Abstract

   M1 macrophage polarization plays a key role in the onset and
   progression of sepsis. Fibroblast growth factor 15 (FGF15) suppresses
   septic inflammation through its FGF receptor 4 (FGFR4); however, the
   underlying mechanisms are largely unclear. In this study, we evaluated
   the anti-inflammatory effects of recombinant FGF15 (rFGF15) in cecal
   ligation and puncture (CLP)-induced septic mice in vivo, as well as
   lipopolysaccharide (LPS)-stimulated mouse bone marrow-derived
   macrophages (BMDMs) and RAW264.7 macrophages in vitro. We observed that
   rFGF15 suppressed M1 macrophage polarization and associated
   inflammatory responses in both CLP-induced septic mice and
   LPS-stimulated BMDMs and RAW264.7 macrophages. Additionally,
   macrophage-depleted CLP mice transplanted with LPS-stimulated BMDMs
   pre-treated with rFGF15 exhibited reduced multi-organ inflammation and
   enhanced survival compared to those receiving LPS-stimulated BMDMs
   without rFGF15 treatment. Mechanistically, FGF15 activated the
   neurofibromin 2 (NF2)-Hippo pathway through FGFR4, leading to the
   inhibition of glycolysis, lactate production, and histone H3K18
   lactylation. This led to reduced expression of interferon regulatory
   factor 7 (Irf7), a key regulator of type I interferon responses. In
   conclusion, FGF15 suppresses M1 macrophage polarization and associated
   inflammatory responses in sepsis by activating the NF2-Hippo pathway,
   thereby inhibiting H3K18 lactylation-driven Irf7 expression. FGF15
   holds promise as a potential innovative therapy for sepsis.

   Subject terms: Inflammasome, Sepsis, Drug delivery

Introduction

   Sepsis is a life-threatening emergency that occurs when the human
   body’s immune response to an infection becomes overreactive and toxic,
   causing damages to vital organs, even death. In 2017 alone, there were
   11 million sepsis-related deaths reported globally, accounting for
   roughly 20% of all deaths worldwide [[36]1]. The management of sepsis
   involves three primary approaches: infection control, hemodynamic
   stabilization, and modulation of the septic immune response. Infection
   control is typically achieved through the timely administration of
   antibiotics, while hemodynamic stabilization is managed with fluid
   resuscitation and the use of vasoactive agents. However, no
   therapeutics targeting specific immune components of the septic
   response have consistently demonstrated improved outcomes in randomized
   clinical trials [[37]2]. The most widely used intervention in this
   category is the administration of glucocorticoids, which has a
   significant but rather limited impact on septic response and mortality.
   The discovery and development of novel therapeutics directed at
   specific components of the septic response are essential to reducing
   patient morbidity and mortality.

   Macrophages serve as the first line of defense in the innate immune
   system, primarily through their role as phagocytes, where they engulf
   and eliminate invading pathogens. They also secret inflammatory
   cytokines to recruit additional immune cells to the site of infection
   and present antigens to activate adaptive immune responses. Macrophages
   exhibit high heterogeneity and plasticity, allowing them to polarize
   into proinflammatory M1-like and anti-inflammatory M2-like phenotypes
   in response to various stimuli [[38]3]. In the early stages of sepsis,
   a predominant M1 macrophage response drives excessive inflammation and
   tissue damage [[39]4]. Researchers are evaluating the therapeutic
   potential of modulating macrophage polarization, with a focus on
   suppressing excessive M1 polarization, as an innovative treatment for
   sepsis [[40]4].

   Recent metabolic studies have revealed that macrophage polarization is
   tightly associated with cellular metabolic reprogramming [[41]5]. In
   sepsis, glycolysis supersedes oxidative phosphorylation (OXPHOS) as the
   primary energy-producing pathway, driving excessive M1 macrophage
   polarization and contributing to uncontrolled septic inflammation and
   organ damage [[42]6, [43]7]. Histone lactylation, a newly discovered
   epigenetic modification that relies on lactate, a byproduct of
   glycolysis, influences macrophage activation and polarization [[44]8,
   [45]9]. Histone lactylation is widely recognized as a crucial regulator
   of sepsis pathogenesis. Specifically, lactylated histone H3 lysine 18
   (H3K18la) was markedly elevated in patients with septic shock and its
   levels were significantly correlated with disease severity and
   prognosis [[46]10]. H3K14la enrichment at promoter regions of
   ferroptosis-related genes drives sepsis-associated pulmonary
   endothelial cell ferroptosis [[47]11]. This study was the first report
   on the role of H3K18la in regulating sepsis-related M1 macrophage
   polarization.

   Fibroblast growth factor 19 (FGF19) and its rodent ortholog FGF15 are
   hormone-like proteins secreted primarily by ileal epithelial cells in
   response to bile acid stimulation. By specifically activating the FGF
   receptor 4 (FGFR4), they regulate hepatic glucose metabolism through
   the intestine-liver crosstalk [[48]12]. FGF15 has been shown to
   activate Hippo, a signaling pathway closely integrated with glucose
   metabolism [[49]13], through neurofibromin 2 (NF2/Merlin) [[50]14].
   Intriguingly, the FGF15/FGFR4 signaling has been implicated in the
   pathogenesis of inflammatory bowel disease by regulating macrophage
   polarization [[51]15]. In 2022, Li and colleagues reported that FGF15
   markedly reduced hepatic inflammation in mice with cecal ligation and
   puncture (CLP)-induced sepsis and significantly prolonged their
   survival [[52]16]; however, the underlying mechanisms remain largely
   unclear. In the present study, we evaluated the effects of recombinant
   FGF15 (rFGF15) on macrophage polarization and multi-organ inflammation
   in mice with CLP-induced sepsis. The underlying mechanisms, including
   the NF2-Hippo signaling pathway, glycolysis, lactate production, and
   histone lactylation, were investigated. The findings from this study
   supported FGF15 as a potential therapeutic intervention for critically
   ill patients.

Methods

Animals

   Male C57BL/6 mice (7–8-week old, weighing 18–22 g) were purchased from
   Hunan Slake Jingda Experimental Animal Co., Ltd. (Changsha, China). The
   animals were acclimated for one week in a specific pathogen-free (SPF)
   environment, maintained on a 12 h light/dark cycle, with free access to
   water and food. All animal protocols were approved by the Changsha
   Hospital of Traditional Chinese Medicine (Changsha No. 8 Hospital)
   committee (the approval number:2023110804) and followed the ARRIVE
   guidelines (Animal Research: Reporting of In Vivo Experiments) and the
   Guide for the Care and Use of Laboratory Animals.

Mouse macrophage isolation

   The C57BL/6 mice were euthanized with intraperitoneal sodium
   pentobarbital (120 mg/kg). The lungs, livers, intestines, kidneys,
   femurs, and tibias were collected. For the isolation of tissue-derived
   macrophages, the lung, liver, intestine, and kidney tissues were
   digested with collagenase and/or trypsin, and the homogenates were
   filtered through a 100 μm membrane and centrifuged at 1000 rpm/min for
   5 min. Macrophages were isolated from the homogenates using Anti-F4/80
   MicroBeads (130-110-443, Miltenyi Biotech, Cologne, Germany) according
   to the manufacturer’s instructions. For the isolation of peritoneal
   macrophages, the peritoneal cavity was repeatedly washed with RPMI 1640
   medium, and the combined lavage fluid was filtered through a 100 μm
   membrane and centrifuged at 1000 rpm/min for 5 min. Macrophage
   colony-stimulating factor (M-CSF) (20 ng/mL) was added to induce
   macrophage differentiation for 7 days. For the isolation of bone
   marrow-derived macrophages (BMDMs), bone marrow cells were harvested
   from the femoral and tibial cavities by repeatedly flushing them with
   RPMI 1640 medium. The combined lavage fluid was filtered through a
   100 μm membrane and centrifuged at 1000 rpm/min for 5 min. Following
   the removal of erythrocytes using red blood cell lysis buffer, the
   remaining cells were induced with 20 ng/mL M-CSF for 7 days to generate
   BMDMs.

Cell culture and interventions

   Mouse BMDMs isolated and induced as above were cultured in RPMI 1640
   medium supplemented with 10% fetal bovine serum (FBS) and 1%
   penicillin-streptomycin (Pen-Strep). RAW264.7 mouse macrophages
   (AW-CCM002, Abiowell, Changsha, China) were maintained in Dulbecco’s
   modified Eagle’s medium (DMEM) containing 10% FBS and 1% Pen-Strep.
   RAW264.7 cells were authenticated through short tandem repeat (STR)
   profiling and tested for mycoplasma contamination. Both BMDMs and
   RAW264.7 cells were maintained at 37 °C in a humidified atmosphere
   containing 5% CO[2].

   All cells were randomly assigned to receive different drug
   interventions. Specifically, following stimulation with 100 μg/mL
   lipopolysaccharide (LPS) for 24 h, the cells were incubated with
   200 ng/mL recombinant FGF15 protein (rFGF15; CSB–EP522052MO, Cusabio,
   Wuhan, China) for 24 h as previously described [[53]16]. To explore the
   functional roles of FGFR4, NF2, and interferon regulatory factor 7
   (Irf7), the cells were transfected with FGFR4-silencing plasmids
   (si-FGFR4; HG-SM008011, HonorGene, Changsha, China), NF2-silencing
   plasmids (si-NF2; HG-SM010898, HonorGene), and Irf7-expressing plasmids
   (oe-Irf7; HG-[54]MO016850, HonorGene), respectively. The cells
   transfected with empty vector (si-NC or oe-NC) served as negative
   control. To explore the functional role of NF2 phosphorylation, the
   cells were transfected with plasmids carrying wild-type NF2 (NF2-WT;
   HG-[55]MO252250, HonorGene) or phospho-null (Y207F) NF2 mutant
   (NF2-Mut; HG-[56]MO252250-Y207F, HonorGene). All cell transfections
   were conducted using lipofectamine 2000 according to the manufacturer’s
   instructions. The sequences of si-FGFR4, si-NF#1, and si-NF#2 were
   GGCCAGATACACAGATATCAT, GAGTTCAACTGCGAGATGAAA, and
   GGGTGATAAATCTCTATCAGA, respectively. To explore the functional role of
   NF2 in repair phenotype, the cells were stimulated with 10 ng/mL
   interleukin-4 (IL-4) and 10 ng/mL IL-13 for 24 h [[57]17], followed by
   incubation with 200 ng/mL rFGF15 protein.

Mouse model of sepsis

   C57BL/6 mice were randomly subjected to cecal ligation and puncture
   (CLP) to induce a condition mimicking human sepsis as previously
   described [[58]10]. In brief, following anesthetization with 50 mg/kg
   sodium pentobarbital, a 1-cm midline incision was made in the abdomen
   to expose the cecum. The cecum was ligated midway between its distal
   pole and base and punctured twice near the distal end with a 20-gauge
   needle. A small amount of fecal matter was gently extruded, after which
   the cecum was repositioned, and the abdominal muscle layer and skin
   were sutured. Sham-operated mice were subjected to the same surgical
   procedure, excluding the CLP. After resting for 2 h, CLP mice randomly
   received tail vein injection of 400 μg/kg rFGF15 or saline alone every
   12 h for 3 days [[59]16]. The sham, CLP, and CLP + rFGF15 groups each
   contained 10 mice. The sample size was calculated using efficacy
   analyses [[60]18]. The mice were sacrificed 12 h after the final rFGF15
   injection using intraperitoneal administration of sodium pentobarbital
   (120 mg/kg). Tissue samples and peripheral blood were collected for
   molecular and histological analysis.

Macrophage depletion and transplantation in vivo

   Four days before the CLP surgery, liposomal clodronate (200 μL) was
   intravenously administered once daily to deplete macrophages [[61]19].
   BMDMs were collected and analyzed with flow cytometry pre- and
   post-CLP. For BMDM transplantation, 3 × 10^6 BMDMs from wild-type donor
   mice were pre-treated with 100 ng/mL LPS or 200 ng/mL rFGF15 for 24 h,
   or transfected with oe-Irf7 or oe-NC for 24 h, alone or in combination
   as specified. The cells were then transplanted into macrophage-depleted
   recipient mice via tail vein injection immediately before the CLP
   surgery. The sham, CLP, CLP + BMDM, CLP + BMDM^LPS,
   CLP + BMDM^LPS+rFGF15, CLP + BMDM ^LPS+rFGF15+oe-NC, and CLP + BMDM
   ^LPS+rFGF15+oe-Irf7 groups each contained 20 mice. Survival was
   assessed every 24 h over a 7-day period. For molecular and histological
   analysis, the mice were sacrificed using intraperitoneal administration
   of sodium pentobarbital (120 mg/kg) 72 h after CLP.

Histological analysis

   Fresh heart, liver, intestine, lung, and kidney tissues were fixed in
   4% paraformaldehyde, embedded in paraffin, and cut into 4 µm sections.
   The sections were stained with hematoxylin (AWI0001a, Abiowell) and
   eosin (AWI0029a, Abiowell), and tissue damage was assessed under an
   optical microscope (BA210T, Motic, Xiamen, China).

Lung edema assessment

   Fresh lung specimens were weighed to obtain wet weight. The specimens
   were subsequently dehydrated in an oven at 60 °C for 72 h and weighed
   to obtain dry weight. The wet/dry ratios were computed to reflect
   pulmonary edema.

Assessment of bacterial loads

   As previously described [[62]20], serial dilutions of mouse peripheral
   blood and peritoneal lavage fluid (PLF) were cultured on soybean-casein
   digest agar medium (146004, Millipore, Billerica, MA, USA) at 37 °C
   overnight, followed by bacterial colony-forming unit (CFU) counting.

Flow cytometry

   The cells were fixed and permeabilized using intracellular fixation and
   permeabilization buffer (eBioscience, San Diego, CA, USA), stained with
   antibodies toward target proteins according to the manufacturer’s
   instructions, and subjected to flow cytometric analysis on a flow
   cytometer (A00-1-1102, Beckman Coulter, Fullerton, CA, USA). The
   antibodies used were as follows (all from eBioscience): CD11b-FITC
   (0.5 µg/test; 11-0112-82), F4/80-APC (0.25 µg/test;12-4801-82),
   CD45-PC7 (25-0451-82), CD86-APC (0.06 µg/test; 17-0862-82), CD206-APC
   (0.25 µg/test; 17-2061-82), CD4-FITC (0.25 µg/test; 11-0041-82),
   IL-17A-PE (0.125 µg/test; 12-7177-81), CD25-APC (0.125 µg/test;
   17-0251-82), and forkhead box protein P3 (Foxp3)-PE (1 µg/test;
   12-5773-82). The M1/M2 and T-helper 17 (Th17)/T-regulatory (Treg) cell
   ratios were determined using the FlowJo software (TreeStar, Ashland,
   OR, USA). BMDMs were characterized by the expression of CD11b, CD45,
   and F4/80.

Western blot analysis

   The mouse tissues and macrophages were lysed using RIPA lysis buffer
   (R0010, Solarbio, Beijing, China). The protein concentrations were
   determined using a bicinchoninic acid (BCA) kit. The proteins were
   separated using sodium dodecyl sulphate (SDS)-polyacrylamide gel
   electrophoresis and transferred to nitrocellulose membranes. After
   blocking with TBST containing 5% skim milk, the membranes were
   incubated with primary antibodies, followed by appropriate
   HRP-conjugated secondary antibodies. Protein bands were visualized
   using an enhanced chemiluminescence (ECL) reagent and detected on a
   ChemiScope6100 system (CLiNX, Shanghai, China). The protein expression
   levels were quantified using ImageJ software (NIH, Bethesda, MD, USA)
   and normalized to β-actin. The antibodies used are presented in
   Supplementary Table [63]1.

Enzyme-linked immunosorbent assay (ELISA)

   Mouse serum was collected by centrifugation of peripheral blood
   (1000 g, 15 min) at 4 °C and stored at −80 °C. Cell lysates, culture
   medium, and serum samples were subjected to ELISA using the following
   commercial kits according to the manufacturer’s instructions: TNF-α
   (KE10002, Proteintech, Rosemont, IL, USA), IL-1β (KE10003,
   Proteintech), IL-6 (KE10007, Proteintech), COX-2 (CSB-E12910m,
   Cusabio), iNOS (CSB-E08326m, Cusabio), IL-10 (KE10008, Proteintech),
   TGF-β (KE10005, Proteintech), BNP (EM0877, Wuhan Fine Biotech Co.,
   Ltd., Wuhan, China), CK-MB (EM0929), and cTnI (JL11280, Jianglai
   Industry Co., Ltd., Shanghai, China).

High Performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS)

   The levels of glucose 6-phosphate (G6P), fructose 6-phosphate (F6P),
   3-phosphoglycerate (3PG), 2-phosphoglycerate (2PG), phosphoenolpyruvate
   (PEP), acetyl coenzyme A (acetyl-CoA), adenosine diphosphate (ADP), and
   adenosine triphosphate (ATP) were determined using HPLC-MS/MS in mouse
   BMDMs treated with vehicle (Control), LPS, or LPS + rFGF15. The
   analysis was conducted on a QTRAP 5500 mass spectrometer (SCIEX)
   coupled to a ExionLC AD UPLC system (SCIEX), which was equipped with a
   ZIC-pHILIC column (5 μm, 2.1 × 100 mm, PN: 1.50462.0001, Millipore)
   maintained at 40 °C. The loading volume was 3 μL for each sample. The
   mobile phases consisted of 25 mM ammonium acetate in 25 mM ammonia
   hydroxide (mobile phase A) and acetonitrile (90/10, v/v) (mobile phase
   B). The gradient program was as follows: 95% B for 1 min, decreased to
   65% B over 14 min, further decreased to 40% B over 16 min and held for
   18 min, increased back to 95% B over 18 min, and finally reduced back
   to 9% B and held for another 23 min. The flow rate was 0.3 mL/min. The
   mass spectra were acquired using a Turbo V ion source operating in
   negative mode, with a spray voltage of −4500 V, a source temperature of
   550 °C, a gas 1 pressure of 50 psi, a gas 2 pressure of 55 psi, and a
   curtain gas pressure of 40 psi. Metabolite peaks were detected in
   single ion monitoring mode, with delustering potentials and collision
   energies optimized using analytical standards. The m/z for the
   metabolites were as follows: 259.0/96.9 for G6P, 259.0/97.0 for F6P,
   184.9/96.9 for 3PG/2PG, 166.9/78.9 for PEP, 808.0/408.0 for acetyl-CoA,
   425.9/133.9 for ADP, and 505.9/408.0 for ATP. Data were collected using
   the Analyst 1.7.1 software (SCIEX). The relative amounts of metabolites
   were calculated using the MultiQuant 3.0.3 software (SCIEX).

   Protein extracts from mouse BMDMs treated with vehicle (Control) or LPS
   were subjected to tryptic digestion. Lactylated peptides were enriched
   following using a lactylation-specific antibody resin enrichment kit.
   Total peptides from each sample were subjected to analysis on a
   nano-UPLC system (Vanquishneo) coupled to a mass spectrometer (Astral)
   equipped with a nano-electrospray ion source. Chromatographic
   separation was performed on a 150 μm ID × 15 cm reversed-phase column
   (EASY-Spray™ HPLC, Thermo Scientific). The mobile phase consisted of
   acetonitrile-water-formic acid, with phase A containing 0.1% formic
   acid in water and phase B containing 0.1% formic acid in 80%
   acetonitrile. Mass spectrometry analysis was performed using
   data-independent acquisition (DIA) in positive ion mode. Full-scan MS1
   spectra were acquired over a range of 380–980 m/z at a resolution of
   240,000, with an isolation window of 2 m/z and normalized collision
   energy (NCE) of 25%. MS2 spectra were collected from 150 to 2000 m/z
   using absolute AGC values of 5E4 and maximum injection times of 3 ms.
   Raw spectral data were processed using the Spectronaut software
   (version 19.0.240604.62635; Biognosys AG).

CUT&Tag assays

   CUT&Tag analysis was performed on mouse BMDMs treated with LPS or LPS +
   rFGF15. Cells were immobilized using concanavalin A-coated magnetic
   beads and permeabilized with the nonionic detergent digitonin. The
   cells were subsequently incubated with anti-H3K18la antibody (1:100;
   PTM-1427RM, PTM BIO) followed by secondary antibody. Protein A/G-Tn5
   (pA/G-Tn5) was introduced to liberate antibody-bound DNA fragments from
   chromatin into extracellular space. DNA was extracted and fragments
   below 700 bp were isolated via magnetic bead-based size selection. PCR
   amplification was conducted followed by quality verification using 1%
   agarose gel electrophoresis. Sequencing libraries were subjected to
   high-throughput sequencing. Raw data underwent quality control through
   the fastp software (version 0.23.1). Clean reads were aligned to the
   reference genome (Ensembl_GRCm39_111). Broad peak scanning was
   implemented using MACS2 to identify chromosomal peak distributions.
   This process ultimately confirmed H3K18la enrichment at the Irf7 loci.

Biochemical assays

   Lactate concentrations in the culture medium were determined using a
   lactic acid test kit (A019-2-1, Nanjing Jiancheng Bioengineering
   Institute, Nanjing, China). Glucose concentrations in cell lysates were
   determined using a glucose uptake colorimetric assay kit (MAK083,
   Sigma-Aldrich, St. Louis, MO, USA). Serum samples were subjected to
   functional marker detection using the following commercial kits. Scr
   (C011-2-1), BUN (C013-2-1), ALT (C009-2-1), TB (C019-1-1), ALB
   (A028-2-1) were purchased from Nanjing Jiancheng Bioengineering
   Institute; PT (BC8081) was purchased from Solarbio (Beijing, China).
   The assays were conducted following the manufacturer’s instructions.

Co-immunoprecipitation (Co-IP)

   The cells were lysed with IP lysis buffer. To explore the interaction
   between FGFR4 and NF2, the cell lysates were incubated with anti-FGFR4
   antibody (1:50; 8562S, CST, Danvers, MA, USA) or IgG (B900610,
   Proteintech) overnight at 4 °C, followed by Protein A/G agarose beads
   for 4 h at 4 °C. The agarose beads were washed thoroughly with IP lysis
   buffer to remove non-specific proteins, and the resulting
   immunoprecipitated protein complexes were subjected to western blot
   analysis.

RNA sequencing (RNA-seq)

   RNA-seq and data analysis were conducted by Shanghai OE Biotech Co.,
   Ltd. RNA-seq was performed on BMDMs treated with vehicle (Control),
   LPS, or LPS + rFGF15. Three independent samples were analyzed for each
   group. Sequence alignment was performed using HISAT2 software, while
   read counts were registered with HTSeq-count. DESeq2 software was used
   for differential analysis and visualization of gene read counts for
   each sample. The raw data can be found in NCBI
   ([64]https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1281990).

Quantitative real-time PCR (qRT-PCR)

   Total RNA was isolated using Trizol reagent (15596026, Thermo
   Scientific, Portsmouth, NH, USA). Following quantification on a Micro
   Drop ultra microspectrophotometer (BIO-DL, Shanghai, China), the RNA
   was reverse transcribed into cDNA using the HiFiScript cDNA Synthesis
   Kit (CW2569M, CWBIO, Taizhou, China). The qRT-PCR was performed using
   UltraSYBR Mixture (CW2601M, CoWin Biosciences) on a QuantStudio1
   Real-Time PCR System (Thermo Scientific). The primers used for
   amplification are listed in Supplementary Table [65]2. The mRNA
   expression levels were determined using the 2^−ΔΔCt method and
   normalized to β-actin.

Chromatin immunoprecipitation (ChIP)

   The enrichment of H3K18la at the promoter regions of Irf7, Ndrg1, Odc1,
   and Rsad2 was analyzed by ChIP in combination with qRT-PCR. In brief,
   3 × 10^6 cells were treated with 1.1% formaldehyde to crosslink
   chromatin proteins to DNA, and the crosslinking reaction was terminated
   by the addition of glycine. To isolate genomic DNA, the cells were
   resuspended in lysis buffer, and the chromatin was fragmented by
   sonication. The size of the DNA fragments was assessed using a 1.5%
   agarose gel. The cell lysates containing fragmented DNA with
   crosslinked chromatin proteins were incubated with anti-lgG
   (6 μg/5 × 10^6 cells; 30000-0-AP, Proteintech), anti-H3K18la
   (6 μg/5 × 10^6 cells; ab40888, Abcam, Cambridge, UK), or anti-H3
   antibody (6 μg/5 × 10^6 cells; ab1791, Abcam). The immunoprecipitated
   DNA-protein complexes were treated with proteinase K, and the purified
   DNA fragments were analyzed by qRT-PCR using specific primers for the
   Irf7, Ndrg1, Odc1, and Rsad2 promoters (Supplementary Table [66]2).

Statistical analysis

   Data are presented as mean ± standard deviation (SD). Statistical
   analysis was performed using GraphPad Prism software (version 9).
   Normality of distribution was assessed using the Shapiro-Wilk test.
   Comparisons among multiple groups were conducted using analysis of
   variance (ANOVA), followed by Tukey’s multiple comparisons. Survival
   was analyzed using the Kaplan-Meier method. All experiments were
   randomized and blindly analyzed to minimize experimental bias. A
   two-tailed p value of less than 0.05 was considered statistically
   significant.

Results

FGF15/FGFR4 signaling inhibits glycolysis in septic macrophages

   To explore the function of FGF5/FGFR4 signaling in septic macrophages,
   we collected macrophages from major organs, peritoneum, and bone marrow
   of CLP-induced septic and sham control mice. Western blot analysis
   revealed a significant downregulation of the phospho-FGFR4
   (p-FGFR4)/FGFR4 ratio in macrophages derived from the liver,
   peritoneum, and bone marrow of CLP septic mice (Fig. [67]1A, [68]B).
   The administration of rFGF15 through tail vein injection restored the
   p-FGFR4/FGFR4 ratio in CLP mice (Fig. [69]1A, [70]B). HPLC-MS/MS
   analysis demonstrated a significant increase in glycolytic metabolites,
   including G6P, F6P, 3PG/2PG, PEP, acetyl-CoA, ADP, and ATP, in mouse
   BMDMs treated with LPS, indicating glycolytic upregulation. Treatment
   with rFGF15 normalized these glycolytic metabolites (Fig. [71]1C). In
   in vitro experiments, LPS stimulation of mouse BMDMs and RAW264.7
   macrophages resulted in increased glucose uptake and lactate production
   (Fig. [72]1D, [73]E), accompanied by upregulated protein levels of key
   glycolytic enzymes, including HK1, HK2, HK3, PFKM, PFKP, PFKL, PFKFB1,
   PFKFB2, PFKFB3, PFKFB4, PKM2, LDHA, LDHB, and LDHC (Fig. [74]S1).
   Additionally, the expression of monocarboxylate transporter 1 (MCT1), a
   rate-limiting protein for lactate transmembrane transport [[75]21], was
   elevated (Fig. [76]1F). Notably, LPS-induced upregulation of glucose
   uptake, lactate production, and glycolytic enzyme and MCT1 expression
   were all reversed by treatment with rFGF15 (Fig. [77]1D–F, Fig.
   [78]S1). These in vivo and in vitro results suggested that FGF15/FGFR4
   signaling inhibits glycolysis in septic macrophages.

Fig. 1. FGF15/FGFR4 signaling inhibits glycolysis in septic macrophage.

   [79]Fig. 1
   [80]Open in a new tab

   A, B Detection of p-FGFR4 and FGFR4 in macrophages derived from the
   intestine, lung, kidney, liver, peritoneum, and bone marrow of Sham,
   CLP, and rFGF15-treated CLP (CLP+rFGF15) mice by western blot analysis.
   A Immunoblot images. B Quantified p-FGFR4/FGFR4 ratios. n = 9,
   *p < 0.05. C Detection of glycolytic metabolites in control,
   LPS-stimulated (LPS), and rFGF15-treated, LPS-stimulated (LPS + rFGF15)
   mouse BMDMs by HPLC-MS/MS analysis. n = 3, *p < 0.05. D, E Glucose
   uptake (D) and lactate production (E) in control, LPS, and LPS + rFGF15
   mouse BMDMs or RAW264.7 macrophages. F Detection of MCT1 protein
   expression in mouse BMDMs and RAW264.7 macrophages treated with
   control, LPS, or LPS + rFGF15 by western blot analysis. n = 6,
   *p < 0.05.

FGF15/FGFR4 signaling downregulates histone lactylation in septic macrophages

   Histone lactylation is a newly discovered epigenetic modification that
   relies on lactate, a byproduct of glycolysis [[81]8], and is associated
   with the severity and mortality of sepsis [[82]10]. LC-MS/MS analysis
   identified a number of BMDM histone lactylations affected by LPS,
   including H4K59la, H2K99la, H4K31la, H1K16la, H2K5la, H3K56la, H1K33la,
   and H2K95la (downregulation), as well as H3K23la, H3K18la, H4K8la, and
   H4K5la (upregulation) (Fig. [83]S2, Supplementary file [84]1). Having
   found that rFGF15 inhibited glycolysis and lactate production in
   LPS-stimulated mouse macrophages, we examined its effects on histone
   lactylation, including the lactylation of all histones and at specific
   histone lactylation sites including H3K23, H3K18, H4K8, and H4K5, in
   septic macrophages in vivo and in vitro. Western blot analysis revealed
   an increase in total lactylated histones, including H3K18la and
   H3K23la, in BMDMs from CLP-induced septic mice. Treatment with rFGF15
   effectively reduced the levels of these lactylated proteins (Fig.
   [85]2A), with more pronounced changes observed for H3K18la in the
   treatment group. In in vitro experiments, LPS stimulation of mouse
   BMDMs and RAW264.7 macrophages resulted in increased levels of total
   lactylated histones and H3K18la. Treatment with rFGF15 reduced the
   levels of these lactylated proteins, an effect that was abolished by
   FGFR4 knockdown (Fig. [86]2B–D). These in vivo and in vitro results
   suggested that FGF15/FGFR4 signaling downregulates histone lactylation
   in septic macrophages.

Fig. 2. FGF15/FGFR4 signaling regulates histone lactylation in septic
macrophages.

   [87]Fig. 2
   [88]Open in a new tab

   A Detection of total lactylated histones (Pan-Kla), H3K18la, H3K23la,
   H3, H4K5la, H4K8la, and H4 in BMDMs of Sham, CLP, and rFGF15-treated
   CLP (CLP + rFGF15) mice by western blot analysis. n = 9, *p < 0.05. B,
   C Detection of Pan-Kla and H3K18la in mouse BMDMs following treatment
   with vehicle (Control), LPS, or LPS + rFGF15 for 6, 12, 24, 48, or 72 h
   by western blot analysis. B Immunoblot images. C Quantified Pan-Kla and
   H3K18la levels. D Detection of Pan-Kla and H3K18la in mouse BMDMs and
   RAW264.7 macrophages following treatment with vehicle (Control), LPS,
   LPS + rFGF15, LPS + rFGF15 + si-NC, or LPS + rFGF15 + si-FGFR4 for 48 h
   by western blot analysis. n = 6, *p < 0.05.

FGF15/FGFR4 signaling inhibits M1 polarization of septic macrophages and
their inflammatory responses

   Next, we investigated the role of FGF15/FGFR4 signaling in septic
   macrophage polarization and inflammatory response. We detected elevated
   serum levels of the pro-inflammatory mediators TNF-α, IL-1β, IL-6,
   COX-2, and iNOS, along with decreased levels of the anti-inflammatory
   mediators IL-10 and TGF-β, in CLP-induced septic mice. These typical
   features of M1-associated inflammation in septic mice were attenuated
   by rFGF15 administration (Fig. [89]3A). In in vitro experiments, LPS
   stimulation of mouse BMDMs and RAW264.7 macrophages resulted in an
   increased ratio of CD86 + M1-like macrophages to CD206 + M2-like
   macrophages (Fig. [90]3B). This LPS-induced M1 polarization was
   associated with cellular inflammation, characterized by upregulated
   pro-inflammatory mediators and downregulated anti-inflammatory
   mediators (Fig. [91]3C). Notably, treatment with rFGF15 restored the
   M1/M2 ratio and mitigated cellular inflammation, effects that were
   reversed by FGFR4 knockdown (Fig. [92]3B, [93]C). These in vivo and in
   vitro results suggested that FGF15/FGFR4 signaling inhibits M1
   polarization of septic macrophages and their inflammatory responses.

Fig. 3. FGF15/FGFR4 signaling inhibits M1 polarization of septic macrophages
and their inflammatory responses.

   [94]Fig. 3
   [95]Open in a new tab

   A Detection of the pro-inflammatory mediators TNF-α, IL-1β, IL-6,
   COX-2, and iNOS and anti-inflammatory mediators IL-10 and TGF-β in
   peripheral blood of Sham, CLP and rFGF15-treated CLP (CLP + rFGF15)
   mice by ELISA. n = 9, *p < 0.05. B Detection of CD86 (a marker of
   M1-like macrophages) and CD206 (a marker of M2-like macrophages) in
   mouse BMDMs and RAW264.7 macrophages following treatment with vehicle
   (Control), LPS, LPS + rFGF15, LPS + rFGF15 + si-NC, or LPS + rFGF15 +
   si-FGFR4 by flow cytometry. C Detection of the pro-inflammatory
   mediators TNF-α, IL-1β, IL-6, COX-2, and iNOS and anti-inflammatory
   mediators IL-10 and TGF-β in mouse BMDMs and RAW264.7 macrophages
   following treatment with vehicle (Control), LPS, LPS + rFGF15, LPS +
   rFGF15 + si-NC, or LPS + rFGF15 + si-FGFR4 by ELISA. n = 6, *p < 0.05.

FGF15/FGFR4 signaling activates the NF2-Hippo pathway in septic macrophages

   FGF15 has been shown to stimulate hepatic FGFR4 to recruit and
   phosphorylate NF2, which in turn activates the Hippo kinases MST1/2,
   suppressing liver tumorigenesis [[96]14]. MOB1 and YAP are key
   components of the Hippo signaling cascade. The Hippo kinases MST1/2
   phosphorylate and activate MOB1, which in turn phosphorylates YAP. This
   leads to YAP’s retention in the cytoplasm, preventing its nuclear
   translocation and the expression of its target genes. Given the close
   integration of the NF2-Hippo pathway with glucose metabolism [[97]13],
   we speculated that rFGF15 inhibited glycolysis in septic macrophages by
   modulating the NF2-Hippo pathway. Indeed, the Co-IP results revealed
   direct association between FGFR4 and NF2 in mouse BMDMs and RAW264.7
   macrophages (Fig. [98]4A). LPS stimulation resulted in reduced levels
   of phospho-NF2 (p-NF2), MST1/2, phospho-MOB1 (p-MOB1), and phospho-YAP
   (p-YAP) (Fig. [99]4B–E), indicating inhibition of the NF2-Hippo
   pathway. Treatment with rFGF15 restored levels of these proteins, an
   effect abolished by FGFR4 knockdown (Fig. [100]4B–E). We induced in
   vitro BMDM polarization toward M2 using IL-4/IL-13, demonstrating that
   IL-4/IL-13 stimulated NF2 and MOB1 phosphorylation without affecting
   MST1, MST2, or YAP phosphorylation. In contrast, rFGF15 treatment
   increased NF2/MOB1 phosphorylation and MST1/MST2 expression while
   maintaining unchanged YAP phosphorylation (Fig. [101]S3A, [102]S3B).
   This indicated the activation of NF2-Hippo signaling in macrophages by
   FGF15. Further Co-IP experiments with the phospho-null (Y207F) NF2
   mutant verified that NF2 phosphorylation was required for FGFR4-NF2
   association and the activation of the Hippo pathway (Fig. [103]4F, G).
   Collectively, these results demonstrated that FGF15 activates FGFR4 in
   septic macrophages, leading to the recruitment and phosphorylation of
   NF2, and subsequent activation of the Hippo pathway.

Fig. 4. FGF15/FGFR4 signaling activates the NF2-Hippo pathway in septic
macrophages.

   [104]Fig. 4
   [105]Open in a new tab

   A Detection of the interaction between FGFR4 and NF2 in mouse BMDMs and
   RAW264.7 macrophages by Co-IP. B, C Detection of p-NF2 and NF2 in mouse
   BMDMs and RAW264.7 macrophages following treatment with vehicle
   (Control), LPS, or LPS + rFGF15 (B) or LPS + rFGF15, LPS + rFGF15 +
   si-NC, or LPS + rFGF15 + si-FGFR4 (C) by western blot analysis. D, E
   Detection of MST1/2, p-MOB1, MOB1, p-YAP, and YAP in mouse RAW264.7
   macrophages (D) and BMDMs (E) following treatment with vehicle
   (Control), LPS, LPS + rFGF15, LPS + rFGF15 + si-NC, or LPS + rFGF15 +
   si-FGFR4 by western blot analysis. F Detection of the interaction
   between FGFR4 and NF2 in mouse RAW264.7 macrophages overexpressing the
   wildtype NF2 (NF2-WT) or phospho-null NF2 mutant (NF2-Mut) following
   treatment with LPS + rFGF15 by Co-IP. G Detection of MST1/2, p-MOB1,
   MOB1, p-YAP, and YAP in mouse RAW264.7 macrophages overexpressing
   NF2-WT or NF2-Mut following treatment with LPS + rFGF15 by western blot
   analysis. n = 6, *p < 0.05.

FGF15/FGFR4 signaling inhibits glycolysis, H3K18 lactylation, M1
polarization, and inflammation of septic macrophages through the NF2-Hippo
pathway

   Having demonstrated that FGF15/FGFR4 signaling activates the NF2-Hippo
   pathway in septic macrophages, we hypothesized that FGF15 inhibits
   glycolysis, thereby reducing lactate production and histone lactylation
   in these cells, through modulation of the NF2-Hippo pathway. To verify
   this hypothesis, we assessed the effects of phospho-null NF2 mutant by
   NF2-Mut transfection (Fig. [106]5A). NF2-Mut abolished the inhibitory
   effects of rFGF15 on LPS-induced glucose uptake, glycolytic enzyme
   expression, lactate production, and H3K18 lactylation in mouse RAW264.7
   macrophages (Fig. [107]5A–D). NF2-Mut also eliminated the inhibitory
   effects of rFGF15 on LPS-induced M1 polarization and cellular
   inflammation (Fig. [108]5E, [109]F). Similar to NF2 mutant, NF2
   knockdown by si-NF2 transfection abolished the inhibitory effects of
   rFGF15 on LPS-induced glucose uptake, lactate production, H3K18
   lactylation, M1 polarization, and cellular inflammation in mouse BMDMs
   and RAW264.7 macrophages (Fig. [110]S4A–[111]S4G). These data
   collectively demonstrated that FGF15/FGFR4 signaling inhibits
   glycolysis, H3K18 lactylation, M1 polarization, and inflammation in
   septic macrophages by activating the NF2-Hippo pathway.

Fig. 5. FGF15/FGFR4 signaling inhibits glycolysis, H3K18 lactylation, M1
polarization, and inflammation of septic macrophages through the NF2-Hippo
pathway.

   [112]Fig. 5
   [113]Open in a new tab

   A–F Assessment of glucose uptake (A), lactate production (B),
   glycolytic enzyme expression (C), H3K18 lactylation (D), M1/M2
   polarization (E), and pro- and anti-inflammatory mediator production
   (F) in mouse RAW264.7 macrophages overexpressing the wildtype NF2
   (NF2-WT) or phospho-null NF2 mutant (NF2-Mut) following treatment with
   LPS or LPS + rFGF15. n = 6, *p < 0.05.

FGF15/FGFR4 signaling suppresses M1 polarization of septic macrophages
through NF2-dependent downregulation of H3K18la-driven Irf7 expression

   Histone lactylation is an epigenetic modification that regulates gene
   expression in macrophages [[114]8, [115]9]. To identify genes
   influenced by histone lactylation in septic macrophages, we assessed
   gene expression in control, LPS-stimulated, and rFGF15-treated,
   LPS-stimulated mouse BMDMs by RNA-seq. A total of four genes were
   identified as differentially expressed genes (DEGs) among the following
   comparisons: control vs. LPS, LPS vs. LPS + rFGF15, and control vs. LPS
   + rFGF15 (Fig. [116]6A). Out of the four DEGs identified by RNA-seq,
   Irf7 and radical S-adenosyl methionine domain containing 2 (Rsad2),
   both of which play key roles in type I interferon response [[117]22],
   were verified by qRT-PCR. LPS induced both Irf7 and Rsad2 mRNA
   expression in mouse BMDMs and RAW264.7 macrophages, effects inhibited
   by rFGF15 treatment (Fig. [118]6B). Subsequent ChIP-qPCR analysis
   demonstrated increased enrichment of H3K18la at the promoter region of
   Irf7, but not Rsad2, in response to LPS stimulation (Fig. [119]6C). To
   identify H3K18la-regulated downstream genes in BMDMs, we performed
   CUT&Tag assays on LPS-stimulated BMDMs treated with or without rFGF15.
   We observed H3K18la enrichment at the Irf7 gene promoter after LPS
   stimulation, which was attenuated by rFGF15 treatment (Fig. [120]6D).
   This was validated through ChIP-qPCR analysis in both mouse BMDMs and
   RAW264.7 macrophages (Fig. [121]6E). Notably, the inhibitory effects of
   rFGF15 on LPS-induced H3K18la enrichment and Irf7 expression were
   attenuated by NF2 knockdown or ectopic expression of phosphor-null NF2
   mutant (Figs. [122]6F, [123]G, S5A–S5C). In addition, Irf7
   overexpression in mouse BMDMs and RAW264.7 macrophages partially
   reversed the inhibitory effects of rFGF15 on LPS-induced M1
   polarization and cellular inflammation (Figs. [124]6H, I, S6A,
   [125]S6B). These data collectively indicated that FGF15/FGFR4 signaling
   suppresses M1 polarization of septic macrophages and their inflammatory
   responses through NF2-dependent inhibition of H3K18la-driven Irf7
   expression.

Fig. 6. FGF15/FGFR4 suppresses H3K18la-driven Irf7 expression and
inflammatory responses in septic macrophages via NF2.

   [126]Fig. 6
   [127]Open in a new tab

   A Venn diagram showing the identification of DEGs between mouse BMDMs
   subjected to specific treatments by RNA-seq. B Detection of Irf7,
   Ndrg1, Odc1, and Rsad2 mRNA expression in mouse BMDMs and RAW264.7
   macrophages following treatment with vehicle (Control), LPS, or LPS +
   rFGF15 by qRT-PCR. C Fold changes in H3K18la enrichment at the promoter
   regions of Irf7, Ndrg1, Odc1, and Rsad2 in mouse BMDMs and RAW264.7
   macrophage in response to LPS stimulation by ChIP-qPCR analysis. D
   Dectection of H3K18la enrichment at the promoter regions of Irf7 in
   mouse BMDMs following treatment with LPS or LPS + rFGF15 using the
   CUT&Tag assay. E Fold changes in H3K18la enrichment at the promoter
   region of Irf7 in mouse BMDMs and RAW264.7 macrophages following
   treatment with vehicle (Control), LPS, or LPS + rFGF15 by ChIP-qPCR
   analysis. F Fold changes in H3K18la enrichment at the promoter region
   of Irf7 in RAW264.7 macrophages overexpressing the wildtype NF2
   (NF2-WT) or phospho-null NF2 mutant (NF2-Mut) following treatment with
   LPS or LPS + rFGF15 by ChIP-qPCR analysis. G Detection of Irf7 mRNA and
   protein expression in RAW264.7 macrophages overexpressing the wildtype
   NF2 (NF2-WT) or phospho-null NF2 mutant (NF2-Mut) following treatment
   with LPS or LPS + rFGF15 by qRT-PCR and western blot analysis. H, I
   Assessment of M1/M2 polarization (H) and pro- and anti-inflammatory
   mediator production (I) in wildtype and Irf7-overexpressing mouse
   RAW264.7 macrophages following treatment with LPS or LPS + rFGF15 by
   flow cytometry and ELISA. n = 6, *p < 0.05.

FGF15 mitigates multi-organ inflammation and improves survival of septic mice
by suppressing Irf7-mediated BMDM activation

   To validate whether our in vitro findings in LPS-stimulated macrophages
   translate to septic mice, we induced macrophage depletion in vivo using
   clodronate liposomes, followed by the transplantation of mouse BMDMs
   subjected to the following treatments: vehicle, LPS, LPS + rFGF15, LPS
   + rFGF15 + oe-NC, or LPS + rFGF15 + oe-Irf7. Intravenous injection of
   clodronate liposomes reduced the proportion of bone marrow macrophages
   (CD11b + F4/80 + ) to approximately 2%, both pre- and post-CLP (Fig.
   [128]S7). Macrophage-depleted mice showed reduced survival after CLP,
   which was restored by BMDM transplantation following the order of LPS +
   rFGF15 ≈ LPS + rFGF15 + oe-NC > LPS + rFGF15 +
   oe-Irf7 > vehicle > LPS ≈ no transplantation (Fig. [129]7A).
   Transplantation of untreated BMDMs effectively reduced bacterial loads
   in both blood and PLF of macrophage-depleted CLP mice. However,
   transplantation of LPS-stimulated BMDMs restored bacterial infection,
   an effect which was alleviated by rFGF15 treatment but worsened by Irf7
   overexpression (Fig. [130]7B). Multi-organ dysfunction represents the
   most severe complication of sepsis. Immune cell infiltration was
   observed in the heart, liver, intestine, lung, and kidney in
   macrophage-depleted CLP mice, which was reduced by BMDM
   transplantation, correlating with survival (Fig. [131]7C). Treatment
   with rFGF15 reduced LPS-induced immune cell infiltration, an effect was
   reversed by Irf7 overexpression (Fig. [132]7C). Pulmonary assessment
   was performed by measuring dry/wet weight ratios. Macrophage-depleted
   CLP mice exhibited increased lung dry/wet ratios, which decreased
   following BMDM transplantation. LPS pretreatment exacerbated pulmonary
   edema, while rFGF15 effectively reduced it; however, this effect was
   reversed upon Irf7 overexpression (Fig. [133]S8A). For renal function,
   Scr and BUN levels - indicators of glomerular filtration capacity -
   were elevated in macrophage-depleted CLP mice. In contrast, BMDM
   transplantation reduced Scr and BUN levels, which increased again after
   LPS treatment. Treatment with rFGF15 attenuated LPS-induced Scr and BUN
   elevations, an effect abolished by Irf7 overexpression (Fig. [134]S8B,
   [135]S8C). Hepatic evaluation included PT and plasma biomarkers,
   including ALB, ALT, and TB. BMDM transplantation improved hepatic
   function by prolonging PT, elevating ALB and TB levels, and reducing
   ALT, effects reversed by LPS pretreatment. Under LPS conditions, Irf7
   overexpression disrupted rFGF15-mediated hepatoprotection (Fig.
   [136]S8D–[137]S8G). Intestinal barrier integrity was assessed through
   tight junction proteins (ZO-1, Occludin, Claudin-1). CLP downregulated
   colonic expression of these proteins in macrophage-depleted mice, which
   was restored by BMDM transplantation. rFGF15 upregulated LPS-suppressed
   tight junction proteins, whereas Irf7 overexpression counteracted this
   effect (Fig. [138]S8H–[139]S8K). Quantitative analysis of serum cardiac
   biomarkers demonstrated that BMDM transplantation reduced accumulations
   of BNP, CK-MB, and cTnI in macrophage-depleted CLP mice. Conversely,
   LPS-upregulated BNP, CK-MB, and cTnI levels were reduced by rFGF15
   administration, effects that were reversed with Irf7 overexpression
   (Fig. [140]S8L–[141]S8N). These in vivo data demonstrated that FGF15
   alleviates multi-organ dysfunction and improves survival of septic mice
   by suppressing Irf7-driven BMDM activation.

Fig. 7. FGF15 protects mice against sepsis-induced multi-organ inflammation
by suppressing Irf7-mediated BMDM activation.

   [142]Fig. 7
   [143]Open in a new tab

   Macrophage-depleted mice were subjected to the following treatments:
   Sham, CLP, CLP + transplantation with BMDMs (CLP + BMDM), CLP +
   transplantation with LPS-stimulated BMDMs (CLP + BMDM^LPS), CLP +
   transplantation with rFGF15-treated, LPS-stimulated BMDMs
   (CLP + BMDM^LPS + rFGF15), CLP + transplantation with rFGF15-treated,
   LPS-stimulated BMDMs transfected with empty vector (CLP + BMDM^LPS +
   rFGF15 + oe-NC), or CLP + transplantation with rFGF15-treated,
   LPS-stimulated BMDMs transfected with Irf7-expressing vector
   (CLP + BMDM^LPS + rFGF15 + oe-Irf7). A Survival analysis. n = 10,
   *p < 0.05. B Bacterial loads in peripheral blood and peritoneal lavage
   fluid (PLF). n = 9, *p < 0.05. C HE-stained images of the heart, liver,
   intestine, lung, and kidney showing immune cell infiltration. D, E
   Detection of Th17 (CD4 + IL17A + ) and Treg cells
   (CD4 + CD25 + FOXP3 + ) in peripheral blood by flow cytometry. D
   Density plots. E Quantified Th17/Treg ratio. n = 6, *p < 0.05. F
   Detection of TNF-α, IFN-γ, IL-6, and IL-17A in peripheral blood by
   ELISA. n = 10, *p < 0.05.

   M1 polarization of macrophages can affect the balance between
   proinflammatory Th17 cells and immunosuppressive Treg cells [[144]23].
   Compared with macrophage-depleted sham mice, macrophage-depleted CLP
   mice exhibited a higher ratio of Th17 cells (CD4 + IL-17A + ) to Treg
   cells (CD4 + CD25 + FOXP3 + ) in peripheral blood, accompanied by
   increased levels of pro-inflammatory cytokines TNF-α, IFN-γ, IL-6, and
   IL-17A (Fig. [145]7D–F). BMDM transplantation reduced the Th17/Treg
   ratio and pro-inflammatory cytokine levels in macrophage-depleted CLP
   mice, mirroring the observed patterns of immune cell infiltration (Fig.
   [146]7D–F). Collectively, these data indicated that FGF15 mitigates
   multi-organ inflammation of septic mice by suppressing Irf7-driven BMDM
   activation.

Discussion

   If left untreated, sepsis can progress rapidly to septic shock, causing
   multiple organ failure and potentially death. A predominant M1
   macrophage response plays a key role in driving early septic
   inflammation and disease progression [[147]4]. In this study, we found
   that rFGF15 mitigated inflammatory damage and dysfunction in major
   organs and improved survival of CLP-induced septic mice. Mechanistic
   investigations in vitro and in vivo demonstrated that, through its
   receptor FGFR4, rFGF15 inhibits H3K18la-driven Irf7 expression through
   NF2-Hippo to suppress septic macrophage M1 polarization and
   inflammation. These findings highlight the pathogenic role of M1
   macrophage polarization in sepsis and support rFGF15 as a potential
   innovative therapy for this life-threatening condition.

   Previous studies have found that FGFR4 deficiency or inactivation in
   mouse BMDMs aggravated LPS-induced M1 polarization and inflammation,
   and myeloid-specific FGFR4 deletion exacerbated colitis in mice
   [[148]15]. RNA-seq in combination with KEGG pathway enrichment analysis
   revealed that FGFR4-deficiency in mouse BMDMs led to
   pentraxins3-mediated activation of the complement cascade through NF-κB
   [[149]15]. In our in vitro experiments, we found that rFGF15 stimulated
   FGFR4 to bind and phosphorylate NF2 in LPS-stimulated mouse BMDMs and
   RAW264.7 macrophages, thereby activating the Hippo pathway and
   subsequently inhibiting glycolysis, lactate production, and the
   lactylation of histone H3K18. This led to reduced Irf7 expression,
   resulting in decreased M1 polarization and associated inflammation. The
   interactions between FGFR4 and p-NF2 and between H3K18la and the Irf7
   promoter were confirmed by Co-IP and ChIP-qPCR analysis, respectively.
   Importantly, our in vivo BMDM transplantation experiments demonstrated
   that rFGF15-mediated inhibition of BMDM M1 polarization provided
   complete protection against CLP-induced sepsis in mice. These findings
   suggest that rFGF15 may serve as an effective early intervention for
   sepsis by modulating macrophage polarization.

   The Hippo pathway is an evolutionarily conserved signaling network that
   regulates numerous biological processes. As such, its dysregulation
   contributes to the pathogenesis of a variety of diseases, such as
   cancer and immune dysfunctions [[150]24, [151]25]. Specifically,
   MST1/2-deficient mice exhibited increased susceptibility to CLP-induced
   sepsis [[152]26]. Mechanistically, MST1 and MST2 promoted ROS
   production and phagocytic activity of mouse BMDMs by regulating
   mitochondrion-phagosome juxtaposition [[153]26]. Moreover, YAP1
   knockout aggravated CLP-induced liver and lung inflammatory injury by
   accelerating ferroptosis [[154]27, [155]28]. In this study, we
   discovered that rFGF15-induced activation of the NF2-Hippo pathway in
   septic macrophages downregulated glycolysis and lactate production,
   leading to a reduction in H3K18 lactylation and H3K18la-driven Irf7
   expression, ultimately attenuating M1 polarization. Our findings
   further underscore the pivotal role of the Hippo pathway in sepsis,
   specifically in controlling gene expression by modulating glycolysis
   and histone lactylation in septic macrophages.

   In the early stage of sepsis, pro-inflammatory stimuli promote
   macrophage polarization toward the M1-like phenotype, which
   predominantly depends on aerobic glycolysis to sustain its energy
   demands for phagocytosis and inflammatory responses [[156]6, [157]7].
   Suppression of glycolysis through inhibition of sphingosine kinase 1
   promoted M2 macrophage polarization and reduced septic inflammation in
   vitro and in vivo [[158]29]. Histone lactylation, an epigenic
   modefication directly linked to glycolysis, has been increasingly
   recognized for its role in regulating macrophage polarization,
   particularly in tumorigenesis [[159]9]. However, how histone
   lactylation affects macrophage polarization in sepsis remains largely
   unclear. In this study, we discovered that septic stimuli enhanced
   glycolysis and histone lactylation in mouse macrophages in vitro and in
   vivo, resulting in increased H3K18la enrichment at the promoter region
   of Irf7, a transcription factor that plays a crucial role in M1
   polarization of septic macrophages [[160]30]. FGF15 protected against
   septic inflammation by inhibiting the signaling cascade of glycolysis,
   histone lactylation, Irf7 induction, and M1 polarization through
   NF2-Hippo activation. In line with previous reports [[161]31, [162]32],
   our findings further supported targeting macrophage polarization as a
   promising therapeutic approach for mitigating sepsis.

   This study had several limitations. In addition to histones,
   lactylation modifications are also widely present in various
   non-histones to coordinate transcription [[163]33]. To our knowledge,
   only a few non-histone proteins - HMGB1 [[164]34], Fis1 [[165]35], and
   Ezrin [[166]36] - have been reported to undergo lactylation in the
   context of sepsis. We did not explore the differential roles of histone
   versus non-histone lactylation in sepsis and the role of non-histone
   proteins in macrophages. Secondly, although we verified that H3K18la
   binds to the promoter region of Irf7 in septic macrophages, we did not
   identify the specific binding site sequences. Thirdly, our mechanistic
   studies investigated only the regulation of glycolysis and histone
   lactylation mediated by the NF2-Hippo pathway. In future studies, we
   plan to identify putative specific H3K18la binding sites within the
   Irf7 promoter region through bioinformatics analysis. We also would
   investigate how mutations at the putative H3K18la-binding sites would
   affect Irf7 expression and septic inflammation. Previous studies have
   shown that FGFR4 deficiency in mouse BMDMs aggravated LPS-induced M1
   polarization and inflammation by activation of the complement cascade
   through NF-κB [[167]15]. The functional diversity of Hippo signaling in
   macrophage activities suggests that it may be involved in different
   cascades [[168]37, [169]38]. Thus, we plan to investigate the effects
   of rFGF15-Hippo on the complement system in septic macrophages in
   future experiments.

   In conclusion, rFGF15 mitigated multi-organ inflammation and improved
   survival of CLP-induced septic mice by suppressing macrophage M1
   polarization. Mechanistically, rFGF15 activated the NF2-Hippo pathway
   through FGFR4 in septic microphages, leading to reduced glycolysis,
   lactate production, H3K18 lactylation, and H3K18la-driven Irf7
   expression. FGF15 holds promise as a potential innovative therapy for
   sepsis.

Supplementary information

   [170]41419_2025_7962_MOESM1_ESM.tif^ (3.7MB, tif)

   Original Western Blot Images in Figure 1-2
   [171]41419_2025_7962_MOESM2_ESM.tif^ (3.5MB, tif)

   Original Western Blot Images in Figure 4-6
   [172]41419_2025_7962_MOESM3_ESM.tif^ (5.9MB, tif)

   Original Western Blot Images in Supplementary Figures
   [173]Supplementary Materials^ (4.7MB, pdf)

Author contributions

   Conceptualization: Xing Li, Bo Li,Ge-Shu Du; Data curation: Zexiang
   Zhu, Yong Tang, Yun Zhou: Formal analysis: Bo Li, Jiayu Li, Xing Li;
   Funding acquisition: Xing Li; Investigation: Bo Li, Jiayu Li, Ge-Shu
   Du; Methodology: Bo Li, Jiayu Li,Ge-Shu Du; Project administration:Xing
   Li, Bo Li; Resources: Zexiang Zhu, Yong Tang, Yun Zhou; Software:
   Zexiang Zhu, Yong Tang, Yun Zhou; Supervision:Xing Li, Bo Li;
   Validation: Bo Li, Jiayu Li; Visualization: Jiayu Li; Writing -
   original draft: Bo Li, Jiayu Li; Writing- review & editing: Xing
   Li,Ge-Shu Du.

Funding

   This work was supported by Natural Science Foundation of Changsha City
   (Kq2208489), Natural Science Foundation of Changsha City Health
   Commission (KJ-A2023012), Natural Science Foundation of Hunan province
   Administration of Traditional Chinese Medicine (A2024048), Natural
   Science Foundation of Hunan Province (2024JJ9525), and
   University-Affiliated Hospitals Joint Fund of Hunan University of
   Chinese Medicine (2024XYLH159).

Data availability

   Data will be made available on request. The raw data can be found in
   NCBI ([174]https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1281990).

Competing interests

   The authors declare no competing interests.

Ethical approval

   All animal protocols were approved by the Changsha Hospital of
   Traditional Chinese Medicine (Changsha No. 8 Hospital) committee (the
   approval number:2023110804) and followed the ARRIVE guidelines (Animal
   Research: Reporting of In Vivo Experiments) and the Guide for the Care
   and Use of Laboratory Animals.

Consent for publication

   All authors have read and approved the publication of the manuscript.

Footnotes

   Edited by Professor Hans-Uwe Simon

   Publisher’s note Springer Nature remains neutral with regard to
   jurisdictional claims in published maps and institutional affiliations.

   These authors contributed equally: Bo Li, Jiayu Li.

Contributor Information

   Geshu Du, Email: dgsmmcc@163.com.

   Xing Li, Email: lixing818216@hotmail.com.

Supplementary information

   The online version contains supplementary material available at
   10.1038/s41419-025-07962-w.

References