their products will consistently up-regulate IL-1α release by keratinocytes directly. In addition, the IL-1α measured in this study was generally biologically inactive. This could have been due to several factors, e.g. the presence of IL-1 receptor antagonist, cytokine inhibitors or loss of bioactivity upon storage of the samples.The association between biologically active IL-1α in comedones with high levels of total micro-organisms demonstrated by Ingham et al. [3] could be explained by an indirect effect of cutaneous micro-organisms in vivo, for example by the stimulation of a third factor which interacts with keratinocytes to release IL-1α. It is also possible that, for reasons as yet unknown, some comedones harbour high numbers of micro-organisms and have high levels of IL-1α due to other factors and that the two are not biologically related.