β-funaltrexamine differentially modulates chemokine and cytokine expression in normal human astrocytes and C20 human microglial cells

Aim: Emerging evidence implicates astrocyte/microglia dysregulation in a range of brain disorders, thereby making glial cells potential therapeutic targets. The novel anti-inflammatory actions of beta-funaltrexamine (β-FNA) are of particular interest. β-FNA is a derivative of naltrexone, and recognized as a selective, irreversible antagonist at the mu -opioid receptor (MOR). However, we discovered that β-FNA has novel anti-inflammatory actions that seem to be mediated through a MOR-independent mechanism. Thus far, we have focused on the acute effects of β-FNA on inflammatory signaling. Methods: The effect of β-FNA treatment on interleukin-1β (IL-1β)-induced inflammatory signaling in normal human astrocytes (NHA) and C20 human microglial cells. Cytokine/chemokine expression was measured using ELISA, and nuclear factor-kappaB (NF-κB) p65 activation was evaluated by immunoblot. Results: IL-1β-induced interferon-γ inducible protein-10 (CXCL10) production in NHA was more sensitive to chronic (3 day) β-FNA as indicated by an approximately 3-fold lower EC50 compared to that observed in acutely treated cells. Chronic β-FNA did not affect IL-1β-induced monocyte chemoattractant protein-1 (CCL2) or IL-6 production in NHA. β-FNA inhibited phosphorylation of NF-κB p65, suggesting that the inhibitory effects may be due in part to reduced NF-κB activation. We showed for the first time that C20 human microglial cells were insensitive to the anti-inflammatory actions of acute β-FNA. Conclusion: β-FNA differentially affects inflammatory cytokine/chemokine expression in human astrocytes and microglia. These findings warrant further investigation into the novel anti-inflammatory actions of β-FNA, with a particular focus on astrocytes. These insights should contribute to the development of strategies to treat brain disorders that involve neuroinflammation.

We, and others, are therefore interested in identifying novel, anti-inflammatory agents that are therapeutically effective in the treatment of neurological disorders. We are interested in the previously identified, novel anti-inflammatory actions of beta-funaltrexamine (β-FNA). As a fumaramate methyl ester derivative of naltrexone, β-FNA is most notably recognized as a selective, irreversible antagonist at the mu-opioid receptor (MOR) [22,23] . In both behavioral and in vitro assays, β-FNA acts initially as a reversible kappa-opioid receptor agonist, and then later results in MOR antagonism [24,25] . As an alkylating agent, β-FNA irreversibly antagonizes MOR by covalently binding at Lys233 on the receptor [23] . However, we discovered that β-FNA also has novel anti-inflammatory actions which seem to be mediated through MOR-independent actions [26][27][28] . For instance, neither naltrexone (a nonselective opioid receptor antagonist) nor D-Phe-Cys-Tyr-D-Trp-D-Arg-Pen-Thr-NH2 (CTAP) inhibits pro-inflammatory-induced CXCL10 expression in human astroglial cells [27,28] . Additionally, we predicted that if the anti-inflammatory actions of β-FNA were due to alkylation, this covalent modification should then remain after washout. Indeed, pretreatment of astroglial cells with β-FNA for 60 min, followed by drug washout prior to stimulating with IL-1β (or tumor necrosis factor α), resulted in inhibition similar to 24 h co-exposure (cytokine stimulus + β-FNA). These findings suggested that β-FNA-induced modifications (i.e., alkylation) are persistent and lead to the disruption of signal transduction. Notably, our in vitro findings also showed that β-FNA reduces inflammatory signaling in astroglia, regardless of whether the stimulus is tumor necrosis factor α, IL-1β or bacterial lipopolysaccharide (LPS). We also determined that β-FNA inhibits LPS-induced proinflammatory cytokine expression in mouse brain (but not in plasma) [29] . Furthermore, treatment with β-FNA reduced LPS-induced sickness behavior in mice suggesting important translational implications [29] .
The primary goal of the present study was to determine the effect of chronic β-FNA treatment on inflammatory signaling in NHA. Additionally, we assessed for the first time the effects of β-FNA on inflammatory signaling in human microglial cells.

Treatment of cells
To determine the dose-dependent effect of chronic (72 h) β-FNA (NIDA reagent supply program; Bethesda, MD, USA) on CXCL10 expression in NHA, cells were initially cultured in growth medium containing 0.04-10 µmol/L β-FNA for 24 h. The medium was then replaced with serum-free medium (SFM) containing β-FNA for an additional 48 h; IL-1β (3 ng/mL; Peprotech, Rocky Hill, NJ, USA) was added to cultures for the final 24 h of the 72 h exposure period. To determine the differential effects of β-FNA on chemokine/ cytokine expression, NHA were treated as described above; however, only a single concentration of β-FNA was used (3 µmol/L; EC 50 for inhibition of CXCL10 expression). To assess the effects of β-FNA on IL-1βinduced activation of (NF)-κB p65, cells were chronically exposed to β-FNA (3 µmol/L) as described above, then stimulated for 30 min with IL-1β (3 ng/mL). IL-1β-induced NF-κB p65 activation was assessed at 30 min after stimulation (when peak activation is observed).
This was our initial investigation into the effects of β-FNA on chemokine/cytokine expression in C20 microglial cells. Thus, we used our acute exposure model. Consistent with the acute model previously used with astrocytes, we used a higher concentration range of β-FNA (3-30 µmol/L). Briefly, C20 microglial cells were serum deprived for 24 h and then treated with β-FNA (3-30 µmol/L) alone or in combination with IL-1β (20 ng/mL) for 24 h in SFM.

Chemokine/cytokine expression
Standard dual-antibody solid-phase immunoassays (ELISA Development Kit, Peprotech) were used for quantitation of cytokines/chemokines in culture medium as previously described [27] . Values were normalized to total protein content, which was determined using the bicinchoninic acid protein assay as previously described [32] .

Statistical analysis
Statistical analyses and figure presentations were performed using PrismTM version 7.04 (GraphPad Inc., San Diego, CA). Dependent measures were analyzed by either one-way or two-way analysis of variance (ANOVA). In those instances where two-way ANOVA was used, stimulus and drug dose were the grouping variables. Data that were > 2 SD from the mean were considered outliers and removed from the analyses. When ANOVA revealed a statistically significant interaction, data were further assessed using a Fisher's LSD test. The data are all presented as mean ± SEM.

Effects of chronic β-FNA on NHA viability
Cell viability was not significantly (P = 0.34) affected by chronic exposure to β-FNA alone or in combination with IL-1β, as revealed by ANOVA [ Figure 3].

Effects of chronic β-FNA on NF-κB activation in NHA
One-way ANOVA and pairwise comparison by a Fisher's LSD test revealed that chronic exposure to β-FNA significantly (P < 0.05) inhibited IL-1β-induced phosphorylation of nuclear NF-κB p65 in NHA [ Figure 4]. The expression of total nuclear NF-κB p65 in NHA was not significantly (P = 0.2) affected by IL-1β or β-FNA alone, or the combination of chronic β-FNA plus stimulation with IL-1β [ Figure 4].

DISCUSSION
Astrocytes and microglia are instrumental in neuroinflammation and both cytokines and chemokines are among the inflammatory molecules released by these cells during neuroinflammation [34][35][36] . Hence, there is substantial interest in targeting neuroinflammation as a therapeutic strategy for selected brain disorders. The therapeutic potential of β-FNA is of particular interest to our group, and the results of this study are in Figure 3. Effects of chronic β-FNA on viability of NHA. Cells were cultured in growth medium containing 3 µmol/L β-FNA for 24 h; the medium was then replaced with serum-free medium containing β-FNA for an additional 48 h. IL-1β (3 ng/mL) was added to cultures for the final 24 h. Cell viability was assessed using the MTT assay. Data are presented as mean ± SEM (n = 8). ANOVA did not reveal any significant differences. MTT: 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; SEM: standard error of the mean; β-FNA: beta-funaltrexamine; NHA: normal human astrocytes; IL-1β: interleukin-1β P Figure 4. Chronic β-FNA inhibits NF-κB activation in NHA. Cells were cultured in growth medium containing 3 µmo/L β-FNA for 24 h; the medium was then replaced with serum-free medium containing β-FNA for an additional 48 h. IL-1β (3 ng/mL) was added to cultures for the final 30 min. Western blot was used to measure levels of phospho-NF-κB p65, NF-κB p65, and β-tubulin in nuclear extracts (a representative blot is presented at the top of the figure). Data are presented as mean ± SEM (n = 5). Differing letters above the bars indicate the means are significantly (P < 0.05) different as determined by ANOVA and subsequent Fisher's LSD. LDS: least significant difference; ANOVA: analysis of variance; SEM: standard error of the mean; IL-1β: interleukin-1β; β-FNA: beta-funaltrexamine; NF-κB: nuclear factor-kappaB; NHA: normal human astrocytes P line with our previous findings and advance our understanding of the anti-inflammatory effects of β-FNA. In our earlier studies, we found that β-FNA inhibition of IL-1β-stimulated CXCL10 expression was at least in part transcriptional, given that both protein and mRNA levels were significantly reduced by β-FNA [37] . We now show that CXCL10 inhibition in human astrocytes is even more sensitive to chronic (3 day) β-FNA as indicated by an approximately 3-fold lower EC 50 compared to that observed in acutely treated cells. Importantly, and consistent with our previous findings [27] , the inhibitory effects of β-FNA are not due to cytotoxicity, as we have shown that even extended exposure to β-FNA does not reduce the viability of human astrocytes. These preclinical findings that astrocyte viability is not compromised are quite important as future therapeutic potential will likely involve extended exposure to this drug.
Interestingly, β-FNA did not inhibit IL-1β-induced expression of either CCL2 or IL-6 in NHA, suggesting a level of selectivity for CXCL10. However, it may be that higher concentrations would inhibit the expression of CCL2 and IL-6. Therefore, further investigation, including a dose-response, is necessary.
Mechanistically, there is still more to learn about the anti-inflammatory effects of β-FNA. Importantly, as stated above and discussed previously in detail, we have determined that the anti-inflammatory effects in astrocytes do not seem to be due to actions at MOR (or other opioid receptors) [26][27][28] . Overall, our previous Figure 5. β-FNA differentially affects chemokine/cytokine expression in C20 human microglial cells. Cells were serum deprived for 24 h and then exposed to β-FNA (3-30 µmol/L) alone or in combination with IL-1β (20 ng/mL) for 24 h. Chemokine/cytokine levels in the medium were measured by ELISA; and viability was determined using the MTT assay. Data are presented as mean ± SEM (n = [8][9][10][11][12] and were analyzed by two-way ANOVA and subsequent Fisher's LSD. *P < 0.05 vs . unstimulated, 0 µmol/L β-FNA; **P < 0.005 vs. IL-1β, 0 µmol/L β-FNA. ELISA: enzyme-linked immunoabsorbant assay; LSD: least significant difference; IL-1β: interleukin-1β; β-FNA: betafunaltrexamine; MTT: 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; SEM: standard error of the mean findings on the acute effects of β-FNA suggest that that the anti-inflammatory actions are likely mediated, at least in part, by disrupting the NF-κB signaling pathway [26][27][28]37] . In this study, we found that NF-κB signaling is also inhibited by chronic exposure to a lower concentration of β-FNA. Interestingly, the overall level of NF-κB p65 in the nucleus was not impacted by β-FNA; rather, phosphorylation of NF-κB p65 was reduced in the presence of β-FNA. Together, these findings suggest that β-FNA acts at a common factor in the signaling pathways activated by these diverse stimuli, in turn implicating the NF-κB signaling pathway. We hypothesize that β-FNA exerts these anti-inflammatory effects via alkylation of one or more lysines of the signaling proteins in the NF-κB pathway, and we are currently testing this hypothesis using in vitro approaches. Further studies are warranted to determine the mechanism by which β-FNA inhibits the phosphorylation of NF-κB p65.
We have also determined that β-FNA exerts anti-inflammatory actions in vivo [29,37] . For instance, LPSinduced neuroinflammation and sickness behavior in mice were attenuated by peripherally administered β-FNA [29] . More specifically, β-FNA inhibited LPS-induced expression of both CXCL10 and CCL2 in the brain, but had no effect on IL-6 levels. Furthermore, β-FNA did not impact plasma levels of CXCL10, CCL2, or IL-6. These in vivo findings are largely in line with our in vitro findings in human astrocytes, except that CCL2 expression in NHA was inhibited by β-FNA. Certainly, the differential sensitivity of CCL2 to β-FNA could be related to species differences or model systems (in vitro vs. in vivo). However, it may also reflect cell type-specific differences in sensitivity to β-FNA. For example, microglial cells (at least IL-1β-stimulated human microglial cells) are not sensitive to the anti-inflammatory actions of β-FNA. Because relatively high concentrations of β-FNA had no effect on chemokine/cytokine expression in C20 microglial cells and because of the observed cytotoxicity, we did not pursue the chronic effects of β-FNA on these cells at this point. However, in future experiments, we expect to assess chronic exposure to lower concentrations of β-FNA. Together, it is conceivable that the protective effects of β-FNA in vivo are largely due to modulatory effects on astrocytes; however, further investigation is needed to clearly establish the cell types affected and mechanisms involved.
In summary, we advanced our understanding of the anti-inflammatory effects of β-FNA by demonstrating that chronic exposure inhibits NF-κB p65 activation and CXCL10 expression in astrocytes more effectively than does acute treatment. We also found that expression of neither CCL2 nor IL-6 in astrocytes is affected by chronic β-FNA. Lastly, we provided evidence of cell type-specific effects of β-FNA, as indicated by the relative resistance of C20 human microglial cells to the anti-inflammatory effects of β-FNA. Further study is warranted and expected to advance the therapeutic potential of β-FNA, or related compounds, in the treatment of brain disorders that involve neuroinflammation.