Abstract

Background & Aims

   The role of infiltrating neutrophils in hepatocellular carcinoma (HCC)
   is modulated by cellular metabolism, specifically lipid homeostasis.
   Throughout the progression of HCC, alterations in lipid metabolism are
   intricately linked with regulation of neutrophil function and the
   release of neutrophil extracellular traps (NETs). However, how much the
   protumor effect of a high-fat diet (HFD) depends on NETs and the
   potential interplay between NETs and other leukocytes in HCC remains
   uncertain.

Methods

   In this study, the molecular mechanism of NET release and the potential
   beneficial effects of PPARα agonists on the HCC microenvironment were
   explored through proteomics, metabolomics, tissue microarray,
   immunofluorescence, flow cytometry, western blot, and dual-luciferase
   reporter gene assays (n = 6 per group).

Results

   Our study demonstrated a notable inhibition of PPARα signaling in HCC.
   Furthermore, the disruption of PPARα-mediated lipid metabolism was
   responsible for the release of NETs. The presence of a HFD was observed
   to induce mitochondrial impairment in neutrophils, leading to the
   activation of cGAS-STING by oxidized mitochondrial DNA (Ox-mtDNA).
   Consequently, this activation triggered the release of NETs containing
   Ox-mtDNA through the enhancement of NLRP3-GSDMD-N in a NF-κB-dependent
   manner. Moreover, the release of NETs within HCC tissues effectively
   isolated cytotoxic leukocytes in the outer regions of HCC.

Conclusions

   Our study not only provides insight into the relationship between lipid
   metabolism disorders and NETs’ tumor-promoting function, but also
   provides an important strategic reference for multi-target or combined
   immunotherapy of HCC.

Impact and implications:

   We have identified PPARα and its agonists as therapeutic targets for
   controlling the neutrophil extracellular traps associated with high
   lipid metabolism. Results from preclinical models suggest that PPARα
   can limit mitochondrial oxidative stress, inhibit cGAS-STING-NF-κB
   signaling, and limit the release of neutrophil extracellular traps,
   thereby increasing the contact of anti-tumor leukocytes and
   hepatocellular cancer cells and limiting tumor growth.

   Keywords: hepatocellular carcinoma, neutrophil extracellular traps,
   PPARα, oxidative stress, cGAS-STING, NF-κB

Graphical abstract

   [41]Image 1
   [42]Open in a new tab

Highlights:

     * •
       HFD induces the release of NETs through Ox-mtDNA.
     * •
       HFD activates the cGAS-STING-NF-κB axis by inducing mitochondrial
       ROS.
     * •
       PPARα inhibits the expression of GSDMD to prevent the release of
       NETs.
     * •
       PPARα promotes cytotoxic leukocyte contact with HCC cells by
       inhibiting release of NETs.

Introduction

   Hepatocellular carcinoma (HCC) is influenced by various factors, with
   fatty liver disease resulting from high-fat intake being a significant
   contributing factor. Disruption of glucose metabolism in normal
   hepatocytes induced by hyperlipidemia resembles the metabolic pattern
   observed in HCC cells.[43]^1 Additionally, a high-fat diet (HFD) can
   induce chronic endoplasmic reticulum stress, exacerbating hepatic
   steatosis and potentially leading to the development of HCC.[44]^2 In
   recent years, the field of tumor immune microenvironment research has
   witnessed a surge, leading to the identification of numerous leukocytes
   as potential targets for regulating HCC progression. Unlike other
   malignant tumors, HCC mainly occurs in a chronic inflammatory
   environment.[45]^3 Neutrophils are numerous and heterogeneous in the
   HCC microenvironment. The timely recruitment of neutrophils is
   controversial in tumorigenesis, showing different functional
   characteristics after being converted into tumor-associated
   neutrophils. The anti-tumor N1 phenotype has the ability to kill tumor
   cells and stimulate T-cell immunity, while the N2 phenotype can promote
   tumor development by inhibiting T-cell responses and upregulating
   angiogenic factors.[46]^4 Therefore, a more precise understanding of
   the subtype characteristics and functions of tumor-associated
   neutrophils in HCC progression is urgent.

   Significant advances have been made in understanding the role of
   neutrophils in the progression of metabolic dysfunction-associated
   steatotic liver disease (MASLD)-associated HCC. The progression from
   steatosis to MASLD in HFD-fed mice is facilitated by the overexpression
   of CXCL1 in the liver, mediated by neutrophil-derived reactive oxygen
   species (ROS) and stress kinases.[47]^5 Additionally, the promotion of
   regulatory T cell (Treg) differentiation through metabolic
   reprogramming of naive CD4^+ T cells by NETs presents a potential
   therapeutic target for alleviating HCC progression in patients with
   MASLD.[48]^6 Previous research has demonstrated the presence of
   elevated levels of NETs in the serum of individuals afflicted with
   hepatitis and HCC.[49]^7 Additionally, a substantial influx of
   neutrophils and the release of NETs have been observed within the
   livers of obese mouse models with HCC.[50]^6 Nevertheless, the
   mechanisms by which a HFD environment triggers the release of NETs, as
   well as the impact of this process on the progression of HCC through
   interactions with leukocytes within the microenvironment, remain
   unexplored in the literature.

   It is worth mentioning that PPARα is closely related to
   MASLD.[51]^8^,[52]^9 PPARα is highly expressed in the liver and is
   widely involved in various biological activities such as energy
   metabolism, oxidative stress, and inflammation. In recent years,
   studies have shown that PPARα hinders fatty liver disease by regulating
   lipid metabolism, improving liver lipid accumulation, accelerating
   fatty acid oxidation, and inhibiting inflammatory responses.[53]^9 It
   also inhibited HCC by increasing antioxidant enzymes and reducing
   oxidative stress.[54]^8 Therefore, whether PPARα can be activated to
   limit the progression of MASLD-related HCC, reduce inflammatory
   responses, and enhance immune infiltration of the tumor
   microenvironment (TME) are key topics of this study.

   The primary constituent of NETs, known as NETs-DNA, has been associated
   with tumor growth, functioning as a chemotactic factor that attracts
   tumor cells to distant metastatic sites.[55]^10 Consequently, the
   targeted manipulation of NETs holds great potential as an effective
   therapeutic approach against tumors. However, the precise origin of DNA
   within NETs remains uncertain, with both mitochondrial DNA (mtDNA) and
   nuclear DNA being considered significant contributors to the reticular
   structure.[56]^11 The ribonucleoprotein immune complex induces the
   upregulation of ROS in low-density granulocytes, leading to oxidative
   stress-induced mitochondrial damage, mtDNA is then released to form
   NETs.[57]^12 Nevertheless, the precise molecular mechanism underlying
   mtDNA extrusion, the specific irritants in the TME that induce
   mitochondrial hypopolarization, and the potential inhibitory effects of
   drugs or small molecule compounds that enhance mitochondrial function
   on NET release remain elusive.

   This study employed an HFD-HCC mouse model to investigate the
   correlation between HFD and NETs, as well as the involvement of NETs in
   HCC. The findings of this study were as follows: 1) HFD led to
   oxidative stress of neutrophils’ mtDNA and contributed to the release
   of NETs, with PPARα playing a regulatory role in this process; 2)
   oxidized mtDNA activated the pro-inflammatory response mediated by
   cGAS-STING signaling, resulting in increased expression of
   NLRP3-GSDMD-N and facilitating the release of the network structure; 3)
   the presence of NETs impeded the infiltration of anti-tumor leukocytes
   into the TME. Our study presented novel evidence elucidating the
   development of HFD-associated NETs and the underlying mechanism by
   which they impeded immune infiltration. Furthermore, we proposed that
   the pharmacological regulation of PPARα could serve as a promising
   strategy for HCC immunotherapy.

Materials and methods

Reagents

   All antibodies, chemicals, small molecule inhibitors, drugs, commercial
   kits and primers used are displayed in [58]Tables S1-S4.

Animals

   C57BL/6J (female, 6 weeks old, 18∼20 g) were purchased from Shanghai
   Model Organisms Center, Inc. Mice were housed four per cage in
   individually ventilated cage systems in specific pathogen-free grade
   animal room at the Animal Center of Fujian Medical University. Mouse
   were kept under a 12/12 h light/dark cycle. The animal experiments in
   this study were approved by the Experimental Animal Ethics Committee of
   Fujian Medical University (approval number IACUC FJMU2023-Y-0841) and
   were carried out according to the Institutional Animal Care and Use
   Committee guidelines.

Statistical analysis

   All data analysis was performed with GraphPad Prism V.8 or SPSS V.19,
   and data are presented as mean ± standard deviation. When comparing two
   samples, the independent sample t test was used if the data were
   normally distributed and the variances were homogeneous. If the data
   were normally distributed with unequal variance, Wilcoxon rank sum test
   was used. European-style clustering was employed to examine and
   evaluate the resemblances among various samples. Flow cytometry data
   were analyzed using FlowJo software V10. p <0.05 was used as the
   criterion for statistical significance.

   Other detailed methods are provided in the supplementary information.

Results

Impaired PPARα signaling was correlated with dysregulation of lipid
metabolism in HCC

   To explore the relative contribution of the etiology or complications
   of HCC in the release of NETs, ELISA was employed to assess the
   correlation between the expression indicator of NETs, specifically
   citrullinated histone H3 (CitoH3), and various clinical indicators
   (such as HBV infection, cirrhosis, liver cyst, focal nodular
   hyperplasia, and MASLD). Notably, our finding indicated that only two
   factors (HBV infection and MASLD) were responsible for the aberrant
   expression of CitoH3 in patients ([59]Fig. 1A). Specifically, HBV
   infection hindered the expression of CitoH3, whereas MASLD enhanced its
   expression ([60]Fig. 1A). Previous research has documented the
   inhibition of NETs by HBV infection.[61]^13 Consequently, we aimed to
   investigate the association between MASLD and NETs. To elucidate the
   link between aberrant lipid metabolism and NETs release, we conducted a
   comparative analysis of the proteome and the metabolome of HCC. The
   proteomic data revealed a significant correlation between impaired PPAR
   signaling and the development of mouse HCC ([62]Fig. 1B). Additionally,
   metabolomics indicated disruptions in various lipid metabolic signals,
   particularly triglycerides, within patient cancerous tissues
   ([63]Fig. 1C). Furthermore, we found that the inhibition of PPAR
   signaling contributed to the dysregulation of lipid metabolism, and
   PPAR signaling was the second most influential pathway ([64]Fig. 1D and
   [65]Table S6). Moreover, our study indicated that, alongside the
   identification of nine differentially expressed proteins through
   proteomics analysis ([66]Figs 1E and [67]S1-9), the expression of an
   additional set of 16 downstream proteins of PPAR signaling, including
   PPARα, was observed to be suppressed in patient cancerous tissues
   ([68]Fig. 1F and [69]Fig. S10-25). However, the expression of another
   21 PPAR signaling pathway-related proteins, including ME1, did not
   exhibit significant changes in patient HCC tissues ([70]Fig. S26).

Fig. 1.

   [71]Fig. 1
   [72]Open in a new tab

   Impaired PPARα signaling was correlated with dysregulation of lipid
   metabolism in HCC.

   (A) ELISA showing the concentration of CitoH3 within patient HCC (n =
   79). (B) KEGG pathway enrichment analysis examining the differentially
   expressed proteins between HCC and para-cancer tissues in the mouse
   models (n = 3). PPAR signaling pathway was marked in red. (C) A network
   map illustrating the differential lipid metabolites between HCC and
   para-cancer tissues in patients (n = 20). The KEGG pathway was
   represented by blue circles, while the red circles indicated the
   metabolites enriched in the corresponding pathway. (D) KEGG pathway
   enrichment analysis investigating the differential metabolites between
   HCC and para-cancer tissues in patients (n = 20). (E) The proteomic
   identification of mice was confirmed through patient tissue microarray,
   revealing differential expression levels of PPAR signaling
   pathway-associated proteins between HCC and para-cancerous tissues in
   patients (n = 20). (F) The expression of the other 16 PPAR signaling
   pathway-related proteins in HCC and para-cancerous tissues was examined
   using tissue microarray (n = 20). (G) KEGG enrichment analysis
   determining the distribution of differentially expressed PPAR signaling
   pathway-related proteins in the three subtypes of PPAR and their
   downstream targets, with repressed genes indicated by the color blue.
   Mean ± SD. Statistical significance evaluated by Student’s t test (A,
   E-F). ∗p <0.05, ∗∗p <0.01. CitoH3, citrullinated histone H3; FNH, focal
   nodular hyperplasia; HCC, hepatocellular carcinoma; KEGG, Kyoto
   Encyclopedia of Genes and Genomes; MASLD, metabolic
   dysfunction-associated steatotic liver disease; NETs, neutrophilic
   extracellular traps.

   There exist three distinct subtypes of PPAR, namely PPARα, PPARβ/δ, and
   PPARγ, each exhibiting distinct functional roles.[73]^8^,[74]^9 Based
   on the patient metabolome data ([75]Fig. 1C), it was evident that HCC
   was characterized by notable alterations in fatty acid synthesis,
   extension, absorption, degradation, and disruption of unsaturated fatty
   acid synthesis. Consequently, we proposed that compromised PPARα
   signaling played a crucial role in HCC. We found that proteins involved
   in PPARα signaling were inhibited in HCC ([76]Fig. 1G). Therefore, low
   PPARα expression associated with high lipid metabolism might be the key
   to inhibiting release of NETs.

PPARα inhibited release of NETs in HCC

   The absence of PPARα hinders the release of pro-inflammatory cytokines,
   thereby impairing the functionality of Tregs and macrophages.[77]^14
   However, the extent to which PPARα is implicated in the functionality
   of neutrophils remains inadequately understood. Consequently, we
   examined the correlation between PPARα and the release of NETs in HCC.
   Patient HCC were categorized into high and low PPARα signaling groups,
   predicated on the expression levels of 25 PPARα-associated genes
   ([78]Fig. 2A). Our study indicated that the release of NETs was
   inhibited in the high PPARα signaling group ([79]Fig. 2B). Moreover,
   the low PPARα signaling group demonstrated a significantly elevated
   occurrence of lymph node metastasis, distant metastasis, and vascular
   invasion ([80]Fig. 2C,D and [81]Fig. S27).

Fig. 2.

   [82]Fig. 2
   [83]Open in a new tab

   PPARα inhibited NET release in HCC.

   (A) HCC tissues obtained from a cohort of 12 patients were stratified
   into two distinct groups, namely high and low PPARα signaling, using
   the expression levels of 25 PPARα-associated genes (n = 6). (B)
   Immunofluorescence investigating the presence of NETs in HCC tissues of
   patients (n = 3). Blue, DAPI; green, CitoH3; red, MPO. The
   co-localization of these three markers indicated NETs. The NETs were
   marked using a white box. (C) Baseline table from TCGA database were
   utilized to conduct a comparative analysis of vascular invasion in HCC
   (n = 205). Red indicates statistical significance. (D) Baseline table
   was employed to compare TNM staging and vascular invasion in patients
   with HCC from our center (n = 24). Red indicates statistical
   significance. (E) Immunofluorescence revealing the presence of NETs in
   tumor tissues of primary HCC mice (n = 6). (F) Immunofluorescence
   visualizing the presence of NETs in HCC tissues from patients (n = 6).
   (G) Immunofluorescence showing the presence of NETs in tumor tissues of
   primary HCC mice (n = 6). (H) Flow cytometry demonstrating the
   expression of NET-related markers in primary HCC-infiltrating
   neutrophils from mice (n = 6). (I) Flow cytometry showing the
   expression of NET-related markers in HCC-infiltrating neutrophils from
   patients (n = 6). Mean ± SD. European-style clustering was employed to
   examine and evaluate the resemblances among various samples (A).
   Statistical significance evaluated by Student’s t test (B, E-I). ∗p
   <0.05, ∗∗p <0.01. CitoH3, citrullinated histone H3; HCC, hepatocellular
   carcinoma; HFD, high-fat diet; KEGG, Kyoto Encyclopedia of Genes and
   Genomes; MASLD, metabolic dysfunction-associated steatotic liver
   disease; NETs, neutrophilic extracellular traps; TCGA, the Cancer
   Genome Atlas; TNM, tumor-node-metastasis.

   We then explored the relationship between hyper-lipid metabolism and
   NETs. We performed cluster analysis based on the expression of 25
   PPARα-associated genes in primary HCC tissues of patients and mice, and
   divided them into high and low PPARα signaling groups ([84]Figs
   S28-29). A higher abundance of NETs was observed in HCC tissues from
   mice fed a HFD or patients with MASLD compared to those from mice on a
   normal diet or patients without MASLD ([85]Fig. 2E,F). Conversely, in
   mice fed a HFD, the release of NETs was suppressed in the group with
   high PPARα signaling compared to low PPARα signaling ([86]Fig. 2E,F).
   In addition, bezafibrate, an agonist of PPARα, or norathyriol, an
   inhibitor, were used to determine the effect of PPARα on the release of
   NETs, and it was found that bezafibrate inhibited the release of NETs,
   while norathyriol did the opposite ([87]Fig. 2G). Futhermore, we
   examined the expression of four markers of NETs,[88]^10 including PAD4,
   mitochondrial ROS, CitoH3, and 8-OHdG in neutrophils infiltrating HCC
   from patients and mice ([89]Fig. S30). We found that HFD or MASLD
   combined with low PPARα signaling resulted in increased expression of
   the four markers ([90]Fig. 2H,I). In comparison to the group consuming
   the HFD/MASLD and low PPARα signaling, the HFD/MASLD and high PPARα
   signaling group exhibited suppressed expression of these markers
   ([91]Fig. 2H,I). Hence, activating PPARα was a feasible strategy to
   inhibit NETs related to high lipid metabolism.

PPARα mitigated HFD-associated NETs through the attenuation of mitochondrial
oxidative stress

   Considering PPARα inhibited the production of mitochondria ROS induced
   by a HFD ([92]Fig. 2H,I), we further investigated whether PPARα hinder
   the release of NETs by suppressing mitochondrial DNA (mtDNA) oxidation.
   We found that overexpression of Ppara downregulated mitochondrial ROS
   levels in neutrophils infiltrating mouse HCC, while knockout
   upregulated them ([93]Fig. 3A). Our findings demonstrated that
   neutrophils generated higher levels of mitochondrial ROS when exposed
   to a HFD, while the opposite effect was observed in the low-fat diet
   (LFD) group ([94]Fig. 3B). The reduction of HFD-induced mitochondrial
   ROS was observed when treated with two PPARα agonists (bezafibrate or
   fenofibrate) ([95]Fig. 3B). Moreover, we found that overexpression of
   Ppara increased mitochondrial membrane potential, while knockout
   decreased it ([96]Fig. 3C). HFD resulted in lower mitochondrial
   membrane potential, whereas low lipidemia led to higher levels
   ([97]Fig. 3D). The reduction in mitochondrial membrane potential
   induced by HFD was suppressed by PPARα agonists, and this effect was
   similarly observed in the groups fed normal-fat diet and LFD
   ([98]Fig. 3D). Given that the loss of mitochondrial membrane potential
   is associated with PKC-mediated mitochondrial ROS,[99]^15 we conducted
   an experiment where we added the PKC inhibitor chelerythrine-chloride
   to neutrophils. Our results showed a reduction in HFD-mediated
   mitochondrial polarization, while chelerythrine-chloride did the
   opposite ([100]Fig. S31A). In addition, we introduced rotenone, an
   inductor of mitochondrial ROS,[101]^16 to neutrophils infiltrating
   HFD-HCC that were treated with PPARα agonists. Our findings indicated
   that PPARα agonists resulted in an increase in the mitochondrial
   polarization that was inhibited by rotenone ([102]Fig. S31B).

Fig. 3.

   [103]Fig. 3
   [104]Open in a new tab

   PPARα mitigated HFD-associated NETs through the attenuation of
   mitochondrial oxidative stress.

   (A,B) MitoSOX™ Green assessing the levels of mitochondrial ROS in mouse
   neutrophils (n = 3). Effects of overexpression or knockout of Ppara
   in vitro (A). Role of HFD, LFD or the PPARa agonists (bezafibrate or
   fenofibrate) in tumor tissues of primary HCC mice (B). (C,D) JC-1
   assessing the extent of mitochondrial membrane potential in mouse
   neutrophils (n = 3). Effects of overexpression or knockout of Ppara
   in vitro (C). Role of HFD, LFD or the PPARa agonists (bezafibrate or
   fenofibrate) in tumor tissues of primary HCC mice (D). (E)
   Immunofluorescence visualizing the presence of NETs in tumor tissues of
   primary HCC mice (n = 6). Blue, DAPI; green, CitoH3; red, MPO. The
   co-localization of these three markers indicated NETs. The NETs were
   marked using a white box. (F) Immunofluorescence showing the presence
   of NETs in tumor tissues of primary HCC mice (n = 6). Mean ± SD.
   Statistical significance evaluated by Student’s t test (A-F). ∗p <0.05,
   ∗∗p <0.01. B, bezafibrate; CitoH3, citrullinated histone H3; F,
   fenofibrate; HCC, hepatocellular carcinoma; HFD, high-fat diet; LFD,
   low-fat diet; NETs, neutrophilic extracellular traps; R, rotenone; ROS,
   reactive oxygen species.

   We further explored the potential of mitochondrial ROS in the release
   of NETs. Rotenone and feprazone, an inhibitor of mitochondrial ROS,
   were used to determine the effect of mitochondrial ROS on the release
   of NETs, and it was found that rotenone induced the release of NETs,
   while feprazone did the opposite ([105]Fig. 3E). Furthermore, our study
   revealed that HFD augmented the release of NETs in HCC ([106]Fig. 3F).
   Notably, this effect was counteracted by PPARα agonists ([107]Fig. 3F).
   To determine whether low levels of mitochondrial oxidative stress were
   necessary for PPARα agonists to inhibit the release of NETs, we
   promoted ROS production with rotenone and found that rotenone further
   exacerbated the release of NETs, while PPARα agonists mitigated its
   effects ([108]Fig. 3F). Therefore, we concluded that PPARα inhibited
   the release of NETs by suppressing the level of mitochondrial ROS in
   neutrophils and increasing mitochondrial membrane potential.

PPARα mitigated HFD-associated NETs by impairing the release of mtDNA

   We observed that overexpression of Ppara inhibited the expression of
   8-OHdG and TOM20 on the cell membrane in neutrophils infiltrating mouse
   HCC, while knockout induced their expression ([109]Fig. 4A).
   Furthermore, our findings indicated that HFD promoted mitochondrial
   mobilization to the cell membrane, whereas LFD had the opposite effect
   ([110]Fig. 4B). However, PPARα agonists mitigated HFD-induced
   mitochondrial mobilization, as observed in both conventionally fed and
   LFD groups ([111]Fig. 4B). To mitigate the potential bias associated
   with two-dimensional fluorescence, we investigated the aggregation of
   mitochondria on the membrane of nonpermeable cells. Ppara knockout was
   found to exhibit strong inducement of mitochondrial outer membrane
   proteins TOM20 and HSP60, and vice versa ([112]Fig. 4C). Within the HFD
   group, we observed the translocation of TOM20 and HSP60 to the surface
   of nonpermeable cells, which was diminished by PPARα agonists
   ([113]Fig. 4D,E and [114]Fig. S32). To eliminate the potential
   occurrence of neutrophils containing mitochondria binding to the cell
   surface, which enhance the staining of TOM20 on the membrane, we
   subjected neutrophils to treatment with DNaseI. The level of TOM20 on
   the cell surface was unaffected by DNaseI, indicating that it was
   unlikely to be a consequence of neutrophils being deposited on the cell
   surface as an external trapping mechanism ([115]Fig. S33). To examine
   the detrimental effects of mitochondrial ROS, we employed 8-OHdG to
   assess whether PPARα agonists mitigate DNA oxidation induced by HFD.
   The HCC-infiltrating neutrophils with high lipid association exhibited
   significant 8-OHdG staining, while the neutrophils associated with LFD
   showed low levels ([116]Fig. 4B). PPARα agonists not only decreased
   8-OHdG expression in HFD-stimulated cells, but also reduced 8-OHdG
   levels to negligible amounts in the hypolipidemia group ([117]Fig. 4B).
   Co-localization of TOM20 with 8-OHdG suggested that the majority of
   oxidation occurs on mtDNA rather than nuclear DNA ([118]Fig. 4B). To
   establish the origin of the released oxidized DNA as being from
   mitochondria, we isolated mitochondria from neutrophils and assessed
   their expression of 8-OHdG. As anticipated, PPARα agonists resulted in
   a reduction of 8-OHdG expression ([119]Fig. 4F). Moreover, this
   inhibitory effect was reversed by rotenone ([120]Fig. 4G). To ascertain
   the impact of PPARα-mediated lipid metabolism on nuclear DNA damage,
   nuclear protein was extracted and analyzed. In the HFD group, it was
   observed that high lipid metabolism induced the expression of 8-OHdG,
   which was inhibited by PPARα agonists ([121]Fig. S34A). Conversely, in
   the LFD group, although the expression of 8-OHdG was suppressed, PPARα
   agonists did not exhibit any protective effects on nuclear DNA
   ([122]Fig. S34A). These findings suggested that PPARα exclusively
   influenced nuclear DNA damage caused by HFD intake, without
   participating in the regulation of nuclear DNA effects associated with
   LFD consumption. It was noteworthy that the protective impact of PPARα
   agonists on nuclear DNA in the HFD group was impeded by the ROS agonist
   GSK5182 ([123]Fig. S34B). Hence, it appeared that HFD induced
   mitochondrial membrane mobilization, and PPARα inhibited this process
   to regulate neutrophil function.

Fig. 4.

   [124]Fig. 4
   [125]Open in a new tab

   PPARα mitigated HFD-associated NETs by impairing the release of mtDNA.

   (A,B) Immunofluorescence illustrating the expression and localization
   of 8-OHdG (green) and TOM20 (red) of neutrophils infiltrated by HCC in
   mice (n = 3). Effects of overexpression or knockout of Ppara in vitro
   (A). Role of HFD, LFD or the PPARa agonists (Bezafibrate or
   Fenofibrate) in tumor tissues of primary HCC mice (B). (C-E) Flow
   cytometry analyzing the expression of TOM20 and HSP60 in nonpermeable
   neutrophils infiltrated by mouse HCC (n = 3). Effects of overexpression
   or knockout of Ppara in vitro (C). Role of HFD, LFD or the PPARa
   agonists (bezafibrate or fenofibrate) in tumor tissues of primary HCC
   mice (D-E). (F,G) Western blot demonstrating mitochondrial 8-OHdG
   expression in infiltrating neutrophils of mouse primary HCC (n = 3).
   Effects of PPARα agonists (F). Role of PPARα agonists and rotenone (G).
   Mean ± SD. Statistical significance evaluated by Student’s t test
   (A-G). ∗p <0.05, ∗∗p <0.01. B, bezafibrate; F, fenofibrate; HCC,
   hepatocellular carcinoma; HFD, high-fat diet; LFD, low-fat diet; mtDNA,
   mitochondrial DNA; NETs, neutrophilic extracellular traps; R, rotenone.

PPARα inhibited cGAS-STING signaling and induced disruption of NETs

   Considering cGAS-STING is a receptor for DNA,[126]^17 we investigated
   the impact of PPARα and lipid metabolism on cGAS-STING signaling. Our
   findings revealed the level of p-STING/STING decreased after Ppara
   overexpression in neutrophils infiltrating mouse HCC, accompanied by a
   robust decline of downstream inflammatory cytokines p-IRF3/IRF3, IFN-α,
   IFN-β, IL-6, TNF-α, and ISG15 ([127]Fig. 5A). Moreover, a HFD led to a
   significant rise in cGAS-STING-dependent type I IFN signaling, while
   the opposite was observed for LFD ([128]Fig. 5A). PPARα agonists
   exhibited notable antagonistic effects on cGAS-STING signaling,
   resulting in the suppression of the cGAS-STING signal to an
   imperceptible level in the LFD group ([129]Fig. 5A). We found that the
   blocking effect of Ppara overexpression on cGAS-STING signaling was
   mitigated by rotenone ([130]Fig. 5B). As expected, rotenone upregulated
   HFD-induced pro-inflammatory signaling, and this positive effect was
   damaged by PPARα agonists ([131]Fig. 5B). In addition, CL656, an
   agonist of STING, and omaveloxolone, an inhibitor, were used to
   determine the effect of STING on the release of NETs, and it was found
   that CL656 induced the release of NETs, while omaveloxolone did the
   opposite ([132]Fig. 5C). Furthermore, our findings indicated that
   elevated levels of STING promoted the release of NETs in HCC, and this
   positive effect was disrupted by PPARα agonists ([133]Fig. 5D). In
   short, PPARα prevented the formation of NETs by inhibiting cGAS-STING
   signaling induced by mitochondrial oxidized DNA.

Fig. 5.

   [134]Fig. 5
   [135]Open in a new tab

   PPARα inhibited cGAS-STING signaling and induced disruption of NETs.

   (A,B) Western blot showing the expression of proteins related to
   cGAS-STING signaling in infiltrating neutrophils of mouse primary HCC
   (n = 3). Effect of overexpression, knockout of Ppara or PPARα agonists
   (A). Role of overexpression of Ppara, PPARα agonists and rotenone (B).
   (C) Immunofluorescence visualizing the presence of NETs in tumor
   tissues of mouse primary HCC mice (n = 6). Blue, DAPI; green, CitoH3;
   red, MPO. The co-localization of these three markers indicated NETs.
   The NETs were marked using a white box. (D) Immunofluorescence
   demonstrating the presence of NETs in tumor tissues of mouse primary
   HCC mice (n = 6). Mean ± SD. Statistical significance evaluated by
   Student’s t test (A-D). ∗p <0.05, ∗∗p <0.01. B, bezafibrate; C, CL656;
   CitoH3, citrullinated histone H3; F, fenofibrate; HCC, hepatocellular
   carcinoma; HFD, high-fat diet; LFD, low-fat diet; NETs, neutrophilic
   extracellular traps; R, rotenone; ROS, reactive oxygen species.

PPARα inhibited NF-κB signaling depending on cGAS-STING signaling

   Given that NF-κB works in conjunction with IRF3 to facilitate the
   production of type I IFN in response to stimulation by double-stranded
   DNA.[136]^18 We hypothesized that the inhibition of NF-κ signaling by
   PPARα was reliant on Ox-mtDNA-cGAS-STING signaling. We found that
   overexpression of Ppara inhibited phosphorylation of IKKα/β and p65 in
   neutrophils infiltrating mouse HCC, while knockdown promoted it
   ([137]Fig. 6A). High levels of IKKα/β and p65 phosphorylation were
   observed in the HFD group, whereas a significant inhibition of their
   phosphorylation was observed in the LFD group ([138]Fig. 6A).
   Additionally, PPARα agonists demonstrated antagonism of the NF-κ
   pathway ([139]Fig. 6A). To ascertain the potential pro-inflammatory
   function of Ox-mtDNA, we subjected HCC-infiltrating neutrophils
   obtained from HFD mice treated with PPARα agonists to rotenone. Our
   findings revealed a significant upregulation of phosphorylation levels
   in IKKα/β and p65 upon rotenone administration, which was impaired by
   overexpression of Ppara or PPARα agonists ([140]Fig. 6B). To ascertain
   the dependence of PPARα inhibition of the NF-κ pathway on cGAS-STING
   signaling, we administered CL656 to neutrophils and observed that
   elevated STING signaling augmented NF-κ signaling, whereas
   overexpression of Ppara or PPARα agonists impeded this favorable impact
   ([141]Fig. 6C). Given that both Ox-mtDNA and cGAS-STING signaling
   contribute to the robust NF-κ signaling induced by heightened lipid
   metabolism, we further investigated whether NF-κ signaling serves as a
   crucial link in the downregulation of NET release by PPARα. To this
   end, NF-κ activator 1, an agonist of NF-κ, and JSH-23, an inhibitor,
   were utilized to determine the effect of NF-κ on the release of NETs,
   and it was found that NF-κ activator 1 induced the release of NETs,
   while JSH-23 did the opposite ([142]Fig. 6D). Moreover, we found that
   elevated levels of NF-κB enhanced the release of NETs in HCC, while
   PPARα agonists disrupted this favorable outcome ([143]Fig. 6E).
   Therefore, PPARα downregulated the release of NETs through the NF-kB
   pathway induced by cGAS-STING signaling.

Fig. 6.

   [144]Fig. 6
   [145]Open in a new tab

   PPARα inhibited NF-κB signaling depending on cGAS-STING signaling.

   (A-C) Western blot showing the expression of proteins related to the
   NF-κB signaling in infiltrating neutrophils of mouse primary HCC (n =
   3). Effect of overexpression, knockout of Ppara or PPARα agonists (A).
   Role of overexpression of Ppara, PPARα agonists and rotenone (B).
   Influence of overexpression of Ppara or PPARα agonists and CL656 (C).
   (D) Immunofluorescence visualizing the presence of NETs in tumor
   tissues of mouse primary HCC (n = 6). Blue, DAPI; green, CitoH3; red,
   MPO. The co-localization of these three markers indicated NETs. The
   NETs were marked using a white box. (E) Immunofluorescence showing the
   presence of NETs in tumor tissues of mouse primary HCC mice (n = 6).
   Mean ± SD. Statistical significance evaluated by Student’s t test
   (A-E). ∗p <0.05, ∗∗p <0.01. B, bezafibrate; C, CL656; CitoH3,
   citrullinated histone H3; F, fenofibrate; HCC, hepatocellular
   carcinoma; HFD, high-fat diet; LFD, low-fat diet; NA, NF-κ activator 1;
   NETs, neutrophil extracellular traps; R, rotenone.

PPARα suppressed NLRP3-GSDMD-N depending on the Ox-mtDNA-cGAS-STING-NF-κ-NETs
axis

   It has been reported that GSDMD is cleaved to form an active N-terminal
   fragment with a size of about 32 kD, which binds to the membrane and
   perforates.[146]^19 Our findings indicated that the expression of NLRP3
   and GDSMD-N induced by HFD was suppressed by PPARα agonists in
   neutrophils infiltrating mouse HCC, but this inhibition was hindered by
   rotenone, CL656, and NF-κ activator 1 ([147]Fig. S35A). The JASPAR
   database predicted that p65 adhered to the promoter of NLRP3
   ([148]Fig. S35B,C), and our study confirmed that p65 bound to the -187
   ∼ -178 “gggaaccccc” of the mouse NLRP3 promoter ([149]Fig. S35D).
   Furthermore, HFD induced p65 binding to the NLRP3 promoter, while PPARα
   agonists inhibited this interaction ([150]Fig. S35E).

PPARα impeded HCC depending on Ox-mtDNA-cGAS-STING-NF-κ-NETs

   We next investigated whether PPARα impeded HCC growth by inhibiting the
   release of NETs in vivo. Our study revealed that HFD resulted in the
   growth of tumors and a higher hemorrhagic ascites rate ([151]Fig. 7).
   Elimination of NETs through Dnase I impeded this cancer-promoting
   effect ([152]Fig. 7A-F). Furthermore, PPARα agonists were found to
   hinder the growth of HCC and hemorrhagic ascites induced by HFD
   ([153]Fig. 7A-F). However, their effectiveness was compromised by PMA,
   which triggers the release of NETs, as well as by rotenone, CL656, and
   NF-κ activator 1 ([154]Fig. 7A-F).

Fig. 7.

   [155]Fig. 7
   [156]Open in a new tab

   PPARα impeded HCC depending on Ox-mtDNA-cGAS-STING-NF-κ-NETs.

   (A) A representative diagram illustrating the growth of HCC in the
   abdominal cavity in a mouse model of peritoneal growth (n = 6). (B) The
   tumor volume for (A) (n = 6). (C) The tumor weight for (A) (n = 6). (D)
   The hemorrhagic ascites for (A) (n = 6). (E) A representative diagram
   showing HCC growth in situ in mice (n = 6). (F) The tumor volume for
   (E) (n = 6). Mean ± SD. Statistical significance evaluated by Student’s
   t test (B–C, F). ∗p <0.05, ∗∗p <0.01. B, bezafibrate; C, CL845; F,
   fenofibrate; HCC, hepatocellular carcinoma; HFD, high-fat diet; NA,
   NF-κ activator 1; NETs, neutrophil extracellular traps; R, rotenone.

NETs inhibited contact between HCC cells and cytotoxic cells

   To investigate the tumor immunomodulatory effect of NETs, we
   investigated the spatial localization of HCC cells, NETs, and cytotoxic
   cells (CD4^+ T cells, CD8^+ T cells and NK cells) within mouse HCC
   tissues. Our findings revealed that NETs increased physical separation
   between the cytotoxic cells and HCC cells ([157]Fig. 8A-D). We observed
   that HFD promoted the release of NETs, isolating cytotoxic cells away
   from HCC cells ([158]Fig. 8A-D). Conversely, PPARα agonists diminished
   the spatial separation between cytotoxic cells and HCC cells by
   inhibiting the release of NETs ([159]Fig. 8A-D). Hence, NETs converted
   the TME from “hot” to “cold” by capturing T cells and NK cells and
   isolating them from tumor cells.

Fig. 8.

   [160]Fig. 8
   [161]Open in a new tab

   NETs inhibited contact between HCC cells and cytotoxic cells.

   (A-C) Immunofluorescence showing the quantitative and spatial
   arrangement of cytotoxic cells, including CD4^+ T cells (A), CD8^+ T
   cells (B), and NK cells (C), along with HCC cells and NETs in tumor
   tissues of mouse primary HCC mice. Blue, DAPI; green, CK18; red,
   CitoH3; yellow, CD8/CD4/NK1.1. The arrows indicated that NETs capture
   cytotoxic cells (n = 6). (D) Quantization graph for A-C (n = 6). (E)
   The mechanism diagram. Mean ± SD. Statistical significance evaluated by
   Student’s t test (D). ∗p <0.05, ∗∗p <0.01. CitoH3, citrullinated
   histone H3; HCC, hepatocellular carcinoma; NETs, neutrophil
   extracellular traps.

Discussion

   Prior research has demonstrated the involvement of mitochondrial
   ROS-mediated cytotoxicity in liver injury induced by HFD.[162]^20
   Through an examination of mitochondria ROS during the release of NETs,
   we have discerned three novel functions of mitochondria in the context
   of HFD. Firstly, the inflammatory nature of a HFD led to mitochondrial
   impairment, while mitochondrial ROS sustained the release of NETs.
   Secondly, Ox-mtDNA, possessing potent pro-inflammatory and interferon
   attributes, activated the DNA receptor cGAS-STING and triggered
   NLRP3-GSDMD-N, thereby augmenting membrane permeability to facilitate
   the release of mtDNA in an NF-κB-dependent fashion. Lastly, the
   inhibition of PPARα signaling in HCC tissues was pronounced, and PPARα
   agonists demonstrated a significant ability to suppress the release of
   oxidative stress-induced NETs. The identification of NETs along with
   the systemic inhibition of the Ox-mtDNA-cGAS-STING-NF-κB-NLRP3-GSDMD-N
   axis, provided robust evidence to support our findings that the
   progression of HCC was exacerbated through NET-mediated mechanisms.

   The correlation between the abundance of neutrophils and the extent of
   steatosis, as well as their association with the development of fatty
   hepatocytes, has been established.[163]^21 The elevated
   neutrophil-to-lymphocyte ratio observed in the peripheral blood of
   individuals is regarded as a non-invasive indicator of both
   non-alcoholic steatohepatitis and the severity of hepatic
   fibrosis.[164]^22 Inhibiting neutrophil infiltration has been shown to
   effectively protect mice from HFD-induced steatohepatitis.[165]^23 In
   this study, it was observed that HFD contributed to the generation of
   mitochondria ROS that induced damage to mtDNA in neutrophils. The
   presence of depolarized mitochondria plays a crucial role in
   facilitating the release of NETs. Significantly, PPARα agonists
   demonstrated effective mitigation of externalized Ox-mtDNA, suppression
   of inflammatory cytokine production, and prevention of membrane damage
   in neutrophils through a mechanism reliant on the
   cGAS-STING-NF-κB-NLRP3-GSDMD-N axis. Furthermore, these agonists
   alleviated the extrusion of mtDNA, a crucial constituent of NETs, into
   surrounding cells.

   The observed increase in neutrophil granuloprotein,[166]^24
   pro-inflammatory cytokines, and type I IFN[167]^25 within the
   microenvironment of HCC closely resembled the induction outcomes
   observed in NETs. Furthermore, it is worth noting that additional
   mechanisms contributing to chronic inflammation in HCC synergistically
   interacted with the dysregulation of neutrophils, thereby intensifying
   the inflammatory response.[168]^26 Consequently, mitigating the innate
   immune response triggered by NETs emerges as a crucial approach to
   attenuate the inflammatory milieu within HCC, diminish the inflammatory
   harm inflicted upon HCC cells, and counteract the malignant conversion
   of hepatocytes as well as the development of metastases. The findings
   of our study indicated that PPARα agonists could decrease the type I
   IFN response and hinder the activation of NF-κB signaling, which was
   responsible for promoting pro-inflammatory responses in HCC associated
   with MASLD. These results provided evidence that the antagonistic
   impact of PPARα on cGAS-STING signaling serves as a mechanism to
   mitigate the release of NETs, which were mediated by disorders in lipid
   metabolism. Additionally, these findings shed light on the interplay
   between PPARα and innate immunity.

   Recent evidence indicated that the pathogenesis of various
   malignancies, including HCC, is influenced by abnormal release of NETs
   and/or impaired clearance.[169]^10^,[170]^11^,[171]^27 This assertion
   was substantiated by mouse studies, which demonstrated that inhibiting
   the release of NETs enhances tumor invasion,[172]^7 macrophage
   polarization,[173]^28 and infiltration of Tregs.[174]^6 Nevertheless,
   the impact of NETs on the functionality of other leukocytes, their
   involvement in tumor-leukocyte interactions, and their potential
   hindrance of cytotoxic leukocyte infiltration remain inadequately
   understood. This study demonstrated that NETs possessed the ability to
   impede intracellular apoptosis by establishing a physical blockade
   between malignant cells and leukocytes via their distinctive web-like
   architecture, thereby capturing and enclosing infiltrating T cells and
   NK cells.

   In conclusion, the deficiency of PPARα in HCC associated with HFD
   triggered the activation of the cGAS-STING-NF-κB-NLRP3-GSDMD-N axis,
   resulting in heightened NET release. The pivotal function of NETs lies
   in impeding the localization and elimination of HCC cells by cytotoxic
   leukocytes. Hence, PPARα agonists to obstruct this axis represent a
   promising therapeutic approach for MASLD-related HCC ([175]Fig. 8E).

Abbreviations

   CitoH3, citrullinated histone H3; HCC, hepatocellular carcinoma; HFD,
   high-fat diet; LFD, low-fat diet; mtDNA, mitochondrial DNA; MASLD,
   metabolic dysfunction-associated steatotic liver disease; NETs,
   neutrophil extracellular traps; ORF, open reading frame; Ox-mtDNA,
   oxidized mitochondrial DNA; ROS, reactive oxygen species; TMA, tissue
   microarray; TME; tumor microenvironment; Treg, regulatory T cell.

Financial support

   This study was supported by the Postdoctoral Supporting Fund of Fujian
   Medical University Union Hospital (2023XH022), the Startup Fund for
   Scientific Research of the Fujian Medical University (2023QH2020), the
   Joint Funds for the Innovation of Science and Technology, Fujian
   Province (2021Y9067), the National Natural Science Foundation of China
   (82372936) and the High-Level Medical Care Construction Foundation of
   Fujian Province ([2021]76).

Authors’ contributions

   BLP, YLC and NHT designed the study. BLP, ZZ and DJY conducted animal
   experiments, processed data and wrote the manuscript. BLP, ZZ, DJY,
   XXZ, YXY and YL conducted the cell experiments and molecular biology
   experiments. HJH and XRC analyzed the pathological and clinical data.
   YLC and NHT supervised the project and reviewed the manuscript. All
   authors read and approved the final manuscript.

Data availability statement

   The raw mass spectrometry data that support the findings of this study
   are deposited (OMIX006828,
   https://ngdc.cncb.ac.cn/omix/select-edit/OMIX006828; OMIX006841,
   [176]https://ngdc.cncb.ac.cn/omix/select-edit/OMIX006841).

Conflict of interest

   The authors of this study declare that they do not have any conflict of
   interest.

   Please refer to the accompanying ICMJE disclosure forms for further
   details.

Footnotes

   Author names in bold designate shared co-first authorship

   Supplementary data to this article can be found online at
   [177]https://doi.org/10.1016/j.jhepr.2024.101228.

Contributor Information

   Yanling Chen, Email: chenyanling@fjmu.edu.cn.

   Nanhong Tang, Email: fztnh@fjmu.edu.cn.

Supplementary data

   The following are the Supplementary data to this article:
   Multimedia component 1
   [178]mmc1.pdf^ (10.3MB, pdf)
   Multimedia component 2
   [179]mmc2.docx^ (99.5KB, docx)
   Multimedia component 3
   [180]mmc3.pdf^ (536.3KB, pdf)
   Multimedia component 4
   [181]mmc4.pdf^ (18.7MB, pdf)

References