Human keratinocyte sensitivity towards inflammatory cytokines varies with culture time

Proliferating keratinocyte cultures have been reported to synthesize higher concentrations of prostaglandin (PG) E than confluent ones. As interleukin-1 (IL-1) stimulates keratinocyte PGE synthesis we investigated whether the degree of confluency of the keratinocyte culture modified the response of the cells to IL-1. It was found that IL-1α (100 U/ml) stimulated PGE2 synthesis by proliferating (7 days in culture) but not differentiating (14 days in culture) keratinocytes. Similar effects were observed using tumour necrosis factor-α. Both arachidonic acid (AA) and the calcium ionophore A23187 stimulated PGE2 synthesis by 7 and 14 day cultures although the increase was greatest when 7 day cultures were used. Our data indicate that there is a specific down-regulation of the mechanism(s) by which some inflammatory cytokines stimulate keratinocyte eicosanoid synthesis as cultured keratinocytes begin to differentiate.


Introduction
Interleukin-1 (IL-lcz and IL-lfl) is a family of pluripotent pro-inflammatory cytokines found in high concentrations in the normal epidermis and synthesized by keratinocytes in vitro. 1-3 IL-1 has been reported to stimulate keratinocyte proliferation in vitro 4'5 and the re-epithelialization of split-skin wounds in human volunteers (personal communication). IL-1 also stimulates the synthesis of keratinocyte prostaglandin (PG) E2, 6 a cyclooxygenase (CO) metabolite of arachidonic acid (AA), which enhances the proliferation of basal keratinocytes in culture and epidermal thymidine uptake in vivo. 7-1 It is possible, therefore, that the stimulation of keratinocyte proliferation reported for IL-1 is mediated via an increase in PGE2 synthesis. However, the absence of inflammatory/proliferative reactions in the normal epidermis indicates that interactions between keratinocytes and IL-1 are tightly controlled. Keratinocyte eicosanoid synthesis varies with its state of differentiationl'2 and proliferating keratinocyte cultures synthesize greater concentrations of PGE2 than confluent ones. 7'13'14 It is possible, therefore, that sensitivity towards IL-1 is a function of the proliferative state/state of differentiation of the target cell and that proliferating and confluent keratinocyte cultures may also show differences in sensitivity towards IL-I in vitro. Using cultured keratinocytes as a model system, alterations in the sensitivity of these ceils towards IL-I with time in culture were investigated, and assayed as changes in IL-1 z stimulated PGE2 synthesis. IL-I was used in these experiments as it is the biologically active form of IL-1 present in the epidermis. However, IL-lz and IL-lfl have been shown to be equally potent in stimulating keratinocyte PGE2 synthesis.
In order to determine the specificity of any changes observed we also monitored the effect of IL-lcz on keratinocyte IL-6 and TNFsecretion and the effect of TNFon keratinocyte PGE2 synthesis.  8.4 ng/ml cholera toxin and 10 ng/ml epidermal growth factor (complete medium) as previously described, is Confluent cells were trypsinized and 10 4 cells/cm 2 re-seeded on a fresh feeder layer in complete medium. The day before the cultures were used, remnants of the feeder layer were removed by vigorous washing. Cultures were sometimes incubated overnight in complete medium before being used in experiments. Such secondary keratinocyte cultures usually reach confluency after 8 days. For the experiments, keratinocytes were used after 7 or 14 days in culture.

Materials and Methods
The release of prostaglandin E2, TNF-o and 1L-6 from keratinocytes" Keratinocyte cultures were incubated (37C, 5% COg) for 3 h or 24 h 6'16 in 1 ml complete medium or DMEM/F12 with or without IL-10.
Portions of the incubation media were stored at -30C for later analysis of PGE2 and IL-6 (kitinstructins) by radioimmunoassay. 7 Analyses for TNF-o were performed immediately after the incubation period using a modification of the L929 cytotoxic assay. 18 L929 cells were cultured in complete RPMI medium (RPMI supplemented with 10% FCS, 2 mM glutamine, 100U/ml penicillin/streptomycin and 5 x 10-s M mercaptoethanol). For use in the cytotoxicity assay 104 viable cells (trypan blue exclusion) in 0.1 ml complete RPMI medium were transferred to 96-well plates and allowed to equilibrate overnight. The next day, 25/1 actinomycin D (final concentration 1/g/ml) was added together with 25 #1 TNF-0 standard (0-500 U/ml final concentration) or 25 1 sample.
The microtitre plates were then incubated overnight, the media were discarded and remaining viable cells were stained with crystal violet (0.2% in 2% ethanol) for 10 min. Microtitre plates were rinsed 5 x with tap water and the stained cells solubilized using 0.1 ml of 1% sodium lauryl sulphate. Absorbance was measured at 595 nm using a Bio-Rad model 3550 microtitre plate reader and the TNF-0 activity in the samples calculated on the basis of the standard curve. Sensitivity of the assay was 0.5 U/ml TNF-o. Reversed phase chromatography was performed using 30% acetonitrile/70% water (pH 2.4, 37C) with a flow of I ml/min for 35 min. The acetonitrile concentration was then increased to 49% over a 13 min period and maintained at that concentration for 40 min. AA was eluted in 100% acetonitrile. Detection of radiolabelled compounds was performed using an on-line Berthold LB 506c radioactivity monitor (Wildbad, Germany) controlled by the HP 1084B terminal.

Results
Basal and IL-1 stimulated synthesis of eicosanoids, IL-6 and TNF-o by 7 and 14 day keratinocyte cultures: Synthesis of PGE2 during the 24 h incubation period directly following removal of the feeder layer was greater than in the subsequent 24 h period, indicating that the washing process stimulated PG formation. PGE2 formation was stimulated by FCS ( Table 1). The basal rate of PGE2 synthesis was greater by 7 day than by 14 day keratinocyte cultures both in  (Fig. 2). However, the stimulatory eect during the first 3h period was not o u. constant whereas the effect during the second period was reproducible ( Table 2). Neither IL-6 (limit of detection of assays 0.1 ng/ml, Amersham and Advanced Magnetics kits) nor TNF-0 (limit of detection of bio-assay 0.5 U/ml) could be detected in the incubation medium of 7 day and 14 day cultures (incubation periods: 3 h and 24 h, with or without FCS, with or without 100 U/ml IL-I). The effect of arachidonic acid and the calcium ionophore A23187 on keratinocyte prostaglandin E2: While AA and A23187 stimulated the synthesis of PGE2 by both 7 and 14 day keratinocyte cultures, PGE2 synthesis (ng/ml) was greater when 7 day cultures were used (Table   3). Whereas the relative stimulatory effect of AA was similar for 7 day and 14 day cultures, the relative increase in PGE2 synthesis when A23187 15  during the 20 min incubation, was small and less than 1% of added radioactivity was associated with metabolites recovered from the incubation medium.
Ther was some variation between replicates and  Figure 3. Preincubating 7 day keratinocyte cultures with IL-10 (1-100U/ml) resulted in a small increase in the synthesis of radiolabelled PGE2 in those cultures incubated with 100U/ml (mean dpm S.D., n 3, control 320 10, IL-10 100 U/ml 560 _q-50; % of total metabolite dpm, control 23% -+-4.0, IL-I 100 U/ml 34% -F 6.0). There were no significant increases in the synthesis of other eicosanoids.  The eect of TNFon keratinocyte eicosanoid synthesis: TNF-(100 U/ml) stimulated the synthesis of PGE2 by 7 day keratinocyte cultures during a 3 h incubation. It had no effect on PG formation by 14 day cultures (Table 4).

Discussion
The sensitivity of human keratinocytes towards IL-10 and TNF-0 in vitro varied with time in culture. It was found that IL-I and TNFstimulated PGE2 synthesis by keratinocytes after 7 days, but not 14 days, in culture. This loss of sensitivity towards IL-10 and TNF-0 was associated with a reduction in both basal and stimulated (AA and A23187) PGE2 formation. Although PGE2 synthesis was enhanced when foetal calf serum was added to the incubation medium, maybe due to the presence of AA in the serum, IL-I and TNF-0 stimulation of PGE2 synthesis was not dependent on the presence of serum or added growth factors. The finding that pre-exposure of keratinocytes to IL-10 was sufficient to stimulate PGE2 synthesis (first and second 3 h incubations) is consistent with reports that IL-I stimulated AA turnover in fibroblasts and keratinocytes is dependent on de novo protein synthesis. 6'16'2-22 Data obtained using radiolabelled AA are in general agreement with reports indicating that keratinocytes are capable of synthesizing different cyclooxygenase (CO) (PGD2, 6-keto PGFI, PGE2) and lipoxygenase (15-HETE, LTB 4 and leukotriene C4) metabolites of AA. 11-13'23-25 However, it appears from our data that the stimulatory action of IL-I could be selective for PGE2 although this has to be confirmed using RIAs. AA stimulated PGE2 synthesis by 7 day keratinocyte cultures, calculated as the relative increase, was the same as for 14 day cells. The differences in AA stimulated PGE2 formation by 7 and 14 day cultures can, therefore, be ascribed to changes in keratinocyte CO concentrations. The stimulatory effect of A23187 on 7 day cultures was, in contrast, greater in both absolute (per mg protein) and relative terms than its effect on 14 day cultures. This finding suggests that intra-cellular, Ca 2+ sensitive, mechanisms controlling the level of free AA are upor down-regulated depending on the degree of differentiation of the culture. Increases in de novo CO and phospholipase A2, a Ca 2+ sensitive enzyme which releases AA from membrane phospholipid stores, 2 have been shown to occur in fibroblasts after exposure to It-1.  It is possible therefore that an increase in the synthesis of these enzymes is coupled to those mechanisms regulating keratinocyte proliferation/differentiation. As IL-I and TNFwere without effect on 14 day cultures it appears that there are also specific alterations in the receptor mediated signal transduction system by which these cytokines stimulate PGE2 formation. Our data are consistent with the concept that, as keratinocyte cultures proliferate and become confluent, there are changes in basal and stimulated PGE2 synthesis which can be attributed to alterations at receptor, second messenger and CO levels.
The present results support previous publications showing that non-confluent keratinocyte cultures release higher concentrations of PGE2 than confluent ones. "'13'4 However data obtained using different isolated epidermal cell populations is contradictory and authors have reported that both difl:erentiated and basal keratinocytes are the most active in metabolizing AA. 11'12 These differences have yet to be reconciled. Concentrations of IL-6 and TNF in keratinocyte incubation media were below the levels of detection of the assays used, for all the experiments performed. This finding is in contrast to other publications reporting that IL-6 release from keratinocytes is enhanced by IL-1. 3'26 Those authors assayed IL-6 using the B9 cell bioassay however and we are not aware of any report where the release of IL-6 from keratinocytes was measured by RIA. It is possible that the different assay systems do not yield compatible results, although IL-6 secretion by human fibroblast was easily detected using the Amersham kit. 1 In view of the lack of effect on secretion of IL-6 or TNF-o, it is unlikely that the stimulatory effect of IL-10 on keratinocyte PGE2 synthesis was mediated via a stimulation of secondary cytokines.
In summary, it has been shown that there are alterations in the sensitivity of keratinocytes towards IL-lo and TNFas keratinocyte cultures become confluent and begin to differentiate. The down-regulation of PGE2 synthesis observed, as keratinocytes begin to dierentiate, occurs at the receptor, second messenger and CO levels. Such changes in the sensitivity of keratinocytes towards inflammatory cytokines could be an important mechanism by which the response of the epidermis to injury (re-epithelialization) or noxious stimuli (hyperplasia) is regulated.