High-Throughput Screening for CEBPD-Modulating Compounds in THP-1-Derived Reporter Macrophages Identifies Anti-Inflammatory HDAC and BET Inhibitors

This study aimed to identify alternative anti-inflammatory compounds that modulate the activity of a relevant transcription factor, CCAAT/enhancer binding protein delta (C/EBPδ). C/EBPδ is a master regulator of inflammatory responses in macrophages (Mϕ) and is mainly regulated at the level of CEBPD gene transcription initiation. To screen for CEBPD-modulating compounds, we generated a THP-1-derived reporter cell line stably expressing secreted alkaline phosphatase (SEAP) under control of the defined CEBPD promoter (CEBPD::SEAP). A high-throughput screening of LOPAC®1280 and ENZO®774 libraries on LPS- and IFN-γ-activated THP-1 reporter Mϕ identified four epigenetically active hits: two bromodomain and extraterminal domain (BET) inhibitors, I-BET151 and Ro 11-1464, as well as two histone deacetylase (HDAC) inhibitors, SAHA and TSA. All four hits markedly and reproducibly upregulated SEAP secretion and CEBPD::SEAP mRNA expression, confirming screening assay reliability. Whereas BET inhibitors also upregulated the mRNA expression of the endogenous CEBPD, HDAC inhibitors completely abolished it. All hits displayed anti-inflammatory activity through the suppression of IL-6 and CCL2 gene expression. However, I-BET151 and HDAC inhibitors simultaneously upregulated the mRNA expression of pro-inflammatory IL-1ß. The modulation of CEBPD gene expression shown in this study contributes to our understanding of inflammatory responses in Mϕ and may offer an approach to therapy for inflammation-driven disorders.


Introduction
Inflammation is a protective host response to pathogens and cell or tissue damage that maintains tissue homeostasis and enables the organism to survive during injury or infection. However, chronic inflammation can have negative effects in the context of inflammatory disorders. Inflammatory responses are mediated by multiple interactions between immune cells and regulated by molecular mediators and their corresponding signaling pathways. However, inflammation is believed to be regulated mainly at the level of gene transcription [1][2][3].
The structural and regulatory characteristics of CEBPD allow the monitoring of CEBPD promoter activity using a gene reporter (e.g., SEAP). The human CEBPD gene has a very simple gene structure containing a distinct promoter and a single exon [36]. Lacking introns, alternative promoters, and alternative polyadenylation sites [36], a single and highly instable CEBPD mRNA is produced (half-life of 35-40 min) [7,37]. Intronless CEBPD pre-mRNA is not processed by the splicing machinery and is translated to a single C/EBPδ protein with one known full-length functional C/EBPδ protein isoform [7,32]. Additionally, C/EBPδ protein displays a short biological half-life ranging between two and four hours in epithelial cells [37] and Mφ [32]. C/EBPδ is predominantly a nuclear protein, it immediately translocates to the nucleus after protein synthesis and does not exist in an inactive form in cytoplasm [7,38].
C/EBPδ protein expression is largely regulated at the level of CEBPD gene transcription initiation [7]. To date, 92 TFs have been confirmed experimentally by chromatin immunoprecipitation sequencing (ChIP-Seq) to bind the human CEBPD promoter [39]. The proximal CEBPD promoter region, approximately 200 base pairs (bp) upstream from transcription start site (TSS), contains most of the characterized TF binding sites [7]. Among the most important TFs regulating C/EBPδ expression is cyclic adenosine monophosphate (cAMP) responsive element-binding protein (CREB) [40,41], specificity protein 1 (SP1) [16,42,43], and a signal transducer and activator of transcription 3 (STAT3) [16,43,44], but also nuclear factor-kappa B (NF-kB) [10,45] and activating transcription factor 3 (ATF3) [10]. However, less robust evidence exists for the position of NF-kB and ATF3 binding sites. Based on the alignment of the mouse cebpd and human CEBPD promoter sequences, we propose NF-kB and ATF3 TF binding sites for the human sequence (Supplementary Materials, Figure S1).
Considering its versatile function during inflammation, both the pharmacological activation and inhibition of CEBPD expression may result in a desired anti-inflammatory effect. However, only a few anti-inflammatory compounds, such as rosmanol [29] and inotilone [29], are known to modulate CEBPD gene expression. Here, we describe the development and application of a phenotypic assay to identify CEBPD-modulating compounds in LPS-and IFN-γ-activated THP-1-derived Mφ.
The high-throughput screening (HTS) of LOPAC ®1280 and ENZO ®774 libraries, containing 2054 compounds in total, identified four hits that showed a pronounced effect on CEBPD transcription activity. The identified hits, two bromodomain and extraterminal domain (BET) inhibitors and two histone deacetylase (HDAC) inhibitors, displayed significant anti-inflammatory effects, downregulating IL-6 and CC-chemokine ligand 2 (CCL2) expression, but differentially affected the mRNA expression of endogenous CEBPD and IL-1ß. In this study, we report that the CEBPD-modulating compounds act as anti-inflammatory agents, as shown in LPS-and IFN-γ-activated THP-1-derived Mφ, and may, therefore, be considered as potential drug candidates for the treatment of inflammation-driven disorders. Secreted alkaline phosphatase (SEAP) is a well-established gene reporter, secreted levels in the conditioned medium of which are proportional to the changes in its intracellular mRNA, protein levels, and cell number [46,47]. SEAP displays a high protein stability with a half-life of over 20 days [48] leading to its accumulation in the cell culture supernatant upon promoter activation and its subsequent expression. It is easy to distinguish the protein from endogenous phosphatases, as SEAP is resistant to inhibition by L-homoarginine and shows high heat tolerance [46,49]. These propensities of the SEAP reporter allow its monitoring for rather long exposure times, which is advantageous, particularly when the time point of reporter activation is unclear. Additionally, weak reporter activation can also be detected, because SEAP reporter and, therefore, the SEAP enzymatic signal accumulates in the supernatants.

Results
First, we generated two HEK293T reporter cell lines by stable vector transfection to characterize the chemiluminescent SEAP assay that we used subsequently as the readout assay. While the CMV::SEAP HEK293T reporter cells stably expressed myc-tagged SEAP (SEAP-myc) under control of the strong constitutive cytomegalovirus (CMV) promoter, the negative control HEK293T reporter cells stably expressed the empty vector backbone without the SEAP gene.
As expected, the SEAP reporter was detectable as enzymatic activity in cellular supernatants of established HEK293T reporter cells (Figure 1a). SEAP assay linearity, sensitivity, and variability were estimated in undiluted and serially-diluted cellular supernatants of HEK293T reporter cells (Figure 1b,c; Table 1). The chemiluminescent SEAP assay was linear (R 2 = 0.99) over four 10-fold dilutions (1 to 10,000) (Figure 1b), demonstrated a broad dynamic range, and displayed a high assay sensitivity, as indicated by a Z'-factor of 0.7 for the 1000-fold diluted sample ( Figure 1c). The Z'-factor is a statistical criterion that takes into account means and signal deviations indicating a clear difference between a positive (SEAP) and a negative (background) signal [50]. An optimal Z -factor is typically set at ≥0.5.
Assay variability is expressed by the coefficient of variation (CV) [51]. The intra-assay CV indicates the variability of signal values between multiple technical replicates of the same sample. The SEAP assay displayed a low intra-assay CV (<10%) for up to 1000-fold diluted samples (Table 1). Assay reproducibility is expressed by the inter-assay CV that indicates the variability of signal values of samples that belong to different individual experiments (biological replicates). The chemiluminescent SEAP assay displayed a low inter-assay CV (< 5%) for up to 1000-fold diluted samples and thus a high assay reproducibility. From these experiments, we concluded that SEAP in combination with the chemiluminescent assay provides a suitable reporter system for the screening of CEBPDmodulating compounds.

Figure 1.
Characterization of chemiluminescent secreted alkaline phosphatase (SEAP) assay. (a-c) SEAP enzymatic activity (a), assay linearity (b), and assay sensitivity (c) were determined in undiluted and up to 10,000-fold diluted cellular supernatants of HEK293T reporter cells, 24 h after cell seeding (mean ± SD, n = 3). Negative control is represented by the cell culture supernatant from empty vector-expressing HEK293T cells. RLU: relative luminescence units. In this study, phorbol 12-myristate 13-acetate (PMA) differentiated and inflammatory-activated THP-1-derived reporter Mϕ constituted the cellular system for the screening of CEBPD-modulating compounds.
Therefore, we generated THP-1 reporter cells by viral transduction of wild type THP-1 cells using viral particles containing the CEBPD::SEAP vector, in which the SEAP reporter gene was placed under the control of the defined CEBPD promoter (Figure 2a). The generated CEBPD::SEAP vector incorporated the proximal promoter region (approximately 330 bp upstream from TSS) containing the already known (SP1, CREB, and APRE) and proposed (NF-kB and ATF3) binding sites of regulatory TFs (Figure 2b). Upon generation, we tested the THP-1 reporter cells for their ability to differentiate into Mϕ. Like wild-type THP-1 cells, THP-1 reporter monocyte-like cells in suspension differentiated into adherent THP-1-derived reporter Mϕ in response to 50 ng/mL PMA treatment for 48 h (Supplementary Materials, Figure S2). , and assay sensitivity (c) were determined in undiluted and up to 10,000-fold diluted cellular supernatants of HEK293T reporter cells, 24 h after cell seeding (mean ± SD, n = 3). Negative control is represented by the cell culture supernatant from empty vector-expressing HEK293T cells. RLU: relative luminescence units. In this study, phorbol 12-myristate 13-acetate (PMA) differentiated and inflammatoryactivated THP-1-derived reporter Mφ constituted the cellular system for the screening of CEBPD-modulating compounds.
Therefore, we generated THP-1 reporter cells by viral transduction of wild type THP-1 cells using viral particles containing the CEBPD::SEAP vector, in which the SEAP reporter gene was placed under the control of the defined CEBPD promoter (Figure 2a). The generated CEBPD::SEAP vector incorporated the proximal promoter region (approximately 330 bp upstream from TSS) containing the already known (SP1, CREB, and APRE) and proposed (NF-kB and ATF3) binding sites of regulatory TFs (Figure 2b). Upon generation, we tested the THP-1 reporter cells for their ability to differentiate into Mφ. Like wildtype THP-1 cells, THP-1 reporter monocyte-like cells in suspension differentiated into adherent THP-1-derived reporter Mφ in response to 50 ng/mL PMA treatment for 48 h (Supplementary Materials, Figure S2). The cell culture medium from THP-1 wild type (wt) and reporter cells was collected 24 h before, during (48 h with PMA), and after (48 h recovery) treatment with 50 ng/mL PMA for SEAP assay (mean ± SD, n = 3). In (d,e) PMA-differentiated THP-1 reporter Mϕ were stimulated with PBS (solvent control), 1 μg/mL or 0.1 μg/mL LPS + 20 ng/mL IFN-γ for up to 6 h for mRNA expression analysis of endogenous CEBPD (d) and reporter CEBPD::SEAP (e) (mean ± SD, n = 3). mRNA expression analysis was performed via qRT-PCR (ΔΔCt method) using RPL37A as an internal reference gene. Differences in CEBPD gene expression levels were analyzed relative to PBS via one-way ANOVA or Kruskal-Wallis test with Dunn's correction for multiple comparisons. * p < 0.05; *** p < 0.001. (f) The cell culture medium from PMAdifferentiated THP-1 reporter Mϕ were collected after treatment with PBS (solvent control), 1 µ g/mL or 0.1 µ g/mL LPS + 20 ng/mL IFN-γ for 1, 3, 6, or 24 h (mean ± SD, n = 3). SEAP enzymatic signal was monitored for 40 min after substrate addition. RLU: relative luminescence units.

A High-Throughput Screening for CEBPD-Modulationg Compounds
We specified the final screening protocol with regard to the semi-automatic screening performance (Figure 3a). In total, we screened 2054 compounds, 1280 from the list of pharmacological active compounds (LOPAC ® ) and 774 from the ENZO ® (Screen-Well ® FDA approved drug library V2) libraries. The cell culture medium from THP-1 wild type (wt) and reporter cells was collected 24 h before, during (48 h with PMA), and after (48 h recovery) treatment with 50 ng/mL PMA for SEAP assay (mean ± SD, n = 3). In (d,e) PMA-differentiated THP-1 reporter Mφ were stimulated with PBS (solvent control), 1 µg/mL or 0.1 µg/mL LPS + 20 ng/mL IFN-γ for up to 6 h for mRNA expression analysis of endogenous CEBPD (d) and reporter CEBPD::SEAP (e) (mean ± SD, n = 3). mRNA expression analysis was performed via qRT-PCR (∆∆Ct method) using RPL37A as an internal reference gene. Differences in CEBPD gene expression levels were analyzed relative to PBS via one-way ANOVA or Kruskal-Wallis test with Dunn's correction for multiple comparisons. * p < 0.05; *** p < 0.001. (f) The cell culture medium from PMA-differentiated THP-1 reporter Mφ were collected after treatment with PBS (solvent control), 1 µg/mL or 0.1 µg/mL LPS + 20 ng/mL IFN-γ for 1, 3, 6, or 24 h (mean ± SD, n = 3). SEAP enzymatic signal was monitored for 40 min after substrate addition. RLU: relative luminescence units.
We then tested the effect of PMA on SEAP secretion into cellular supernatants of the THP-1 reporter cells, collected before, during, and after PMA treatment. As expected, SEAP enzymatic activity was at background levels in cellular supernatants of undifferentiated THP-1 reporter cells (Figure 2c, 24 h before PMA addition), since CEBPD gene displays low basal expression in monocytes [41]. PMA-activated CEBPD promoter activity and consequently SEAP secretion were detected in cellular supernatants of THP-1 reporter cells collected during (48 h with PMA) and after (48 h recovery) PMA treatment (Figure 2c).
We aimed to perform a HTS for CEBPD-modulating compounds in the cellular context of inflammation. Therefore, we activated non-polarized PMA-differentiated THP-1 reporter Mφ (M0) to a pro-inflammatory M1 phenotype state with LPS-and IFN-γ to evaluate the inflammation-induced CEBPD gene expression. Treatment with 20 ng/mL IFN-γ and 0.1 or 1 µg/mL LPS significantly activated the mRNA expression of endogenous CEBPD up to 4.5-fold, 6 h after treatment (Figure 2d).
As a last step, we tested whether the CEBPD::SEAP construct is able to adequately report the endogenous CEBPD mRNA expression. In fact, we found that the SEAP reporter mRNA expression under the control of the CEBPD::SEAP construct closely mirrored the mRNA expression of the endogenous CEBPD with even similar fold changes at 3 and 6 h after treatment (Figure 2e). In the supernatant, however, activation of the CEBPD::SEAP reporter, due to the additional protein translation, transport and secretion, showed a lag in detecting the response. As a consequence, the most prominent reporter signal was measured 24 h post-treatment ( Figure 2f). In total, these observations suggested that the generated THP-1 reporter cell line was appropriate for the intended screening assay, as the cell line still possessed its differentiation and polarization capabilities, compared to the wild type THP-1 cells and the SEAP reporter activation matched the activation of CEBPD.

A High-Throughput Screening for CEBPD-Modulationg Compounds
We specified the final screening protocol with regard to the semi-automatic screening performance ( Figure 3a). In total, we screened 2054 compounds, 1280 from the list of pharmacological active compounds (LOPAC ® ) and 774 from the ENZO ® (Screen-Well ® FDA approved drug library V2) libraries.
The HTS for CEBPD-modulating compounds included seven 384-well plates: four with LOPAC ® and three with ENZO ® screening compounds. Here, we show one plate containing hit compounds from the LOPAC ® ( Figure 3b) and one plate containing hit compounds from the ENZO ® (Figure 3c) libraries. The primary screening data for all plates from LOPAC ® and ENZO ® libraries are summarized in supplementary Tables S1 and S2, respectively.
We performed the primary screening twice (read 1 and read 2). For each screening plate, we displayed enzymatic SEAP signals from read 1 (x-axis) in correlation to corresponding signals from read 2 (y-axis) (example shown in Figure 3b To identify hit compounds, we characterized each compound-caused SEAP signal (representatively shown in Figure 3b,c; gray dots) according to its difference to an average signal of the corresponding M1 solvent control (representatively shown in Figure 3b,c; red lines). We calculated the means (m) and standard deviations (SD) of SEAP signals from M1 solvent control for each read and plate. After, we added up to three-fold SDs to the corresponding average signals of M1 solvent control (Figure 3b,c; black lines), for each read and plate. A compound, for which the SEAP signal was clearly different (>3 SDs) from the average SEAP signal of the corresponding M1 solvent control in both reads, we considered as a hit. Here, we identified three hit compounds: I-BET151, Ro 11-1464, and suberoylanilide hydroxamic acid (SAHA).  I-BET151 and Ro 11-1464 are members of the family of BET inhibitors, while SAHA is an HDAC inhibitor [52]. We also included TSA in the hit compound list, as it reproducibly activated SEAP secretion and is also an epigenetically active HDAC inhibitor [53].
As both LOPAC ® and ENZO ® libraries also contain well-characterized cytotoxic compounds, we performed a cell viability test (CellTiter-Glo ® Assay) during read 2. Notably, cells treated with compounds such as lasofoxifene tartrate, sanguinarine chloride, auranofin, or mitoxantrone hydrochloride exhibited decreased cell viability as well as decreased SEAP activity, pointing out cell viability as a confounding factor of the assay (Supplementary Materials, Tables S1 and S2). We repeated the cell viability test for the four identified hits, I-BET151, Ro 11-1464, SAHA, and TSA, and confirmed that the increase in SEAP activity was not associated with an increase in cell viability or proliferation (Supplementary Materials, Figure S3). On the contrary, I-BET151 and TSA even reduced cell viability, but by far not as much as lasofoxifene tartrate, sanguinarine chloride, auranofin, or mitoxantrone hydrochloride.
In the HTS, we observed that the correlation of the signals from read 1 and 2 was rather weak, negatively affecting assay robustness and sensitivity. Exemplarily calculated CV (range: 27.7 to 48.9%) for the M0 and M1 DMSO control of the two plates and the corresponding Z -factor (−0.4; −1.3) demonstrated the limitations of the assay to identify inhibitory compounds (Figure 3b,c). We, therefore, restricted the hit selection to compounds which displayed a prominent activatory effect in both reads, taking into account that further potential CEBPD-modulating compounds would be missed by that selection.

Analysis of Hit Compound Effects on Gene Expression
Following HTS, we analyzed the identified hits according to their ability to modulate the mRNA expression of CEBPD and other inflammation-related genes. We observed compound-mediated changes in the mRNA expression of endogenous CEBPD, reporter CEBPD::SEAP, IL-6, IL-1ß, and CCL2.  (Table 2). Ro 11-1464 caused approximately 9-fold suppression in the mRNA expression of IL-6 and CCL2 (Table 3). Whereas Ro 11-1464 showed no effect on IL-1ß mRNA expression (Figure 5f; red triangles), this was more than doubled (Table 2)  . mRNA expression analysis was performed via qRT-PCR (ΔΔCt method) using RPL37A as an internal reference gene. Fold change in mRNA expression is displayed relative to M0 control, set as 1. Changes in mRNA expression were analyzed relative to M1 DMSO control via unpaired t-test with (for normally distributed data and different SDs) and without (for normally distributed data and equal SDs) Welch's correction. ** p < 0.005; *** p < 0.001; **** p < 0.0001. Table 2. Fold change (mean ± SD) for mRNA expression in I-BET151-pre-treated target cells compared to M0 controls (compare Figure 4).  . mRNA expression analysis was performed via qRT-PCR (∆∆Ct method) using RPL37A as an internal reference gene. Fold change in mRNA expression is displayed relative to M0 control, set as 1. Changes in mRNA expression were analyzed relative to M1 DMSO control via unpaired t-test with (for normally distributed data and different SDs) and without (for normally distributed data and equal SDs) Welch's correction. ** p < 0.005; *** p < 0.001; **** p < 0.0001. Table 2. Fold change (mean ± SD) for mRNA expression in I-BET151-pre-treated target cells compared to M0 controls (compare Figure 4).   Table 3. Fold change (mean ± SD) for mRNA expression in Ro 11-1464-pre-treated target cells compared to M0 controls (compare Figure 5).

Condition 4 h, Mean ± SD
CEBPD::SEAP M1 DMSO 5.5 ± 1.5 . mRNA expression analysis was performed via qRT-PCR (∆∆Ct method) using RPL37A as internal reference gene. Fold change in mRNA expression is displayed relative to M0 control, set as 1. The Ro 11-1464-mediated changes in mRNA expression were analyzed relative to M1 DMSO control via unpaired t-test with (for normally distributed data and different SDs) and without (for normally distributed data and equal SDs) Welch's correction. n.s.: not significant; *** p < 0.001; **** p < 0.0001. Table 3. Fold change (mean ± SD) for mRNA expression in Ro 11-1464-pre-treated target cells compared to M0 controls (compare Figure 5).   (Table 4). HDAC inhibitor TSA suppressed the mRNA expression of IL-6 over 130-fold and that of CCL2 over 17-fold (Table 4). Interestingly, as observed for I-BET151, the mRNA expression of IL-1ß was slightly but significantly upregulated by the two HDAC inhibitors, 4 h after LPS + IFN-γ treatment (Figure 6f; red and orange triangles). Table 4. Fold change (mean ± SD) for mRNA expression in SAHA-pre-treated and TSA-pre-treated target cells compared to M0 controls (compare Figure 6).  (Figure 6e; red and orange triangles) relative to the M1 DMSO control. Thus, SAHA reduced the mRNA expression of IL-6 over 110-fold and that of CCL2 over 15-fold (Table  4). HDAC inhibitor TSA suppressed the mRNA expression of IL-6 over 130-fold and that of CCL2 over 17-fold (Table 4). Interestingly, as observed for I-BET151, the mRNA expression of IL-1ß was slightly but significantly upregulated by the two HDAC inhibitors, 4 h after LPS + IFN-γ treatment (Figure 6f; red and orange triangles). . mRNA expression analysis was performed via qRT-PCR (∆∆Ct method) using RPL37A as an internal reference gene. Fold change in mRNA expression is displayed relative to M0 control, set as 1. The compound-mediated changes in mRNA expression were analyzed relative to M1 DMSO control via ordinary one-way ANOVA (for normally distributed data and equal SDs), Brown-Forsythe and Welch ANOVA test with Dunnett's correction (for normally distributed data and different SDs), or Kruskal-Wallis test with Dunn s correction for multiple comparisons (for non-normally distributed data). * p < 0.05; ** p < 0.005; *** p < 0.001.
In this study, we used the in vitro-differentiated and LPS + IFN-γ-activated THP-1 Mφ reporter cell line to screen for CEBPD expression-modulating compounds. CEBPD encodes the C/EBPδ TF, functioning as a key regulator of inflammatory gene transcription in Mφ [8][9][10]. We identified I-BET151, Ro 11-1464, SAHA, and TSA as showing prominent and reproducible effects on the SEAP reporter, in terms of its mRNA expression, as well as its secretion into the cell supernatant, the primary readout of the reporter assay. How-ever, subsequent gene expression analysis of the endogenous CEBPD revealed that BET inhibitors I-BET151 and Ro 11-1464, as expected, upregulated CEBPD expression, whereas HDAC inhibitors SAHA and TSA, in contrast, downregulated CEBPD expression to the baseline level.
Compounds of both inhibitor classes display epigenetic activity by the direct regulation of gene transcription [64]. Opposite effects of hits on the mRNA expression of the endogenous CEBPD and reporter CEBPD::SEAP were, therefore, highly unexpected. As a possible explanation, HDAC inhibitors may, coincidently, show an activatory cisand a stronger inhibitory trans-effect on the CEBPD promoter. The activatory cis-effect may result in the upregulated gene expression of both endogenous CEBPD and reporter CEBPD::SEAP. The domination of the inhibitory trans-effect over the cis-activation may, however, require distal CEBPD promoter regions, which are absent in the gene reporter setting of CEBPD::SEAP, as only the proximal 332 bp promoter region was used (Figure 2b). BET inhibitors may exclusively control the activatory cis-mechanisms.
On the other hand, BET inhibitors may activate CREB-, SP1-, and STAT3-related signaling pathways, resulting in the elevated gene expression of the endogenous CEBPD and reporter CEBPD::SEAP, as the defined CEBPD promoter contains only the most proximal and well-known TF binding sites. HDAC inhibitors may inhibit the mRNA transcription of the endogenous CEBPD through the activation of the ATF3 TF binding site, which is crucial for CEBPD gene transcription inhibition in Mφ [10]. As the defined CEBPD promoter may contain a disrupted and, therefore, non-functional ATF3 binding site (Supplementary Materials, Figure S1), the mRNA expression of reporter CEBPD::SEAP could not be inhibited by HDAC inhibitors.
This study is the first report on the ability of I-BET151 to regulate the mRNA expression of CEBPD in general and that of IL-6, IL-1ß, and CCL2 in THP-1 Mφ. I-BET151 significantly suppressed the mRNA expression of IL-6 and CCL2, but upregulated that of IL-1ß, demonstrating simultaneous anti-and pro-inflammatory effects in LPS + IFN-γ-stimulated THP-1 reporter Mφ.
I-BET151 may suppress CCL2 mRNA expression by the inhibition of LPS-or IFN-γsignaling mediated by the corresponding TLR4 and Janus kinase (JAK) and STAT pathways. Thus, I-BET151 represses the transcription of STAT target genes and TLR4-induced IFN responses in IFN-γ-activated human monocytes [71]. The inhibited CCL2 gene expression may also result from the I-BET151-modulated suppression of positive transcription elongation factor b (pTEFb) leading to a reduced transcription elongation [71]. Further, the gene expression of CCL2 may be epigenetically repressed by I-BET151 via the inhibition of BRD4, as shown for other BET inhibitors, IBET and JQ1, in activated Mφ [72,73].
The upregulation of IL-1ß mRNA expression by I-BET151 was highly unexpected, as I-BET151 significantly reduced IL-1ß secretion in bone marrow-derived macrophages (BMDM) [72] and IL-1ß gene expression in GFs in the context of inflammatory periodontitis [70]. Presumably, the IL-1ß mRNA expression may be upregulated at higher concentrations of I-BET151 used in this study or due to cell-specific gene expression regulation of IL-1ß [74].

BET Inhibitor Ro 11-1464
Ro 11-1464 was first reported to stimulate the production of high-density lipid protein apoA-I, a well-known anti-inflammatory agent [75], in cultured liver cells Hep G2 [76]. Ro 11-1464 also binds BETs, but displays weak inhibitory potential [77]. To date, the known therapeutic potential of Ro 11-1464 is mostly based on its action as a potent apoA-I inducer and thus, as a potential anti-atherosclerotic agent [78].
This study is the first report to show that Ro 11-1464 affects the gene expression of CEBPD, IL-6, CCL2, and IL-1ß in general and to show an anti-inflammatory effect in THP-1 Mφ. Ro 11-1464 significantly reduced the mRNA expression of IL-6 and CCL2, but had no effect on IL-1ß mRNA expression in LPS + IFN-γ-stimulated THP-1 Mφ.
Ro 11-1464 may suppress IL-6 and CCL2 mRNA expression by BRD4 inhibition or by the apoA-I-mediated blunting of TLR4 and INF-γ signaling pathways. Thus, silencing of BRD4 results in the reduced secretion of IL-6 in LPS-stimulated Mφ [72,79]. Endogenous apoA-I decreases the expression of LPS-induced pro-inflammatory genes through a selective dampening of TLR4 signaling in Mφ [80]. ApoA-I also decreases CCL2 mRNA, CCL2 protein synthesis, and secretion in a dose-dependent manner by the reduction in nuclear NF-kB-p65 subunit [81].

HDAC Inhibitor SAHA
SAHA (vorinostat) is a potent (IC 50 < 86 nM [82]) class I and class II HDAC inhibitor that displays anti-inflammatory properties detected in human primary Mφ [83] and in rodent models of inflammation [84]. It is a synthetic compound that was approved by the Food and Drug Administration (FDA) for the treatment of cutaneous T-cell lymphoma (CTCL) in the United States of America in 2006 [85].
This study is the first report to show that SAHA regulates the mRNA expression of CEBPD in general, and that of IL-6, CCL2, and IL-1ß in THP-1 Mφ. SAHA reduced the mRNA expression of IL-6 and CCL2, but slightly upregulated that of IL-1ß, demonstrating its simultaneous anti-and pro-inflammatory action in LPS + IFN-γ-stimulated THP-1 Mφ.
IL-6 is a target gene of NF-kB [87], while IL-1ß promoter contains a functional NF-kB binding site [74], and IL-1ß mRNA generation depends on the non-impaired function of STAT3 in LPS-stimulated Mφ [88]. Thus, SAHA may affect the gene expression of IL-6 and IL-1ß by the control of acetylation-sensitive transcription factors, such as NF-kB [89,90] and STAT3 [91].
The Lys122 and Lys123 acetylation of NF-kB reduces the DNA binding affinity of the p65 NF-kB subunit, resulting in NF-kB nuclear export [89]. In LPS-stimulated THP-1 cells, SAHA reduced the nuclear accumulation of NF-kB accompanied by the suppressed secretion of pro-inflammatory cytokines, including IL-6 [92]. SAHA-mediated sustained Lys122 and Lys123 acetylation of NF-kB may, therefore, have resulted in reduced transcription of IL-6.
On the other hand, the HDAC3 deacetylation of acetylated p65 NF-kB subunit causes the IkBα-dependent nuclear export of NF-kB and the suppressed expression of its target genes [90]. The enhanced expression of IL-1ß may result from SAHA-mediated HDAC3 inhibition that can restore NF-kB nuclear localization.
STAT3 acetylation by CBP/p300 enhances its DNA binding and transactivation, while STAT3 deacetylation by HDAC1-3 inhibits the transcription of its target genes [93]. The sustained STAT3 acetylation may, therefore, result in an enhanced STAT3 transcriptional activity leading to the upregulation of IL-1ß mRNA transcription.
CCL2 gene expression is activated by the interaction of HDAC3 with SP1 and c-Jun [94] and HDAC11 with PU.1 [95]. SAHA may downregulate CCL2 gene expression by HDAC3 and HDAC11 inhibition, preventing their interaction with the relevant TFs.

HDAC Inhibitor TSA
TSA is a natural compound from Streptomyces spp., which displays not only antifungal, antibiotic, but also HDAC inhibitory activity in mammalian cells [96]. TSA functions as a potent (IC 50 of 1.9-2.9 nM [97]) class I and class II HDAC inhibitor [98] causing the altered expression of various genes in the context of inflammation [99] and cancer [100].
This study is the first report of TSA to regulate gene the expression of CEBPD, IL-6, CCL2, and IL-1ß in THP-1 Mφ. TSA suppressed the mRNA expression of IL-6 and CCL2, but slightly upregulated that of IL-1ß, demonstrating anti-and pro-inflammatory action in LPS + IFN-γ-stimulated THP-1 Mφ.
The anti-inflammatory effect of TSA is consistent with the literature, as it also reduced pro-inflammatory gene expression and cytokine secretion in murine BMDMs [101,102]. TSA inhibits IFN-γ production as well as increases LPS-depressed acetylation and decreased the LPS-induced phosphorylation of NF-kB-p65 in U-937 cells [103]. TSA also suppresses NF-kB-p65 DNA binding activity and TLR4 protein expression in LPS-activated RAW264.7 [104]. The suppressed expression of IL-6 may, therefore, result from the TSAmediated prevention of TLR4-dependent NF-kB-p65 binding and interruption of IFN-γ signaling.
In Mφ, PU.1 functions as a master TF, activating CCL2 expression [102] and regulating Mφ differentiation at mRNA and protein levels [105]. TSA may downregulate CCL2 mRNA expression by the suppression of PU.1 gene expression, as seen in multiple murine Mφ cell lines, including RAW264 [105].
As the IL-1ß gene displays an SP1 TF binding site in its promoter [74], TSA may upregulate IL-1ß mRNA expression by the regulation of the SP1 TF acetylation status. Thus, TSA increases SP1 acetylation at Lys703, leading to the elevation of its DNA binding activity to target promoters, as seen in human Jurkat T cells [106].

The Role of C/EBPδ TF in Gene Expression Regulation
C/EBPδ TF acts as an integrator of cellular responses in a cell context-dependent manner and is believed to be regulated mainly at the level of CEBPD gene transcription initiation. Thus, the modulation by the hit compounds of C/EBPδ TF may contribute to the regulation of IL-6, CCL2, and IL-1ß gene expression in inflammatory-activated THP-1 Mφ.
C/EBPδ TF can display activatory or inhibitory transcriptional activity, resulting in the up-or downregulation of its target genes [7]. While C/EBPδ TF can directly target IL-6 [10] and IL-1ß [29,74], it participates in the pro-inflammatory induction of CCL2 [107]. However, there is no experimental confirmation of direct C/EBPδ TF binding to the CCL2 promoter.
On the other hand, the HDAC inhibitor-mediated abolition of CEBPD gene expression may also contribute to the suppressed expression of IL-6 and CCL2. Thus, C/EBPδ-null Mφ display a reduced induction of IL-6 in response to several TLR ligands [6,114,115] and decreased IL-6 plasma levels contributing to a reduction in endotoxin-induced systemic inflammation in C/EBPδ-null mice [116].
In this study, we showed that epigenetically active BET and HDAC inhibitors are able to modulate CEBPD gene expression and to elicit an anti-inflammatory effect in LPS + IFN-γ-activated THP-1-derived Mφ. Macrophage C/EBPδ has been proposed to be a possible target for the treatment of inflammation-driven rheumatoid arthritis [29]. The anti-inflammatory effects of BET and HDAC inhibitors are widely suggested to be novel approaches to the therapy of inflammation-linked or -driven disorders, including atherosclerosis, multiple sclerosis, or sepsis [117,118]. However, further research is needed to determine which mechanisms of action (MOA) of each of the four identified hits might be the most promising for potential therapeutics for epigenetic CEBPD modulation in vivo.

Cloning
The empty and CMV::SEAP vectors were cloned to characterize the enzymatic chemiluminescent SEAP assay. The empty vector, containing enhanced green fluorescent protein (eGFP) reporter that was linked via T2A self-cleavage peptide to a puromycin resistance gene (G2P), was generated first. To do so, the G2P insert was amplified via PCR from a pX335A-G2P template and ligated into the BsaBI and BstBI (NEB, Ipswich, MA, USA) double-digested pcDNA3.1(-) backbone following a heat-shock transformation in SCS110 bacteria (Agilent, Santa Clara, CA, USA) and positive clone selection using 100 µg/mL ampicillin (Carl Roth, Karlsruhe, Germany). A CMV::SEAP vector was generated using the empty vector as a backbone. An insert encoding SEAP reporter (GenBank acc# U89937) was generated via XhoI and HindIII-HF (NEB) restriction digestion of SEAP-myc encoding plasmid and ligated into the correspondingly double-digested empty vector following heat-shock transformation in DH-5α bacteria and positive clone selection using 100 µg/mL ampicillin (Carl Roth).
The CEBPD::SEAP vector was cloned to generate THP-1 reporter cells for HTS. An insert encoding the CEBPD promoter was generated via PCR from the bacterial artificial chromosome (BAC) clone CH17-293N3 (BACPAC services, CHORI [119], Emeryville, CA, USA) that encodes a 200,000 bp fragment of the human chromosome 8 containing the CEBPD gene. The insert encoding the SEAP reporter was generated via PCR using the CMV::SEAP vector as a template. The second PCR combined two inserts to the CEBPD::SEAP construct that was double digested by Acc65I and XhoI (NEB), following its ligation into the correspondingly digested pSEW-eGFP backbone and positive clone selection using 100 µg/mL ampicillin (Carl Roth).
The HEK293T and THP-1 reporter cells generated were tested for mycoplasma contamination using a mycoplasma detection kit (Lonza, Basel, Switzerland) after cell sorting.

Generation of HEK293T Reporter Cell Lines
Wild type HEK293T cells were transfected with the transfection reagent Lipofectamine ® 2000 (Invitrogen™, Thermo Fischer Scientific, Oberhausen, Germany) according to the manufacturer's protocol. Lipofectamine-plasmid DNA complexes (2.5 µg DNA and 9 µL Lipofectamine 2000 per well) were added to wild type HEK293T cells, in 6-well format with 450,000 cells per well, 48 h after cell seeding. The transfected HEK293T cells were cultured in selective DMEM GlutaMAX™ medium (Thermo Fisher Scientific) containing 10 µg/mL puromycin (Carl Roth, Karlsruhe, Germany), 24 h post-transfection. Stable transfected HEK293T cells expressing eGFP were selected via fluorescence-activated cell sorting (FACS, FACSAria TM Cell Sorter, BD Biosciences, San Jose, CA, USA) approximately three weeks after transfection.

Generation of THP-1 Reporter Cell Line
The CEBPD::SEAP THP-1 reporter cell line was generated via viral transduction. Viral particles, containing CEBPD::SEAP vector, were produced using HEK293T wild type cells, transfected with jetPRIME ® transfection reagent (Polyplus transfection ® , New York, NY, USA) in 10 cm 2 -dishes with 2 million cells per dish. A single reaction with 0.5 mL volume comprised 20 µL jetPRIME ® transfection reagent, 2 µg vector DNA, 1.5 µg PSPAX2 plasmid, 0.5 µg PMD2.G plasmid, and JetPrime ® reaction buffer ad 500 µL. The day after transfection, cellular supernatant of transfected HEK293T cells was collected and centrifuged at 500× g for 5 min. Next, viral particles were concentrated by incubation of sterile-filtered viruscontaining supernatant with a polyethylene glycol (PEG) virus precipitation kit (Abcam, Cambridge, England, United Kingdom) according to the manufacturer's protocol, at 4 • C for 24 h. The concentrated viral particles were obtained by centrifugation at 3200× g for 30 min at 4 • C. The virus-enriched pellet was aliquoted and stored at −80 • C. A fresh aliquot of viral particles was added to 1 million wild type THP-1 cells and incubated at 37 • C in a humidified atmosphere of 5% CO 2 and 95% air for 24 h. Virally transduced cells were cultured for approximately two weeks before positive eGFP-expressing cell clones were selected via FACS (FACSAria TM Cell Sorter, BD Biosciences, San Jose, CA, USA).

In Vitro PMA-Induced Differentiation
Phorbol 12-myristate 13-acetate (PMA) was purchased from Sigma Aldrich (St. Louis, MO, USA). THP-1 wild type and THP-1 reporter cells were differentiated to Mφ by treatment with 50 ng/mL PMA for 48 h with an additional recovery step in fresh PMA-free cell culture medium for 72 h, as described previously [120]. During HTS, THP-1 reporter cells were PMA-differentiated in a bulk format in T175 flasks with 10 million cells per flask in 30-45 mL of cell culture medium.

Reagents and Compound Libraries
The list of pharmacologically active compounds (LOPAC ® ) library contains 1280 experimentally validated small molecules that belong to diverse chemical classes. The ENZO ® SCREEN-WELL ® FDA-approved drug library V2 contains 774 FDA drug compounds belonging to various indication classes. LOPAC ® and ENZO ® compound libraries were kindly provided by Fraunhofer ITMP in Hamburg. Screening compounds were maintained in DMSO at 10 mM and stored on 384-well plates (Echo Qualified 384-well plates, Labcyte, San Jose, CA, USA) at −20 • C. The composition of the LOPAC ® and ENZO ® libraries is shown in supplementary Table S1 and Table S2, respectively. I-BET151, Ro 11-1464, trichostatin A (TSA), and suberoylanilide hydroxamic acid (SAHA) were purchased from Sigma Aldrich (St. Louis, MO, USA). Stock solutions of I-BET151 (10 mM), Ro 11-1464 (10 mM), TSA (6.6 mM), and SAHA (10 mM) were prepared in DMSO (Sigma Aldrich) and stored at −20 • C. Each of the conditioned media was freshly prepared before treatment.

High-Throughput Screening
HTS for CEBPD-modulating compounds was performed at Fraunhofer ITMP in Hamburg. First, THP-1 reporter cells were differentiated to Mφ in a bulk format for five days, detached using Accutase ® solution (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at 37 • C, centrifuged at 300× g for 5 min, and seeded in a 384-well format with 10,000 cells per well in 50 µL RPMI 1640 cell culture medium without phenol red (Thermo Fisher Scientific, Oberhausen, Germany). The next day, cell culture medium was robotically removed, and 20 µL of fresh cell culture medium was manually added. PMA-differentiated THP-1-derived Mφ were pre-treated by screening compounds (end concentration of 10 µM), TSA control compound (end concentration of 0.5 µM), and DMSO solvent control (0.001% v/v), which were added robotically (Echo 550 Liquid Handler, Beckman Coulter, Inc., Brea, CA, USA), for 1 h at 37 • C. Next, pre-treated cells were stimulated with 20 µL of M1 conditioned medium (0.1 µg/mL LPS + 20 ng/mL IFN-γ) and incubated in a humidified atmosphere of 5% CO 2 and 95% air for 24 h. The final concentrations of screening compounds and controls were restored after the addition of the M1 conditioned medium. The next day, cellular supernatants were robotically (JANUS Mini Platform, Perkin Elmer, Waltham, MA, USA) collected for a chemiluminescent SEAP assay (performed in 384-well format). LOPAC ® and ENZO ® libraries were screened twice (read 1 and read 2) in a 384-well format. Read 1 was conducted in phenol-red containing medium, and read 2 was conducted in phenol-red-free medium, resulting in higher net signal values. During read 2, the remaining cells were used for the cell viability assay using the CellTiter-Glo ® Kit (Promega, Madison, WI, USA), which was performed according to the manufacturer s protocol. In short, 100 µL of freshly prepared CellTiter-Glo ® reagent was added to the cells, mixed for 2 min on an orbital shaker, and incubated for 10 min at room temperature. The luminescence was recorded on an EnSpire plate reader with 0.25 s/well integration time.

SEAP Assay
The SEAP assay protocol was conducted according to the Phospha-Light™ SEAP Kit (Applied Biosystems™, Thermo Fischer Scientific, Oberhausen, Germany). Cellular supernatants from wild type and reporter cells were collected 24 h after cell seeding (for stable transfected HEK293T cells) or 24 h after LPS + IFN-γ-treatment (for THP-1 reporter cells). Cellular supernatants were heat-inactivated at 65 • C for 30 min, cooled to room temperature, and diluted 1:1 (v/v) in a supplied dilution buffer. For assay in a 96-well format used during SEAP assay characterization, a single reaction mixture composed of 50 µL of assay buffer and 50 µL of a diluted sample was incubated for 5 min at room temperature. The enzymatic reaction was started by addition of 50 µL of a substrate buffer. For assay in a 384-well format used during HTS, a single reaction mixture composed of 10 µL of assay buffer and 10 µL of diluted sample was incubated for 5 min at room temperature. The enzymatic reaction was started by addition of 10 µL of a substrate buffer. Samples were measured at least in technical triplicates. The chemiluminescent SEAP signal was monitored using the multimode Plate Reader EnSpire plus (0.1 s/well) plate reader (Perkin Elmer, Waltham, MA, USA) for up to 40 min immediately after substrate buffer addition.

SEAP Assay Characterization Parameters
The chemiluminescent SEAP assay was characterized according to assay linearity, sensitivity, and variability in a 96-well format using non-and serially diluted cellular supernatants of CMV::SEAP-expressing HEK293T reporter cells. The intra-assay coefficient of variation was calculated as follows: CV = 100 * (standard deviation technical replicates / mean technical replicates ). The inter-assay coefficient of variation was calculated as follows: CV = 100 * (standard deviation biological replicates / mean biological replicates ). Z -factor values were calculated using cellular supernatants of CMV::SEAP-expressing (sample) and empty vector-expressing (negative control) HEK293T reporter cells. The Z -factor was calculated as follows: Z = 1 − [3 * (standard deviation sample + standard deviation negative control ) / absolute value (mean sample − mean negative control )] [50].

qRT-PCR Analysis
For gene expression analysis, THP-1 reporter cells were PMA-differentiated in a bulk format, seeded in a 24-well format with 120,000 cells per well in 0.5 mL of RPMI 1640 cell culture medium without phenol red, and treated as described for HTS. Total RNA was collected 4 h after stimulation with 0.1 µg/mL LPS + 20 ng/mL IFN-γ and extracted using an RNeasy Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer's protocol. Lysis buffer was supplemented with 1:100 (v/v) ß-mercaptoethanol (Applichem, Darmstadt, Germany), and sample DNA was digested on-column using an RNase-free DNase kit (Qiagen). Up to 1 µg of the total RNA was reverse transcribed using a First Strand cDNA Synthesis Kit (Thermo Scientific, Waltham, MA, USA) with random primers according to the manufacturer s instructions. The qRT-PCR analysis was carried out in a 10-µL reaction mixture containing 10 ng reverse-transcribed RNA, SYBR ® Select Master Mix (Thermo Fischer Scientific, Oberhausen, Germany), and 0.35-0.5 µM of gene-specific primer pair (Microsynth, Balgach, Switzerland). The primer pairs used in this study are listed in Table 5. Table 5. Sequences of primers used for quantitative reverse transcriptase polymerase chain reaction. The reaction protocol was as follows: initial incubation at 95 • C for 10 min; denaturation at 95 • C for 15 s; annealing and extension at 60 • C for 1 min and 40 cycles. In each PCR run, the melting curve analysis was performed to ensure the homogeneity of the PCR product. The qRT-PCR analysis was conducted in a 384-well format, at least in technical duplicates on an Applied Biosystems TM QuantStudio TM 12K Flex system (Thermo Fischer Scientific, Waltham, MA, USA). The mRNA expression levels of selected genes were determined using the comparative Ct method (∆∆Ct) and normalized to an internal reference gene ribosomal protein L37a (RPL37A) [121]. Fold changes in the test group (M1 DMSO or M1 compound) compared to the M0 control group were calculated as 2 −∆∆Ct , where ∆∆Ct = ∆Ct mean (test) − ∆Ct mean (M0 control) = [Ct mean (gene, test) − Ct mean (reference, test)] − [Ct mean (gene, M0 control) − Ct mean (reference, M0 control)].

Live-Cell Imaging
Phase contrast (transmitted light) and fluorescent pictures of CEBPD::SEAP-expressing THP-1 reporter cells were taken by the ZOE Fluorescent Cell Imager (Bio-Rad, Hercules, CA, USA).

Conclusions
In conclusion, this study identified the BET inhibitors I-BET151 and Ro 11-1464 as well as the HDAC inhibitors SAHA and TSA from LOPAC ® and ENZO ® libraries as potent CEBPD-modulating compounds in LPS-and IFN-γ-stimulated THP-1 reporter Mφ. The identified hit compounds also showed pronounced and compound-specific effects on the expression of IL-6, IL-1ß, and CCL2 inflammation-linked genes. The results suggest that CEBPD modulation results in an anti-inflammatory effect in THP-1 Mφ and can be considered as an approach to the treatment of inflammation. Further studies are needed to characterize the compound-specific mechanism of action on CEBPD expression regulation and the clinical efficacy of epigenetic CEBPD-modulation in vivo.