Abstract

   PGC1α‐Related Coactivator (PRC) is a transcriptional coactivator
   promoting cytokine expression in vitro in response to mitochondrial
   injury and oxidative stress, however, its physiological role has
   remained elusive. Herein we investigate aspects of the immune response
   function of PRC, first in an in vivo thioacetamide (TAA)‐induced mouse
   model of drug‐induced liver injury (DILI), and subsequently in vitro in
   human monocytes, HepG2, and dendritic (DC) cells. TAA treatment
   resulted in the dose‐dependent induction of PRC mRNA and protein, both
   of which were shown to correlate with liver injury markers. Conversely,
   an adenovirus‐mediated knockdown of PRC attenuated this response,
   thereby reducing hepatic cytokine mRNA expression and monocyte
   infiltration. Subsequent in vitro studies with conditioned media from
   HepG2 cells overexpressing PRC, activated human monocytes and
   monocyte‐derived DC, demonstrated up to 20% elevated expression of
   CD86, CD40, and HLA‐DR. Similarly, siRNA‐mediated knockdown of PRC
   abolished this response in oligomycin stressed HepG2 cells. A putative
   mechanism was suggested by the co‐immunoprecipitation of Signal
   Transducer and Activator of Transcription 1 (STAT1) with PRC, and
   induction of a STAT‐dependent reporter. Furthermore, PRC co‐activated
   an NF‐κB‐dependent reporter, indicating interaction with known major
   inflammatory factors. In summary, our study indicates PRC as a novel
   factor modulating inflammation in DILI.

   Keywords: cytokine expression, drug‐induced liver injury, hepatic
   inflammation, PPRC1, PRC
     __________________________________________________________________

Abbreviations

   ALT
          alanine aminotransferase

   AST
          aspartate aminotransferase

   DC
          dendritic cells

   DILI
          drug‐induced liver injury

   NF‐κB
          nuclear factor kappa‐light‐chain‐enhancer of activated B cells

   PRC
          PGC1α‐related coactivator

   STAT1
          signal transducer and activator of transcription 1

   TAA
          thioacetamide

   TASO
          thioacetamide S‐oxide

1. INTRODUCTION

   PGC1α‐Related Coactivator (PRC, PPRC1), together with PGC1α and PGC1β
   (Peroxisome proliferator‐activated receptor gamma coactivator 1‐alpha
   and beta), forms a small family of transcriptional coactivators which
   control mitochondrial biogenesis and energy expenditure.

   In general, transcriptional coactivators regulate broad networks of
   genes[46] ^1 by activating the gene expression through interactions
   with nuclear receptors.[47] ^2 Indeed, PRC was originally identified as
   a transcriptional coactivator linking cellular proliferation to
   mitochondrial biogenesis.[48] ^3 Later it was demonstrated that
   oxidative stress in U2OS cells induces PRC, which subsequently
   coactivates the expression of pro‐inflammatory cytokines.[49] ^4 ,
   [50]^5 However, PRC function in vivo has not been well described, as
   all mouse homozygous knockouts die during the preimplantation
   stage.[51] ^6

   Inflammation of the liver is a manifestation of a variety of diseases,
   including viral infection, nonalcoholic steatohepatitis, and
   drug‐induced liver injury (DILI). The pro‐inflammatory cascade that
   follows hepatocyte death, whereby immune cells react to the apoptotic
   bodies and to the cytokines released from moribund hepatocytes, is a
   hallmark of this condition. The degree of immune system activation, for
   example, the number and type of infiltrating cells, can influence the
   severity of the injury as monocytes and neutrophils are rapidly
   recruited to the injury site and could either exacerbate the ongoing
   tissue damage or alternatively drive the resolution of liver
   injury.[52] ^7 , [53]^8

   An interrupted immune response modifies DILI. For instance, mouse
   knockout of the Ccl2 gene, which is a potent monocyte chemoattractant,
   or of the Tnfα receptor (Tnfr1/55), protects mice from carbon
   tetrachloride‐(CCl[4])induced liver injury.[54] ^9 , [55]^10 In
   contrast, acetaminophen‐induced hepatotoxicity was exacerbated in
   Tnfr1/55 ‐/‐ mice.[56] ^11 Such differences indicate that the
   beneficial or detrimental effect of immune system activation in DILI is
   multifactorial, such as the dose, genetic background of the in vivo
   model, and the mechanism of action of the toxicant.

   Oxidative stress is a common feature both in disease and DILI. In fact,
   a disproportionate number of drugs that have been removed from the
   market due to DILI have been shown to interfere with some aspect of
   mitochondria,[57] ^12 which is a primary intracellular source of
   reactive oxygen species (ROS). Similarly, oxidative stress is a
   characteristic response to several hepatic toxicants, including:
   Thioacetamide (TAA,14), Methapyrilene,[58] ^13 Diclofenac,[59] ^14
   CCl[4],[60] ^15 Bromobenzene,[61] ^16 2‐Acetylaminofluorene (2‐AAF,[62]
   ^17 ), Lipopolysacharide (LPS,[63] ^18 ), and Aflatoxin B1.[64] ^19
   Indeed, interrogation of the public transcriptomics rat liver dataset
   from the Toxicogenomics Informatics Project (TG‐GATE,
   [65]http://toxico.nibiohn.go.jp [66]^20 ), revealed that PRC mRNA
   induction in vivo, especially within the first 24 hours, correlates
   with treatment with the aforementioned compounds (Figure [67]1A,B).
   Surprisingly, acetaminophen treatment in these rat studies did not
   result in any significant increase in PRC mRNA expression.

FIGURE 1.

   FIGURE 1
   [68]Open in a new tab

   Prc mRNA expression profiles in rats and mice treated with hepatic
   toxicants and effect of adenovirus‐mediated Prc knockdown on liver
   injury in TAA‐treated mice. (A) rat and (B) mouse hepatic Prc mRNA
   microarray expression profiles in animals treated with hepatic
   toxicants as indicated. Each square represents a group treated with
   different dose (mg/kg) and sacrificed at a different time point. Only
   statistically significant induction is marked (red squares,
   FDR ≤ 0.05), see[69] ^20 for experimental details. Fold changes are
   omitted. (C‐G) Male mice were infected with scramble or shPRC
   adenovirus (3 × 10^12 VP/kg) for 48 h and treated with 100 mg/kg TAA
   for 24 h before sacrifice. (C) Plasma AST and ALT levels in animals
   from two independent experiments, one tailed t test, n = 7‐9, *P < .05,
   SD (D) Hematoxylin‐eosin stained representative sections of the liver
   from the animals treated as indicated. (E and F) Western blots of
   hepatic protein probed with indicated antibodies. Vinculin was used as
   a loading control and quantification of Vinculin normalized PRC and
   Cyp2E1 levels are shown. ANOVA, Tukey′s test, n = 3‐4, * denotes
   statistical significance to saline injected scramble‐Ad group while #
   relates to scramble‐Ad or shPRC‐Ad treated with TAA, ***P < .001,
   ##P < .01, SD (G) Histopathological evaluation of livers from two
   independent experiments, scoring was done as described in the Materials
   and methods section. Statistical significance was not calculated

   While it is clear that oxidative stress can activate NF‐κB
   signaling,[70] ^21 the factors initiating the inflammatory cascade in
   hepatocytes remain poorly described. Considering the effect PRC has on
   cytokine expression in response to ROS and mitochondrial damage in U2OS
   cells, we hypothesized that inducible PRC may regulate hepatic
   inflammation in DILI.

   This study provides the first evidence that PRC induced by hepatic
   toxicants and metabolic inhibitors can control cytokine expression and
   activation of the immune system, both in vivo and in vitro.

2. MATERIALS AND METHODS

2.1. Materials

   Thioacetamide (TAA), Oligomycin, DMSO, and CCCP were purchased from
   Sigma‐Aldrich. Vinculin #V9264, STAT1 #SAB4300326, and CYP2E1 HPA009128
   were obtained from Sigma‐Aldrich, while p100/p52 #06‐413 from Millipore
   (Billerica, MA, USA) and p105/p50 #D4P4D from Cell Signaling (Danvers,
   MA, USA). PRC antibody was described previously.[71] ^5 Anti‐mouse:
   CD45 (30‐F11), CD11c (HL3), Ly6c (AL‐21), Ly6g (1A8), MHCII I‐A/I‐E
   (M5/114.15.2), CD19 (ID3), CD8 (53‐6.7), CD4 (RM4‐5), TCRβ (H57.597),
   CD11b (M1/70) were purchased from BD Biosciences and F4/80 (BM8) from
   eBioscience. Anit‐human: CD14‐v500, CD1a‐FITC, HLA‐DR‐APC‐H7,
   CD86‐BV421, CD83‐PE were acquired from BD Biosciences and CD40‐APC from
   eBiosciences.

2.2. TASO synthesis

   Hydrogen peroxide (30% w/w, 580 µL, 6.32 mmol) was added dropwise to a
   solution of TAA (500 mg, 6.66 mmol) in MeOH (10 mL) at 0°C. The
   reaction was stirred at 0°C. NMR showed 89% conversion with respect to
   TAA after 1 h. Subsequently, after 2 hours, the solvent was evaporated
   (the water bath temperature was set to 30°C). The residue was purified
   by automated flash chromatography on a Biotage® KP‐SIL 25 g column. A
   gradient from 10% to 25% of MeOH in EtOAc over 12 CV was used as mobile
   phase. The product was collected using the wavelength 230 nm. The
   pooled fractions were evaporated (the water bath temperature was set to
   30°C). The compound obtained, thioacetamide S‐oxide (521 mg, 86%), was
   found to be a white solid which was stored at –20°C. ^1H NMR (500 MHz,
   DMSO) δ 1.95 (s, 3H), 8.14 (s, 1H), 8.96 (s, 1H). ^13C NMR (126 MHz,
   DMSO) δ 13.8, 189.9.

2.3. Pathology

   Liver samples were fixated in buffered 10% formaldehyde. After
   fixation, sections of the tissues were dissected out and blocked in
   paraffin. From the paraffin blocks 5 micron thick sections were made,
   mounted on glass slides and further deparaffinized and dehydrated
   according to standard protocols. Slides were stained with routine
   hematoxylin‐eosin and cover slipped before histological examination.
   Degrees of hepatocellular hypertrophy, hepatocellular necrosis and
   lobular inflammation were scored semi‐quantitatively on a five‐degree
   scale from 0 (none) to 4 (severe). Digital photomicrographs were
   obtained with an Olympus SC50 camera mounted on a Zeiss Axioskop 40
   microscope.

2.4. Protein arrays and cell culture cytokine measurement

   Cytokine levels in cell culture supernatants from HepG2 cells
   overexpressing PRC for 48 hours were measured with V‐PLEX Cytokine
   Panel (Rockville, MD, USA) or Human CCL5/RANTES Quantikine ELISA Kit
   (R&D Systems, Inc, Minneapolis, MN, USA) according to the
   manufacturer's protocol.

2.5. Differentiation of human monocyte‐derived DC

   Peripheral blood mononuclear cells (PBMCs) from healthy donors were
   separated from peripheral blood by Ficoll‐Paque PLUS gradient (GE
   healthcare) centrifugation. CD14+ monocytes were subsequently isolated
   via negative selection using the Human Monocyte isolation kit II
   (Miltenyi Biotec) according to the manufacturer's protocol. For DC
   differentiation, purified monocytes were incubated in RPMI 1640 medium
   (Life Technologies) supplemented with 10% heat‐inactivated FBS,
   2 mmol/L L‐glutamine, 1 mmol/L pyruvate, 100 U/mL
   penicillin/streptomycin (all from Thermo Fisher Scientific), 100 ng/mL
   human rGM‐CSF, and 40ng rIL‐d on day 7. Subsequently, cells were
   cultured in two time diluted media from HepG2 cells overexpressing PRC
   for 48 hours (RMPI 1640:DMEM 1:1; FCS 10%, P/S 1%). After 24 hours
   cells were collected an expression of selected surface markers measured
   as follow: Fc‐blocked cells (BD Biosciences) were stained with
   appropriate anti‐human antibodies in PBS (2% FCS, 5 mmol/L EDTA). 7‐AAD
   (BD Biosciences) was added to exclude death cells. Samples were
   acquired with FACS Canto and analyzed (FACS Diva software, BD
   Biosciences).

2.6. Plasmids and adenoviral vectors

   Ha and Flag tags (N‐ and C‐terminal end, respectively), were added to
   PRC sequence by PCR reaction using FL‐PRC/pSV‐SPORT[72] ^3 as a
   template and following primers: Fwd.
   5′CTCGAGATGTACCCATACGATGTTCCAGATTACGCTGCGGCGCGCCGGGGACGG3′,
   Rev.5′GGATCCTTACTTATCGTCGTCATCCTTGTAATCCCTCCTGAGGTTCTTCT3′. PCR product
   was cloned into pCDNA3 vector between XhoI and BamH1 sites. PRC‐pcDNA3
   was sequenced and no amino acid substitution, frameshift, or deletions
   were detected. Purified, in vivo grade, Scramble (Ad‐GFP‐U6‐scrmb‐shRNA
   (#1122N)) and shPRC (Ad‐GFP‐U6‐mPPRC1‐shRNA (shADV‐269267))
   adenoviruses were obtained from Vector Biosystems Inc Reporter vector
   pGL4.32[luc2P/NF‐κB‐RE/Hygro] was purchased from Promega and expression
   levels measured with ONE‐Glo™ Luciferase Assay System (Promega). ISRE
   Cignal Reporter system was obtained from Qiagen and expression measured
   with Dual‐Glo® Luciferase Assay System (Promega).

2.7. Next generation sequencing and subsequent analysis

   Quality of total isolated hepatic RNA was tested with Standard
   Sensitivity RNA Analysis Kit (Advanced Analytical Technologies) and
   library prepared with TruSeq® Stranded mRNA HT (Illumina). Libraries
   were validated with Standard Sensitivity NGS Fragment Analysis Kit 1‐
   6,000bp (Advanced Analytical Technologies), pooled in equimolar
   concentrations, diluted and denatured according to Illumina guidelines.
   Samples were sequenced with NextSeq 500/550 using High Output v2 kit
   (Illumina) to average 9 million reads. Base calling were done by
   bcl2fastq from Illumina. Reads were mapped to mouse genome version mm10
   with HiSat2 v.2.0.1‐beta and counts generated using featurecounts
   (version 1.4.4) and TPMs were generated using Sailfish v. 0.7.6.
   Normalized counts were analyzed with Array Studio software (OmicSoft,
   Cary, NC, USA) and raw p‐values and False Discovery Rates (FDR)
   probability calculated. The set of genes with fold changes ≥ ±2 and
   p‐raw ≤0.05 were used for MetaCore pathway enrichment analysis.

2.8. Cell culture, compound treatment, and transfection

   HepG2 (#HB‐8065) cells were obtained from ATCC and cultured in 1 g/L
   glucose DMEM, 5% Fetal Calf Serum (FCS), and % penicillin/streptomycin
   (P/S, Thermo‐Fisher Scientific). Twenty‐four hours after plating and
   counting with NC‐3000 (Chemometec), medium was replaced with DMEM
   (1 g/L glucose, no FCS, no P/S) Supplemented with compounds as
   indicated. After an additional 24 hours cells were collected.
   Alternatively, 24 hours after plating, cells were transfected with
   control endoribonuclease prepared siRNA (esiGFP or esiLUC) or esiPRC
   using Xtreme siRNA transfection reagent (Sigma‐Aldrich) according to
   the manufacturer′s protocol. After 24 hours oligomycin or CCCP were
   added and after additional 24 hours cells and culturing media were
   collected. For immunoprecipitation and reporter experiments cells were
   plated in DMEM medium (1 g/L glucose, 5%FCS and 1%P/S), transfected as
   indicated with Lipofectamine 3000 (Thermo‐Fisher Scientific) and
   collected 48 hours later. pcDNA3 or pGAL vectors were used either as
   controls or to normalize DNA load.

2.9. Animal handling

   Ten‐week‐old female (TAA dose study) or 8‐week‐old male (adenovirus
   experiments) male C57BL/6N mice (Charles River, Germany), were fed with
   regular chow diet and had free access to tap water during all times.
   Temperature was maintained between 21°C and 22°C with relative humidity
   of 50 ± 5%, 20 air changes per hour and standard day/night cycle. Mice
   were housed in non‐sterile open cages (Makrolon III; Scanbur)
   containing hardwood bedding with nesting material (Papyrus, Mölndal,
   Sweden), gnawing sticks, and a shelter (Brogaarden). Female mice were
   group‐housed, while male mice were single‐housed. Animals were
   acclimatized for at least 5 days before the procedure. Mice were not
   fasted before a challenge and all procedures were performed during
   light cycle.

2.10. Animal studies

   All experiments were conducted using protocols approved by the regional
   animal welfare committee (Göteborg Djurförsöksetiska Nämnd) in
   accordance with Swedish regulatory requirements. Mice were randomized
   the day before the experiment and placed in restrainers during tail
   vein injections with Viral Particles (VP). After 48 hours mice were
   intraperitoneally injected with 100 mg/kg TAA or saline and sacrificed
   24 hours later. Alternatively, uninfected mice were intraperitoneally
   injected with different doses of TAA or saline and sacrificed 24 hours
   later. The animals were terminated by exsanguination through cardiac
   bleed under isoflurane (Attane vet 1000 mg/g, Apoteket Sweden) and
   oxygen anesthesia. Blood was collected in Li‐Heparin coated tubes,
   centrifuged and ALT and AST measured with ABX Pentra ALT/AST CP
   reagents (Horiba Abx Sas). Liver samples from caudate, right medial and
   left lateral lobe were snap frozen in liquid nitrogen. Liver samples
   were homogenated with a Liver Dissociation Kit and gentleMACS™
   Dissociator (Miltenyi Biotec) according to the manufacturer′s protocol.
   Isolated cells were resuspended in PBS (2% FBS/5 mmol/L EDTA),
   incubated with the FcR blocking antibody (2.4G2) for 5 minutes,
   followed by the Live/Dead Fixable Aqua Dead Cell Stain Kit
   (Thermo‐Fisher Scientific). Cells were stained with appropriate
   antibodies (Online materials and methods), acquired with BD LSRFortessa
   flow cytometer and analyzed with FlowJo (Tree Star) software. For
   histopathology studies, liver samples were fixated in buffered 10%
   formaldehyde. After fixation, sections of the tissues were dissected
   out and blocked in paraffin. From the paraffin blocks 5 μm thick
   sections were made, mounted on glass slides and further deparaffinized
   and dehydrated according to standard protocols. Slides were stained
   with routine hematoxylin‐eosin and cover slipped before histological
   examination. Degrees of hepatocellular hypertrophy, hepatocellular
   necrosis, and lobular inflammation were scored semi‐quantitatively on a
   five‐degree scale from 0 (none) to 4 (severe). Digital photomicrographs
   were obtained with an Olympus SC50 (Tokyo, Japan) camera mounted on a
   Zeiss Axioskop 40 (Darmstadt, Germany) microscope.

2.11. Protein isolation and western blots

   Cells were lysed with RIPA or IP buffer (Sigma‐Aldrich). Liver samples
   were homogenized in RIPA with 5mm Stainless steel beads and TissueLyze
   homogenizer (Qiagen). Nuclear and cytosolic fractions were isolated
   with CellLytic Nuclear Extraction kit (Sigma‐Aldrich).
   Immunoprecipitation was performed with ANTI‐FLAG M2 Affinity Gel
   (Sigma‐Aldrich) according to the manufacturer's protocol.
   Alternatively, Protein G immunoprecipitation kit (Sigma‐Aldrich) and
   appropriate antibodies (p100/p52 or p105/p50) were used. All buffers
   were supplemented with Complete ULTRA proteasome inhibitor
   (Sigma‐Aldrich). Protein concentration was measured with Pierce BCA
   Protein Assay Kit (Thermo‐Fisher Scientific). After isolation, proteins
   were separated on 4%‐15% SDS‐PAGE gel (Bio‐Rad). Native PRC was
   transferred onto 0.2 µmol/L nitrocellulose membranes (Bio‐Rad) using
   Hoefer chamber (Holliston, MA, USA, 400 mA (constant), 73 minutes, Tris
   Glycine, 0.2% SDS). TransBlot Turbo system (Bio‐Rad) was used in
   experiments with recombinant PRC. Membranes were incubated in TBS
   Odyssey Blocking Buffer (Li‐Cor) and probed with appropriate antibodies
   diluted in blocking buffer Supplemented with 0.01% Tween20
   (Sigma‐Aldrich). Afterward blots were incubated with anti‐rabbit IRDye
   800CW and anti‐mouse IRDye 680RD 600 antibodies (Li‐Cor) diluted
   1:20 000 (0.02% Tween20). Blots were developed using Odyssey CLx
   scanner and analyzed in Image Studio software (Li‐Cor).

3. RESULTS

3.1. PRC knockdown protects from TAA‐induced liver damage in mice

   TAA is a hepatotoxic compound, used as a molecular tool to induce liver
   damage and fibrosis. As PRC mRNA expression was robustly activated by
   TAA in rat liver (Figure [73]1A, File [74]S1, TG‐GATE,
   [75]http://toxico.nibiohn.go.jp [76]^20 ;), we explored in detail a
   mouse model of acute hepatic damage in the context of PRC induction. As
   expected, TAA‐induced Alanine Aminotransferase (ALT) and Aspartate
   Aminotransferase (AST) levels, hepatic cytokine expression, and was
   associated with monocyte and neutrophil infiltration in a
   dose‐dependent manner. These markers of ongoing liver inflammation and
   damage correlated with increased mRNA and protein levels of PRC
   (Figure [77]S1A‐E).

   Hepatocytes are the main target of adenoviral transduction during
   systemic delivery[78] ^22 and are a major source of hepatic
   cytokines.[79] ^23 , [80]^24 To investigate the role of PRC induction
   in response to a TAA insult, we reduced the hepatic PRC mRNA levels
   with an adenoviral vector carrying a specific shRNA sequence. In order
   to distinguish the effects of PRC knockdown from general adenoviral
   infection, we studied the effect of the adenovirus alone,[81] ^25 and
   as expected, elevated expression of Tnfα and Ccl5 were observed in
   infected mice, indicating ongoing inflammation independent of shRNA
   insertion (Figure [82]S1F). However, in contrast to animals with a
   higher viral load (6 × 10^12 VP/kg), treatment with 3 × 10^12 VP/kg had
   no discernible effect on ALT and AST levels, suggesting negligible
   liver injury at this viral load (Figure [83]S1G).

   Subsequently, we investigated the effect of PRC knockdown in
   TAA‐induced liver damage. As indicated by elevated levels of the liver
   injury biomarkers ALT and AST, TAA (100 mg/kg) was indeed hepatotoxic
   (Figure [84]1C), with histopathological evaluation establishing that
   injury was limited to Zone 1 and located around central veins
   (Figure [85]1D). The shPRC‐Ad vector did not significantly affect
   constitutive expression of PRC protein, but clearly attenuated
   TAA‐mediated induction (Figure [86]1E). PRC knockdown modestly reduced
   the pathology scores for hepatic necrosis and inflammation, although
   there was no apparent effect on the hypertrophic response
   (Figure [87]1G). In keeping with the histopathological evaluation, ALT
   and AST levels were lower in animals treated with shPRC‐Ad and TAA
   (Figure [88]1C).

   TAA is bioactivated and metabolized to thioacetamide‐S‐oxide (TASO) by
   Cyp2E1.[89] ^26 Consistent with previous observations (Figure [90]S1A),
   Cyp2E1 expression was reduced by TAA, but not affected by shPRC‐Ad
   (Figure [91]1F), indicating that the observed changes were not related
   to transcriptional downregulation of Cyp2E1 by PRC and subsequent
   inhibition of TASO formation.[92] ^21

   In summary, this set of data indicates that infection with shPRC‐Ad
   attenuates hepatic inflammation and necrosis and indicates that PRC
   contributes to TAA‐induced liver injury.

3.2. PRC knockdown reduces hepatic cytokine expression and leukocyte
infiltration

   To specifically study the effect of PRC knockdown on the immune system
   we measured hepatic cytokine and chemokine expression. The shPRC‐Ad
   infected animals exhibited a reduced expression of several hepatic
   cytokines and chemokines in response to the compound, for example,
   Il1β, Tnfα, Ccl2, Cxcl1, and Cxcl10 mRNA levels were reduced by 55%,
   31%, 37%, 48%, and 46%, respectively (Figure [93]2A). Neither Nf‐κB1 or
   Nf‐κB2 expression (Figure [94]2A), or p50 and p52 protein levels
   (unpublished data) were affected by PRC knockdown.

FIGURE 2.

   FIGURE 2
   [95]Open in a new tab

   Adenovirus‐mediated Prc knockdown attenuates hepatic inflammation in
   TAA‐treated mice. (A) Animals were treated as described in Figure [96]3
   and relative mRNA levels of selected genes were measured with qPCR. B)
   Relative number of monocytes and neutrophils in the liver presented as
   percentage of CD45+ cells. Combined results from two independent
   experiments are shown. ANOVA, Tukey′s test, (A) n = 3‐4, B) n = 7‐9 *
   denotes statistical significance to saline injected scramble‐Ad group
   while # relates to scramble‐Ad or shPRC‐Ad treated with TAA, *P < .05,
   **P < .01, ***P < .001, #P < .05, ## P < .01, ###P < .001, SD

   Furthermore, we investigated the functional aspect of cytokine and
   chemokine regulation by measuring the number of monocytes and
   neutrophils in the liver: In agreement with the mRNA expression
   profile, the shPRC‐Ad reduced both monocyte and (to a lesser degree)
   neutrophil infiltration in TAA‐injured liver (Figure [97]2B, Gating
   profile Figure [98]S2A). However, numbers of eosinophils, dendritic
   cells and CD4+ T lymphocytes were not affected (data not shown).
   Similarly, expression of activation markers on monocytes (CD86, PD‐L1,
   CD40, and CD80) were not altered by shPRC‐Ad (Figure [99]S2B).

   Together, these results indicate PRC involvement in transcriptional
   regulation of inflammatory gene expression in mouse liver and suggest
   that PRC modulates cytokine expression during TAA‐mediated liver injury
   that potentially could affect the monocyte and neutrophil infiltration.

3.3. PRC controls HepG2 cytokine secretion and mediates human DC and monocyte
activation

   In order to study the potential role of PRC in the regulation of
   cytokine expression in human hepatocytes, HepG2 cells were transiently
   transfected with a PRC expression vector for 48 hours and the cell
   culture supernatant analyzed for a panel of cytokines. As demonstrated
   in Figure [100]3A, PRC overexpression induced IL8, IFNγ, TNFα, IL6, and
   IL10 in cell culture supernatants, while robustly reducing CCL5 levels
   (3.3‐fold change). These data indicate that increased PRC expression is
   sufficient to modulate the expression of pro‐ and anti‐inflammatory
   cytokines also in the absence of toxicants.

FIGURE 3.

   FIGURE 3
   [101]Open in a new tab

   PRC controls cytokine expression and dendritic cell (DC) and monocyte
   activation by hepatocytes. (A) Levels of selected cytokines in culture
   supernatant were analyzed using MSD platform (V‐Plex) or ELISA (CCL5)
   in culturing media of HepG2 cells transfected with pGAL or PRC for
   48 h. * denotes statistical significance to pGAL, ANOVA Dunnett's
   multiple comparisons test, n = 5‐8 *P < .05, **P < .01, ***P < .001, SD
   (B) Human monocytes and human monocyte derived dendritic cell were
   cultured for 24 h in supernatants from HepG2 cells overexpressing PRC
   for 48 h. Expression of activation markers: CD86, CD40, HLA‐DR, and
   CD83 was measured with flow cytometry (C) Quantification of Mean
   Fluorescence Intensity (MFI) from panel B. Statistical significance to
   pcDNA3 control samples is shown, Two tailed t‐test, *P < .05,
   **P < .01, ***P < .001, n = 7‐8. For this and all following box and
   whiskers graphs: The box always extends from the 25th to 75th
   percentiles, the whiskers go down to the smallest value and up to the
   largest, the line in the middle of the box is plotted at the median.
   Representative of two independent experiments with cells derived from
   two separate donors is shown. (D) PRC levels in HepG2 cells transfected
   with esiGFP or esiPRC for 24 h and treated with 0.04 µmol/L oligomycin
   for additional 24 h. Normalized quantification of PRC protein is shown
   E) Human monocyte derived dendritic cells and monocytes were cultured
   for 24 h in supernatants from HepG2 cells transfected with esiGFP or
   esiPRC1 for 24 hours and treated for additional 24 h with 0.04 µmol/L
   oligomycin. MFI is shown. Panels D and E: * denotes statistical
   significance to esiGFP control samples. # represents statistical
   significance between esiGFP and esiPRC samples treated with oligomycin.
   ANOVA, Tukey's test, F (n = 3) Representative of two independent
   experiments; panel E (n = 7‐8). Representative results of two
   independent experiments with cells derived from two separate donors is
   shown

   To further demonstrate the physiological consequences of PRC‐dependent
   changes in cytokine secretion in hepatic cells, we isolated monocytes
   from the blood of healthy donors that were differentiated into DC for
   7 days (Figure [102]S3F). Subsequently, a mixture of undifferentiated
   monocytes (CD14^+CD1a^‐) as well as DC (CD14^+CD1a^+) were cultured in
   medium collected from HepG2 cells that had 48 hours prior been
   transfected with either a control vector, or PRC expression vector.
   Both monocytes and DC (Figure [103]3B,C) treated with culture
   supernatants from cells overexpressing PRC displayed higher expression
   of the surface activation markers CD86, CD40 and HLA‐DR, but not CD83.

   Next, we investigated PRC response to TAA and its active
   metabolite––TASO––as well as oligomycin and carbonyl cyanide
   m‐chlorophenyl hydrazine (CCCP). While TAA proved ineffective, and its
   active metabolite TASO demonstrated only limited activation of PRC mRNA
   and protein in HepG2 cells, robust PRC induction was seen with
   oligomycin and CCCP, (Figure [104]S4A‐E) as demonstrated previously in
   U2OS cells.[105] ^5 This observation prompted us to use specific
   mitochondrial inhibitors as PRC inducers in order to investigate the
   direct effect of PRC in hepatocytes.

   Subsequently, we investigated the effect of PRC knockdown in HepG2
   cells cultured either with oligomycin (Figure [106]3D) or CCCP
   (Figure [107]S3G). PRC esiRNA effectively blunted the oligomycin and
   CCCP‐mediated induction of PRC (Figure [108]3D and Figure [109]S3G,
   respectively). Conditioned medium collected from oligomycin‐treated
   HepG2 cells‐induced CD86 expression on both DC and monocytes
   (Figure [110]3D), while medium of CCCP‐treated HepG2 cells only
   affected monocytes (Figure [111]S3H). Importantly, this response was
   attenuated in both monocytes and DC when the PRC induction had been
   blunted by PRC esiRNA. CD83 was reduced by oligomycin and even further
   attenuated by esiPRC but remained unchanged in CCCP samples
   (Figure [112]S3I,J). CD40 and HLA‐DR were not affected, while all
   markers were robustly upregulated in LPS treated samples (unpublished
   data). Taken together, these data indicate that induction of PRC
   expression is necessary for a functional pro‐inflammatory program also
   in human hepatoma cells that can translate into the activation of
   monocytes and DC.

3.4. PRC modulates NF‐κB and STAT1 signaling

   Several pro‐inflammatory pathways converge in NF‐kB signaling,
   prompting us to test whether it is co‐activated by PRC. A PRC
   expression vector was co‐transfected with a luciferase reporter under
   control of tandem NF‐κB elements. PRC‐induced reporter gene expression
   in a dose‐dependent manner consistent with transcriptional
   co‐activation (Figure [113]4A). The NF‐κB pathway includes two routes:
   canonical, mediated by p105/p50 (NF‐κB1), and; noncanonical, mediated
   by p100/p52 (NF‐κB2). Both p105/p50 and p100/p52 must be cleaved to p50
   and p52 and translocated to the nucleus to become transcriptionally
   active. We did not observe increased nuclear translocation of p50 or
   p52 after PRC overexpression, despite modestly elevated levels of p50
   (Figure [114]4B). Likewise, we did not detect p50 or p52 in PRC
   immunoprecipitates (unpublished data). While a direct interaction
   between PRC and elements of NF‐κB pathway could not be demonstrated,
   the data suggests transcriptional co‐activation of NF‐κB.

FIGURE 4.

   FIGURE 4
   [115]Open in a new tab

   PRC modulates NF‐κB and STAT signaling. (A) HepG2 cells were
   co‐transfected with a luciferase reporter gene under the control of
   NF‐κB responsive elements and either pcDNA3 or PRC. Luciferase
   intensity was measured 48 h later. Representative of three independent
   experiments. ANOVA, Tukey's test, n = 11, ***P < .001, ###P < .001, SD.
   Symbols denote statistical significance to respective pcDNA3 control (*
   0.9 pg/cell, # 3.6 pg/cell). (B) HepG2 cells were transfected with
   either pGAL or PRC vectors for 48 h, total cell lysates separated on
   SDS‐gel and expression of selected proteins assessed with indicated
   antibodies. (C) HepG2 cells were co‐transfected with ISRE reporter
   vector together with pcDNA3 or PRC and luciferase expression evaluated
   48 h later. Two tailed t‐test. (D) STAT1 levels in HepG2 cells
   overexpressing PRC for 48 h. (E) PRC was immunoprecipitated from HepG2
   cells overexpressing PRC for 48 h and probed with indicated antibodies.
   (F) RNA expression profiles from GDS4271 data set, Id_ref: Prc
   203737_s_at, NF‐κB1 209239_at, NF‐κB2 209636_at, Stat1 200887_s_at. Two
   tailed t test, ****P < .0001, ***P < .001, *P < .05.Hollow square
   indicates an outlier (ROUT, Q = 5%). Values below graphs represent
   Pearson's correlation measures between Prc expression and selected gene
   in either fibrosis or inflammation category with corresponding
   statistical significance (two‐tailed)

   Owing to PRC’s role as a transcriptional coactivator, we expected it to
   modulate a relatively large set of genes, as such, we performed RNA
   sequencing (RNAseq) of the mouse hepatic RNA isolated from the previous
   in vivo experiments to identify possible PRC‐dependent pathways. As
   expected, expression of several NRF2 target genes (Nqo1, Srx1, Txnrd1)
   were attenuated by shPRC‐Ad (Figure [116]S4A,B), in agreement with the
   proposed PRC function as a regulator of oxidative stress response.
   Interestingly, expression of Stat1 and Stat2 transcriptional factors
   were downregulated by shPRC‐Ad in the RNAseq dataset, a result
   confirmed by qPCR analysis of the same samples (Figure [117]S4C).
   Furthermore, several STAT downstream targets for example, Mx2, Oas1a,
   Isq15, Ifit2 (FDR ≤ 0.05), were downregulated by shPRC‐Ad (File
   [118]S4). Also, expression of Interferon regulatory factor 9 (Irf9),
   which together with Stat1 and Stat2 forms an Interferon‐stimulated gene
   factor 3 (Isgf3) complex, was reduced. In addition, Interferon
   regulatory factor 7 (Irf7) and several Interferon Sensitive Genes
   (ISG), for example, Interferon‐induced protein with tetratricopeptide
   repeats 1 and 44 (Ifit1 and Ifit44), were attenuated by PRC knockdown
   with robust statistical significance (File [119]S4), indicating that
   the coactivator is involved in STAT signaling.

   Next, we investigated the PRC‐STAT1 interaction in more detail in human
   cells. Initially, we co‐transfected HepG2 cells with PRC and a reporter
   vector under the control of Interferon‐Stimulated Response Element
   (ISRE). We observed that PRC induced an ISRE‐mediated response (35%,
   Figure [120]4C), despite, paradoxically, simultaneously reducing STAT1
   levels by 50% (Figure [121]4D). Interestingly, PRC
   co‐immunoprecipitated with STAT1 (Figure [122]4E), strongly suggesting
   a direct interaction between both molecules, even during conditions
   when STAT signaling is not activated.

   As a final point, we searched publicly available GEO datasets for human
   hepatic diseases characterized by significantly increased Prc levels.
   Data from infants with biliary atresia characterized by an inflammatory
   signature, revealed 2.13‐fold higher Prc mRNA expression levels when
   compared to patients with a fibrotic profile (GDS4271[123] ^27 ).
   Moreover, Prc levels correlated positively with elevated NF‐κB1 mRNA
   (r = 0.6956) in patients with the inflammatory molecular profile, and
   to lesser extent with those in the fibrotic category (r = 0.4487,
   Figure [124]4F). NF‐κB2 levels followed a similar pattern in infants
   with liver inflammation (r = 0.4522), while Stat1 expression was
   reduced and correlated negatively with Prc expression (r = −0.4906).

4. DISCUSSION

   In this study we have shown that PRC is involved in stimulating a
   pro‐inflammatory response in hepatic cells during DILI. Our data
   indicates that PRC regulates the in vivo expression of several key
   pro‐inflammatory cytokines, including Tnfα, Il‐1β, and Il16, as well as
   the chemokines Ccl2, Ccl5, Cxcl1, and Cxcl10, responsible for
   attracting or activating neutrophils and monocytes.[125] ^7 , [126]^8
   Consistent with these observations, adenovirus‐mediated PRC knockdown
   attenuated monocyte infiltration and, to a lesser extent, neutrophil
   recruitment to TAA‐damaged liver. Moreover ALT, AST, and
   histopathological changes were reduced by shPRC‐Ad, indicating that the
   cytokine signaling it inhibits ameliorates TAA‐induced necrosis.
   Together, these results suggest that PRC regulates a toxicant‐induced,
   pro‐inflammatory program that aggravates the immune response and might
   contribute to corresponding hepatic response.

   Mouse models do not replicate the human immune system in all aspects,
   for example, mice do not express IL8, which is instead functionally
   replaced by Cxcl1, Cxcl2, and Cxcl5. In addition, a response to TAA in
   vivo is pleiotropic and is not only limited to PRC‐mediated responses.
   Consequently, to establish whether the gene expression response to PRC
   induction was associated with a functional pro‐inflammatory response in
   isolation from toxicants, and in a human context, we performed an
   overexpression experiment in HepG2 cells. Human IL8 is a strong
   neutrophil attractant[127] ^28 that together with IL6, IL10, IFNγ, and
   TNFα, was shown to be robustly induced by PRC‐overexpression in HepG2
   cells. In contrast, CCL5 was strongly reduced in human hepatoma cells,
   with the opposite response in TAA‐treated mice. The variety of
   cytokines regulated by PRC indicate that they share a common, upstream
   transcriptional factor. Our results indicate that PRC activates the
   NF‐κB pathway in human HepG2 cells, while the cytokine response to TAA
   suggests a similar mechanism in the mouse model. This overall finding
   is in contrast with another study describing inhibition of NF‐κB
   signaling by PRC in human umbilical vein endothelial cells.[128] ^29
   However, we were unable to demonstrate a direct interaction between PRC
   and p50 or p52 (transcriptionally functional elements of the NF‐κB
   system), nor demonstrate their enhanced nuclear translocation after PRC
   overexpression, despite elevated levels of p50. As murine p50 or p52
   levels were not affected by PRC knockdown in vivo, a more complex
   mechanism is likely involved that warrants further investigation.

   Interestingly, shPRC‐Ad attenuated the expression of Stat1 and Stat2 in
   mouse liver, both of which regulate expression of several chemokines.
   For example, STAT1 drives the transcription of CXCL10 and CCL2,[129]
   ^30 and regulates CCL5.[130] ^31 We suspect that this reduced
   expression of STAT downstream targets is beneficial in mitigating
   TAA‐induced liver injury. Our in vitro results indicate that PRC
   interacts directly with STAT1 in HepG2 cells and can promote the
   transcription of genes under control of the ISRE element. STAT1,
   together with STAT2 and IRF9, forms an ISGF3 complex which is then only
   transcriptionally active in the context of ISRE.[131] ^32 Incomplete
   activation or missing components of the complex during sterile (ie, not
   virally infected) transfection in HepG2 might explain the modest
   reporter induction observed, however, it does demonstrate that PRC is
   capable of activating ISRE genes independent of interferon signaling or
   viral infection. Interestingly, PRC overexpression reduced STAT1
   levels, indicating a negative feedback loop. Clearly, the role of PRC
   in STAT signaling, especially during viral infection deserves an
   independent study. Nevertheless, our data, together with NF‐κB results,
   strongly suggests that PRC is involved in the regulation of at least
   two major immune signaling pathways. This notion was further
   strengthened by the fact that Stat1, NF‐κB1, and NF‐κB2 expression
   correlated with PRC levels in infants with inflammatory signature of
   biliary atresia.

   Although our results from HepG2 cells and in vivo experiments indicate
   that PRC promotes inflammation, we could not reject a hypothesis that
   reduced levels of CCL5 or other, as yet unidentified, factors may
   counteract or even reverse this effect. Therefore, we evaluated the net
   effect of hepatic cytokines regulated by PRC on the effector immune
   system cells. Both monocytes and DC displayed increased expression of
   surface activation markers: CD86, CD40, and HLA‐DR, indicating that the
   response mediated by PRC is functional beyond hepatocytes and might
   activate lymphoid cells. Loss‐of‐function experiments combined with
   oligomycin and CCCP treatment further support this theory. Overall, our
   results indicate that the inflammatory program regulated by PRC in mice
   has a potential equivalent in humans.

   The exact mechanisms involved in PRC induction are unknown[132] ^33 and
   identifying factors regulating PRC expression, including specific
   nuclear receptors, were not in the scope of this study. It is also
   clear that not all hepatoxicants induce PRC, and that the toxicological
   mechanism of action likely contributes to the changes in PRC expression
   in response to different toxicants. In addition, as the action of the
   immune system differs depending on the model and the drug, as
   highlighted by the opposing liver toxicity outcomes of CCl4 and
   acetaminophen treatment in Trnf1/55 KO mice, the role of PRC is likely
   to be different depending on the context. Nonetheless, mitochondrial
   oxidative stress, induced by for example, CCCP, has previously been
   shown to increase PRC levels in vitro.[133] ^3 , [134]^4 , [135]^5 This
   data has now been replicated in HepG2 cells and supported by further
   results with oligomycin, suggesting a general mechanism. TAA also
   directly inhibits the mitochondrial respiratory chain,[136] ^34 ,
   [137]^35 and, given that all compounds that have been identified to
   induce PRC mRNA expression, also compromise mitochondrial function,
   suggests that in vivo activation of the pro‐inflammatory program
   following PRC induction relates to oxidative stress. However, the
   direct mechanism regulating PRC levels remains to be investigated.

   In conclusion, in this work we investigated the functional aspects of
   PRC and demonstrated that PRC mediates inflammation during DILI in
   mice, and in human hepatoma cells treated with mitochondrial toxicants.
   The activation of human DC and monocytes indicated that the
   inflammatory cascade initiated by PRC is functional in hepatocytes and
   beyond, and could contribute to various liver pathologies.

AUTHOR CONTRIBUTIONS

   M. Buler, U. Andersson designed research; M. Buler, U. Andersson, T.
   Naessens, M. Söderberg, Y. Morias analyzed data; M. Buler, T. Naessens,
   J. Mattsson, Y. Morias, L. Kärrberg, P.Thulin performed research, TS
   Svensson, R. C. Scarpulla contributed new reagents or analytic tools,
   M. Buler, U. Andersson, P. Robbins, B. Glinghammar wrote the paper.

Supporting information

   Fig S1
   [138]Click here for additional data file.^ (1.1MB, pptx)

   Fig S2
   [139]Click here for additional data file.^ (509.3KB, pptx)

   Fig S3
   [140]Click here for additional data file.^ (548.3KB, pptx)

   Fig S4
   [141]Click here for additional data file.^ (319.6KB, pptx)

   Table S1
   [142]Click here for additional data file.^ (154KB, pdf)

   File S1
   [143]Click here for additional data file.^ (36.3KB, xlsx)

   File S2
   [144]Click here for additional data file.^ (1.1MB, docx)

   File S3
   [145]Click here for additional data file.^ (219KB, xls)

   File S4
   [146]Click here for additional data file.^ (1.4MB, xlsx)

   Supplementary Material
   [147]Click here for additional data file.^ (171.6KB, pdf)

   Buler M, Naessens T, Mattsson J, et al. The regulatory role of
   PGC1α-related coactivator in response to drug-induced liver injury.
   FASEB BioAdvances. 2020;2:453–463. 10.1096/fba.2020-00003

REFERENCES