Department of Pharmacology, Physiology and Therapeutics, University of North Dakota School of Medicine and Health Sciences, Grand Forks, North Dakota, USA
Address correspondence and reprint requests to Thad A. Rosenberger, University of North Dakota, School of Medicine and Health Sciences, Department of Pharmacology, Physiology, and Therapeutics, 501 North Columbia Road, Room 3742, Grand Forks, ND 58203, USA.
Acetate supplementation increases brain acetyl-CoA and histone acetylation and reduces lipopolysaccharide (LPS)-induced neuroglial activation and interleukin (IL)-1β expression in vivo. To determine how acetate imparts these properties, we tested the hypothesis that acetate metabolism reduces inflammatory signaling in microglia. To test this, we measured the effect acetate treatment had on cytokine expression, mitogen-activated protein kinase (MAPK) signaling, histone H3 at lysine 9 acetylation, and alterations of nuclear factor-kappa B (NF-κB) in primary and BV-2 cultured microglia. We found that treatment induced H3K9 hyperacetylation and reversed LPS-induced H3K9 hypoacetylation similar to that found in vivo. LPS also increased IL-1β, IL-6, and tumor necrosis factor-alpha (TNF-α) mRNA and protein, whereas treatment returned the protein to control levels and only partially attenuated IL-6 mRNA. In contrast, treatment increased mRNA levels of transforming growth factor-β1 (TGF-β1) and both IL-4 mRNA and protein. LPS increased p38 MAPK and JNK phosphorylation at 4 and 2–4 h, respectively, whereas treatment reduced p38 MAPK and JNK phosphorylation only at 2 h. In addition, treatment reversed the LPS-induced elevation of NF-κB p65 protein and phosphorylation at serine 468 and induced acetylation at lysine 310. These data suggest that acetate metabolism reduces inflammatory signaling and alters histone and non-histone protein acetylation.
cytoplasmic polyadenylation element-binding protein
extracellular signal-regulated kinase
histone H3 at lysine 9
c-Jun N-terminal kinase
mitogen-activated protein kinase
nuclear factor-kappa B
transforming growth factor-beta1
tumor necrosis factor-alpha
Neuroinflammation involves an innate immune response that is advantageous with regard to normal brain physiology. However, uncontrolled neuroinflammation is detrimental and associated with numerous neurological pathologies. The physiological functions of cytokines include regulating cell growth, differentiation, and body temperature regulation (Hopkins and Rothwell 1995; Rothwell and Hopkins 1995). Cytokines also have a pleiotropic function in neuroimmune communication between neuroglia, neurons, and endothelium involving both injury resolution and progression (Suzuki et al. 2009). Excessive pro-inflammatory cytokines, tumor necrosis factor-alpha (TNF-α), interleukin-(IL)1β, and IL-6 are implicated in the pathogenesis of numerous neuroinflammatory diseases (Denes et al. 2010; Merson et al. 2010; Qian et al. 2010; Helmy et al. 2011; Johnston et al. 2011). Transforming growth factor-beta1 (TGF-β1), IL-4, and IL-10 collectively share anti-inflammatory features that counteract the pro-inflammatory cytokines and provide control over the neuroinflammatory response. Anti-inflammatory cytokines are generally involved in tissue repair, enhancing neuronal survival, and down-regulating the expression of pro-inflammatory cytokines (Ledeboer et al. 2000; Vitkovic et al. 2001).
Acetate supplementation increases brain acetate levels (Mathew et al. 2005) as well as the metabolically active intermediate acetyl-CoA in normal animals (Reisenauer et al. 2011). In a rat model of neuroinflammation, acetate supplementation attenuates lipopolysaccharide (LPS)-induced neuroglial activation and the loss of cholinergic immunoreactivity (Reisenauer et al. 2011). Acetate supplementation is also neuroprotective in a rat model of head trauma (Arun et al. 2010a) and a tremor model of Canavan's disease (Arun et al. 2010b). To understand the mechanism underlying the neuroprotective and anti-inflammatory effects of acetate supplementation, we examined the effect that acetate metabolism has on histone hyperacetylation, which is associated with anti-inflammatory and neuroprotective responses (Langley et al. 2005; Adcock 2007). A single oral dose of glyceryl triacetate, used to induce acetate supplementation, results in site- and time-specific histone hyperacetylation in the brains of normal animals (Soliman and Rosenberger 2011). In addition, long-term acetate supplementation in a rat model of neuroinflammation induces site-specific brain histone hyperacetylation, reverses LPS-induced hypoacetylation of histone H3 at lysine 9 (H3K9), and suppresses IL-1β expression (Soliman et al. 2012).
Microglia, the primary immune cells in the brain, transform into phagocytic cells upon changes in the structural or functional integrity of the brain and produce a wide range of inflammatory cytokines (Streit et al. 1999; Hanisch and Kettenmann 2007; Ransohoff and Perry 2009; Lehnardt 2010). The BV-2 mouse microglia cell line, immortalized through oncogenes carrying retrovirus, exhibit morphological, functional, and phenotypical properties similar to primary microglia (Blasi et al. 1990; Bocchini et al. 1992). Therefore, BV-2 microglial cells are commonly used as an alternative to primary microglia to study various microglial responses and interactions (Petrova et al. 1999; Woo et al. 2003; Rojanathammanee et al. 2011).
Mitogen-activated protein kinases (MAPK) are key regulators of the biosynthesis of pro-inflammatory cytokines TNF-α, IL-6, and IL-1β and are hence potential therapeutic targets in inflammatory and autoimmune diseases (Pearson et al. 2001; Kumar et al. 2003). MAPK include p38, c-Jun N-terminal kinase (JNK), and extracellular signal-regulated kinase (ERK), whose activities are crucial for normal immune and inflammatory responses. The activation of these kinases is implicated in a number of biological processes including cell differentiation and survival, and the response to stress. Specific lysine acetylation activates MAPK phosphatase-1, which subsequently dephosphorylates and deactivates MAPK signaling (Cao et al. 2008), which provides a link between acetylation and phosphorylation as a regulator of inflammation. Another signaling complex that regulates inflammatory responses and is altered by acetylation is nuclear factor-kappa B (NF-κB), which is most commonly a heterodimer of p65 and p50. In the cytoplasm, NF-κB is associated with inhibitors of kappa B (IκB), which mask the nuclear export motif. Upon stimulation by pro-inflammatory cytokines, B- and T-cell receptor signaling, and viral and bacterial toxins, NF-κB is released from IκB and translocates to the nucleus where it binds DNA sequences and alters the transcriptional activity of genes involved in inflammatory responses and cell survival (Chen and Ghosh 1999). p65, but not p50, binds transcriptional coactivators p300 and CREB-binding protein (Perkins et al. 1997). The p65 subunit of NF-κB can be modified by acetylation at certain lysine residues with variable functional outcomes (Chen et al. 2001; Kiernan et al. 2003; Huang et al. 2010).
No reports are available that describe a decline in brain acetate levels in response to LPS or other neurological pathologies. Consequently, rather than replenishing endogenous acetate stores, we propose that acetate supplementation acts to increase intracellular levels of acetyl-CoA as an inducer of metabolic and molecular processes that ultimately result in the reduction in inflammatory phenotype. To test this hypothesis, we measured the ability of acetate treatment to alter inflammatory signaling in LPS-challenged microglia. We found that acetate treatment effectively reversed the LPS-induced H3K9 hypoacetylation and increases in the pro-inflammatory cytokine protein, but not mRNA levels. Furthermore, treatment up-regulated the mRNA levels of the anti-inflammatory cytokine TGF-β1, and both the protein and mRNA levels of IL-4. Because MAPK and NF-κB signaling are also associated with the neuroinflammatory responses, we expanded our study to quantify the effect acetate treatment had on these signaling pathways. In this regard, treatment attenuated p38 and JNK phosphorylation at 2 h, and not 4 h, and increased phosphorylated ERK1/2 at 4 h only in the presence of LPS. In addition, acetate treatment returned LPS-mediated increases in p65 protein levels and phosphorylation at serine 468 to control levels, and induced p65 hyperacetylation at lysine 310. These data suggest that in LPS-challenged microglia, acetate metabolism can modulate inflammatory signaling and shift cytokine balance toward a more anti-inflammatory state.
Materials and methods
LPS (Escherichia Coli 055:B5) was purchased from Sigma (St. Louis, MO, USA), antibodies against total histone H3, acetylated H3K9, phosphorylated p38 (Thr180/Tyr182), p38, phosphorylated JNK (Thr183/Tyr185, Thr221/Tyr223), phosphorylated ERK1/2 (Th202/Tyr204, Thr185/Tyr187), and ERK1/2 were from Millipore (Billerica, MA, USA), and anti-JNK and NF-κB p65 antibodies were purchased from Cell Signaling Technology Incorporated (Danvers, MA, USA). Rabbit polyclonal antibodies to IL-1β, IL-6, TNF-α, TGF-β1, IL-4, IL-10, and acetyl-CoA synthetase were from Abcam (Cambridge, MA, USA) and all western blot supplies and a goat anti-rabbit horse radish peroxidase (HRP)-linked antibody and iScript cDNA synthesis kits were purchased from Bio-Rad Laboratories (Hercules, CA, USA). Reverse and forward IL-1β, IL-6, TNF-α, IL-4, IL-10, TGF-β1, and β-actin primers from SA Biosciences (Frederick, MD, USA), FastStart Universal SYBR Green Master from Roche Applied Science (Indianapolis, IN, USA) from Bio-Rad, TRIzol® reagent from Life Technologies (Grand Island, NY, USA), Dulbecco's modified Eagle's medium–F-12 media from Invitrogen (Grand Island, NY, USA), and buffering reagents and other chemicals from EMD Biosciences (Gibbstown, NJ, USA).
Primary and BV-2 microglial cell cultures
Primary microglia were derived from C57BL/6 mouse brains as described previously (Rojanathammanee et al. 2011). The BV2 microglia were obtained from Dr. Gary E. Landreth (Cleveland, OH, USA) and maintained until used as described previously (Dhawan et al. 2012). Cells were plated in six-well dishes and allowed to replicate until 90% confluence, (1.1 × 106 cells/dish). Prior to stimulating the cells (3 h), the media was changed to serum-free media. Plates were divided into four different groups; a group treated with 12 mM NaCl as a control group, another group treated with 12 mM sodium acetate, a third group treated with both 6.25 ng/mL LPS and 12 mM NaCl, and a fourth group treated with both 6.25 ng/mL LPS and 12 mM sodium acetate (n = 6 per group for BV-2 cells and n = 5 per group for primary microglia). The concentration of acetate used in this study is based on studies to determine the maximal amount of acetate that did not lead to significant cell death over a 24-h exposure period, compared to cells grown in serum-free media. After a single oral gavage of GTA (5.8 g/kg), brain acetate levels rise to 8 μM/g tissue at 1 h, and then decline to 6 and 2 μM/g tissue at 2 and 4 h, respectively (Mathew et al. 2005). However, the metabolically active molecule in this process is not acetate, but rather acetyl-CoA which reaches a maximum of 5.7 μg/g brain at 30 min and remains constant out to 4 h in vivo (Reisenauer et al. 2011). The cellular concentration of acetyl-CoA is controlled metabolically by acetyl-CoA synthetases 1 and 2, and not by cellular levels of free acetate (Fujino et al. 2001; Ariyannur et al. 2010). Therefore, our rationale for using the highest tolerable acetate concentration was not to mimic maximal tissue concentrations of acetate but rather to maximize, over a 4-h period, cellular levels of acetyl-CoA in an effort to identify metabolic and the inflammatory pathways that are modulated downstream of the formation of acetyl-CoA. For dose–response studies, plates were divided in six different groups treated with LPS in the following concentrations: 25, 12.5, 6.25, 3.125, 1.56, or 0 ng/mL (n = 3). After 4 h, the media was collected and stored at −20°C, and the cells were lysed in either TRIzol® reagent for quantitative real-time polymerase chain reaction (qrt-PCR) analysis or ice-cold RIPA lysis buffer (150 mM sodium chloride, Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 50 mM Tris, pH 8.0) for western blot analysis and stored at −80°C until used.
Western blot analysis
Gel electrophoresis and protein transfer was performed as described previously (Soliman and Rosenberger 2011; Soliman et al. 2012). The antibody concentrations used were 1 : 500 for total H3 and acetyl-CoA synthetase, 1 : 1000 for each of acetylated H3K9, IL-1β, IL-6, TNF-α, TGF-β1, IL-4, IL-10, p38, phosphorylated p38, JNK, phosphorylated JNK, ERK1/2, phosphorylated ERK1/2, and all NF-κB antibodies, and 1 : 3000 for α-tubulin. All western blot data are expressed as the ratio of the optical density of the respective protein to the optical density of α-tubulin, except acetylated H3K9 (normalized to total histone H3), and phosphorylated MAPK p38, JNK, and ERK1/2 (normalized to total MAPK p38, JNK, and ERK1/2, respectively).
Quantitative real-time polymerase chain reaction
mRNA extraction, quantification, and cDNA synthesis, and amplification were performed as described previously (Soliman et al. 2012). The expression of IL-1β, IL-6, TNF-α, IL-4, IL-10, and TGF-β1 transcripts amplified was normalized to the expression of β-actin.
Lactate dehydrogenase assay
Cellular release of lactate dehydrogenase (LDH) used to measure cell viability was measured using a commercial non-radioactive assay kit (Clontech Inc., Mountain View, CA, USA), according to the manufacturer's guidelines. Absorbance measurements were taken at 490 nm.
One-way analysis of variance (anova) followed by Tukey's post hoc test using GraphPad InStat software (Version 3.06 for Windows, San Diego, CA, USA). Results are expressed as means ± SD with significance set at p ≤0.05.
Optimizing the duration of acetate treatment and LPS concentration
In rat brain, H3K9 acetylation is reduced by 50% in a rat model of neuroinflammation and is returned to control levels with acetate supplementation (Soliman et al. 2012). To determine the duration of acetate treatment required to achieve a similar H3K9 hyperacetylation pattern in vitro, we treated BV-2 microglia with 12 mM sodium acetate for 1, 2, and 4 h (Fig. 1a and b). Cell lysates were used for western blot analysis to measure acetylated H3K9, total histone H3 (Fig. 1a). We found that acetate treatment increased H3K9 acetylation by 2 h which remained elevated out to 4 h (Fig. 1b). To insure protein expression following treatment, we used 4 h as the treatment duration for all the experiments with the exception of change in MAPK phosphorylation where additional time points were used. To determine the optimal LPS concentration required to produce the same percentage of H3K9 hypoacetylation found in vivo, we treated BV-2 microglia for 4 h using a serial dilution of LPS ranging between 0 and 25 ng/mL (Fig. 1c and d). Cell lysates were used for western blot analysis to measure acetylated H3K9, total histone H3, and the pro-inflammatory cytokines pro-IL-1β, IL-6, and TNF-α, which were detected as protein bands corresponding to the molecular weights of 17, 17, 32, 25, and 23, respectively (Fig. 1c). LPS reduced H3K9 acetylation and increased the pro-inflammatory cytokine levels in a concentration-dependent manner (Fig. 1d–g). On the basis of these data, we used the LPS concentration 6.25 ng/mL because this concentration resulted in a 50% reduction in H3K9 acetylation, similar to that found in vivo (Soliman et al. 2012), and also increased protein levels of all the pro-inflammatory cytokines measured.
Acetate treatment reverses LPS-induced H3K9 hypoacetylation without inducing cytotoxicity in primary microglia
To determine the ability of acetate treatment to reverse LPS-induced H3K9 hypoacetylation in microglia similar to that found in the rat (Soliman et al. 2012), we treated primary mouse microglia with LPS 6.25 ng/mL for 4 h in the presence and absence of 12 mM sodium acetate, with 12 mM NaCl treatment as control. Using whole cell lysates for western blot analysis, we found that primary microglia express acetyl-CoA synthetase (ACS), which converts acetate to acetyl-CoA, as protein bands corresponding to the molecular weight of 79 kDa (Fig. 2a). The expression level of ACS was not different between groups (Fig. 2b). Furthermore, acetate treatment increased H3K9 acetylation by 1.7-fold, whereas LPS reduced H3K9 acetylation by 50% (Fig. 2c). Acetate treatment, similar to that found in vivo, effectively increased H3K9 acetylation to control levels in the presence of LPS (Fig. 2c). Cell viability assays showed no difference in cell survival between groups (Fig. 2d). These data indicate that acetate treatment in vitro reverses LPS-induced H3K9 hypoacetylation in microglia similar to that found in vivo (Soliman et al. 2012).
Acetate treatment reverses LPS-induced increases in the pro-inflammatory cytokine proteins, but not mRNA, in primary microglia
To determine the ability of acetate treatment to reverse pro-inflammatory cytokine production in vitro similar to that found in vivo (Soliman et al. 2012), cell lysates were analyzed using western blot to probe for IL-1β, IL-6, and TNF-α (Fig. 3a). We found that LPS increased pro-IL-1β, IL-6. and TNF-α by about 4-, 1.5-, and 2.5-fold, respectively, which were returned to control levels with acetate treatment (Fig. 3b, d and f). In parallel studies, we found that LPS increased the mRNA levels of all the pro-inflammatory cytokines measured, but were not altered by acetate treatment (Fig. 3c and g) with the exception of IL-6 mRNA, which was attenuated three-fold (Fig. 3e). These data demonstrate that this in vitro system reproduces the findings from the animal model, and that acetate treatment decreases pro-inflammatory cytokine levels possibly by disrupting mRNA translation or by increasing protein turnover.
Acetate treatment reverses LPS-induced H3K9 hypoacetylation in BV-2 microglia
We examined H3K9 acetylation in BV-2 microglia (Fig. 4a) under the same experimental conditions used with primary mouse microglia to confirm that both cell types respond similarly. First, we confirmed that BV-2 microglia express ACS; which was not different between groups (Fig. 4b). Furthermore, we found that acetate treatment increased H3K9 acetylation by 1.8-fold, and reversed the LPS-induced 50% reduction in H3K9 acetylation (Fig. 4c), similar to that found in primary microglia cultures. Furthermore, like the primary microglia cultures, treatment did not alter cell viability (Fig. 4d).
Acetate treatment reverses the LPS-induced increases in pro-inflammatory cytokine protein, but not mRNA, in BV-2 microglia
We proceeded to determine the effect of acetate treatment and LPS on pro-inflammatory cytokine proteins (Fig. 5a) and mRNA levels in BV-2 microglia under the same experimental conditions used with primary microglia to confirm that both cell types respond similarly in this regard. We found that LPS increased pro-IL-1β, IL-6, and TNF-α production by 25-, 1.5-, and 8-fold, respectively, which were returned to control levels with acetate treatment (Fig. 5b, d, and f). In parallel, we found that LPS increased the mRNA levels of the same pro-inflammatory cytokines, similar to that found with the primary microglia cultures, and were not altered by acetate treatment (Fig. 5c and g) with the exception of IL-6, which was attenuated two-fold (Fig. 5e). Therefore, the inflammatory response of BV-2 microglia toward LPS and acetate treatment is similar to primary microglia.
Acetate treatment increases the expression of anti-inflammatory cytokines in BV-2 microglia
Anti-inflammatory cytokines are an integral part of the inflammatory response to minimize the potential of the pro-inflammatory cytokines to produce neuronal damage. We determined the effect of acetate treatment on expression levels of the anti-inflammatory cytokine proteins TGF-β1, IL-4, and IL-10 (Fig. 6a). Acetate treatment did not alter the protein levels of TGF-β1 or IL-10 (Fig. 6b and f), however, IL-4 was increased by 1.3-fold (Fig. 6d). In parallel, we found that acetate treatment increased TGF-β1 mRNA by two-fold (Fig. 6c) and IL-4 mRNA by 11- and 4-fold, depending on the group (Fig. 6e). Conversely, LPS increased IL-10 protein and mRNA by 1.4- and 16-fold, respectively. Acetate treatment returned IL-10 protein to control levels and attenuated IL-10 mRNA by eight-fold (Fig. 6f and g). The possible reasons why increases in mRNA levels are not paralleled by increased protein levels may involve mRNA stability or reflect the short treatment duration. Regardless, these data suggest that acetate treatment modulates pro- and anti-inflammatory cytokine release in BV-2 microglia toward a more anti-inflammatory state.
Acetate treatment and LPS alter MAPK phosphorylation in a time-dependent manner in BV-2 microglia
Because MAPK signaling can be inhibited by the acetylation of MAPK phosphatase-1, which induces deacetylation and deactivation of MAPK (Cao et al. 2008), we measured the effects of acetate treatment on LPS-induced MAPK phosphorylation at 0.5, 1, 2, and 4 h. The rationale for including multiple time points is that other studies reported MAPK activation by LPS at much earlier time points than 4 h (Schumann et al. 1996; Kraatz et al. 1998). Whole cell lysates were used for western blot analysis, and phosphorylated p38, p38, phosphorylated JNK, JNK, phosphorylated ERK1/2, and ERK1/2 were detected as protein bands corresponding to the molecular weights of 38, 38, 46, 54 and 46, and 42 and 46 kDa, respectively (Fig. 7a). At 0.5 and 1 h, neither LPS nor acetate treatment had an effect on the levels of phosphorylated MAPK (Fig. 7b–d). At 2 h, acetate treatment reduced the level of phosphorylated p38 as compared to LPS, and LPS increased JNK phosphorylation by five-fold, which was attenuated 2.5-fold with acetate treatment (Fig. 7b and c). At 4 h, LPS increased phosphorylated p38 and phosphorylated JNK by two-fold, and was not altered upon acetate treatment; however, treatment did increase the level of phosphorylated ERK1/2 by two-fold only in the presence of LPS (Fig. 7b–d).
Acetate treatment alters LPS-induced increases in NF-κB p65 protein levels and phosphorylation at serine 468 in BV-2 microglia
Because NF-κB signaling is altered by acetylation of p65 (Chen et al. 2001; Kiernan et al. 2003; Huang et al. 2010) and has a prominent role in the regulation of inflammatory and immune responses, we tested the effect of acetate treatment on LPS-induced changes in p65 protein levels, phosphorylation, and acetylation after 4 h of treatment. Whole cell lysates were used for western blot analysis, and total p65, phosphorylated p65 at serine 536, phosphorylated p65 at serine 468, and acetylated p65 at lysine 310 were detected as protein bands corresponding to the molecular weight of 65 kDa (Fig. 8a). LPS increased the total protein level of p65 by 1.5-fold, which returned to control levels with acetate treatment (Fig. 8b). While neither acetate treatment nor LPS altered the level of phosphorylated p65 at serine 536, LPS did increase the levels of phosphorylated p65 at serine 468 by two-fold, which was reduced to control levels with acetate treatment (Fig. 8c and d). In addition, acetate treatment increased p65 acetylation at lysine 310 by 3.5-fold (Fig. 8e). These data suggest that acetate metabolism alters the LPS-induced p65 response in microglia, and that the anti-inflammatory effect of acetate treatment can potentially be attributed to acetylation of non-histone targets.
In this study, we demonstrated that acetate treatment reversed the LPS-induced reduction in H3K9 acetylation and decreases pro-inflammatory cytokines in microglia in vitro. Moreover, acetate treatment increased the transcription of the anti-inflammatory cytokines, TGF-β1 and IL-4, suggesting that acetate-induced histone modulation may influence more strongly the expression of anti-inflammatory cytokines in this model, considering histone hyperacetylation is conventionally linked to increased gene expression. We also demonstrated the time-dependent effects of LPS and acetate treatment on MAPK activation. In addition, acetate treatment reduced LPS-induced increases in total NF-κB p65 protein level, serine 468 phosphorylation, and increased its acetylation at lysine 310. These data suggest that acetate metabolism can modulate cytokine balance in microglia, which correlates to increases in both histone and non-histone protein acetylation.
The differential effect of acetate treatment on mRNA and protein levels suggests that the reduction in pro-inflammatory cytokines may be because of a disruption in mRNA translation rather than gene transcription or pro-inflammatory cytokine turnover. Translation involves the interaction of mRNA with various subsets of proteins which, we speculate, may be regulated by acetylation. For example, nuclear mRNA binds to nuclear proteins that transport mRNA to the cytosol. Some of these proteins repress translation by interfering with the binding of mRNA to ribosomal subunits (Wells 2006). Similarly, the integrity of mRNA is modulated by mRNA-stabilizing proteins (Kohn et al. 1996). It is possible that acetylation may alter the expression and/or activity of mRNA-binding and/or -stabilizing proteins. Of particular interest is cytosolic polyadenylation element-binding protein (CPEB) expressed both in neuroglia and neurons, which prevents the formation of the translation initiation complex and represses translation (Mendez and Richter 2001; Theis et al. 2003). CPEB is regulated by phosphorylation (Atkins et al. 2004), however, the effect that acetylation has on its activity remains unknown. Furthermore, the eukaryotic initiation factor 5A (eIF5A), which regulates initiation and elongation, contains a polyamine–lysine conjugated amino acid ‘hypusine’ that is essential to its activity (Zanelli et al. 2006; Gregio et al. 2009; Saini et al. 2009) and is inactivated following acetylation by spermidine/spermine acetyltransferase 1 (Lee et al. 2011). In addition, acetylation by a histone acetyltransferase PCAF leads to eIF5A accumulation in the nucleus that prevents translocation to the cytosol and in turn disrupts translation (Ishfaq et al. 2012). All of which suggests that acetylation may be involved in the regulation of mRNA translation.
Acetate treatment may also reduce pro-inflammatory cytokine levels, but not mRNA by increasing protein turnover. A number of histone acetyltransferases possess intrinsic ubiquitin-conjugating activity and are associated with ubiquitin transferases in multiprotein complexes that stimulate degradation (Sadoul et al. 2008). Furthermore, acetylation of the translation elongation factor (E2F1) (Galbiati et al. 2005) and the hypoxia-inducible factor 1α (HIF-1) at lysine 532 enhances their ubiquitination and degradation (Jeong et al. 2002). Thus, it is plausible that non-histone protein acetylation may alter mRNA translation and the turnover of pro-inflammatory cytokines in activated microglia.
An increase in pro-inflammatory cytokine production is generally considered deleterious based on their involvement in a wide number of neurological and non-neurological disorders. As an example, coculture of primary rat cortical neurons with LPS-activated microglia results in neuronal death, which can be largely blocked using the naturally occurring IL-1 receptor antagonist IL-1ra (Li et al. 2003). Not surprisingly, suppression of pro-inflammatory cytokines is associated with improved behavioral and cognitive endpoints in animal models of neurodegenerative diseases (Hu et al. 2007; Lloyd et al. 2008). On the other hand, IL-4, IL-10, and TGF-β1 share features of anti-inflammatory and neuroprotective actions that can be attributed to down-regulating glial production of pro-inflammatory cytokines and/or attenuating their secondary release. IL-4 reduces the production of inflammatory mediators, including inducible nitric oxide (NO) synthase, TNF-α, IL-1β, cyclooxygenase 2, and macrophage chemoattractant protein-1 by activated microglia in vivo and in vitro (Furlan et al. 2000; Ledeboer et al. 2000). In addition, TGF-β has a neuroprotective effect by regulating Bad (pro-apoptotic) and Bcl-2 and Bcl-x1 (anti-apoptotic) proteins (Dhandapani and Brann 2003). Furthermore, anti-inflammatory cytokines reduce the expression levels of the pro-inflammatory cytokines in LPS-stimulated microglial–astroglial cocultures (Ledeboer et al. 2000). Endogenous and exogenous TGF-β1 and β2 suppress the production of NO but not IL-1β, IL-6, or TNF-α and exogenous IL-4 down-regulates NO, IL-6, and TNF-α, but not IL-1β (Ledeboer et al. 2000). Our findings showing that LPS stimulation up-regulated IL-10 is not counterintuitive, because stimulation of an inflammatory response can lead to up-regulation of both conventional pro-inflammatory and anti-inflammatory mediators as a biological self-checking mechanism. In this regard, IL-10 inhibits the LPS-induced increase in IL-1β and TNF-α (Sawada et al. 1999) and IL-10 release by LPS-stimulated microglia increases simultaneously with TNF-α (Seo et al. 2004). The multiplicity of receptors, signaling cascades, cellular and subcellular targets, and various experimental designs all demonstrate the complexity of how anti-inflammatory cytokines can regulate the transcription and/or translation of the pro-inflammatory cytokines.
Lysine acetylation is a common post-translational modification that occurs on both histones as well as non-histone proteins. Histone acetylation is conventionally linked to an increase in gene expression. Non-histone targets of acetylation include cytoskeletal proteins and transcription and nuclear import factors. Acetylation of these targets have many functional consequences including altering subcellular localization, DNA binding, transcriptional activity, protein–protein interaction, and protein stability (Glozak et al. 2005; Sadoul et al. 2008). MAPK signaling is inducible by pro-inflammatory cytokines and also regulates their transcription and translation. For example, MAPK signaling regulates the production of IL-8 in response to IL-1 and osmotic shock (Shapiro and Dinarello 1995), and regulates the production of IL-6 in response to TNF-α (Beyaert et al. 1996). Animals with genetic deletion of one of the MAPK accessory proteins show diminished IL-6 and TNF-α production in response to LPS stimulation (Kotlyarov et al. 1999). Because a MAPK phosphatase is activated by acetylation, which inhibits MAPK signaling, we studied whether acetate treatment alters MAPK phosphorylation (activation) at different time points. We found that the effect of LPS on MAPK phosphorylation was time dependent, as was the ability of acetate treatment to reduce LPS-induced p38 and JNK phosphorylation. LPS increased phosphorylated p38 at 4 h and phosphorylated JNK at 2 and 4 h, whereas acetate treatment reduced phosphorylated p38 and JNK only at 2 h, but not 4 h. We did not observe an increase in MAPK activation at 0.5 or 1 h unlike other studies (Schumann et al. 1996; Kraatz et al. 1998). However, this may be because of our using a lower concentration of LPS or may demonstrate a cell-type-specific response. While the therapeutic effect of acetate supplementation is demonstrated in the in vivo studies (Reisenauer et al. 2011), these results further strengthen our understanding of the possible therapeutic mechanism(s) involved in modulating cytokine expression by increasing acetate metabolism. Therefore, because the effect of acetate treatment on the LPS-induced MAPK p38 phosphorylation is transient, the effect of acetate treatment on cytokine release may be because of the convergence of multiple pro- and anti-inflammatory signaling mechanisms.
NF-κB is acetylated on p65, which modulates nuclear translocation, DNA binding, and transcriptional activity (Chen et al. 2001, 2002; Huang et al. 2010). In this study, we found that acetate treatment induced p65 hyperacetylation at lysine 310. This is of interest because p65 interacts with histone deacetylases (HDAC) 1, 2, and 3, but only HDAC3 deacetylates p65 (Chen et al. 2001; Kiernan et al. 2003), which is down-regulated with acetate supplementation (Soliman et al. 2012). Therefore, the effect that acetate metabolism has on HDAC3 expression may help to explain the hyperacetylation of p65 at lysine 310 observed in this study. The acetylation of p65 may be associated with anti-inflammatory outcomes as it represses transcriptional activity, reduces binding to κB-DNA, and facilitates its interaction with IκB that increases p65 export to the cytoplasm. Because acetylated p65 accumulates in the cytoplasm suggests that post-activation turn-off of NF-κB-dependent transcription is regulated, at least in part, by acetylation (Kiernan et al. 2003). However, β-amyloid toxicity increases hyperacetylated p65 at lysine 310 in microglia, which is reversed by SIRT1 over-expression and stimulation (Chen et al. 2005). This suggests that changes in the activity and expression of the of the sirtuins and class I HDAC can differentially modulate NF-κB-mediated inflammatory phenotype, possibly as a result of differing inflammatory stimulation or differing intercellular regulation points. Alternately, acetate treatment-induced p65 hyperacetylation in the presence of LPS may be linked to pro-inflammatory signaling that is generally outweighed by the other anti-inflammatory mechanisms. Regardless, the functional consequences of post-translational modification of p65 are diverse and specific to the modification and the residue involved (Huang et al. 2010). Future studies are necessary to determine the impact that acetylation of p65 has on NF-κB functionality in this model.
As histone acetylation is conventionally associated with enhanced gene expression (Strahl and Allis 2000), we speculate that the increase in H3K9 acetylation may be instrumental in up-regulating the transcription of anti-inflammatory cytokines, as found in this study. We chose to focus on H3K9 because of the effect that neuroinflammation and acetate supplementation have on its acetylation state as opposed to H4K8 and H4K16, which are hyperacetylated during acetate supplementation but not altered by neuroinflammation (Soliman et al. 2012). This is further supported by other reports implicating H3K9 hypoacetylation in neuroinflammation, and microglial activation (Zhang et al. 2008; Govindarajan et al. 2011; Silva et al. 2012). Our data also demonstrate a correlation between acetate treatment-induced inhibition of pro-inflammatory cytokine release and hyperacetylation of H3K9 and p65 at lysine 310. H3K9 can also be modified by methylation where methylated H3K9 is associated with gene repression, contrary to acetylated H3K9 that is associated with active gene expression (Rice and Allis 2001). In this regard, the enrichment of methylated H3K9 at the promoter region of opioid receptors is linked to decreased opioid receptor transcription in mice fed a high-fat diet (Vucetic et al. 2011). Similarly, genome-wide mapping demonstrates that an increase H3K9 acetylation corresponds with areas of transcription activity (Shin et al. 2012). H3 methylation is more predominant in areas of enriched acetylated H4, unlike methylated H4, which is more evident in less acetylated chromatin regions (Annunziato et al. 1995). Thus, it is possible that H3K9 hyperacetylation may alter the expression and/or activity of effector proteins involved in translation, which may help to explain the decrease in pro-inflammatory cytokines in the absence of a reduction in their mRNA levels.
In conclusion, we describe microglial-specific responses to acetate treatment, where modulation of cytokine balance is attributable to a reduction in pro-inflammatory cytokine levels and induction of anti-inflammatory cytokine transcription, all of which correspond to a reversal of LPS-induced changes in the acetylation of histone and non-histone proteins with acetate treatment. To better understand the contribution that histone versus non-histone acetylation has in the control of cytokine balance, it will be necessary to determine the differential impact that an increase in histone acetylation has on pro- and anti-inflammatory cytokine transcription.
This publication was made possible by Grant Number 5P20RR017699 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH). We thank Drs. Othman Ghribi and Joyce Ohm for their technical support and generous use of their equipment. The authors declare no conflict of interests.