BET Inhibition Upregulates SIRT1 and Alleviates Inflammatory Responses

Control of histone acetylation is a part of the epigenetic mechanism that regulates gene expression and chromatin architecture. The members of the bromodomain and extra terminal domain (BET) protein family are a group of epigenetic readers that recognize histone acetylation, whereas histone deacetyl- ases such as sirtuin 1 (SIRT1) function as epigenetic erasers. We observed that BET inhibition by the specific inhibitor JQ1 upregulated SIRT1 expression and activated SIRT1. Moreover, we observed that BET inhibition functionally reversed the pro-inflammatory effect of SIRT1 inhibition in a cellular lung disease model. SIRT1 activation is desirable in many age-related, metabolic and inflammatory diseases; our results suggest that BET protein inhibition would be beneficial in treatment of those conditions. Most importantly, our findings demonstrate a novel mechanism of SIRT1 activation by inhibition of the BET proteins.

Controlo fh istone acetylation is ap arto ft he epigenetic mechanism that regulates gene expression and chromatin architecture. The members of the bromodomain and extra terminal domain (BET) protein family are ag roup of epigenetic readers that recognizeh istonea cetylation, whereas histoned eacetylases such as sirtuin 1( SIRT1) functiona se pigenetic erasers. We observedt hat BET inhibition by the specific inhibitor JQ1 upregulated SIRT1 expression and activated SIRT1. Moreover,w e observedt hat BET inhibition functionally reversed the pro-inflammatory effect of SIRT1 inhibition in ac ellular lung disease model.S IRT1 activation is desirable in many age-related, metabolic and inflammatory diseases;o ur resultss uggest that BET protein inhibition would be beneficiali nt reatment of those conditions. Most importantly,o ur findings demonstrate an ovel mechanism of SIRT1 activation by inhibition of the BET proteins.
BET proteins (BRDT and BRD2-4) activate transcription, whereas sirtuins have the oppositee ffect, gene silencing. [1] BET proteins are important for cell-cycle control and they have been linked to the developmento fan umber of extremely aggressive tumors. [2] Sirtuins have been associated with the elongation of life-span, and the activationo fs irtuin 1( SIRT1) has caught particular interest world-wide. [3,4] Both BET proteins and SIRT1 have been connected to several age-related, inflammatory and metabolic diseases,t hus making these epigenetic regulators interesting targets for drug development. During the past years, we and others have developed variousS IRT1 inhibitors, and several inhibitors for BET proteins have been reported. [1,[5][6][7][8] As the functions of both BET proteins and SIRT1 depend on the histonea cetylation status,w eexplored whether inhibition of the BET proteins by as elective inhibitor and gene silencing could affect human SIRT1. After the finding of SIRT1 upregulation and activation, we analyzed the biological effects of this pathway in A549 cells, al ung disease cell line where SIRT1 is known to have beneficial anti-inflammatory effects. [9] BET inhibition by JQ1 (1)w as used to explore the interplay between BET proteins and human SIRT1 in different cell types.
Next we studied whether the deacetylase activity of SIRT1 increases with increased expression. The acetylation level of p53, aS IRT1 substrate, was analyzed by western blotting of humanc ell lysates. JQ1 enhanced p53 deacetylation (indicative of SIRT1 activation) in all tested cell lines ( Figure 1C). JQ1 did not affect the activity of SIRT1 in enzymatic activity assays (SIRT1 activity with 400 nm JQ1 was 99.3 % AE 2.1 %o fc ontrol). This indicates that JQ1 is not ad irect activatoro fS IRT1 in vitro. Visuali nspection at 24 hs howed that JQ1 treatments did not elicit any changes in cell morphology, and cell number was not affected by treatment, as determined by protein content in the wells (data not shown). In addition, flow cytometry analysis of cellular DNA content revealed that treatment with 400 nm JQ1 for 24 hd id not modifyt he cell cycle or induce apoptosis in A549 cells( Figure S1 in the Supporting Information).
In order to induce inflammation, A549 (adenocarcinomic alveolare pithelial cells) were treated with lipopolysaccharide (LPS) for 24 h, and IL-8 secretion wasq uantified as an indicator of the inflammatoryr esponse. BET inhibition by JQ1 treatment prevented LPS-inducedi nflammation, whereas SIRT1 inhibition by the specific inhibitor EX527 (2)e nhanced the inflammatory response. JQ1 treatment wasa ble to reverse the inflammationenhancing effect of SIRT1 inhibition ( Figure 2A). Furthermore, LPS treatment increased ROS production whereas treatments with EX527 or JQ1 had no significant effect on ROS generation ( Figure 2B). Gene silencing by siRNA transfectionsw as used in order to assess the contributions of BRD2 and BRD4 in the JQ1-evoked responses. Four commercial siRNAs were tested for silencing efficiencyo fB RD2 and BRD4g enes, and the most efficient siRNAs were chosen for further experiments( see Figure S2). Silencing of BRD2 abolished (and silencing of BRD4 diminished) the inflammation-enhancing effect of SIRT1 inhibition withouta ffecting basal or LPS-stimulated IL-8 secretion ( Figure 3A). BRD2 silencingw as also found to reduce the stimulatory effect of LPS on ROS generation ( Figure 3B); BRD4 silencing did not have as ignificant effect on ROS levels (Figure 3B).
Because of the numerous beneficial effects of SIRT1, ways and means of its activation have been eagerly hunted. [3,4] Several selectiveS IRT1 inhibitors are availablef or research, but the effectiveness of direct SIRT1 activators is more controversial. [10][11][12][13][14][15] Also, initial enthusiasm following the originalr eport [16] on the physiological effects of AROS as an SIRT1 activatorh as faded. [15][16][17][18] Therefore, it seems that the most promising way to activate SIRT1 would be by ac ellular pathway.I nt he search for an ovel regulation mechanism of SIRT1, we investigated the interplayb etween BET proteins and SIRT1. JQ1 upregulates SIRT1 protein levels in severalh uman and mousec ell lines. In line with these results, upregulation of SIRT1 mRNA has been previously detected in microarrays from JQ1-treated lymphocytes. [19] JQ1 also enhanced the deacetylation of tumor suppressor protein p53, aw ell-established SIRT1 substrate, [20] in human cells. Because JQ1 did not directly modulate SIRT1 activity in aF luor de Lysi nv itro activity assay,t he observed increase in p53 deacetylation by JQ1 treatment was most likely aresult of increased SIRT1 expression. This suggests that BET protein inhibition not only exposes more acetylated sites for the deacetylase activity of SIRT1, but also upregulates SIRT1 and thus enhances the cellular activity of SIRT1. This will serve as as tarting point for developing novel strategies of ChemBioChem 2015, 16,1997 -2001 www.chembiochem.org SIRT1 activation, which would be desirable in many age-related and metabolic diseases.
Human cell line A549 was chosen to investigate the biological responses evoked by JQ1-induced BET inhibition/SIRT1a ctivation. A549 cells have been extensively used to study alveolar immunotoxicity. [21] They have been used in the study of both chronic inflammation associated with chronic obstructive pulmonaryd isease, asthma, or cigarettes moke exposure, and acute inflammation associated with influenza, tuberculosis, or pneumonia. They are also often used to model non-small-cell lung cancer,whichist he most common form of lung cancer. [22] As chronic inflammation is associated with the pathogenesis of lung cancer, [23] the inflammatory response in A549 cells is important with regard to most the severe lung diseases. Mimicking the intrusion of microbial pathogens by exposure of A549 cells to bacterial LPSh as been found to stimulate ROS production, and to evoke an inflammatory response, as demonstrated by NF-kB activation and increased interleukin-8( IL-8) production. [24] Importantly,a ctivation of SIRT1 by exposure to resveratrol has been found to significantly dampen the inflammatory responses and reverse the effects of LPS in A549 cells, whereas inhibition of SIRT1 was found to result in opposite effects. [25] Therefore, we used IL-8 and ROS production as parameters re-flectingt he functiono fe pithelial defense mechanismsd uring the JQ1 treatment.
SIRT1 is an effective inhibitor of inflammatory signaling. [24,27] In contrast, BRD2a nd BRD4 are essential for inflammation, and JQ1 treatment has been found to dampen the inflammatory response. [28,29] We found that SIRT1 inhibition enhanced the LPS-inducedi nflammatory response, whereas BET inhibition had the opposite effect. Furthermore, BET inhibition by JQ1 was able to alleviate the inflammatoryr esponse evoked by SIRT1 inhibition. Our resultsa re in agreement with the previously published resultsa bout anti-inflammatory effects of SIRT1 activationa nd BET inhibition. However,o ur study does not offer final confirmation that the reversal of the pro-inflammatorye ffect of EX527 by JQ1 occurs primarily through SIRT1 upregulation. Silencing of BRD2 resulted in ac omplete (and silencing of BRD4 in ap artial) reversal of the inflammation-enhancing effect of SIRT1 inhibition. This indicates that both BRD2 and BRD4 are involved in the SIRT1-mediated pathway in A549 cells. The relative importance of these BET proteins on inflammatory signaling has previously been shown to be celltype specific. [30,31] In ar ecent paper JQ1 was found to inhibit ROS production, [32] whereas SIRT1 inhibition by an onspecific sirtuin inhibi-  ChemBioChem 2015, 16,1997 -2001 www.chembiochem.org tor nicotinamide wasp reviously reported to increase ROS production. [25] In our hands, BET inhibition by JQ1a nd SIRT1 inhibition by EX527 had no significant effect on ROS generation. These discrepancies could be explained by the use of different cell types, ROS measurement and exposure times, and by the use of as elective SIRT1 inhibitor in our study.B RD2 silencing was found to reduce the stimulatory effect of LPS on ROS generation. As neither BRD4 silencing nor JQ1 treatment had significant effects on ROS levels, BRD2 silencing is most likely to have itse ffect on ROS generation through an unknown pathway that does not involve recognition of histone acetylation.
Our resultsp resent an ovel pathway for SIRT1 upregulation and activation in multiple human and mouse cell types by BET bromodomain inhibitorJ Q1 (Figure 4). The inhibitor blockst he interaction of BET proteins with acetylated lysines (Ac in Figure 4), thus allowing the removal of acetylation modifications by SIRT1. This influences the transcription of genes that are induced during inflammatory responses. The inhibition of the bindingo fB ET bromodomains can further influence neighboring histone acetylation modifications. We have also shown that SIRT1 upregulation and bromodomain inhibition have anti-inflammatory effects in al ung disease model, human A549 alveolar epithelial cells, and that both BRD2 and BRD4 bromodomain proteins are involved in these processes. Our study reveals for the first time af unctional connectionb etween BET proteins and SIRT1. The BET bromodomains have gained substantial attentionf or the treatment of human cancers and several inhibitors have already been developed. By combining two classes of epigenetic regulators-BET bromodomainswith sirtuins-novel strategies to regulate gene activity are now emerging.
The effect of JQ1 on the enzymatic activity of SIRT1 was measured with Fluor de LysS IRT1 fluorometric drug discovery activity assay kit (BML-AK555, Enzo Life Sciences, Farmingdale, NY) with substrates BML-KI177 (50 mm,E nzo) and NAD + (500 mm). Fluorescence readings were obtained in aVictor 1420 Multilabel Counter (Perkin-Elmer) with excitation at 360 nm and emission at 460 nm. The assay was repeated twice. siRNA gene silencing was done by using Lipofectamine 2000 transfection reagent (LifeTechnologies) according to the manufacturer's instructions to transfect A549 cells seeded in 24-well plates (30 000 cells per well) with FlexiTube siRNAs (40 pmol, Qiagen). Four siRNAs were tested for silencing efficiencyo fe ach gene (BRD2 and BRD4). AllStars Negative Control siRNA (Qiagen) was used as anegative control. Other treatments were started 24 hp ost-transfection. IL-8 determination from cell culture medium was done with an ELISA kit (RAB0319, Sigma-Aldrich). ROS production was assessed by using the fluorescent indicator 2',7-dichloro-dihydrofluorescein diacetate (H 2 DCFDA) as previously described. [34] Data shown in Figures 2a nd 3a re from as ingle experiment measured in triplicate and analyzed with one-way ANOVAa nd Tukey's post hoc test in GraphPad Prism. For clarity,o nly as ingle level of significance is shown (p < 0.05). The experiments were repeated at least three times with similar results.