Histone deacetylase 5 regulates the inflammatory response of macrophages

Modifying the chromatin structure and interacting with non-histone proteins, histone deacetylases (HDAC) are involved in vital cellular processes at different levels. We here specifically investigated the direct effects of HDAC5 in macrophage activation in response to bacterial or cytokine stimuli. Using murine and human macrophage cell lines, we studied the expression profile and the immunological function of HDAC5 at transcription and protein level in over-expression as well as RNA interference experiments. Toll-like receptor-mediated stimulation of murine RAW264.7 cells significantly reduced HDAC5 mRNA within 7 hrs but presented baseline levels after 24 hrs, a mechanism that was also found for Interferon-γ treatment. If treated with lipopolysaccharide, RAW264.7 cells transfected for over-expression only of full-length but not of mutant HDAC5, significantly elevated secretion of tumour necrosis factor α and of the monocyte chemotactic protein-1. These effects were accompanied by increased nuclear factor-κB activity. Accordingly, knock down of HDAC5-mRNA expression using specific siRNA significantly reduced the production of these cytokines in RAW264.7 or human U937 cells. Taken together, our results suggest a strong regulatory function of HDAC5 in the pro-inflammatory response of macrophages.


Introduction
Gene transcription is regulated by multiple mechanisms including modifications of the DNA itself or of changes in DNA-associated histones. Histone acetyltransferases mediate acetylation of the e-amino groups of lysine and histone deacetylases (HDAC) remove these acetyl groups. The eponymous function of the latter resulting in chromatin compaction inhibits transcription factor binding [1]. HDAC also modify non-histone proteins often involved in crucial functions like DNA repair and replication, metabolism or cellular signalling like p53, hypoxia-inducible factor 1a, STAT3 or p65 [2].
HDAC are evolutionarily highly conserved proteins. HDAC5 belongs to class II of the classical family sharing a zinc-dependent catalytic domain [3]. Regulated by phosphorylation of serine residues at the N-terminus of the enzyme, class II HDAC are able to shuttle between nucleus and cytoplasm [4,5]. This capacity to translocate enables the interaction with cytoplasmic non-histone proteins. The enzymatic activity is effectively blocked by HDAC inhibitors, representing excellent tools to evaluate HDAC function in vitro or in vivo. Originally approved as anti-cancer agents [6,7], low dose HDAC inhibitors were proven also anti-inflammatory, leading to several clinical trials e.g. for systemic-onset juvenile idiopathic arthritis [8]. This concept is strongly supported by experimental data indicating an anti-inflammatory effect for HDAC inhibitors in murine models of chronic inflammation such as rheumatoid arthritis [9], hepatitis [10] and various models of colitis [11] or inflammation-associated cancer [12]. At cellular level, CD4 + T cells, dendritic cells and/or macrophages are shown to be major players [13].
Macrophages connect innate and adaptive immune responses by phagocytosis of extracellular pathogens and antigen presentation via MHC II. While anti-inflammatory macrophages prevent overwhelming immune responses by secretion of IL-10, pro-inflammatory macrophages enhance the immune response by secretion of pro-inflammatory cytokines such as tumour necrosis factor (TNF)a or monocyte chemotactic protein (MCP)-1 [14,15]. Thus, macrophages are critical in immune homeostasis and macrophage dysfunction strongly associates to the pathogenesis of chronic inflammatory conditions including inflammatory bowel diseases (IBD) [16]. Suggesting that HDAC in macrophages in general rather convey a pro-inflammatory function, secretion of pro-inflammatory cytokines such as TNFa, IL-1a, IL-1b and IFNc by human mononuclear cells and murine bone marrow-derived macrophages was suppressed in the presence of HDAC inhibitors [17,18].
As for the contribution of single HDAC in complex processes, increased expression of HDAC5 in human alveolar epithelial cells after infection with Legionella pneumophila indicates a central role fort HDAC5 in host defense [19]. Our group revealed a pro-inflammatory function of HDAC5 by a specific knock-down in CD4 + T cells (unpublished data). For HDAC3 in macrophages in general and in IBD regarding the intestinal epithelial cells in particular, an important role in the pro-inflammatory response could be demonstrated [20,21]. These studies prove, that single HDAC are able to act upon activation in adult (immune) cells, especially in macrophages, and are not restricted to developmental or cancer biology.
Addressing the potential role of HDAC5 in chronic inflammation, we here specifically asked whether HDAC5 expression and function directly affect the response of macrophages to pro-or anti-inflammatory stimuli. Making use of murine and human macrophage cell lines, we studied the effect of HDAC5 over-expression and knock-down on nuclear factor kappa B (NF-jB)-dependent regulation of cytokine and chemokine expression.

Cell lines
The murine macrophage cell line RAW264.7 (ATCC TIB71; American Type Culture Collection, Manassas, VA) and the human histiocytic lymphoma cell line U937 (ACC5; DSMZ -German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany) were cultured according to the provider's instructions and routinely checked for mycoplasma contamination by mycoplasm-specific PCR (e-Myco TM plus; Intron Biotechnology, Sangdaewon-Dong, South Korea).

Quantification of mRNA expression
RNA (500 ng) prepared from defined cell numbers (Qiagen) was reversely transcribed (Life Technologies, Carlsbad, CA, USA). Complementary DNA was subjected to quantitative PCR (qPCR) using SYBR-Green (Life Technologies) on a StepOne Plus Realtime PCR System (Applied Biosystems, Darmstadt, Germany) using primer pairs specific for murine or human HDAC5 and GAPDH mRNA (Table 1). HDAC5 expression was determined in relation to GAPDH as housekeeping gene.

Quantification of secreted cytokines and chemokines
Supernatants of RAW264.7 and U937 cells were analysed for TNFa, IL-6 or IL-10 by ELISA; MCP-1 of RAW264.7 by cytometric bead array (all BD Bioscience, Heidelberg, Germany) according to the manufacturer's instructions.
Western-blot analysis RAW264.7 or U937 cells were lysed in 20 ll 50 mM Tris-HCl (pH 8.0) containing 20% SDS and 5% b-mercaptoethanol. Murine and human HDAC5 were detected by a rabbit-derived polyclonal antibody (#2082; Cell Signaling Technology, Danvers, MA, USA) and a rabbit-specific secondary antibody (DAKO, Hamburg, Germany). The HA-tag was assessed using a murine monoclonal antibody (clone 6E2; Cell Signaling Technology) and a mouse-specific secondary antibody (DAKO). b-actin served as control for equal protein loading and was detected by a murine monoclonal antibody (AC15; Sigma-Aldrich) and a mouse-specific second antibody (DAKO). Secondary antibody binding was identified by chemiluminescence and protein sizes were approximated relative to a pre-stained molecular weight marker (Thermo Fisher Scientific, Waltham, MA, USA). Densitometric analysis was performed using ImageJ software (NIH, Bethesda, MD, USA).

In vitro mutagenesis
Single base pairs of murine HDAC5 were exchanged by oligonucleotidedirected mutagenesis using plasmid DNA as a template as described earlier [24]. Sequence-specific primers (Table 1) were designed to delete the adenosine in the start codon (DA in ATG; pHDAC5-FS) or to mutate the histidine residue at position 885 to phenylalanine (H885->F885; pHDAC5-F885) [4]. Mutated sequences were confirmed by DNA sequencing from the expression plasmid (Seqlab, G € ottingen, Germany).
Luciferase reporter assay RAW264.7 transfected with pGL4.10 were used with the Luciferase Reporter Gene Detection Kit (Sigma-Aldrich, Munich, Germany) according to the manufacturer's instructions. Light emission measured in relative light units was assessed for 2 sec. in relation to a luciferase standard curve using a luminometer (Berthold Technologies, Bad Wildbad, Germany).

Statistical analysis
Statistical significance was determined by Kruskal-Wallis test to compare multiple groups or Mann-Whitney test to compare two groups using GraphPad Prism 5 software for Windows (GraphPad Software, San Diego, CA, USA).

Expression of HDAC5 in macrophages
The presence of HDAC5 built the prerequisite for all experiments and so we assessed the baseline expression of the enzyme in murine RAW264.7 and human U937 cells. The HDAC5-specific Western blot analyses showed bands of about 130 kD in both cell lines (Fig. 1A). As confirmed in the knock down experiments (see below, Fig. 4) the upper band represents HDAC5, the lower band is an unspecific staining of the polyclonal anti-HDAC5 antibody. This was confirmed at mRNA level for the RAW264.7 cells (Fig. 1B). As the question was whether pro-inflammatory activation influenced HDAC5 expression levels, RAW264.7 cells were also stimulated with LPS. After 24 hrs, the HDAC5 mRNA expression was comparable to the baseline (Fig. 1C). Kinetics over 7 hrs revealed a significant decrease of HDAC5 expression by about 70% (Fig. 1D), which appeared to revert to the level of non-activated cells after 24 hrs (Fig. 1C). Looking at the short-term kinetics within the first 60 min. after activation, a very early significant up-regulation of the HDAC5 mRNA expression in response to LPS is demonstrated (Fig. 1E). Corresponding results could be observed using CpG as a toll-like receptor-(TLR) 9 agonist ( Fig. 1F and G). The down regulation after LPS stimulation after 7, 10 and 14 hrs was confirmed on a protein level via western blot (Fig. 1H).
Since TNFa belongs to the genes up-regulated early on after LPS stimulation, first the pattern of HDAC5 mRNA expression after stimulation with TNFa was investigated. As shown in Figure 2, the presence of TNFa did not affect HDAC5 mRNA expression over the 7-hr time-period investigated. To assess the milieu dependent expression of HDAC5 mRNA, RAW264.7 cells were subjected to defined chemotaxis-inducing, anti-inflammatory, polarizing or pro-inflammatory conditions for 5 to 7 hrs. While the presence of TNFa, M-CSF, TGFb, MCP-1 or a combination of IL-4 with IL-13 had no impact on the HDAC5 expression, stimulation with the pro-inflammatory cytokine IFNc did profoundly suppress HDAC5 mRNA expression similar to LPS ( Fig. 2A). But different form LPS treatment, IFNc-treatment did not lead to an upregulation of HDAC5 expression within the first 60 min. of stimulation, neither did stimulation with MCP-1 or IL-4 + IL-13 (Fig. 2B).
These findings of an early HDAC5 regulation pattern in macrophages in the presence of a bacteria-derived pro-inflammatory stimulus led to the question of the functional impact.

Functional impact of HDAC5 over-expression
Over-expression of HDAC5 was established in RAW264.7 cells by transiently transfecting murine full-length HDAC5 under the control of a strong promoter. The kinetics as assessed at transcriptional level showed significantly more than 10-fold elevated levels of HDAC5 mRNA after 3 hrs that constantly declined within 9 hrs (Fig. 3A) and returned to starting level within 24 hrs (data not shown). Independent from the presence of specific mRNA, high levels of HDAC5 protein were detected for at least 24 hrs (Fig. 3B). Having established the cell model, the effect of HDAC5 expression on LPS induced cytokine secretion was examined (Fig. 3C). Over-expressed HDAC5 resulted in a significant increase of TNFa, MCP-1 as well as IL-10 secretion, whereas the expression of IL-6 remained unchanged. There was also no change in the expression of the costimulatory molecules and activation markers CD80 and CD86 (data not shown). Excluding HDAC5independent effects caused by the plasmid itself, a frame shift mutant disabling HDAC5 expression did not change the expression levels of TNFa (Fig. 3D). The same result was achieved by replacing the histidine residue in position 885 for phenylalanine in the transiently overexpressed HDAC5 leading to a functional deficiency of the protein.
Addressing the signalling pathways recruited for TNFa production, pHDAD5 was co-transfected with a NF-jB reporter plasmid into RAW264.7 cells (Fig. 3E). Suggesting that HDAC5 might be directly involved in regulating the expression of this cytokine, a profound, about twofold increase in the activation of NF-jB could be detected in parallel to the HDAC5 over-expression.

Functional impact of HDAC5 knock down
The next step addressed the functional impact of HDAC5 by knocking down the enzyme using RNA interference. Transfection of RAW264.7 cells with the HDAC5-specific siRNA resulted in a profound suppres- sion of HDAC5 mRNA which remained stable for at least 48 hrs (Fig. 4A). The HDAC protein also was not found 48 hrs after transfection of the specific siRNA but was high in the control cells (Fig. 4B). Comparable expression pattern were found for HDAC5 mRNA (Fig. 4C) and protein (Fig. 4D) if the human monocyte cell line U937 was also transfected with specific siRNA. As for the functional impact of the successful knock down, the cytokine and chemokine expression in response to LPS stimulation of transfected RAW264.7 and U937 cells was studied. Confirming the findings from the transient over-expression of HDAC5 described above, LPS stimulation in the HDAC5 knock down cells was followed by a robust suppression of TNFa and MCP-1, whereas IL-10 and IL-6 remained unchanged (Fig. 5A). Again no change in the expression of CD80 or CD86 could be detected (data not shown). The knockdown of HDAC5 had no impact on NF-jB activity as shown by co-transfection of the siRNAs with the NF-jB reporter plasmid (Fig. 5B). The profound reduction in TNFa expression in murine cells was confirmed with the human cells, when siHDAC5-treated U937 cells were analysed for the expression of this cytokine (Fig. 5C).
To verify the specificity of the HDAC5 effect on LPS-induced cytokine expression RAW264.7 cells were co-transfected with the plasmid for HDAC5 over-expression and the HDAC5-specific siRNA. Tendencies of lowered expression of pro-inflammatory TNFa as well as of MCP-1 indicated a specific down-regulating effect of the HDAC5 specific siRNA compared to the control siRNA, even upon plasmid-driven high expression of HDAC5 (Fig. 5D).
Effects regarding the production of TNFa and MCP-1 after knockdown of HDAC5 that were consistent with findings from the over-expression experiments emphasized the conclusion that HDAC5 regulates the pro-inflammatory capacity of macrophages.

Discussion
Anti-inflammatory effects of general HDAC inhibition has been proven in various in vitro and in vivo models but the role of single HDAC in defined cell populations remains to be defined. Focusing on the role of HDAC5 in macrophages, we showed that HDAC5 expression was down-regulated by pro-inflammatory LPS or IFNc. HDAC5 over-expression was associated with a significant increase of TNFa and MCP-1 paralleled by an activation of the NF-jB pathway. Supporting the concept of direct involvement of HDAC5 in the pro-inflammatory capacity of macrophages, HDAC5 knockdown by siRNA resulted in the suppression of cytokine production.
Establishing an impact of HDAC5 in innate immune cells and cell lines as suitable models, HDAC5, 7 and 9 are up-regulated during U937 differentiation into macrophages [25]. Down-regulation of pro-inflammatory cytokine production by extracts of edible blue-green algae with generally decreased HDAC expression including HDAC5 and parallel inhibition of NF-jB activity in RAW264.7 cells and in murine bone marrow-derived macrophages link the anti-inflammatory effect to this central pathway [26].
The time-dependent decrease of HDAC5 mRNA after onset of LPS stimulation was concordant to earlier studies [26,27,28], but this immediate effect ceased within 24 hrs. Additionally to LPS, only IFNc or CpG oligonucleotides affected the HDAC5 mRNA levels, which connects this modulation of HDAC5 expression to defined pro-inflammatory conditions and to the cross-talk between IFNc signalling and the Toll-like receptor pathway, activated by LPS via TLR4 or CpG via TLR9 [29]. IFNc sensitizes macrophages for TLR ligands [30,31] and STAT1 activity, induced by TLR4 and TLR9 signalling, mediates IFNc-dependent gene regulation [32,33]. These processes are accompanied by  chromatin remodelling, where HDAC act as negative feedback regulators for LPS-inducible genes like Cox2 [28]. HDAC are also required for signalling via STAT1, where the activity is regulated by protein acetylation [34,35]. We suggest that in macrophages especially HDAC5 is crucial to control inflammation caused by bacterial pathogen-associated molecular pattern including LPS and CpG or bypassed by IFNc. Although TNFa is involved in intracellular localization and activity of HDAC e.g. by regulating shuttling of HDAC3 and HDAC4 into the nucleus [36] and diminishing HDAC1 activity by increased protein turnover [37], TNFa did not represent a potential autocrine feedback mechanism for HDAC5 to LPS stimulation. This held also true for other cytokines released by macrophages or known for affecting macrophages like TGFb, IL-6 or MCP-1.
Enhanced induced NO synthase production upon HDAC5 over-expression in murine RAW264.7 cells or a reduced p65 activation underline the involvement in the defense of pathogens by macrophages but so far does not fully define the possible role for HDAC5 [38,39]. Knockdown of HDAC5 in alveolar epithelial cells, astrocytes and preadipocytes revealed no relation to inflammatory responses of the respective cell types [40][41][42]. In our hands, transient over-expression of full-length HDAC5 in macrophages was associated with significantly augmented levels of pro-inflammatory TNFa and MCP-1 as well as of anti-inflammatory IL-10, while IL-6 levels remained unaffected. This was strengthened by the finding that knockdown of HDAC5 significantly lowered TNFa and MCP-1 levels. Assigning this functional impact to HDAC5, inactivating frame-shift mutation had no proinflammatory effect. Since a point mutation in the HDAC5 domain crucial for the formation of matrix-associated deacetylase bodies that are necessary in transcriptional regulation [23] also did not led to increased TNFa and MCP-1 secretion, we conclude that the integrity of this HDAC domain is indeed manda-  tory for the immune regulatory function of HDAC5 in macrophages. While in recent years, macrophages have been classified as M1 (pro-inflammatory) and M2 (anti-inflammatory) phenotypes, we do not propose a role of HDAC5 in macrophage polarization. In our system, not only 'M1 cytokines' like TNF and MCP-1 are affected by HDAC5 expression, but also IL-10 as classical anti-inflammatory cytokine.
HDAC inhibitors provide evidence that all HDAC can affect the NF-jB signalling recruited in the regulation of immune responses of macrophages [12,43]. HDAC activity is also known as prerequisite to activate DNA polymerase II by NF-jB [44]. Studying dependency of NF-jB activity from HDAC5 levels in an approved NF-jB reporter system, confirmed NF-jB signalling to be involved in the modulating activity of HDAC5 in macrophages.
Our study reveals a direct regulatory function of HDAC5 in the inflammatory response of macrophages. TLR4 or TLR9 stimulation induced immediate up-followed by fast down-regulation of HDAC5 expression. Defining HDAC5 expression levels by over-expression or by knockdown resulted in a concordant change mainly in the pro-inflammatory capacity of the macrophages with HDAC5 to be considered as part of the NF-jB activation process in a TLR-mediated pro-inflammatory response. Furthermore, it seems that the HDAC5 expression level at the time of activation (or shortly after) defines the following cytokine response. Hence, the early up-regulation of HDAC5 ( Fig. 1E and G) might be responsible for the cytokine release, while the following down-regulation for several hours could represent a compensatory regulation. Although the exact feedback mechanism remains unclear, autocrine cytokine signalling is not likely to be involved.
Taken together, the present study sheds new light on the role of individual HDAC in the activation of immune cells and appoints HDAC5 as a potential therapeutic target in diseases with massive macrophage activation like modulated as in autoimmune diseases or inflammation-associated cancer.