RPMI 1640 medium, penicillin, streptomycin, and puromycin were purchased from Hyclone (Logan, UT), while fetal bovine serum (FBS) was obtained from Atlanta Biologicals (Lawrenceville, GA). Sodium butyrate and bacterial lipopolysaccharide (LPS) from Escherichia coli O55:B5 was procured from MilliporeSigma (St. Louis, MO), and BIX01294, A-366, UNC1999, SGI-1027, 5-azacytidine (AZA), vorinostat, and mocetinostat ( Figure 1 ) were acquired from Cayman Chemical (Ann Arbor, Michigan). Sodium butyrate and LPS were dissolved in RPMI 1640, while all other chemicals were dissolved in dimethyl sulfoxide (DMSO). In all subsequent cell culture experiments, cells treated with an equal volume of RPMI 1640 or DMSO were used as negative controls to the cells treated with individual compounds or their combinations.

Two different chicken macrophage cell lines, HTC (24) and HD11 (25), was kindly provided by Dr. Narayan C. Rath and Dr. Hyun Lillehoj of USDA-ARS, respectively, and maintained in RPMI 1640 containing 10% heat-inactivated FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin and subcultured every 3-4 days. After overnight seeding at 1.5 × 10⁶ cells/well in 6-well cell culture plates, cells were treated with different concentrations of each compound for 24 h or with the indicated concentration of each compound for various lengths of time for dose-response and time-course experiments, respectively. To study the synergy between HMTi/DNMTi and HDACi, different concentrations of an HMTi or DNMTi were added to the cells in the presence or absence of an HDACi for 24 h, followed by RNA isolation and HDP gene expression analysis as described below. HTC cells were also stimulated with butyrate, BIX01294, or their combination for 24 h, followed by 3-h stimulation with 10 ng/mL LPS and subsequent expression analysis.

A stable HTC/AvBD9-luc stable cell line was generated by cloning a 2.0-Kb AvBD9 gene promoter fragment into a lentiviral luciferase reporter vector, pGreenFire1-mCMV-Puro (System Biosciences, Palo Alto, CA) and then transfected into HTC cells as we described (22). The stable cell line was maintained in RPMI 1640 containing 10% heat-inactivated FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 0.5 µg/mL puromycin and subcultured every 3-4 days. To measure transcriptional activation of the AvBD9 gene promoter in response to different epigenetic compounds, cells were seeded at 4 × 10⁴ cells/well in white 96-well tissue culture plates and stimulated in duplicate with different compounds individually or in combination for 24 h. An equal amount of a solvent was applied as a negative control. Luciferase activity was measured using Steady-Glo^® Luciferase Assay System (Promega, Madison, WI) on an L-Max II luminescence microplate reader (Molecular Devices, Sunnyvale, CA). Cell viabilities were assessed by adding alamarBlue Reagent (ThermoFisher, Waltham, MA) to cell culture 4 h before luciferase assay. Relative luciferase activity was determined for each well after normalization to the cell viability as we described (22).

Chicken PBMCs were isolated from EDTA-anticoagulated blood collected from 2- to 4-week-old broiler chickens through gradient centrifugation using Histopaque 1077 (MilliporeSigma) as previously described (22, 26). PBMCs were seeded at 3 × 10⁷ cells/well in 6-well cell culture plates and stimulated with different concentrations of an HMTi or DNMTi with or without 1 mM butyrate for 24 h, followed by RNA isolation and HDP gene expression analysis as described below.

Total RNA was extracted using RNAzol (Molecular Biology Research Center, Cincinnati, OH) according to the manufacturer’s protocol. Gene expression was performed using reverse transcription-quantitative PCR (RT-qPCR) as we described (27–29). Reverse transcription was performed with 300 ng of total RNA and Maxima^® First Strand cDNA Synthesis Kit (ThermoFisher Scientific, Pittsburgh, PA). The cDNA was then diluted 4-fold with RNase-free water prior to real-time PCR using QuantiTect SYBR^® Green PCR Kit (Qiagen, Valencia, CA) in 10-μL reactions using primers specific for HDP, barrier function, and inflammatory cytokine genes as described (27, 29, 30). PCR was run in CFX Connect Real-Time PCR System (Bio-Rad, Hercules, CA) using the following program: 95°C for 10 min, followed by 40 cycles of 94°C for 15 sec, 55°C for 20 sec, and 72°C for 30 sec. The specificity of PCR reactions was confirmed by the melting curve analysis. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene was used as the reference for data normalization. The fold changes in gene expression in the cells treated with individual compounds or in combinations were calculated relative to the cells treated with an equal volume of RPMI 1640 or DMSO using the ΔΔCt method (31).

Data analysis and visualization were implemented in GraphPad Prism (GraphPad Software, La Jolla, CA). Results were presented as means ± standard error of the mean (SEM) and compared among treatments using one-way analysis of variance and post hoc Tukey’s test. Statistical significance was considered if P < 0.05.

We previously identified a number of epigenetic compounds including HDACi, HMTi, and DNMTi that are capable of inducing HDP gene expression using luciferase reporter cell lines driven by HDP gene promoters (21, 22). HDACi are well-known HDP inducers (10, 16). To further evaluate whether newly-identified HMTi (BIX01294 and UNC1999) and a DNMTi (SGI-1027) can induce HDP gene expression in both concentration- and dose-dependent manners, chicken HTC/AvBD9-luc reporter cell line was initially treated with different concentrations of each compound for 24 h. BIX01294 and UNC1999 showed an obvious concentration-dependent activation of the AvBD9 gene promoter in the reporter cell line as indicated by increases in relative luciferase activity, but SGI-1027 was a weak activator ( Figure 2A ). BIX01294 and UNC1999 were further confirmed to induce AvBD9 mRNA expression in chicken HTC macrophage cell line in a concentration-dependent fashion peaking with a 175- and 259- fold increase in response to 20 µM of each compound, respectively, while 10 µM SGI-1027 gave a 51-fold induction ( Figure 2B ). Because it triggered a robust AvBD9 mRNA induction with each compound, 10 µM was used in subsequent time-course experiments. All three compounds also showed a time-dependent induction of AvBD9 mRNA expression in HTC cells peaking at 24 h ( Figure 2C ). Therefore, cells were treated for 24 h for all downstream analyses of HDP gene expression.

Butyrate, an HDACi, is well-known to induce HDP gene expression in different animal species including humans and chickens (26, 32). Sodium butyrate at 2 mM gives optimal HDP gene induction in chicken HTC and HD11 cells and further synergizes strongly with several other HDP-inducing compounds such as forskolin (27, 30), sugars (28), and cyclooxygenase-2 inhibitors (33). To explore a possible synergy between butyrate and an HMTi or DNMTi, we treated HTC/AvBD9-luc reporter cells with different concentrations of BIX01294, UNC1999, or SGI-1027 in the presence or absence of 2 mM butyrate for 24 h. A marked dose-dependent increase in the synergy in luciferase activity was observed between butyrate and BIX01294 ( Figure 3A ) as well as between butyrate and UNC1999 ( Figure 3B ). However, no synergy between butyrate and SGI-1027 was observed in the luciferase reporter cell line ( Figure 3C ). HTC cells were used to confirm whether the synergy occurs between butyrate and an HMTi or DNMTi to induce AvBD9 mRNA expression. A concentration-dependent increase in AvBD9 mRNA expression was indeed observed between butyrate and BIX01294, UNC1999, or SGI-1027. For example, 20 µM BIX01294 and butyrate individually induced AvBD9 expression by 207- and 27-fold, respectively; however, a combination of 20 µM BIX01294 and butyrate gave a remarkable 11,165-fold increase in AvBD9 expression, which is an additional 54-fold increase over 20 µM BIX01294 alone ( Figure 4A ). Similarly, 20 µM UNC1999 enhanced AvBD9 expression by 324-fold, but a UNC1999/butyrate combination gave a 4,470-fold increase, which reflected an additional 14-fold increase over UNC1999 alone ( Figure 4B ). SGI-1027 at 10 µM augmented AvBD9 expression by 51-fold, but further induced AvBD9 by 411-fold when combined with butyrate, which is an additional 8-fold increase over 10 µM SGI-1027 alone ( Figure 4C ). AZA, another well-known DNMTi (34), was also examined and confirmed its synergy with butyrate in a concentration-dependent manner ( Figure 4D ). It is noted that the AvBD9-inducing synergy between butyrate and an HMTi tended to be stronger than between butyrate and a DNMTi.

Moreover, among those HDP genes that are detectable in HTC cells, all were synergistically induced in response to both BIX01294 and butyrate, although different magnitudes of induction were observed with different HDP genes. For example, AvBD2 was only upregulated by approximately 2- to 3-fold by a combination of butyrate and BIX01294, while AvBD4, AvBD8, AvBD10, and AvBD14 were induced by greater than 100-fold in response to butyrate and at least one concentration of BIX01294, with the remaining HDP genes showing an intermediate response ( Figure 5 ).

To confirm whether the HDP-inducing synergy between butyrate and an HMTi is preserved in a different cell line, chicken HD11 macrophages were stimulated with butyrate in the presence or absence of BIX01294 or UNC1999 for 24 h. Consistent with the results in HTC cells, a drastic synergy in AvBD9 induction was observed between butyrate and BIX01294 and between butyrate and UNC1999 ( Figure 6A ). Similarly, AvBD1 ( Figure 6B ), AvBD4 ( Figure 6C ), and AvBD10 ( Figure 6D ) were also synergistically induced albeit at lower magnitudes. On the other hand, CATH2 mRNA was minimally regulated by individual compounds or their combinations in HD11 cells ( Figure 6E ). To further verify a possible HDP-inducing synergy between butyrate and BIX01294 in primary cells, chicken PBMCs were isolated and treated with BIX01294 in the presence or absence of butyrate for 24 h. BIX01294 alone induced AvBD9 in a dose-dependent manner, and the BIX01294/butyrate combination showed an obvious synergy ( Figure 6F ).

To further examine whether other structurally different HDACi also synergize with BIX01294 in AvBD9 induction, chicken HTC cells were treated with different concentrations of BIX01294 with or without an HDACi (vorinostat, butyrate, and mocetinostat) ( Figure 1 ). A dramatic synergy in AvBD9 induction was observed between BIX01294 and any of three HDACi. For example, 5 μM BIX01294 in combination with vorinostat, butyrate, and mocetinostat gave an additional 69-, 26-, and 23-fold increase in AvBD9 gene expression over BIX01294 alone, respectively ( Figure 7A ). In addition to AvBD9, most other HDP genes were also synergistically induced in response to a combination of BIX01294 and any of the three HDACi ( Figure 7B ). To study whether other HMTi also synergize with HDACi in HDP induction, HTC cells were treated with UNC1999 or A-366 ( Figure 1 ) in the presence or absence of butyrate or mocetinostat. Indeed, a marked synergy was seen between any two-compound combination of an HMTi (UNC1999 or A-366) and an HDACi (butyrate or mocetinostat) ( Figure 7C ).

Butyrate was shown to synergize in HDP gene induction with SGI-1027 and AZA ( Figures 4C, D ). We sought to further examine a possible synergy between a DNMTi and a structurally different HDACi. Unsurprisingly, a drastic synergy was observed between mocetinostat and either DNMTi. For example, 5 μM SGI-1027, 5 μM AZA, and 2.5 μM mocetinostat induced AvBD9 gene expression by 4-, 3-, and 152-fold, respectively, while a combination of 5 μM SGI-1027/mocetinostat and 5 μM AZA/mocetinostat gave 747- and 1,925-fold increases in AvBD9 expression ( Figure 7D ).

Butyrate is known to enhance mucosal barrier function by inducing the genes for mucin 2 (MUC2) and tight junction proteins such as claudin 1 (CLDN1) and tight junction protein 1 (TJP1) while suppressing inflammation (35). To examine the impact of the butyrate/BIX01294 combination on barrier function and inflammation, HTC cells were treated with butyrate and BIX01294 individually or in combination for 24 h, followed by a 3-h stimulation with LPS. To our surprise, although BIX01294 had a minimum activity in increasing CLDN1 expression, a marked synergy was observed in response to a combination of BIX01294 and butyrate, and the synergy was further enhanced in response to subsequent LPS treatment ( Figure 8A ). However, MUC2 ( Figure 8B ) and TJP1 (data not shown) were minimally affected in HTC cells by the compounds or their combination with or without LPS, while it is desirable that LPS had no impact on the AvBD9-inducing synergy ( Figure 8C ).

Interestingly, although both butyrate and BIX01294 had no or weak activity in inducing gene expressions of inflammatory cytokines IL-1β ( Figure 8D ) and IL-8 ( Figure 8E ), the combination tended to have a synergistic effect on triggering the synthesis of both cytokines. While butyrate suppressed LPS-induced IL-1β expression, BIX01294 had an opposite effect, and a combination of butyrate and BIX01294 led to a significant reduction in IL-1β expression ( Figure 8D ). Both butyrate and BIX01294 potentiated LPS-induced IL-8 expression individually, and the IL-8 expression level remained elevated in response to the combination ( Figure 8E ). It is important to note that CLDN1 was also synergistically induced by butyrate and UNC1999 or by mocetinostat and UNC1999 ( Figure 8F ), suggesting that a combination of HDACi and HMTi is synergistic in augmenting CLDN1 gene expression; however, MUC2 and TJP1 genes were minimally regulated by HDACi and HMTi (data not shown).