Sirtuin 3 relieves inflammatory responses elicited by lipopolysaccharide via the PGC1α-NFκB pathway in bovine mammary epithelial cells

Excessive inflammation in bovine mammary endothelial cells (BMEC) due to mastitis leads to disease progression and eventual culling of cattle. Sirtuin 3 (SIRT3), a mitochondrial deacetylase, downregulates pro-inflammatory cytokines in BMEC exposed to high concentrations of nonesterified fatty acids by blunting nuclear factor-κB (NFκB) signaling. In nonruminants, SIRT3 is under the control of PGC1α, a transcriptional cofactor. Specific aims were to study (1) the effect of SIRT3 on inflammatory responses of lipopolysaccharide (LPS)-challenged bovine mammary epithelial cells (bovine mammary alveolar cells-T, MAC-T) models, and (2) the role of PGC1α in the attenuation of NFκB signaling via SIRT3. To address these objectives, first, MAC-T cells were incubated in triplicate with 0, 50, 100, 150, or 200 μg/mL LPS (derived from Escherichia coli O55:B5) for 12 h with or without a 2-h incubation of the NFκB inhibitor ammonium pyrrolidine dithio-carbamate (APDC, 10 μ M ). Second, SIRT3 was over-expressed using adenoviral expression (Ad-SIRT3) at different multiplicity of infection (MOI) for 6 h followed by a 12 h incubation with 150 μg/mL LPS. Third, cells were treated with the PGC1α agonist ZLN005 (10 μg/ mL) for 24 h and then challenged with 150 μg/mL LPS for 12 h. Fourth, cells were initially treated with the PGC1α inhibitor SR-18292 (100 μ M ) for 6 h followed by a 6-h culture with or without 50 MOI Ad-SIRT3 and a challenge with 150 μg/mL LPS for 12 h. Data were analyzed using one-way ANOVA with subsequent Bon-ferroni correction. Linear and quadratic contrasts were used to determine dose-responses to LPS. There were linear and quadratic effects of LPS dosage on cell viability. Incubation with 150 and 200 μg/mL LPS for 12 h decreased cell viability to 78.6 and 34.9%, respectively. Compared with controls, expression of IL1B , IL6 , and TNFA was upregulated by 5.2-, 5.9, and 2.7-fold with 150 μg/mL LPS; concentrations of IL-1β, IL-6, and tumor necrosis factor-α (TNF-α) in cell medium also increased. Compared with the LPS group, LPS+APDC increased cell viability and reversed the upregulation of IL1B , IL6 , and TNFA expression. However, mRNA and protein abundance of SIRT3 decreased linearly with increasing LPS dose. Ad-SIRT3 infection (50 MOI) reduced IL1B , IL6 , and TNFA expression and also their concentrations in cell medium, and decreased pNFκB P65/NFκB P65 ratio and nuclear abundance of NFκB P65. The PGC1α agonist increased SIRT3 expression, whereas it decreased cytokine expression, pNFκB P65/NFκB P65 ratio, and prevented NFκB P65 nuclear translocation. Contrary to the agonist, the PGC1α inhibitor had opposite effects, and elevated the concentrations of IL-1β, IL-6, and TNF-α in cell medium. Overall, data suggested that SIRT3 activity could attenuate LPS-induced inflammatory responses in mammary cells via alterations in the PGC1α-NFκB pathway. As such, there may be potential benefits for targeting SIRT3 in vivo to help prevent or alleviate negative effects of mastitis.


INTRODUCTION
Mastitis is one of the most frequent and destructive diseases in the dairy industry globally (De Vliegher et al., 2012). It reduces milk yield and quality and compromises health status and welfare. Escherichia coli is among the most predominant pathogens that causes clinical mastitis (Green et al., 2004) and severe inflammatory reactions that often culminate in sepsis (Burvenich et al., 2007). Lipopolysaccharide, a major coliform endotoxin, is commonly used in studies with bovine mammary epithelial cells (BMEC) to study the pathophysiology of mastitis in vitro (Loor et al., Sirtuin 3 relieves inflammatory responses elicited by lipopolysaccharide via the PGC1α-NFκB pathway in bovine mammary epithelial cells 2011). Upon LPS challenge, BMEC activate signaling pathways and trigger production and secretion of pro-inflammatory cytokines such as tumor necrosis factor-α (TNF-α), IL-1β, and IL-6. These cytokines will, in turn, amplify the inflammatory response by means of autocrine or paracrine mechanisms with clinical severity being assessed through changes in concentrations of these pro-inflammatory cytokines (Blum et al., 2000). The control of TNF-α, IL1-β, and IL-6 synthesis is regulated via nuclear factor-κB (NFκB) signaling (Hirano, 2021). In this regard, blocking or attenuating signaling through the NFκB pathway may bring about relief from excessive inflammatory symptoms due to mastitis.
In nonruminants, Sirtuin 3 (SIRT3), a mitochondrial NAD + -dependent deacetylase, is a key regulator of cellular energy metabolism and redox homeostasis (Cao et al., 2022). For instance, its activity can help attenuate lipolysis and protect cells from oxidative damage, but high concentrations of nonesterified fatty acids (NEFA) can suppress its expression (Zeng et al., 2019). Such an effect helps explain the marked downregulation of SIRT3 expression in the liver and mammary gland of cows with fatty liver and ketosis, underscoring its potential role in the onset and progression of these diseases. The fact that SIRT3 attenuated the overactivation of NFκB signaling in NEFA-challenged BMEC in vitro suggested its activity can elicit a protective effect against inflammation (Liu et al., 2021a).
Peroxisome proliferator-activated receptor-γ coactivator 1α (PGC1α), a transcriptional coactivator, is a key regulator of mitochondrial biogenesis and function in nonruminants (Koh and Kim, 2021). As a mitochondrial protein, SIRT3 expression is subject to control by PGC1α activity. Recent data have demonstrated that PGC1α also participates in the regulation of inflammatory responses via crosstalk with NFκB (Álvarez-Guardia et al., 2010). Thus, it is likely that PGC1α participates in the SIRT3-mediated suppression of NFκB signaling in BMEC exposed to LPS. We hypothesized that increases in SIRT3 activity would reduce inflammatory responses elicited by LPS through alterations of the PGC1α-NFκB pathway. We tested this hypothesis in LPS-challenged bovine mammary alveolar cells (MAC-T).

MAC-T Culture
We followed the cell culture protocol described in one of our previous studies (Sun et al., 2020). The MAC-T cells were thawed and cultured in Dulbecco's modified Eagle's medium (DMEM). The DMEM consisted of Nutrient Mixture F-12 medium (BL305A, Biosharp), 10% fetal bovine serum (FBS, FB15015, Clark Bioscience), and 1% (vol/vol) antibiotic-antimycotic mix (SV30010, penicillin 100 U/mL, streptomycin 100 μg/ mL; Hyclone). The concentration of glucose in DMEM was 17.505 mM. No acetate was included in the incubation medium. Cells were cultured at 37°C under 5% CO 2 , and further immunoblotted with antibodies against Cytokeratin 18 (10830-1-AP, Proteintech), an epithelial cell marker, for identification purpose. Then after 2 to 3 passages, cells were seeded into a cell culture plate in DMEM supplemented with 10% FBS and the antibiotic-antimycotic mix. Cell medium was replaced every 24 h.

LPS Treatment and Cell Viability Assay
Cell viability assay was done according to instructions in a commercial cell counting kit (CCK-8, CA1210, Solarbio). The CCK-8 kit contained a WST-8 reagent. The WST-8 would be transformed into orange-yellow colored formazan by mitochondrial dehydrogenases in alive cells only. The number of alive cells would be proportional to the orange-yellow color. The MAC-T cells were seeded into a 96-well plate at a density of 5 × 10 3 cells per well and cultured for 24 h. Cells were then incubated with 0, 50, 100, 150, or 200 μg/mL LPS (E. coli O55:B5, L8880, Solarbio; LPS was dissolved in cell medium) for 12 h (n = 3). Afterward, 10 μL of WST-8 reagent was added for a 3-h incubation. Absorbance was measured at 450 nm. Cell viability was set at 100% for the blank control. Intra-assay and interassay coefficients of variation (CV) of the CCK-8 kit was ≤5 and <10%, respectively.

Incubation with NFκB Inhibitor
The MAC-T cells were seeded at 1 × 10 6 into 6-well plates and allowed to grow to about 80% confluence. Cells were then treated with 10 μM NFκB inhibitor ammonium pyrrolidine dithiocarbamate (APDC, HY-18738, MedChemExpress) for 2 h, followed by incubation with 150 μg/mL LPS for 12 h (n = 3). Cells were then collected for further analysis.

Adenovirus Infection
The MAC-T cells were infected with SIRT3 overexpression adenovirus (Ad-SIRT3). Both Ad-SIRT3 and the control adenovirus vector (Ad-GFP) were products from Hanbio as in our previous study (Liu et al., 2021a). The titer of adenovirus before administration was >1 × 10 10 pfu/mL. The MAC-T were seeded into cell culture flasks (25 cm 2 , Corning) at a density of 1.5 × 10 6 cells. When cell confluence reached 50%, MAC-T cells were infected with Ad-SIRT3/Ad-GFP at 25 multiplicity of infection (MOI), 50 MOI, and 100 MOI for 6 h in basic DMEM (no FBS supplementation). Cells were then incubated in complete DMEM with 10% FBS for another 30 h. Subsequently, they were subject to LPS treatment (150 μg/mL) for 12 h (n = 3). Cells were collected by trypsin digestion and stored at −80°C for further use within 1 wk.

Incubation with PGC1α Inhibitor
At about 50% confluence, cells were first incubated with PGC1α inhibitor SR-18292 (100 μM, HY-101491, MedChemExpress) for 6 h, then with or without 50 MOI Ad-SIRT3 infection for 6 h, followed by treatment with 150 μg/mL LPS for 12 h (n = 3). Cells were harvested by trypsin digestion and stored for further analysis.

Quantitative Reverse Transcription PCR Assay
The minimum information for publication of quantitative real-time PCR experiment (MIQE) guidelines were followed for qPCR assays (http: / / rdml .org/ miqe .html). The RNA from MAC-T cells was extracted using Trizol (15596026, Invitrogen) according to the supplier's instructions. RNA concentration and quality were determined using K5500 MicroSpectrophotometer (Beijing Kaiao Technology Development Ltd.) and electrophoresis (1% agarose gels). The OD260/OD280 ratio of the RNA extracts ranged from 1.8 to 2.0. RNA (1 μg) was reverse-transcribed to cDNA using an Evo M-MLV Mix Kit with gDNA Clean for qPCR (AG11728, Accurate Biology) in accordance with the manufacturer's instructions. The mRNA abundance was detected using an SYBR Green Premix Pro Taq HS qPCR Kit (AG11728, Accurate Biology) with a 7500 Real-Time PCR System (Applied Biosystems). The reaction system contained 10 μL of 2 × SYBR Green Pro Taq HS Premix, 0.4 μL of forward primer (final concentration at 0.2 μM), 0.4 μL of reverse primer (final concentration at 0.2 μM), 1 μL of cDNA templates, and 7 μL of RNase Free dH 2 O. Conditions were as follows: initial denaturation at 94°C for 30 s, 40 cycles of amplification (denaturation at 94°C for 5 s, annealing at 60°C for 30 s, and extension at 60°C for 1 min), and extension at 72°C for 5 min. Primer sets are listed in Supplemental Table S1 (https: / / figshare .com/ articles/ figure/ Supplementary _material _to _JDS2022 _22114 _pdf/ 21563778; Liu, 2022), and were synthesized by the Beijing Genomics Institute. The cycles-to-threshold values of GAPDH and ACTB did not change upon different treatments (Supplemental Figure S1, https: / / figshare .com/ articles/ figure/ Supplementary _material _to _JDS2022 _22114 _pdf/ 21563778; Liu, 2022). Thus, the relative expression of target genes was normalized to GAPDH and ACTB. Relative expression of target genes was determined by the 2 −ΔΔCT method. The qRT-PCR experiments were performed in triplicate.

Statistical Analysis
All data are expressed as the means ± standard error of the mean (SEM). Statistical analyses were done using GraphPad Prism 8 (GraphPad InStat Software). Linear and quadratic contrasts were run to determine the dose-response of increasing LPS on SIRT3 mRNA protein abundance, cell viability and expression of IL1B, IL6, and TNFA. Data from qRT-PCR and western blotting (WB) were normally distributed and analyzed using the one-way ANOVA with subsequent Bonferroni correction. A P < 0.05 was significant.

Cell Viability and Production of Pro-inflammatory Cytokines in LPS-challenged MAC-T Cells
Immunostaining indicated the cells used in the present study were pure mammary epithelial cells (Supplemental Figure S2, https: / / figshare .com/ articles/ figure/ Supplementary _material _to _JDS2022 _22114 _pdf/ 21563778; Liu, 2022). Cell viability was in both linear and quadratic manner with 50 to 200 μg/mL LPS treatment for 12 h ( Figure 1A, P < 0.001). A dosage of 50 or 100 μg/mL LPS increased cell viability compared with blank controls. Of note, a dosage of 150 μg/mL LPS led to a cell viability of 78.6 ± 3.0%. Compared with blanks, LPS incubation increased IL1B, IL6, and TNFA expression in a linear manner ( Figure 1B-D, P < 0.001). Consistent with results of mRNA abundance, concentrations of IL-1β, IL-6, and TNF-α in cell culture supernatant were linearly increased with LPS dosage ( Figure 1E-G, P < 0.001). A dosage of 150 μg/ mL LPS increased concentrations of IL-1β, IL-6, and TNF-α to 11.1 ± 0.2, 3.6 ± 0.2, and 31.6 ± 0.6 ng/mL, respectively. This dosage of LPS resulted in peak concentrations of these pro-inflammatory cytokines. Given the moderate cell death during obvious inflammatory responses, a dosage of 150 μg/mL LPS was chosen for subsequent experiments.

Effect of SIRT3 in Synthesis of Pro-inflammatory Cytokines and NFκB Signaling
SIRT3 mRNA expression was linearly downregulated with increasing LPS dosage ( Figure 3A, P < 0.001). Consistent with this, protein abundance of SIRT3 was downregulated with increasing LPS dosage ( Figure 3B and C, linear P < 0.001). Our approach to induce SIRT3 overexpression in MAC-T cells with Ad-SIRT3 infection was successful ( Figure 3D, P < 0.001). Infection with Ad-SIRT3 downregulated mRNA expression of IL1B, IL6, and TNFA compared with Ad-GFP infection (Figure 3E-G, P < 0.001). Western blot analysis indicated that the pNFκB P65 band was lighter in Ad-SIRT3 infection group compared with the Ad-GFP infection group ( Figure 3H). There was a significant decrease of pNFκB P65/NFκB P65 ratio upon Ad-SIRT3 infection in comparison with the Ad-GFP infection ( Figure 3J, P < 0.001). In addition, Ad-SIRT3 infection at different MOI revealed a decreased pattern of pNFκB P65 levels compared with Ad-GFP infection ( Figure 3J). Compared with the Ad-GFP + LPS group, Ad-SIRT3 + LPS treatment reduced concentrations of IL-1β, IL-6, and TNF-α in cell culture supernatant ( Figure 3K-M, P < 0.001).
As the immunofluorescence analysis indicated, LPS treatment induced the nuclear translocation of NFκB P65 compared with blank controls, whereas Ad-SIRT3 + LPS treatment suppressed the nuclear translocation of NFκB P65 compared with Ad-GFP + LPS ( Figure  4A). Quantification of fluorescence intensity of NFκB P65 confirmed the lower cellular content of NFκB P65 in Ad-SIRT3 + LPS compared with Ad-GFP + LPS ( Figure 4B, P < 0.001).

DISCUSSION
Sirtuin 3 antagonizes inflammatory responses resulting from high NEFA through attenuation of NFκB signaling in BMEC (Liu et al., 2021a). It remains unclear what roles, if any, SIRT3 plays on LPS-induced inflammatory responses in BMEC. The upstream events that drive the blockage of NFκB signaling by SIRT3 also are unknown. The present study revealed that LPS treatment suppressed SIRT3 expression, and SIRT3 overexpression downregulated expression of pro-inflammatory  cytokines via blocking NFκB signaling. In addition, PGC1α was capable to control the deactivation of NFκB signaling by SIRT3. Thus, the results support our hypothesis highlighting a role of SIRT3 on the attenuation of LPS-induced inflammatory responses in MAC-T via the PGC1α-NFκB axis.
Challenge of BMEC with LPS is the most common tool to study inflammatory responses to a mastitis pathogen in vitro. Our results indicated that a low dosage of LPS promoted cell proliferation, whereas a high dosage reduced cell viability. However, the starting LPS dosage that led to an obvious reduction in cell viability in the present study was higher compared with previous studies (Sun et al., 2015;Liu et al., 2021b). It is possible that differences in the source of LPS used across studies accounts for the apparent discrepancies. The typical LPS contains a tripartite structure comprised of lipid A, core oligosaccharide, and O antigen (Caroff and Karibian, 2003). Modifications of these structures, which happens frequently in gram-negative bacteria, results in different cellular responses as well as different immunomodulatory properties. For example, LPS treatment caused upregulation of pro-inflammatory cytokine expression, such as IL1B, IL6, and TNFA (Silva et al., 2018;Sun et al., 2020;Xia et al., 2020). Concentrations of these pro-inflammatory cytokines increased in cell medium as well. We confirmed these effects in the present study. Upregulation of synthesis and secretion of these pro-inflammatory cytokines would further bring about severe consequences, for instance, pyroptosis and apoptosis. Of note, compared with blank controls, an LPS dosage of 200 μg/mL did not evoke significant upregulation of IL1B, IL6, and TNFA expression (P > 0.05). Furthermore, IL-1β, IL-6, and TNF-α concentrations in cell medium only increased slightly. This might have been due to the fact that LPS caused severe cell death at this dosage. Because 150 μg/mL LPS gave rise to moderate cell death and marked increase in pro-inflammatory cytokines synthesis, this dosage was chosen for subsequent assays.
Synthesis of pro-inflammatory cytokines is under the control of the NFκB pathway (Oeckinghaus and Ghosh, 2009). In a resting state, the NFκB complex, most commonly the P50-P65 heterodimer, is secluded in the cytoplasm via protein-protein interactions with its inhibitor IκBα. Upon exposure to LPS or other stimuli, the NFκB dimers dissociate from IκBα and translocate into the nucleus (Oeckinghaus and Ghosh, 2009). They act as a transcription factor and modulate expression of inflammatory cytokines such as TNFA, IL6, and IL1B. The inducible phosphorylation of NFκB P65 is critical for transcriptional activity. Thus, the observation that LPS induced phosphorylation of NFκB P65, along with the blunted response in LPS challenged MAC-T on pro-inflammatory cytokine synthesis and cell viability when cells were incubated with ADPC, confirmed that NFκB signaling is vital for LPS-induced inflammatory injury in BMEC. These data also confirmed a previous transcriptomics study in which NFκB signaling was the most prominent pathway in LPS-treated primary BMEC .
Given the crucial role of NFκB signaling on LPSinduced inflammatory injury, attenuation of this pathway may result in a therapeutic effect (Khan et al., 2020). Despite the suppression of SIRT3 expression at high NEFA levels, we previously showed that SIRT3 inhibited activation of NFκB in NEFA-challenged BMEC (Liu et al., 2021a). Thus, restoration of SIRT3 activity might be a therapy. The overexpression of SIRT3 by Ad-SIRT3 infection confirmed that it can attenuate NFκB activation in LPS-challenged BMEC. Targeting NFκB via SIRT3 has been studied in multiple tissues of rodents and in human cells. For example, honokiol, a SIRT3 agonist, resulted in reduction of NFκB P65 phosphorylation in an experimental murine kidney injury model (Quan et al., 2020). The SIRT3 blockage via 3-TYP increased p-NFκB P65/NFκB P65, and inhibited αCyperone-related attenuation on depressive behaviors in mice subjected to chronic mild stress (Xia et al., 2020). In human fibroblast-like synoviocytes, SIRT3 overexpression decreased p-NFκB/NFκB, whereas SIRT3 knock down caused an elevation of p-NFκB/ NFκB (Bao et al., 2020). Those authors demonstrated that SIRT3 directly binds with NFκB and suppresses NFκB activation (Bao et al., 2020). In rat H9c2 cells, activation of SIRT3 with resveratrol induced the translocation of NFκB P65 from the nucleus to the cytoplasm and protected cells from oxidative stress-mediated cell death . Restoration of SIRT3 activity further reduced pro-inflammatory cytokine synthesis and ameliorated cell viability in LPS-treated MAC-T. Our results suggested that SIRT3 has the potential to be a therapeutic target for cow mastitis. In fact, there have been applications of SIRT3 on inflammatory diseases in rodent models. For instance, SIRT3 diminished inflammatory injury in lungs (Kurundkar et al., 2019), vascular tissues (Dikalova et al., 2020), cardiac tissues (Palomer et al., 2020), macrophages , proximal tubular cells (Koyama et al., 2011), and liver (Chen et al., 2021).
The fact that the PGC1α agonist promoted mRNA and protein abundance of SIRT3, and the PGC1α inhibitor caused the opposite effect highlighted that SIRT3 expression is regulated by PGC1α in BMEC. This was further supported by the decrease in phosphorylation of NFκB P65 and its nuclear translocation in response to the PGC1α agonist. Direct evidence that deactivation of NFκB signaling by SIRT3 was mediated by PGC1α arose from data with the PGC1α inhibitor demonstrating decreased SIRT3 mRNA and protein abundance. These effects were accompanied by overactivation of NFκB signaling, such as increased phosphorylation of NFκB P65 and its nuclear translocation.
Previous work with rodents indicated that PGC1α activity is closely related to inflammatory processes. For instance, PGC1α levels were negatively associated with severity of inflammation (Fontecha-Barriuso et al., 2019;Pérez et al., 2019), and it regulated levels of inflammatory cytokines in murine kidney and pancreas. Consistent with this, dairy cows with fatty liver had lower PGC1α protein abundance while experiencing a robust inflammatory response . In nonruminant models, PGC1α regulates synthesis of pro-inflammatory cytokines via physical interaction between PGC1α and the P65 NFκB subunit (Rius-Perez et al., 2020). In that way, PGC1α prevents the transcriptional activity of NFκB toward its target genes, including those encoding proinflammatory cytokines, such as constitutive binding of PGC1α to P65 and pP65 prevented upregulation of IL-6 synthesis (Pérez et al., 2019). Because NFκB is a redox-responsive transcription factor, PGC1α-dependent transcriptional regulation of antioxidant genes is likely to modulate NFκB activation during inflammation. Hence, oxidative stress is considered to mediate the PGC1α-NFκB interaction (Rius-Perez et al., 2020).
Activity of SIRT3 can affect oxidative stress through deacetylation of antioxidant enzymes such as super-oxide dismutase 2 and catalase (Singh et al., 2018). These enzymes are responsible for quenching cellular reactive oxygen species (ROS), and elevated levels of ROS will upregulate SIRT3 transcription, thus, may enhance deacetylation of SOD2 and serve as a negative feedback loop for controlling oxidative stress (Merksamer et al., 2013). Given its role in modulation of oxidative stress, the fact that SIRT3 is involved in the interaction between PGC1α and NFκB underscores its potential role in the regulation of inflammatory responses. However, cautions should be taken when using these in vitro data for interpretation of the actual in-cow situation because cell culture system cannot perfectly mimic the in vivo conditions. For instance, the glucose concentration in our DMEM was much higher than that in blood of cows. Using high glucose was to stimulate the cell proliferation, and in turn the high glucose would affect the lactogenesis in MAC-T, such as inhibiting the synthesis of casein (Heo et al., 2017). Also, no acetate was included in this DMEM while blood acetate represents a major substrate for de novo fatty acid synthesis in mammary glands of dairy cows (Zhao et al., 2021). Acetates were shown to inhibit NFκB activation in BMEC challenged with the mastitis pathogen Staphylococcus aureus by suppressing (Wei et al., 2017). Futures studies are essential to measure SIRT3 abundance in mammary gland tissue in cows with mastitis, as well as the status of the PGC1α-NFκB axis.

CONCLUSIONS
The present study revealed that LPS suppressed SIRT3 expression, and SIRT3 restoration alleviated inflammatory responses elicited by LPS via the PGC1α-NFκB pathway in mammary cells. These results highlight a functional role for SIRT3 on the attenuation of LPS-induced inflammatory responses, providing a basis for further in vivo investigations on the alleviation of mastitis by targeting SIRT3.