Regulation by anti-inflammatory cytokines (IL-4, IL-10, IL-13, TGFβ)of interleukin-8 production by LPS- and/ or TNFα-activated human polymorphonuclear cells

The capacity to down-regulate the production of IL-8 by LPS-activated human polymorphonuclear cells (PMN) has been demonstrated for IL-4, IL-10, and TGFβ. We compared their relative capacities and further extended this property to IL-13. We report a great heterogeneity among individuals related to the responsiveness of PMN to the IL-4 and IL-13 inhibitory effects while their response to the IL-10 effect was homogenous. The inhibitory activities were observed at the transcriptional level. IL-8 induction by TNFα was, unlike its induction by LPS, resistant to the inhibitory effects of IL-10, IL-4, IL-13 and TGFβ. Furthermore, IL-10 and IL-4 inhibitory activity were less effective when TNFα was acting synergistically with LPS to induce IL-8 production by PMN. LPS-induced cell-associated IL-8, detected in the PMN cultures, could be marginally inhibited by IL-4 and IL-10. Altogether, our data demonstrate that IL-13 is able to inhibit LPS-induced IL-8 production by human PMN, although IL-10 remains the most active anti-inflammatory cytokine. Despite the capacity of IL-4, IL-10, and IL-13 to limit the production of TNFα-induced IL-8 in a whole blood assay, none was able to inhibit this production when studying isolated human polymorphonuclear cells.


Introduction
Inflammation is characterized by a complex interaction of proand anti-inflammatory mediators. A favourable balance leads to a recovery of homeostasis while an unfavourable balance, associated with an exacerbated release of proinflammatory cytokines, leads to deleterious effects. Individual background seems to govern this equilibrium. In man, for example, it is known that a great heterogeneity exists between individuals in terms of their levels of production of TNFcz by LPS-stimulated monocytes. Similarly, the capacity to produce the Th2 cytokines, including the anti-inflammatory IL-4 and IL-IO cytokines, rather than the Thlderived cytokines which include the pro-inflammatory interferonq,, seems to be genetically determined. 2 Furthermore, the sensitivity of target cells to cytokine signals constitutes another 3 level of heterogeneity. The capacity of IL-4, IL-1 O, IL-13 and TGF to inhibit the synthesis of pro-inflammatory cytokines by activated monoces/macrophages has been well estab- 4 7 lished.-However, the mononuclear phagocytes are not the unique source of IL-1, TNFcz, 334 Mediators of Inflammation Vol 5 1996 IL-6 and IL-8, and many other cell types can contribute to the inflammatory process as a consequence of their ability to produce proinflammatory cytokines. Upon activation, polymorphonuclear cells (PMN) are a major source of mediators, including eicosanoids, proteolytic enzymes and free radicals, all of which contribute to the exacerbation of inflammatory processes. In vivo models of infection have clearly established that PMNs do contribute to tissue injury and organ dysfunction and are involved in the development of fatal complication in sepsis syndrome.  Their capacity to be activated by cytokines following the release of bacterial derived substances has been illustrated in in vivo models of neutrophil depletion. 11 '12 Indeed, many substances, including opsonized yeast, bacterial-derived productsusuch as endotoxin (LPS, lipopolysaccharide)mand cyt0kinesmsuch as IL-1, IL-2 and TNFzmare able to trigger MN, leading to a notable production of IL-8.13-IL-8, being a potent chemokine and activator of PMN, has emerged in recent years as a pivotal factor of inflammation. The presence of IL-8 in vivo has been clearly associated 7 19 with neutrophil recruitment and the sever-  [23][24][25][26][27][28] In addition, experiments using RPMI-1640 medium supplemented with L-glutaanti-IL-8 antibodies have clearly demonstrated mine, antibiotics and 5% heat inactivated northe contribution of IL-8 in the inflammatory mal human serum (a pool of sera from healthy process. [29][30][31] In the present study we have volunteers). 0.5 ml per well in 24-well multidish compared the relative capacities of IL-4, IL-IO, plates of PMN suspension were incubated for IL-13 and TGF to interfere with IL-8 produc-18-24 h at 37C in a 5% CO2 incubator. The tion by activated human PMN. This is the first cell viability after 24 h ranged from 64% to 84%. report on the individual heterogeneity related Stimuli were added at the beginning of the to the inhibitory activities of IL-4 and IL-13. In culture in a volume not exceeding 10 >1. At the addition, we demonstrated that these cytokines end of the culture, the supernatants were are devoid of any activity when TNF was harvested, centrifuged 10 min at 300 g and employed as the triggering agent, and showed 15C and kept at -20C before cytokine assessthat the inhibitory activity of IL-IO on LPSments. In some experiments, at the end of the induced IL-8 was significantly hampered by the culture period, the cells were lysed by adding presence of TNFa.
the pellet was washed and centrifuged once 5 min at 300 g. Contaminating erythrocytes Ik-8 ELISA were lysed following a 5 min incubation at 4C of the cell pellet resuspended in 5 ml of lysis IL-8 ELISA was performed as previously buffer (NH4C1-8.32 g/l; NaHCO-0.84 g/l; described 28 using a monoclonal anti-human IL-8 Na4EDTA 43.2 mg/1). Lysis was stopped by antibody and a rabbit polyclonal anti-IL-8 antiadding a large excess of RPMI and the cells body kindly provided by N. Vita (Sanofi Rewashed and centrifuged for 10 min at 200 g. cherche, Labge, France). The viability of polymorphonuclear cells (PMN) was assessed by numerating the cells in 0.1% Northern blot analysis cosine. A non-specific esterase staining was performed to evaluate the monocyte contamina-Total cellular RNA from 20-30 106 PMN was tion which never exceeded 0.5%. extracted using the guanidine isothiocyanate method. 32 5-10 g total RNA per wells were separated on 1% agarose gels containing 6.6% In vitro culture formaldehyde, transferred to Hybond-N + mem-For whole blood assays, blood was diluted 1:5 branes (Amersham) and hybridized with 32p_ in RPMI-1640 medium (Bio-Whittaker, USA) labelled IL-8 and -actin probes. The human IL-8 supplemented with L-glutamine (300 mg/1) and cDNA probe, generously provided by Dr J.

Statistical analysis
Results are expressed as mean 4-SEM. Statistical analysis was performed using the Wilcoxon signed-rank test.

Results
Inhibition of IL-8 production by IL-13 compared with other antiinflammatory cytokines As shown in Fig. 1A, IL-4, IL-IO, and IL-13 are able to inhibit in a dose-dependent fashion the IL-8 production in a whole blood assay triggered by LPS. The results are the mean of experiments performed with five different donors and were homogeneous. This was not the case when these cytokines were studied with isolated human polymorphonuclear cells (see below, the individual responsiveness). A dose-response curve of the inhibitory effect of IL-4, IL-10 and IL-13 on LPS-induced IL-8 production by activated human PMN from one given donor is shown in Fig. lB. IL-IO, IL-4 and IL-13 were able, in a dose-dependent fashion, to inhibit the release of IL-8 by LPS-activated PMN. In the whole blood assay as well as with the isolated PMN, no further inhibitory effects could be obtained with 30 ng/ml of cytokines, compared with 10 ng/ml, and subsequently, a fixed cyt0kine concentration of 10 ng/ml was used. Neither IL-IO nor IL-13 had any effect on the 'spontaneous' IL-8 production by human PMN (262 -t-64 pg/ml and 240 4-41 pg/ml, respectively vs 191 4-39 pg/ml; n 12). IL-4 had a weak but significant capacity to limit the spontaneous production of IL-8 (103 4-20 pg/ml; p < 0.01) while TGF amplify it (588 + 234 pg/ ml; p 0.05).
Inhibitory activities of IL-IO, IL-4, IL-13 and TGF were compared on PMN activated by either LPS or TNF(z. As shown in Fig. 2A, the studies on the inhibition of LPS-induced Ib8 production performed with the cells from different donors, revealed that IL-IO at 10 ng/ml was very efficient whereas similar amounts of Ib4 were less active (81% inhibition vs 37% inhibition; p < 0.01). IL-13 and TGF, with respectively 16% and 12% inhibition, were very weakly inhibitory, although this was statistically significant in view of the large number of experiments performed (n 12).
When TNF(z was employed as the stimulating agent, IL-IO had a high inhibitory activity, and IL-4 and IL-13 had similar capacity to counteract the production of IL-8 by whole blood cells, while TGF was devoid of any inhibitory activity (data not shown). In contrast to the results obtained in whole blood assays, none of the anti-inflammatory cytokines were able to counteract the IL-8 production induced by TNF0t (Fig. 2B).
Synergy between LPS and TNF: effect of anti-inflammatory cytokines Since cells are most likely exposed to many triggering signals during infection by Gram negative bacteria, it was interesting to investigate the capacity of the anti-inflammatory cytokines to reduce the IL-8 production stimulated by both LPS and TNF(z. The results are shown in Fig. 2C  was noticed when PMN were studied (arithmetic sum 5007 4-514 pg/ml vs experimental data 6225 -+-651 pg/ml, n 17, p < 0.02). In whole blood cells, IL-10 and IL-4 were particularly efficient in inhibiting IL-8 production (data not shown). On the contrary with PMN, only IL-10 dramatically inhibited IL-8 production, while IL-4 was weakly inhibitory, and IL-13 and TGF were not significantly inhibitory. However, the inhibitory activity of IL-IO, which was higher than 80% when LPS was used alone, dropped to only 45% (for the same concentration of 10 ng/ ml) when both activators were employed.  (Table 1). To further evaluate the proportion of the cytokine linked to its receptor on PMN, IL-8 was measured following treatment of cells with a glycine-HC1 buffer. One-fifth of the cell-associated IL-8 was recovered with such a treatment (data not shown).
IL-4, IL-10 and IL-13 but not TGF could also reduce the presence of cell-associated IL-8 in the PMN cultures (Table 2). However, the reduction was less pronounced than when studying released IL-8 and did not reach statistical significance. When the results were individually plotted (Fig.  3), it was noticed that the inhibitory effect of IL-10 was very similar between donors (range: from 65% to 90% inhibition) whilst the effect of IL-4 varied greatly depending on donor (range: from 0% to 78% inhibition). Similarly, IL-13-

Discussion
The intensity of the inflammatory response is the result of a complex balance between proand anti-inflammatory mediators. Among the set of cytokines, IL-1, IFN/, and TNF( orchestrate the inflammatory process, whilst IL-4, IL-10, IL-13, and TGFI3 have been characterized as antiinflammatory agents. This latter group is able to reduce the production of the pro-inflammatory 4 7 cytokines by monocytes/macrophages-and also to interfere with some of their activities. In the present study we have compared the capacity of these anti-inflammatory cytokines to inhibit the production of IL-8 by isolated polymorphonuclear cells stimulated by LPS and/or TNFc. Human polymorphonuclear cells, the main target cells for IL-8, can be activated to release IL-8 following stimulation by LPS and/or TNFc. These results suggest that, following their recruitment in inflammatory tissues and local activation by bacterial-derived products or cytokines, PMN can perpetuate the chemoattractant process and cellular activation. We observed that the production of IL-8 by LPS-activated human polymorphonuclear cells could be down-regulated by IL-IO, IL-4, IL-13 and to a lesser extent by TGF. This is in agreement with the results of Kasama et al. 33 and Cassatella et al., 34 who reported the capacity of IL-10 to limit IL-8 production, and with those of Wertheim et al. 35 who obtained similar results with IL-4. We observed that IL-4 was less efficient than IL-10 in agreement with the report of Wang et +t. In addition, when comparing the efficiency of the various antiinflammatory cytokines, we have clearly shown a profound variability of PMN sensitivity to IL-4 and IL-13 between individuals. In contrast, Ibl0 was uniformly active. These results probably do not reflect differences at the IL-4 receptor levels since there was no correlation between the levels of IL-8 and IL-lra released in response to LPS plus IL-4. 7 In preliminary experiments, it seemed that the individual responsiveness to IL-4 might vary with time since the percentage of inhibition was not reproducible with the cells prepared from the same donor at different time intervals. Although IL-13 shares many activities with IL-4, in most experiments IL-13 appeared less active than IL-4. The observed inhibitory activity was also observed at the transcriptional level. In a recent report, Girard et al. 8 showed that IL-13 itself could induce IL-8 production by human PMN. We never observed such induction at the protein level (n 12 different donors).
We are unable to explain such discrepancy, the only differences noted being the origin of the recombinant cytokines and the culture conditions. Further experiments should be performed to address this issue.
The upand down-regulation of TNF-induced IL-8 production by PMN has been rarely studied. 34 Interestingly, none of the tested antiinflammatory cytokines were able to inhibit the production of IL-8 by PMN when TNFc was used as the stimulus. This observation seems to be linked to an intrinsic property of PMN since IL-10, IL-4 and IL-13 were able to significantly inhibit the IL-8 production by whole blood cells activated by TNFcz (data not shown). However, IL-4 and IL-IO can modulate some PMN functions when activated by TNFcz. Indeed, we have previously shown that IL-lra production by TNFcz-activated PMN could be up-regulated by these cytokines. 37 As shown in Table 3, the down-regulation of IL-8 production by PMN by the so-called anti-inflammatory cytokines, as well as the up-regulation of IL-lra production is dependent on the nature of the triggering signal and these modulations are not correlated. Thus, the modulation of the cytokine production by IL-4, IL-IO or IL-13 is dependent on the nature of the stimuli. In addition, the nature of the activated cells play a crucial role as illustrated by the fact that IL-4 and IL-10 which are well known to inhibit the LPS-induced IL-8 production by monocytes-macrophages, 4-7 enhance it when acting on endothelial cells. 39 We have also investigated the simultaneous action of LPS andTNFc on IL-8 production, particularly since PMN in infectious sites may well be exposed to both agents. In these experimental conditions, only IL-10 retained its ability to inhibit the IL-8 production by activated PMN, whereas IL-4 was far less active and IL-13 had no more inhibitory activity. However the inhibitory effect of IL-IO was far less pronounced than when LPS was used alone as a stimulus (48-58% vs 83% inhibition).
A significant amount of cell-associated IL-8 was also observed in the PMN culture, which regularly exceeded that found in the supernatants. The presence of IL-8 in lysates of activated PMN might be explained by the presence of receptor-bound IL-8 and by the presence of IL-8 in the intracellular compartments, resulting from secretory mechanisms and/or following the rapid internalization of the cytokine bound to its receptors. 4 As shown by Kuhns and Gallin, the accumulation of IL-8 within PMN is localized to a subcellular fraction of heterogenous light membranous organelles. 41 In addition, the authors suggested that IL-8 accumulation was under translational control.
The fact that the anti-inflammatory cytokines had a minor effect on the cell-associated form of IL-8 suggests that most of its detection corresponds to the rapid internalization of surrounding IL-8 which is still present in the culture medium, independently of the presence of IL-4, IL-IO or IL-13.
In conclusion this is the first report on the ability of IL-13 to down-modulate the production of IL-8 by LPS-activated PMN in comparison with the other anti-inflammatory cytokines. In addition, we showed an individual responsiveness to the inhibitory activity of IL-4 and IL-13 on IL-8 production by LPS-activated PMN, whereas IL-10 was uniformly active. It appeared that IL-IO was the most active cytokine when LPS was employed as the stimulating agent, while its inhibitory property could not be observed when TNFcz was the activator and was minimized when TNFc was used together with LPS. This late observation may well be relevant to some in vivo situations where recruited PMN encountered in situ various stimulating signals.