Abstract

Objective

   Peroxiredoxin 1 (PRDX1) is a peroxidase and guards against oxidative
   stress by scavenging intracellular peroxides, whereas it also has been
   shown to stimulate inflammatory response by functioning as a chaperone
   protein. The potential in vivo link between PRDX1's peroxidase activity
   and its pro-inflammatory activity remains elusive.

Methods

   We generated peroxidase-dead PRDX1 variant mice by mutating its
   peroxidatic cysteine at 52 (Cys52) to serine, here referred to as
   PRDX1^Cys52Ser. Trx-TrxR-NADPH coupled activity assay was applied to
   evaluate the peroxidase activity of global PRDX in PRDX1^Cys52Ser
   variant mice. PRDX1^Cys52Ser mice and their wild-type littermates were
   subjected to western diet or methionine and choline deficient diet
   feeding. NASH phenotypes were assessed through different analyses
   including physiological measurements, immunohistochemical staining, and
   quantitative PCR (qPCR). RNA sequencing, qPCR and western blotting were
   used to reveal and validate any changes in the signaling pathways
   responsible for the altered NASH phenotypes observed between WT and
   PRDX1^Cys52Ser variant mice.

Results

   PRDX1^Cys52Ser variant mice showed impaired global PRDX peroxidase
   activity and reduced susceptibility to diet-induced NASH and liver
   fibrosis. Mechanistically, PRDX1 Cys52Ser variant suppressed NF-κB
   signaling and STAT1 signaling pathways that are known to promote
   inflammation and NASH.

Conclusion

   The peroxidatic Cys52 of PRDX1 is required for its pro-inflammatory
   activity in vivo. This study further suggests that PRDX1 may play dual
   but opposing roles in NASH.

   Keywords: PRDX1, NASH, Inflammation, Cys52, NF-κB, STAT1

Highlights

     * •
       A peroxidase-dead PRDX1 variant mouse strain was generated by
       mutating its peroxidatic Cys at 52 (Cys52) to Ser (PRDX1^Cys52Ser).
     * •
       PRDX1^Cys52Ser mice were resistant to WD-induced NASH and liver
       fibrosis.
     * •
       PRDX1^Cys52Ser mice showed significantly reduced NF-kB and STAT1
       signaling activities and their relevant pathways in the liver.
     * •
       PRDX1 Cys52Ser variant also conferred resistance to MCD-induced
       NASH by suppressing NF-kB and STAT1 signaling activities.

1. Introduction

   Non-alcoholic steatohepatitis (NASH) is a more advanced form of
   non-alcoholic fatty liver disease (NAFLD) as the most common chronic
   liver disease worldwide and may progresses to cirrhosis and
   hepatocellular carcinoma [[35]1]. The exact etiology of NASH remains
   complex or even to be debated [[36]2]; however, several prominent
   events have been generally considered to be pathogenically driving
   NASH, which include lipotoxicity, inflammation and oxidative stress
   [[37][1], [38][2], [39][3], [40][4], [41][5]]. Thus, targeting these
   individual events could be potential therapeutic strategies for NASH
   treatment [[42]2].

   Peroxiredoxin 1 (PRDX1) belongs to a mammalian peroxidase family
   composed of six members that scavenge the majority of intracellular
   peroxides [[43][6], [44][7], [45][8], [46][9]]. PRDX1 is a typical
   2-Cys peroxidase containing a conserved cysteine residue at 52 (Cys52,
   known as peroxidatic cysteine) and a resolving cysteine residue at 173
   (Cys173) [[47]9]. When PRDX1 scavenges hydrogen peroxide (H[2]O[2]),
   oxidized Cys52 forms a disulfide-bond with Cys173 from another PRDX1
   unit, consequently forming an inactive homo-dimer that can be further
   reduced back to the active monomer by thioredoxin (Trx) [[48]9,[49]10].
   By this means, PRDX1 continuously exerts its peroxidase function to
   eliminate toxic H[2]O[2].

   A large body of evidence has well established the protective roles of
   PRDX1 through its peroxidase activity in guarding against oxidative
   stress in a variety of human diseases including cancer [[50]11], ageing
   [[51]12], atherosclerosis [[52]13], and liver injury [[53]14].
   Surprisingly, however, PRDX1 has also been reported to promote
   inflammatory response and aggravate liver injury by functioning as a
   secreted chaperone protein [[54]15,[55]16]. Moreover, PRDX family
   members have been shown to initiate post-ischemic inflammation in the
   brain though their chaperon function [[56]17], which is thought to be
   independent of their peroxidase activity [[57]9,[58]18]. Nonetheless,
   one previous in vitro study suggested that PRDX1's pro-inflammatory
   effect is dependent of its redox status [[59]19]. Thus, the
   relationship between PRDX1's peroxidase activity and pro-inflammatory
   activity remains unclear.

   Here, we aimed to investigate a potential connection between PRDX1's
   peroxidase activity and its pro-inflammatory effect by generating
   PRDX1^Cys52Ser variant mice. These variant mice showed impaired
   peroxidase activity as expected and alleviated diet-induced NASH
   phenotypes by suppressing NF-κB signaling and STAT1 signaling pathways.
   Our study provides in vivo evidence that peroxidatic Cys52 of PRDX1 is
   prerequisite for its pro-inflammatory activity and further suggests
   that PRDX1 plays dual but opposing roles in NASH.

2. Methods

2.1. Animals

   All mice used in the study contain C57BL/6J background, are housed at
   an animal facility core with stable temperature (22 °C), humidity and a
   12/12 h light/dark cycle, and have free access to food and water.
   C57BL/6J mice were obtained from GemPharmatech (Nanjing, China).
   PRDX1^Cys52Ser variant mice were generated by GemPharmatech (Nanjing,
   China) through CRISPR–Cas9 technique. The gRNA sequences used for
   generation of PRDX1^Cys52Ser variant mice: CCACAGAAGCGCCAATCACT,
   CCAAGTGATTGGCGCTTCTG.

   Eight-week-old mice (male or female) were fed a western diet (WD,
   Research Diets, D09100310, 40 kcal% Fat, 20 kcal% Fructose and 2%
   Cholesterol) for 20 weeks or a methionine and choline deficient diet
   (MCD, MolDietes, M0421) for 5 weeks to induce nonalcoholic
   steatohepatitis (NASH).

   To evaluate hepatic NF-κB signaling activity, 8-week-old male WT and
   PRDX1^Cys52Ser variant mice on normal chow diet were intraperitoneally
   injected with LPS (10 mg/kg body weight) for 15 min before their livers
   were harvested for WB analysis.

   All animal procedures were approved by the Animal Care and Use
   Committee at Chu Hsien-I Memorial Hospital & Tianjin Institute of
   Endocrinology, Tianjin Medical University. Both male and female mice
   were used for NASH phenotypic and biochemical analyses.

2.2. Primary mouse hepatocyte isolation and culture

   Primary mouse hepatocytes were isolated from the livers of 6-week-old
   male WT and PRDX1^Cys52Ser mice on chow diet as previously described
   [[60]20]. In brief, after anesthesia with isoflurane, the mouse liver
   was perfused sequentially with Krebs buffer containing EGTA and Krebs
   buffer containing CaCl[2] and collagenase type IV through the portal
   vein. The liver was dissected and hepatocytes were collected with
   ice-cold RPMI 1640, followed by filtration through a nylon filter.
   Afterward, primary hepatocytes were washed with ice-cold RPMI 1640 for
   3 times to remove the remaining collagenase, evaluated for their
   viability (usually more than 85%) by trypan blue exclusion test,
   counted and seeded in collagen-coated plates. Primary hepatocytes were
   cultured and maintained in RPMI 1640 medium containing 10% FBS, 50 U/ml
   penicillin and 50 μg/ml streptomycin.

   To evaluate the NF-κB signaling activity, primary mouse hepatocytes
   from WT and PRDX1^Cys52Ser mice were incubated with Hanks' balanced
   salt solution (HBSS) for 30 min, followed by LPS (500 ng/ml)
   stimulation for 15 min. To assess the expression of pro-inflammatory
   genes, primary mouse hepatocytes from WT and PRDX1^Cys52Ser mice were
   cultured in serum-free RPMI 1640 and treated with LPS (100 ng/ml) for
   6 h [[61]21].

2.3. Physiological measurements

   Body weight and food intake measurements were performed as previously
   described [[62]22]. Promethion High-Definition Multiplexed Respirometry
   System (Sable Systems, North Las Vegas, NV, USA) was applied to monitor
   mouse metabolism including oxygen consumption rate (VO[2]), production
   rate of carbon dioxide (CO[2]) (VCO[2]) and locomotion activities.

2.4. Measurement of serum aspartate aminotransferase (AST) and alanine
aminotransferase (ALT)

   30 min after blood collection without addition of anticoagulant, the
   blood samples were centrifuged to collect serum. Serum levels of AST
   and ALT were analyzed with an automatic blood biochemical analyzer
   (AU680, Beckman Coulter).

2.5. Histological staining

   Histological staining of mouse liver samples was prepared as previously
   described [[63]23]. Briefly, mouse liver samples were fixed in
   freshly-made 4% paraformaldehyde (PFA) solution for 48 h, followed by a
   thorough washing with 1×PBS solution. Liver samples were embedded in
   paraffin and then sectioned for H&E, Sirius Red, or α-SMA staining.

   For H&E and Sirius Red staining, specific kits from Solarbio (Beijing,
   China) were used according to the manufacturer's instructions. For
   α-SMA staining, a rabbit primary antibody against α-SMA (14395-1-AP,
   Proteintech), a polymer-HRP anti-rabbit secondary antibody and DAB
   (3,3′-diaminobenzidine) stain (PK10006, Proteintech) were employed.

   For Oil Red O staining (SHBL 1039, Sigma), fresh liver samples were
   immediately frozen and embedded in OCT (SAKURA), and later used for
   preparation of cryosection (8 μm).

   An Olympus microscope equipped with different objective lenses was used
   to capture images.

2.6. Measurement of PRDX peroxidase activity

   The peroxidase activity of global hepatic PRDX was measured as
   previously described [[64]24] with some modifications. In brief, liver
   samples were homogenized in 50 mM HEPES-NaOH buffer (pH 7.0) including
   0.2% Triton X-100 and protease inhibitors. To measure the global PRDX
   peroxidase activity, 200 μM NADPH, 3 μM recombinant Trx, and 1.5 μM
   TrxR were added to 100 μg of protein sample. The reaction was initiated
   by addition of H[2]O[2] (50 μM) at 30 °C, immediately followed by
   detection of absorbance at 340 nm (A[340]) for 15 min. In the meantime,
   the background activity was assessed in liver lysate with only NADPH,
   but without Trx and TrxR.

   For initial rate calculation, a smooth curve was drawn through A[340]
   readings, and the initial NADPH consumption rate (initial rate) was
   calculated by performing a simple linear regression using A[340]
   readings during the first five minutes. The global PRDX peroxidase
   activity was calculated by subtracting the background activity (initial
   rate) from total activity (initial rate).

2.7. Quantitative PCR (qPCR)

   qPCR was performed and quantified as previously described [[65]22]. In
   brief, mouse liver tissues were homogenized in TRIzol (Invitrogen) and
   total RNA was extracted. 1 μg of RNA was used for reverse
   transcription, followed by qPCR using QuantStudio 3 Real-Time PCR
   System (Applied Biosystems). The relative expression levels of interest
   genes were quantified by 2^−ΔΔCt method using 36b4 as the reference
   gene. The qPCR primers used in the study was listed in [66]Table S1.

2.8. Western blotting (WB)

   Protein samples were prepared as previously described [[67]22,[68]25].
   In brief, liver tissues were homogenized in 1× lysis buffer containing
   1% deoxycholic acid, 10 mM Na[4]P[2]O[7], 1% Triton 100, 100 mM NaCl,
   5 mM EDTA, 50 mM Tris–HCl and 0.1% SDS. Protein concentrations were
   determined with a BCA Protein Assay kit (23228; 23224, Thermo Fisher
   Scientific, Rockford, IL, USA). In general, a total of 20 μg of protein
   was loaded in SDS-PAGE and transferred to PVDF membranes. The membranes
   were first blocked with 5% nonfat milk in 1XTBST, and then incubated
   with primary antibodies against PRDX1 (8499, Cell Signaling Technology,
   Danvers, MA, USA), Peroxiredoxin-SO3 (ab16830, abcam), phospho-NF-κB
   p65 (3033, Cell Signaling Technology), NF-κB p65 (8242, Cell Signaling
   Technology), Phospho-STAT1 (9167, Cell Signaling Technology), STAT1
   (9145, Cell Signaling Technology), and Actin (A5441, Sigma) overnight
   at 4 °C, followed by incubation with secondary antibodies for 1 h at
   RT. Finally, an ECL detection system was used to develop signals.

2.9. Measurement of malondialdehyde (MDA) levels

   To measure lipid peroxidation levels, we employed a commercial MDA
   detection kit (Beyotime Biotechnology, China) to measure MDA levels
   following the manufacturer's instructions. Briefly, we mixed liver
   lysate with MDA detection buffer and heated the mixture at 100 °C for
   10 min, followed by the centrifuge at 1000 × g for 10 min. The
   supernatant was collected and then used for detection of absorbance at
   535 nm (A[535]) by a fluorometer (BioTek). Finally, MDA concentrations
   were normalized by protein concentrations.

2.10. HKPerox-Red staining in liver sections

   Liver samples were frozen and embedded in OCT compound (SAKURA) in
   liquid nitrogen. Liver sections were prepared at the thickness of 8 μm
   for HKPerox-Red staining [[69]26]. Briefly, liver sections were
   incubated with 5 μM HKPerox-Red in PBS (0.1% DMF (Macklin) and 100 mM
   CCl[3]CN (Macklin)) for 10 min at room temperature. An Olympus
   fluorescence microscope with 40× objective lens was used to capture
   images. Image J was used to quantify the fluorescence intensity of
   images, and the data were processed by GraphPad Prism 9.

2.11. Transcriptome analysis

   Liver samples were collected and stored in RNAlater (Ambion, TX, USA).
   Total RNA was isolated and purified for transcriptome analysis by
   LC-Bio Technology (Hangzhou, China). In brief, RNA was isolated with
   TRIzol reagent and RNA quality was evaluated with a Bioanalyzer 2100
   (Agilent, CA, USA). cDNA library was created using high-quality RNA. In
   the end, an Illumina Novaseq 6000 was applied to carry out 2x150bp
   paired-end sequencing.

   Bioinformatics analysis was completed by LC-Bio Technology (Hangzhou,
   China). Brief, fastp ([70]https://github.com/OpenGene/fastp), HISAT2
   ([71]https://ccb.jhu.edu/software/hisat2), StringTie
   ([72]https://ccb.jhu.edu/software/stringtie), Gffcompare
   ([73]https://github.com/gpertea/gffcompare/) and R package edgeR
   ([74]https://bioconductor.org/packages/release/bioc/html/edgeR.html)
   were used for various analyses. For Kyoto Encyclopedia of Genes and
   Genomes (KEGG) pathway enrichment analysis and gene-set enrichment
   analysis (GSEA), DAVID software ([75]https://david.ncifcrf.gov/) and
   GSEA4.1.0 software ([76]http://www.gsea-msigdb.org/gsea/index.jsp) were
   used, respectively.

2.12. Quantification and statistical analyses

   All data in this study were presented as means ± SEM. Statistical
   analyses were carried out with unpaired Student's t test (Excel), or
   two-way ANOVA followed by Bonferroni's test (GraphPad Prism 9).
   p < 0.05 was considered statistically significant. We applied GraphPad
   Prism 9 and Adobe Illustrator 2020 to generate and prepare figures.

3. Results

3.1. Generation of peroxidase-dead PRDX1 variant (PRDX1^Cys52Ser) mice

   To investigate the in vivo role of PRDX1's peroxidase activity in the
   control of inflammation, we generated a peroxidase-dead PRDX1 variant
   (PRDX1^Cys52Ser) mouse strain through CRISPR-Cas9 approach by mutating
   the peroxidatic Cys52 to serine residue ([77]Figure 1A) as confirmed by
   DNA sequencing analysis ([78]Figure 1B). Since Cys52 is the key site
   mediating PRDX1's peroxidase activity [[79]9], we assessed the
   peroxidase activity of global PRDX in the liver of PRDX1^Cys52Ser mice
   using a classical Trx-TrxR-NADPH coupled activity assay
   [[80]24,[81]27], where NADPH consumption rates indirectly reflect the
   peroxidase activity of global PRDX. As revealed by this activity assay,
   PRDX1^Cys52Ser variant mice showed significantly reduced global PRDX
   peroxidase activity compared with their wild-type (WT) littermates
   ([82]Figure 1C).

Figure 1.

   [83]Figure 1
   [84]Open in a new tab

   Generation and validation of PRDX1^Cys52Ser variant mice. (A) Schematic
   showing the generation of PRDX1^Cys52Ser variant mice, where the
   peroxidatic cysteine at 52 (Cys52) of PRDX1 was mutated to serine (Ser)
   via CRIPR-Cas9 system. (B) Validation of homozygous PRDX1^Cys52Ser
   variants via mouse genome DNA sequencing. (C) Global hepatic PRDX
   peroxidase activity in WT and PRDX1^Cys52Ser variant mice.
   Eight-week-old male mice on a chow diet were used to measure the
   peroxidase activity of global PRDX via a classical Trx-TrxR-NADPH
   coupled assay. Background activity was measured in samples without Trx
   and TrxR. The initial NADPH consumption rate (initial rate) was
   calculated by performing a simple linear regression using absorbance at
   340 nm (A[340]) readings during the first five minutes. The global PRDX
   peroxidase activity was calculated by subtracting the background
   activity (initial rate) from total activity (initial rate). WT mice
   (n = 6), PRDX1^Cys52Ser mice (n = 5). ∗p < 0.05 by student's t-test.
   (D) Hyperoxidation of PRDX in the various tissues detected by WB. A
   commercial antibody recognizing the hyperoxidized form (SO3) of PRDX
   was used to detect any changes in PRDX hyperoxidation between WT and
   PRDX1^Cys52Ser mice. The arrow denotes the potential hyperoxidized
   PRDX1 according to its molecular weight (22 kDa).

   Some PRDX members (PRDX1-3) could become hyperoxidized at their
   conserved peroxidatic cysteine owing to slow formation of disulfide
   between the peroxidatic cysteine and the resolving cysteine during a
   catalytic cycle [[85]9]. Using a commercial antibody
   (Anti-Peroxiredoxin-SO3, ab16830, abcam) that is supposed to recognize
   the hyperoxidized forms of PRDX [[86]28], we observed that PRDX1
   hyperoxidation was drastically reduced in the liver of PRDX1^Cys52Ser
   mice ([87]Figure 1D). Collectively, these results confirm that
   PRDX1^Cys52Ser variant mice are enzymatically inactive.

3.2. PRDX1^Cys52Ser variant mice were protected from western diet
(WD)-induced NASH and liver fibrosis

   We sought to evaluate the functional role of PRDX1 Cys52Ser in NASH by
   feeding both male and female WT and PRDX1^Cys52Ser mice a WD that is
   rich in fat and fructose as a common strategy to induce NASH
   experimental models [[88]3]. Twenty20 weeks after WD feeding, compared
   with WT mice, PRDX1^Cys52Ser mice showed similar body weight
   ([89]Figs. S1A and S1B), food intake ([90]Fig. S1C) and energy
   expenditure ([91]Fig. S1D). We also found no difference in the ratio of
   either fat mass or lean mass over body weight ([92]Figs. S1E and S1F).
   Interestingly, however, PRDX1^Cys52Ser mice significantly reduced liver
   weight ([93]Figure 2A, B) and the ratio of liver weight over body
   weight ([94]Figs. S1G and S1H).

Figure 2.

   [95]Figure 2
   [96]Open in a new tab

   PRDX1^Cys52Ser variants are protected from WD-induced NASH and liver
   fibrosis. (A and B) Liver weights of male (A) or female (B) WT and
   PRDX1^Cys52Ser mice after WD feeding. Eight-week-old mice were fed a WD
   for a total of 20 weeks before their blood and liver samples were
   collected for subsequent analyses including serum measurement,
   histological staining, WB and qPCR. Male WT mice (n = 9), male
   PRDX1^Cys52Ser mice (n = 7); female WT mice (n = 5), female
   PRDX1^Cys52Ser mice (n = 4). ∗p < 0.05, ∗∗p < 0.01 by Student's t test.
   (C and D) Serum levels of aspartate aminotransferase (AST) and alanine
   aminotransferase (ALT) in male (C) or female (D) WT and PRDX1^Cys52Ser
   mice after WD feeding (as in A and B). Male WT mice (n = 9), male
   PRDX1^Cys52Ser mice (n = 7); female WT mice (n = 5), female
   PRDX1^Cys52Ser mice (n = 4). ∗p < 0.05, ∗∗p < 0.01 by Student's t test.
   (E and F) Representative images from 3 mice per group showing Oil Red O
   staining of liver tissues from male (E) or female (F) WT and
   PRDX1^Cys52Ser mice after WD feeding (as in A and B). Scale bars,
   50 μm. (G and H) Representative images from 3 mice per group showing
   Sirius Red staining of liver tissues from male (G) or female (H) WT and
   PRDX1^Cys52Ser mice after WD feeding (as in A and B). Scale bars,
   50 μm. (I and J) mRNA levels of pro-inflammatory or fibrotic genes in
   the liver of male (I) or female (J) WT and PRDX1^Cys52Ser mice after WD
   feeding (as in A and B). Male WT mice (n = 9), male PRDX1^Cys52Ser mice
   (n = 7); female WT mice (n = 5), female PRDX1^Cys52Ser mice (n = 4).
   ∗p < 0.05, ∗∗p < 0.01 by Student's t test. See also [97]Figs. S1 and
   S2. All data are presented as means ± SEM.

   Serum levels of alanine aminotransferase (ALT) and aspartate
   aminotransferase (AST) that are indicative of liver injury [[98]29],
   were significantly decreased in PRDX1^Cys52Ser mice compared with WT
   mice ([99]Figure 2C, D), suggesting alleviated liver injury in
   PRDX1^Cys52Ser mice. We next carried out a series of histological
   staining in the liver including H&E, Oil-Red-O, Sirius Red and smooth
   muscle alpha-actin (α-SMA) to reflect the severity of NASH and liver
   fibrosis. Clearly, PRDX1^Cys52Ser mice reduced steatosis
   ([100]Figs. S2A and S2B), lipid accumulation ([101]Figure 2E, F), and
   liver fibrosis ([102]Figure 2G, H, [103]Figs. S2C and S2D).

   Consistent with the observed NASH phenotypes, the expression levels of
   some pro-inflammatory (e.g. Mcp-1, Cdllb, Cdllc, Tnf-α, Il-1β) and
   fibrotic (e.g. Pdgfa, Pdgfb, Pdgfra, Col1a1, Col3a1, Icam1) genes were
   significantly reduced in the liver of PRDX1^Cys52Ser mice
   ([104]Figure 2I, J). Interestingly, we observed no differences in the
   levels of hepatic H[2]O[2] ([105]Fig. S2E), as detected by HKPerox-Red,
   a selective H[2]O[2] probe [[106]26], and hepatic malondialdehyde (MDA)
   ([107]Fig. S2F), a marker for lipid peroxidation [[108]4], between male
   WT and PRDX1^Cys52Ser mice after WD, suggesting that PRDX1 peroxidase
   mutation does not alter the extent of oxidative stress in the liver.
   Taken together, these data indicate that PRDX1 Cys52Ser variant
   mitigates WD-induced NASH and liver fibrosis.

3.3. Downregulated NF-κB and JAK-STAT signaling pathways in PRDX1^Cys52Ser
mice

   To understand the signaling pathways accounting for alleviated NASH and
   liver fibrosis in PRDX1^Cys52Ser mice, we performed RNA sequencing
   (RNA-Seq) of liver samples from male WT and PRDX1^Cys52Ser mice after
   WD. Heatmaps showed a significant decrease of genes related to
   inflammation ([109]Fig. S3A), fibrosis ([110]Fig. S3B), and
   specifically to NF-κB signaling ([111]Figure 3A) and JAK-STAT signaling
   ([112]Figure 3B) pathways in PRDX1^Cys52Ser mice. Consistently, both
   gene-set enrichment analysis (GSEA) and Kyoto Encyclopedia of Genes and
   Genomes (KEGG) showed an enrichment of differentially expressed genes
   related to NF-κB signaling or JAK-STAT signaling pathway
   ([113]Figure 3C, D, [114]Fig. S3C). In addition, volcano plots
   indicated the significantly downregulated genes related to NF-κB
   signaling or JAK-STAT signaling pathway in PRDX1^Cys52Ser mice
   ([115]Figs. S3D and S3E).

Figure 3.

   [116]Figure 3
   [117]Open in a new tab

   Reduced NF-κB signaling and STAT1 signaling pathways in the liver of
   PRDX1^Cys52Ser mice after WD feeding. (A) Heatmap showing the
   significantly decreased expression of genes related to NF-κB signaling
   pathway in the liver of male PRDX1^Cys52Ser mice after WD.
   Eight-week-old male WT and PRDX1^Cys52Ser mice were fed a WD for 20
   weeks before their liver tissues were collected for RNA-Seq and
   subsequent bioinformatics analyses. (B) Heatmap showing the
   significantly decreased genes related to JAK-STAT signaling pathway (as
   in A). (C and D) Gene-set enrichment analysis (GSEA) showing an
   enrichment of differentially expressed genes related to NF-κB signaling
   (C) or JAK-STAT signaling (D) pathway. (E) Expression of genes related
   to NF-κB or JAK-STAT signaling pathways in the liver (as in A). (F and
   G) Western blotting and quantitative analysis of hepatic NF-κB and
   STAT1 signaling activities in male (F) or female (G) WT and
   PRDX1^Cys52Ser mice after 20 weeks of WD. n = 4 mice per group.
   ∗p < 0.05, ∗∗p < 0.01 by Student's t test. (H) WB of hepatic NF-κB
   signaling activity in WT and PRDX1^Cys52Ser mice treated with LPS and
   quantification. Eight-week-old male mice on chow diet were
   intraperitoneally injected with LPS (10 mg/kg) for 15 min before their
   liver samples were collected for WB analysis. Veh, vehicle; LPS,
   lipopolysaccharide. WT mice (n = 3 for Veh treatment; n = 4 for LPS
   treatment), PRDX1^Cys52Ser mice (n = 3 for Veh treatment; n = 4 for LPS
   treatment); ∗p < 0.05 by one-way ANOVA followed by Bonferroni's test.
   See also [118]Fig. S3. All data are presented as means ± SEM.

   Using qPCR, we further verified the significantly reduced expression of
   some genes related to NF-κB signaling or JAK-STAT signaling pathway in
   PRDX1^Cys52Ser mice ([119]Figure 3E). Taken together, these data
   provide evidence supporting that NF-κB and JAK-STAT signaling pathways
   were downregulated in the liver of PRDX1^Cys52Ser mice.

3.4. PRDX1 Cys52Ser variant reduced the signaling activity of hepatic NF-κB
and STAT1

   Numerous studies have well established NF-κB as a central player in
   stimulating inflammatory response [[120]30], and one previous study has
   shown that activation of STAT1 and STAT3 promotes obesity-associated
   NASH and hepatocellular carcinoma, respectively [[121]31]. Our RNA-Seq
   and qPCR results have indicated a significant decrease in NF-κB and
   JAK-STAT signaling pathways in PRDX1^Cys52Ser mice. Therefore, we next
   evaluated the signaling activities of NF-κB and STAT1. The
   phosphorylation levels of both NF-κB and STAT1 were significantly
   reduced in the liver of PRDX1^Cys52Ser mice after WD ([122]Figure 3F,
   G).

   We next sought to test for a causal link between PRDX1 Cys52Ser variant
   and reduced inflammation. First, we treated 8-week-old male WT and
   PRDX1^Cys52Ser mice on normal chow diet with lipopolysaccharide (LPS).
   As shown in [123]Figure 3H, LPS-induced NF-κB p65 phosphorylation in
   the liver was significantly reduced in PRDX1^Cys52Ser mice compared
   with WT mice. Second, using primary hepatocytes isolated from chow-fed
   6-week-old male WT and PRDX1^Cys52Ser mice, we observed that
   LPS-induced NF-κB p65 phosphorylation ([124]Fig. S3F) and expression of
   some pro-inflammatory genes (Il-6, Cxcl1 and Mcp-1) ([125]Fig. S3G)
   were significantly reduced in PRDX1^Cys52Ser hepatocytes compared with
   WT hepatocytes. These results collectively demonstrate that PRDX1
   Cys52Ser variant inhibits the NF-κB signaling activity and the
   pro-inflammatory response in the liver.

3.5. PRDX1 Cys52Ser variant conferred resistance to methionine and choline
deficiency diet (MCD)-induced NASH by suppressing NF-kB and STAT1 signaling
pathways

   MCD feeding is another common strategy to induce NASH experimental
   models without obesity and insulin resistance [[126]3]. We tested the
   potential effect of PRDX1 Cys52Ser variant on MCD-induced NASH. After 5
   weeks of MCD feeding, PRDX1^Cys52Ser mice showed comparable body weight
   ([127]Figs. S4A and S4B), liver weight ([128]Figs. S4C and S4D), and
   the ratio of liver weight over body weight ([129]Figs. S4E and S4F),
   but significantly lower serum levels of AST and ALT than WT mice
   ([130]Figure 4A, B).

Figure 4.

   [131]Figure 4
   [132]Open in a new tab

   PRDX1 Cys52Ser variant confers resistance to MCD-induced NASH and liver
   fibrosis by suppressing NF-κB and STAT1 signaling activities. (A and B)
   Serum levels of AST and ALT in male (A) or female (B) WT and
   PRDX1^Cys52Ser mice after MCD feeding. Eight-week-old mice were fed a
   MCD for 5 weeks before their blood and liver samples were collected for
   subsequent analyses. Male WT mice (n = 6), male PRDX1^Cys52Ser mice
   (n = 5); female WT mice (n = 6), female PRDX1^Cys52Ser mice (n = 7).
   ∗p < 0.05, ∗∗p < 0.01 by Student's t test. (C and D) mRNA levels of
   pro-inflammatory or fibrotic genes in the livers of male (C) or female
   (D) WT and PRDX1^Cys52Ser mice (as in A and B). Male WT mice (n = 6),
   male PRDX1^Cys52Ser mice (n = 5); female WT mice (n = 6), female
   PRDX1^Cys52Ser mice (n = 7). ∗p < 0.05, ∗∗p < 0.01 by Student's t test.
   (E and F) Representative images from 3 mice per group showing Oil Red O
   staining of liver tissues from male (E) or female (F) WT and
   PRDX1^Cys52Ser mice (as in A and B). (G and H) Representative images
   from 3 mice per group showing Sirius Red staining of liver tissues from
   male (G) or female (H) WT and PRDX1^Cys52Ser mice (as in A and B).
   Scale bars, 50 μm. (I and J) Activities of hepatic NF-κB and STAT1
   signaling in the liver of male (I) or female (J) WT and PRDX1^Cys52Ser
   mice after MCD feeding (as in A and B). n = 4 mice per group.
   ∗p < 0.05, ∗∗p < 0.01 by Student's t test. See also [133]Fig. S4. All
   data are presented as means ± SEM.

   The expression of some genes related to inflammation and fibrosis
   (Mcp-1, F4/80, Cd11b, Col1a-1 and Col3a1) was significantly reduced in
   the liver of PRDX1^Cys52Ser mice ([134]Figure 4C, D), suggesting the
   alleviated NASH phenotype. Indeed, histological staining including H&E,
   Oil Red O, Sirius Red and α-SMA staining showed a drastic decrease in
   hepatic lipid accumulation ([135]Figure 4E, F, [136]Figs. S4G and S4H)
   and liver fibrosis ([137]Figure 4G, H, [138]Figs. S4I and S4J) in
   PRDX1^Cys52Ser mice.

   We further observed that the signaling activities of STAT1 and NF-κB
   were significantly reduced in the liver of PRDX1^Cys52Ser mice
   ([139]Figure 4I, J). Collectively, these results have demonstrated that
   PRDX1 Cys52Ser variant confers resistance to MCD-induced NASH and liver
   fibrosis by suppressing hepatic NF-κB and STAT1 signaling pathways.

4. Discussion

   In this study, we tried to address an interesting question regarding
   the reported pro-inflammatory effect of PRDX1 that simultaneously shows
   potent antioxidant activity as a peroxidase. By generating
   PRDX1^Cys52Ser variant mice, we had an opportunity to investigate how
   PRDX1 enzymatic function impacts its pro-inflammatory activity. Our
   results have clearly demonstrated that inflammation and liver fibrosis
   are markedly reduced in PRDX1^Cys52Ser mice on either WD or MCD. Thus,
   our study provides in vivo evidence supporting a link of PRDX1's
   peroxidase activity with its pro-inflammatory activity. In line with
   our study, one previous in vitro study suggested the dependence of
   PRDX1's pro-inflammatory effect on its peroxidatic or resolving
   cysteine mediating its peroxidase activity [[140]19].

   How PRDX1 Cys52Ser variant suppresses inflammation mechanistically
   remains elusive yet. Previous studies have demonstrated that
   extracellular PRDX1 stimulates inflammatory responses by binding to
   TLR4 [[141]32], and mutation of peroxidatic Cys52 of PRDX1 inhibits its
   extracellular release as a chaperone protein in response to TNF-α
   [[142]19], suggesting that PRDX1 Cys52 is required for its chaperone
   function in promoting inflammatory response. In support of an
   inflammation-promoting role for PRDX1 Cys52, we found that expression
   of various pro-inflammatory genes and NF-κB signaling activity were
   markedly reduced in PRDX1^Cys52Ser variant mice. It should be noted
   that these previous studies were mainly carried out in cultured cells.
   Whether or not PRDX1 Cys52Ser variant blocks its extracellular release
   in vivo remains to be investigated. In addition, we also postulate that
   PRDX1 Cys52Ser variant may inhibit the pro-inflammatory response by
   interacting with some molecules involved in the pro-inflammatory
   signaling pathways. Yet, future studies are required to fully
   understand how PRDX1 Cys52Ser variant suppresses the activation of
   NF-κB signaling mechanistically.

   Some studies have suggested that the chaperone function of PRDX depends
   on their oligomerization (e.g., dimerization, decamerization) or
   hyperoxidation [[143]18,[144]19,[145]33]. The peroxidatic Cys52 is
   critical for PRDX1 dimerization or hyperoxidation [[146]9].
   Consistently, we observed reduced PRDX1 hyperoxidation and inflammation
   in PRDX1^Cys52Ser mice, demonstrating that the pro-inflammatory
   activity of PRDX1 is dependent of its Cys52.

   Inflammation and anti-oxidation are two opposing forces in the
   pathogenesis of NASH, and PRDX1 possesses these two opposing
   activities, implicating a complex role for PRDX1 in NASH. Although we
   demonstrate that PRDX1 Cys52 is required for its pro-inflammatory
   effect, whether or not PRDX1's pro-inflammatory activity influences its
   peroxidase activity is unknown. Overall, our study suggests that in the
   physiological settings, inactive PRDX1 dimer is reduced by Trx
   efficiently and sufficiently to active PRDX1 monomer that protects
   against oxidative stress by functioning as a potent peroxidase; whereas
   in the pathogenesis of NASH, dimerized or hyperoxidized PRDX1 due to
   overwhelming oxidants might not be reduced by Trx efficiently,
   consequently promoting inflammation by functioning as a chaperone
   protein.

5. Conclusions

   PRDX1 Cys52Ser variant confers resistance to NASH and liver fibrosis in
   mice by reducing NF-κB and STAT1 signaling pathways, suggesting that
   PRDX1 exerts dual but opposing roles in NASH.

Author contributions

   XX conceived and supervised the project, analyzed and interpreted the
   data, and wrote the manuscript. ZB performed physiological measurements
   of mice, WB, qPCR and histological staining of liver samples. WY did
   histological staining of liver samples, isolated primary mouse
   hepatocytes, and measured PRDX peroxidase activity as well as serum
   levels of AST and ALT. RL performed qPCR analysis. MT conducted WB and
   qPCR analyses. XS discussed and analyzed the data. CL discussed and
   interpreted the data.

Declaration of competing interest

   The authors declare that no conflict of interest exists.

Acknowledgements