The crucial role of RIG-I in upregulating neutrophil CXCR2 expression through the MAPK signaling pathway in viral pneumonia Authors: Yuxin Wang, Qi Shi, Yaxin Pan, Xiaoyu Sun, Huayang Wang, Yi Zhang Published: 2026 ABSTRACT BackgroundDysregulated neutrophilic inflammation is a hallmark of severe COVID-19, but the mechanisms driving human peripheral neutrophils (HPNs) recruitment remain incompletely defined. C-X-C motif chemokine receptor 2 (CXCR2), a key chemokine receptor on HPNs, is essential for HPNs recruitment during viral pneumonia. However, the role and underlying regulatory mechanisms of CXCR2 on HPNs during viral pneumonia remain unclear.MethodsWe enrolled 66 COVID-19 patients (30 mild, 33 severe) and 31 healthy controls (HCs). The content of HPNs and the CXCR2 expression on the HPNs were tracked using flow cytometry, RT-qPCR, and ELISA. Transcriptomic sequencing identified differential pathways in the patients. In vitro, HPNs and lung epithelial cells (A549, BEAS-2B) were stimulated with the viral mimic poly(I: C) and treated with pathway inhibitors (CXCR2: SB225002; RIG-I: BX795; MEK: PD98059). In vivo, a poly(I: C)-induced acute lung injury mouse model was used to confirm the roles of RIG-I and CXCR2 in lung injury progression.ResultsSevere patients exhibited a higher HPN percentage and CXCR2 expression. Higher CXCR2 expression was correlated with poor prognosis and elevated inflammatory cytokines (TNF-α, IL-6, and IL-8). In vitro, poly(I: C) triggered lung epithelial cells to produce CXCR2 ligands (CXCL1/2/8), increase pro-inflammatory factors, and promote the HPNs migration to epithelial cells. However, CXCR2 antagonists could reduce the HPNs recruitment and inflammatory factors in HPNs supernatants. Mechanistically, poly(I: C)-mediated activation of intracellular RIG-I upregulated CXCR2 expression on HPNs via the MAPK pathway. In vivo, inhibiting CXCR2, RIG-I, or depleting neutrophils significantly alleviated poly(I: C)-induced inflammatory lung injury, inflammation, and neutrophil infiltration.ConclusionOur study identifies the RIG-I–MAPK–CXCR2 axis as a novel pathogenic signaling cascade that drives dysregulated neutrophilic inflammation in COVID-19, providing a basis for novel therapeutic strategies against respiratory viral infections.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12931-026-03775-3. Introduction Coronavirus disease 2019 (COVID-19), which first appeared in December 2019, is induced by severe acute respiratory syndrome coronavirus-2 (SARS-COV-2) [1]. According to the World Health Organization’s, SARS-COV-2 has caused approximately 779 million infections and more than seven million people death to date, which resulting in an unprecedented economic and social disruption [2]. Viral pneumonia caused by pathogens such as coronaviruses, influenza viruses, and respiratory syncytial viruses (RSV) also imposes a considerable burden on global health [3–5]. Acute respiratory distress syndrome (ARDS) is one of the most serious complications and the leading cause of death associated with viral pneumonia, with mortality rates reaching 70% [6, 7]. Studies on the mechanisms of ARDS have revealed that an imbalanced immune response to viral pathogenicity can lead to severe lung injury characterized by excessive cytokine production and a massive influx of immune cells [8]. As the most predominant peripheral immune cells and the first line of defense against invading pathogens [9, 10], human peripheral neutrophils (HPNs) play a crucial role not only in antiviral processes but also in determining the severity of the inflammatory response [11]. For example, in children infected with RSV and patients with severe influenza, increased numbers of HPNs in the lungs have been observed [12, 13]. HPNs persist substantially in the lungs even after antiviral treatment, and a decrease in viral load leads to continued HPNs-mediated airway damage [14]. In severe COVID-19, upregulated HPNs were also observed [15]. Taken together, HPNs serve a central role in viral-induced inflammation and tissue damage. In response to inflammatory stimulation, HPNs migrate from the circulating blood to the infected tissues mediated by chemokines [16]. CXCR2 is a primary chemokine receptor for HPNs that interacts with a variety of chemokines such as CXCL1-3 and CXCL5-8 [17]. Activation of CXCR2 by CXCL8 triggers signaling pathways that induce HPNs function, including degranulation, cell recruitment, integrin activation, migration, and cytokine secretion [18]. Viral infection alters the activity of the CXCR2 signaling pathway in HPNs by both altering chemokine secretion and modulating CXCR2 expression. For instance, chronic hepatitis B virus infection [19] or dengue virus infection [20] led to an increased number of HPNs with low CXCR2 expression, whereas significant upregulation of CXCR2 expression was observed in dsRNA (poly(I: C))-induced viral lung injury models [21]. However, the specific mechanisms influencing CXCR2 expression remain unclear. Retinoic acid-inducible gene I (RIG-I) is a pivotal cytosolic pattern recognition receptor for activating antiviral immune responses [22–24]. RIG-I expression is low in normal cells. However, upon viral infection, RIG-I can recognize and bind to viral RNA, inducing the release of interferons and other cytokines, which subsequently promote HPNs recruitment and inflammation [23, 25]. Moreover, although traditionally considered short-lived effector cells reliant on TLR signaling, emerging evidence indicates that HPNs express functional intracellular viral sensors, including RIG-I [26]. Upon sensing viral RNA, RIG-I in HPNs can initiate potent inflammatory responses [26]. HIV-1-derived single-stranded RNA40 activates HPNs and induces production of proinflammatory cytokines and reactive oxygen species (ROS) through the RIG-I pathways [27]. Moreover, positive correlations between RIG-I expression and CXCR2-related chemokines have been observed in in some cells (such as porcine endothelial cells and human bronchial epithelial cells) [28, 29]. However, the specific relationship between RIG-I and CXCR2 has yet to be clarified. In the present investigation, we aimed to uncover the roles and mechanisms related to HPNs during SARS-CoV-2-infected pneumonia, with a focus on the pathway regulated by RIG-I-dependent CXCR2 expression. This study elucidates a novel immunopathological pathway in viral pneumonia by revealing RIG-I-mediated regulation of CXCR2 in HPNs, thereby providing a potential therapeutic target for neutrophil-focused interventions. Materials and methods Patients and healthy controls (HC) This study was conducted in accordance with the Declaration of Helsinki and approved by the Human Research Ethics Committee of Qilu Hospital of Shandong University [KYLL-202306-057]. All participants provided informed consent for sample collection and subsequent analysis. COVID-19 was confirmed by RT-qPCR at the Qilu Hospital of Shandong University. Exclusion criteria included ongoing immunosuppressant treatment, pregnancy or lactation, a history of severe health conditions such as end-stage renal or hepatic failure, heart failure or arrhythmia, and prolonged corrected QT interval (> 450 ms on electrocardiogram) [30]. Patients were classified based on the severity of their disease according to the officially confirmed Chinese criteria [31]. In brief, mild cases were identified as those with a respiratory rate (RR) of less than 30 breaths per min and an oxygen saturation level above 93% while breathing ambient air at rest; while severe cases were identified as those who met at least one of the following criteria: (1) RR of 30 breaths per min or greater; (2) Oxygen saturation level of 93% or lower while breathing ambient air at rest; (3) Ratio of arterial blood oxygen pressure (PaO2) to fractional inspired oxygen (FiO2) of 300 mmHg or less; (4) Progressive clinical symptoms worsen, and pulmonary imaging shows progression of lesions greater than 50% within 24 to 48 h; (5) The presence of shock, multiple organ dysfunction requiring ICU care, or respiratory failure necessitating mechanical ventilation. In total, 30 mild COVID-19 patients, 33 severe COVID-19 patients, and 31 HC participants were involved [32]. Blood samples (5 ml EDTA) from COVID-19 patients were collected on the first day of hospitalization and during the hospital stay. Table 1 presents detailed information about the participants. Table 1Clinical data of COVID-19 patientsMild patients(n = 30)Severe patients(n = 33)P valueAge(y)72.00(64.50, 80.00)75.00(66.50, 79.50)0.565Gender(n,%) Male14(46.7%)23(69.7%)0.064 Female16(53.3%)10(30.3%)Days since symptom onset26.00(10.00, 30.00)15.00(10.00, 21.00)0.087Days since hospitalization12.00(8.00,17.25)12.00(5.50,19.00)0.756NEU%81.05(71.68, 88.98)89.80(83.90, 94.00)0.002CRP14.59(3.32, 53.39)65.23(26.89,134.33)< 0.001ESR25.00(12.75, 56.00)28.00(11.00,63.00)0.764IL-620.76(7.27, 35.62)a58.17(27.3, 417.40)< 0.001IL-871.50(61.77, 90.65)b190.00(88.10, 298.93)c0.002IL-107.68(6.11, 25.95)d27.92(11.77, 78.00)e0.003TNFα12.20(9.34,18.23)f21.83(13.08, 34.60)g0.01CXCR2 MFI200.50(158.25, 253.75)330.00(298.00, 405.00)< 0.001an = 26; bn = 9; cn = 19; dn = 13; en = 21; fn = 24; gn=18 Table 1Clinical data of COVID-19 patientsMild patients(n = 30)Severe patients(n = 33)P valueAge(y)72.00(64.50, 80.00)75.00(66.50, 79.50)0.565Gender(n,%) Male14(46.7%)23(69.7%)0.064 Female16(53.3%)10(30.3%)Days since symptom onset26.00(10.00, 30.00)15.00(10.00, 21.00)0.087Days since hospitalization12.00(8.00,17.25)12.00(5.50,19.00)0.756NEU%81.05(71.68, 88.98)89.80(83.90, 94.00)0.002CRP14.59(3.32, 53.39)65.23(26.89,134.33)< 0.001ESR25.00(12.75, 56.00)28.00(11.00,63.00)0.764IL-620.76(7.27, 35.62)a58.17(27.3, 417.40)< 0.001IL-871.50(61.77, 90.65)b190.00(88.10, 298.93)c0.002IL-107.68(6.11, 25.95)d27.92(11.77, 78.00)e0.003TNFα12.20(9.34,18.23)f21.83(13.08, 34.60)g0.01CXCR2 MFI200.50(158.25, 253.75)330.00(298.00, 405.00)< 0.001an = 26; bn = 9; cn = 19; dn = 13; en = 21; fn = 24; gn=18 Clinical data of COVID-19 patients an = 26; bn = 9; cn = 19; dn = 13; en = 21; fn = 24; gn=18 Flow cytometry analysis The blood (100 µL) was washed with phosphate-buffered saline (PBS), incubated with Red Blood Cell Lysis Buffer (Cat: R1010, Solarbio, Beijing, China) for 20 min, and centrifuged at 450 × g for 10 min, and the supernatant was completely aspirated. The samples were stained with anti-human CD11b (Clone: ICRF44 )-FITC (Cat: 982614, BioLegend, San Diego, CA, USA), CXCR2 (Clone: 5E8/CXCR2)-APC (Cat: 320710, BioLegend), and CXCR2-FITC (Cat: 320704, Biolegend) antibodies. Neutrophils were identified as CD11b+ cells with high side scatter (SSC) characteristics. The percentage of HPNs was calculated as (absolute neutrophil count/total white blood cell count) × 100%, based on flow cytometric quantification. For examining ROS and CXCR2 expression in HPNs under distinct stimulatory conditions, after exposure to various stimulants, HPNs were stained with ROS (Cat: D3861, (Thermo Fisher Scientific, Waltham, MA, California, USA) and CXCR2. Isotype controls were analyzed simultaneously. The collected samples were processed using a FACSCalibur flow cytometer (BD Biosciences), and the data were analyzed using FlowJo software (FlowJo). The detailed gating strategy is provided in Fig. S1. Immunofluorescence staining HPNs isolated from peripheral blood were fixed in a cold solution of acetone and methanol (1:1 ratio) and subsequently permeabilized using 0.1% Triton X-100. The fixed cells were then treated with a rabbit anti-CXCR2 antibody (Cat: 3935, Cell Signaling Technology Danvers, MA, USA) followed by a secondary incubation with goat anti-rabbit IgG H&L (Alexa Fluor® 488, Cat: ab150077, Abcam, Cambridge, UK). Nuclei were stained using DAPI at a concentration of 10 µg/mL (Cat: C0065, Solarbio). The fluorescence was observed using an Olympus IX81 inverted fluorescence microscope equipped with an Olympus DP73 digital camera. Cell isolation and culture HPNs were isolated from HC and COVID-19 patients using a Human Peripheral Blood Neutrophil Isolation Kit (Cat: P9040, Solarbio). The purity of the HPNs, as assessed by flow cytometry, exceeded 80%. These isolated HPNs were then cultured in RPMI 1640 medium enriched with 10% fetal bovine serum (FBS)(Gibco Invitrogen Corp, Carlsbad, CA, USA), 10,000 U/ml of penicillin, and 10,000 µg/ml of streptomycin (Cat: CSP006, Zhong Qiao Xin Zhou biotechnology, Shanghai, China) for subsequent studies. To evaluate how serum from HC and COVID-19 patients affects ROS levels in HPNs, 2 × 106 cells from both groups were cultured in RPMI 1640 medium supplemented with 20% serum from either HC or patients for a 24-h period, respectively. In some experiments, HPNs (2 × 106) were pre-treated with the CXCR2 antagonist SB225002 (10 µM, Cat: S7651, Selleck, Houston, TX, USA), RIG-I inhibitor BX795 (10 µM, Cat: S7651, MedChemExpress, Monmouth Junction, NJ, USA), or MEK inhibitor PD98059 (10 µM, Cat: HY-12028, MedChemExpress) for 1 h before being stimulated with poly(I: C) (20 µg/ml) for either 1 h (for intracellular signaling pathway study) or 20 h (for cytokine levels analysis). The human bronchial epithelial cell (BEAS-2B, servicebio, STCC10202G) were cultured in DMEM enhanced with 15% FBS, 10,000 U/ml of penicillin, and 10,000 µg/ml of streptomycin. The alveolar epithelium-like cell (A549, zqxzbio, ZQ0003) were cultured in RPMI 1640 medium containing 10% FBS, 10,000 U/ml of penicillin, and 10,000 µg/ml of streptomycin. Both BEAS-2B and A549 cells (2 × 105) were then stimulated with poly(I: C) (20 µg/ml) for 24 h. All cells were incubated at 37 °C with 5% CO2. For TLR7/8 agonist stimulation, A549 and BEAS-2B cells (2 × 10⁵) were treated with ssRNA40 (2 µg/ml, Cat: tlrl-lrna40, Invivogen, San Diego, CA, USA) for 24 h. For RIG-I-specific agonist stimulation, HPNs (2 × 106) were treated with 3p-hpRNA (1 µg/ml, Cat: tlrl-hprna, Invivogen) for 4 h (for mRNA analysis). Chemotaxis assay To assess the chemotactic cpacity of HPNs in vitro, we employed a Transwell migration assay. HPNs (1 × 105) from both HC and COVID-19 patients were placed in the upper compartment of 24-well inserts with 3 μm pores (Corning Life Sciences, Corning, NY, USA) in 100 µL of RPMI 1640 medium. RPMI 1640 medium supplemented with/without 20 µg/ml of poly(I: C) or CXCR2 antagonist SB225002 (10 µM, Cat: S7651, Selleck), without serum. The lower chamber contained conditioned medium (600 µL) from BEAS-2B cells as a source of chemotactic factors. For experiments using TLR7/8 agonist stimulation, BEAS-2B cells were pre-treated with ssRNA40 (2 µg/ml, Cat: tlrl-lrna40, Invivogen) for 24 h. After a 4-h incubation period at 37 °C, cells that had migrated to the lower surface of the filter were fixed using 10% formalin and stained with crystal violet. Enzymes linked immunosorbent assay (ELISA) Quantification of inflammatory factors in the culture medium of HPNs was carried out using different ELISA kits, specifically for Human TNFα (Cat: KE00154, Proteintech, Wuhan, Hubei, China), Human IL-1β(Cat: KE00021, Proteintech), Human IL‐6 (Cat: KE00139, Proteintech), Human IL‐8 (Cat: KE00275, Proteintech), and Human IL‐10(Cat: KE00170, Proteintech). Apoptosis assay Apoptosis was evaluated using the Annexin V-FITC/PI staining kit (Cat: A211-01, Vazyme, Jiangsu, China). After 24 h of culture with the HPNs supernatant, 5 × 105 BEAS-2B cells were collected and stained with Annexin V-FITC for 15 min, followed by propidium iodide (PI) staining for 5 min. The proportion of apoptotic cells was measured using the FACSCanto II flow cytometer (BD Biosciences). RNA extraction and RT-qPCR Total RNA was extracted from A549 cells, BEAS-2B cells, and human circulating HPNs using TRIzol Reagent (Thermo Fisher Scientific, Waltham, MA, California, USA). Complementary DNA (cDNA) was synthesized via reverse transcription. RT-qPCR analyses were conducted using a LightCycler 2.0 Instrument (Roche). β-actin was employed as the internal control, and relative gene expression levels were determined using the ΔΔCt quantification method. The sequences of the primers were listed at (Table 2). Table 2The primers used in this studyNamePrimer sequences (5’ to 3’)CXCR2Forward: CAGTTACAGCTCTACCCTGCCReverse: CCAGGAGCAAGGACAGACCCCCXCL1Forward: CGCCCAAACCGAAGTCATAGReverse: TCAGGAACAGCCACCAGTGAGCXCL2Forward: TTCACAGTGTGTGGTCAACATReverse: TCTCTGCTCTAACACAGAGGGACXCL8/IL-8Forward: TCTGCAGCTCTGTGTGAAGGReverse: TGAATTCTCAGCCCTCTTCAATNF-αForward: ATGAGCACTGAAAGCATGATCReverse: GAGGGCTGATTAGAGAGAGGTCIL-1βForward: TCGCCAGTGAAATGATGGCTReverse: GGTCGGAGATTCGTAGCTGGIL-6Forward: CACACAGACAGCCACTCACCReverse: GCTCTGGCTTGTTCCTCACTIL-10Forward: AAAGAAGGCATGCACAGCTCReverse: TCGAAGCATGTTAGGCAGGTβ-actinForward: GAAGAGCTACGAGCTGCCTGReverse: CAGACAGCACTGTGTTGGCGA Table 2The primers used in this studyNamePrimer sequences (5’ to 3’)CXCR2Forward: CAGTTACAGCTCTACCCTGCCReverse: CCAGGAGCAAGGACAGACCCCCXCL1Forward: CGCCCAAACCGAAGTCATAGReverse: TCAGGAACAGCCACCAGTGAGCXCL2Forward: TTCACAGTGTGTGGTCAACATReverse: TCTCTGCTCTAACACAGAGGGACXCL8/IL-8Forward: TCTGCAGCTCTGTGTGAAGGReverse: TGAATTCTCAGCCCTCTTCAATNF-αForward: ATGAGCACTGAAAGCATGATCReverse: GAGGGCTGATTAGAGAGAGGTCIL-1βForward: TCGCCAGTGAAATGATGGCTReverse: GGTCGGAGATTCGTAGCTGGIL-6Forward: CACACAGACAGCCACTCACCReverse: GCTCTGGCTTGTTCCTCACTIL-10Forward: AAAGAAGGCATGCACAGCTCReverse: TCGAAGCATGTTAGGCAGGTβ-actinForward: GAAGAGCTACGAGCTGCCTGReverse: CAGACAGCACTGTGTTGGCGA The primers used in this study RNA-sequencing and functional annotations HPNs were isolated from HC and COVID-19 patients using the TRIzol reagent (Thermo Fisher Scientific). The total RNA were assessed using a Bioanalyzer 2100 and RNA 6000 Nano Kit (Agilent Technologies). Only high-quality RNA samples with an RIN greater than 7.0 were selected for sequencing library construction. The constructed sequencing library was subjected to transcriptome sequencing using the Illumina NovaSeq 6000 platform (LC-Bio Technology Co., Ltd. Zhejiang, China). Differential expression analysis was performed using DESeq2 software (http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html). Genes were considered differentially expressed if they met the criteria of a p-value less than 0.05 and an absolute fold change of 2 or greater. These differentially expressed genes were subsequently analyzed for enrichment using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (http://www.kegg.jp). Western blot analysis HPNs were lysed using 150 µL RIPA lysis buffer (Cat: P0013D; Beyotime Inc, Jiangsu, China) supplemented with phosphatase inhibitors and a complete ultra-protease inhibitor (Cat: P1005, Beyotime). The total protein concentration was quantified using a bicinchoninic acid (BCA) Protein Assay Kit (Cat: P0010, Beyotime). Equal quantities of proteins were subjected to SDS-PAGE and transferred onto PVDF membranes (IPVH00010, Millipore Corporation, Billerica MA, USA). The membranes were blocked with a solution of 5% bovine serum albumin in TBST (0.1% Tween 20) before being incubated with primary antibodies including anti-human β-actin (1:1000, Cat: 66009-1-Ig, Proteintech), RIG-I(1:1000, Cat: NBP2-45678, Novus, Colorado, USA), MEK (1:1000, Cat: ET1602-3, Huabio, Zhejiang, China), phosphorylated MEK (1:1000, Cat: #9154, Cell Signaling Technology (CST)), ERK1/2 (1:1000, Cat: ET1601-29, Huabio), phosphorylated ERK1/2 (1:1000, Cat: ET1610-13, Huabio). After washing with TBST, the membranes were incubated with an HRP-conjugated secondary antibody (at a 1:10000 dilution, Proteintech) for 1 h at room temperature. Protein bands were detected using an ECL chemiluminescent detection system (Amersham Imager 600, GE Healthcare Life Sciences, and ChemiDoc™ Touch Imaging System, Bio-Rad). Mouse experiments Thirty-five male inbred C57BL/6 mice (6–8 weeks old) were obtained from the Vital River Company and housed in a specific pathogen-free environment at the Model Animal Research Center of Shandong University. All experiments were approved by the Animal Experimentation Ethics Committee of Shandong University and strictly adhered to established guidelines. In the poly(I: C)-induced acute lung injuery model, twenty mice were randomly divided into five groups (N = 4), Control, poly(I: C), poly(I: C)+ SB225002, poly(I: C)+ BX795, and poly(I: C)+ anti- Ly6G. The virus induced acute lung injury model was established according to the protocol described in a previous study [30]. Briefly, mice were intranasally administered poly(I: C) (10 mg/kg/day) to establish an acute lung injury model. The control group was intranasally conducted an equivalent volume of vehicle (PBS). Moreover, poly(I: C)-induced peneumonia model mice were treated with a daily intraperitoneal injection of either SB225002 (10 mg/kg), BX795 (500 µg/kg), or anti-mouse Ly6G (0.625 mg/kg) for five days. 12 h after the final dose, all mice were thoroughly evaluated. To further validate the involvement of TLR7/8 signaling, an additional acute lung injury model was established using the TLR7/8 agonist R848. Fifteen mice were randomly divided into three groups (N = 5): Control, R848, and R848 + SB225002 group. Mice in R848 groups were intranasally administered R848 (2.5 µg/g body weight, Cat: tlrl-r848-1, Invivogen) in a volume of 50 µL PBS, following a previously described protocol [33]. All other procedures were performed as described for the poly(I: C)-induced acute lung injury model. Histopathological and immunologic analyses of lung injury and neutrophil activation After euthanizing the mice with CO2, the lungs and trachea were excised and fixed in 4% paraformaldehyde for paraffin embedding. The whole lung was subsequently sectioned at 5 μm and stained with hematoxylin and eosin (H&E) kits (Beyotime lnc). The H&E-stained slides were digitized using a light microscope (Leica DM6000B; Leica Microsystems Inc., Buffalo Grove, IL, USA). A standardized lung injury scoring system was used to analyze tissue injury quantitatively, according to the criteria of a previous study [34]. The scoring system parameters included the presence of HPNs in the alveolar space and interstitial space, hyaline membranes, proteinaceous debris filling the airspace, and alveolar septal thickening. Slides were immunostained with anti-MPO (1:500, 22225-1-ap, Proteintech) to determine the accumulation and activation status of HPNs. Images were acquired on a VS200 (OLYMPUS, Japan) and processed using ImageJ software (version 1.44p, National Institutes of Health, USA). The ratio of wet lung to dry lung (W/D ratio) After the mice were sacrificed, the upper lobe of the left lung was removed, and its wet weight was noted. The lung tissue was subsequently dried in an incubator at 80 °C for 24 h to assess the dry weight. The W/D ratio was determined by dividing wet weight by dry weight. Acquisition of Bronchoalveolar Lavage Fluid (BALF) To obtain BALF, a 20-gauge catheter was inserted into the trachea and flushed with 800µL of cold PBS three times. The BALF underwent centrifugation at 1,500 rpm for 10 min at a temperature of 4 °C. After centrifugation, the supernatant was collected, and the total protein content was measured using a BCA protein quantification kit (Cat: 20201ES76, Yeasen, Shanghai, China). The cell pellet remaining from the BALF was added with a red cell lysis buffer (Beyotime Inc.), resuspended in PBS, and subsequently used for cell count analysis using a hemocytometer (ThermoFisher). The concentrations of cytokines in the BALF supernatants were analyzed using ELISA kits, according to the manufacturer’s guidelines. Statistics Each experiment was carried out a minimum of three times with independent biological replicates, and all results are expressed as mean ± standard deviation (SD). Normal distribution was evaluated for all the data. Statistical comparisons were performed by one-way Analysis of Variance (ANOVA) (for normally distributed data) or Kruskal-Wallis test (for non-normally distributed data). Statistical significance between two groups was determined by Student’s paired t-test (for normally distributed data) and Mann-Whitney U test (for non-normally distributed data). Pearson’s correlation coefficient was used to assess the relationship between the two variables. The specific statistical tests used in each experiment are described in the figure captions. Statistical significance was defined as NS (not significant), *P < 0.05, **P < 0.01, and ***P < 0.001. Results CXCR2 expression on HPNs is correlated with disease severity in COVID-19 patients To investigate the potential roles of HPNs in COVID-19, 31 healthy controls (HCs) and 66 COVID-19 patients (30 mild and 33 severe, Table 1) were involved in this study [32]; and the percentage of HPNs among these groups were compared. As shown in Fig. 1A, the percentage of HPNs within total leukocytes was significantly higher in COVID-19 patients than in HC at the time of admission, and the percentage of HPNs in the severe COVID-19 group was higher than that in the mild COVID-19 group. Moreover, the changes in the HPNs percentage in patients with mild and severe COVID-19 during hospitalization were also monitored. During hospitalization, the percentage of HPNs in patients with severe COVID-19 was higher than that in patients with mild patients (Fig. 1B-C). Fig. 1HPNs are abundant and CXCR2 expression levels are higher in patients with severe COVID-19. A The percentage of HPNs within total leukocytes of HCs (n = 31), mild (n = 30) and severe (n = 33) COVID-19 patients at admission. B The percentage of HPNs of mild (n = 30) and severe (n = 33) COVID-19 patients during hospitalization, each column represented one patient and each row depicted different hospitalization days, and the color in each cell represented the percentage of HPNs. C The average percentage of HPNs in mild (n = 30, blue) and severe (n = 33, red) COVID-19 patients during hospitalization. The dashed line indicated the average percentage of HPNs from HCs. D The ΔMFI value of CXCR2 on HPNs of mild (n = 30) and severe (n = 33) COVID-19 patients at admission. E (Left) Representative images of immunofluorescence stained with CXCR2 in HPNs of mild and severe COVID-19 patients. Nuclei were stained with DAPI (4′, 6-diamidino-2-phenylindole), displayed in blue. Scale bars, 125 μm. (Right) Quantitative analysis of the ΔMFI of CXCR2. F The ΔMFI value for CXCR2 on HPNs from mild (n = 30) and severe (n = 33) COVID-19 patients during hospitalization. Each column represented one patient, each row represented different hospitalization days and each cell represented the ΔMFI value for CXCR2 on HPNs. G The average ΔMFI value of CXCR2 on HPNs of mild (blue) and severe (red) COVID-19 patients during hospitalization. H Kaplan–Meier curve for overall survival of COVID-19 patients during hospitalization, stratified by ΔMFI value of CXCR2 on HPNs (n = 33 for CXCR2 low expression and n = 30 for CXCR2 high expression; cutoff value = 296.5). Data were presented as mean ± SD. For (A), Kruskal-Wallis test for multiple groups was used (non-normal distribution). For (D, E), Mann-Whitney U test for two-group comparisons was used (non-normal distribution). For (H), the log-rank test was used. *P < 0.05, **P < 0.01, and ***P < 0.001 HPNs are abundant and CXCR2 expression levels are higher in patients with severe COVID-19. A The percentage of HPNs within total leukocytes of HCs (n = 31), mild (n = 30) and severe (n = 33) COVID-19 patients at admission. B The percentage of HPNs of mild (n = 30) and severe (n = 33) COVID-19 patients during hospitalization, each column represented one patient and each row depicted different hospitalization days, and the color in each cell represented the percentage of HPNs. C The average percentage of HPNs in mild (n = 30, blue) and severe (n = 33, red) COVID-19 patients during hospitalization. The dashed line indicated the average percentage of HPNs from HCs. D The ΔMFI value of CXCR2 on HPNs of mild (n = 30) and severe (n = 33) COVID-19 patients at admission. E (Left) Representative images of immunofluorescence stained with CXCR2 in HPNs of mild and severe COVID-19 patients. Nuclei were stained with DAPI (4′, 6-diamidino-2-phenylindole), displayed in blue. Scale bars, 125 μm. (Right) Quantitative analysis of the ΔMFI of CXCR2. F The ΔMFI value for CXCR2 on HPNs from mild (n = 30) and severe (n = 33) COVID-19 patients during hospitalization. Each column represented one patient, each row represented different hospitalization days and each cell represented the ΔMFI value for CXCR2 on HPNs. G The average ΔMFI value of CXCR2 on HPNs of mild (blue) and severe (red) COVID-19 patients during hospitalization. H Kaplan–Meier curve for overall survival of COVID-19 patients during hospitalization, stratified by ΔMFI value of CXCR2 on HPNs (n = 33 for CXCR2 low expression and n = 30 for CXCR2 high expression; cutoff value = 296.5). Data were presented as mean ± SD. For (A), Kruskal-Wallis test for multiple groups was used (non-normal distribution). For (D, E), Mann-Whitney U test for two-group comparisons was used (non-normal distribution). For (H), the log-rank test was used. *P < 0.05, **P < 0.01, and ***P < 0.001 Next, we measured the expression intensity of CXCR2 (represented as MFI) on the HPNs of patients with mild and severe COVID-19. At the time of admission, the expression intensity of CXCR2 on the HPNs of mild patients was lower than that on the HPNs of severe patients (Fig. 1D and E). During hospitalization, the CXCR2 expression intensity on the HPNs of patients with severe disease was also generally higher than that in patients with mild disease (Fig. 1F-G). Following, to investigate the correlation between CXCR2 expression and the patient prognosis, all the patients were further generated into two groups: low (< 296.5) and high (≥ 296.5) CXCR2 expression according to the cut-off value derived from the ROC curve of CXCR2 expression intensity (MFI = 296.5). The results indicated that the prognosis of the patients with high CXCR2 expression was worse than that of the group with low CXCR2 expression (Fig. 1H). To further characterize the relationship between CXCR2 expression and patient outcomes, we compared baseline characteristics between deceased and surviving patients. As shown in Table S1, patients who died exhibited significantly higher baseline CXCR2 expression than those who survived (P < 0.001). Table S2 further shows that among patients with high CXCR2 expression, 28 of 33 (84.8%) died during hospitalization, compared with only 6 of 30 (20.0%) in the low expression group. These data confirm the presence of survivor bias, as later time points predominantly reflect data from surviving patients. Collectively, these results suggest that both HPNs percentage and CXCR2 expressions on HPNs are altered during the progression and treatment of COVID-19 disease, and that CXCR2 expression levels may serve as an indicator of COVID-19 status and prognosis. To assess the potential confounding effect of symptom onset-to-admission days, we performed additional analyses. When we selected a subset of mild and severe patients with matched symptom onset-to-admission days, the differences in HPNs percentage and CXCR2 expression between the two groups remained significant (Fig. S2A). Moreover, multivariate logistic regression analysis identified disease severity, but not symptom onset-to-admission days, as an independent risk factor for elevated CXCR2 expression (Fig. S2B). These findings suggest that the observed differences in HPNs percentage and CXCR2 expression are primarily driven by disease severity rather than the timing of hospitalization. Correlation between CXCR2 expression and inflammatory factors in COVID-19 patients COVID-19 is an immune-related disease, and its occurrence is coupled with the release of a large quantities of inflammatory factors—a finding that has also been confirmed in our research (Fig. 2A). First, the serum levels of TNFα, IL-6, IL-8, IL-10 and C-reactive protein (CRP) were higher in the severe group than in the mild group (Fig. 2A). Then, considering that HPN enrichment and CXCR2 expression levels were altered in COVID-19 patients with different degrees of disease severity, we detected whether CXCR2 expression on HPNs was correlated with serum inflammatory markers and cytokines. As expected, correlation analysis indicated a significant positive correlation between CXCR2 expression levels and erythrocyte sedimentation rate (ESR) (weak correlation, r = 0.221, P = 0.029), CRP (weak correlation, r = 0.207, P = 0.007), TNFα (strong correlation, r = 0.638, P = 0.014), IL-6 (strong correlation, r = 0.623, P = 0.003), and IL-8 (strong correlation, r = 0.609, P = 0.007) (Fig. 2B). Conversely, a significant negative correlation was observed between CXCR2 expression levels and IL-10 (moderate correlation, r = 0.485, P = 0.041) (Fig. 2B). These findings indicated the closely correlation of CXCR2 expression and inflammatory factors in COVID-19 disease. Fig. 2The correlation between CXCR2 expression intensity and inflammatory factors in COVID-19 patients. A The levels of ESR, CRP, TNFα, IL-6, IL-8, and IL-10 in the peripheral blood of COVID-19 patients at admission. B Correlation analysis between ΔMFI of CXCR2 and the following inflammatory parameters: ESR levels, CRP levels, serum concentrations of TNFα, IL-6, IL-8, and IL-10 in COVID-19 patients. Data were presented as mean ± SD. For (A), Student t-test for two groups was used (normal distribution). For (B), Spearman correlation coefficient was used (non-normal distribution). NS (not significant), *P < 0.05, **P < 0.01 and ***P < 0.001 The correlation between CXCR2 expression intensity and inflammatory factors in COVID-19 patients. A The levels of ESR, CRP, TNFα, IL-6, IL-8, and IL-10 in the peripheral blood of COVID-19 patients at admission. B Correlation analysis between ΔMFI of CXCR2 and the following inflammatory parameters: ESR levels, CRP levels, serum concentrations of TNFα, IL-6, IL-8, and IL-10 in COVID-19 patients. Data were presented as mean ± SD. For (A), Student t-test for two groups was used (normal distribution). For (B), Spearman correlation coefficient was used (non-normal distribution). NS (not significant), *P < 0.05, **P < 0.01 and ***P < 0.001 CXCR2 inhibition ameliorates inflammatory responses and HPN functions in viral mimic poly(I: C) stimulation in vitro The above clinical investigation demonstrated a significant association between CXCR2 expression and the inflammatory response of the COVID-19 patients. To further prove the underlying mechanism based on CXCR2 expression in vitro, the alveolar epithelium-like cell line (A549) and bronchial epithelial cell line (BEAS-2B) were stimulated with poly(I: C), an analog of dsRNA that mimics the molecular pattern associated with viral infection. The results implied that the expression of the main ligands of CXCR2, including chemokines CXCL1, CXCL2, and CXCL8/IL-8, in both cell lines was significantly upregulated at the mRNA level after stimulation with the viral mimic poly(I: C) (Fig. 3A), indicating that virus-infected respiratory epithelial cells might be a potent source of chemotactic signals for neutrophil recruitment. Moreover, a Transwell migration assay was conducted to assess the chemotactic capacity of HPNs toward BEAS-2B cell. Briefly, HPNs derived from COVID-19 patients and HCs were placed in the upper compartment with 3 μm pore inserts; the lower chamber contained conditioned medium from BEAS-2B cells, which was either supplemented with either poly(I: C), treated with the CXCR2 antagonist SB225002, or left unstimulated as a control (Fig. 3B). HPN migration was evaluated after a defined co-culture period. The results indicated that HPNs from COVID-19 patients exhibited enhanced chemotaxis toward the conditioned medium of BEAS-2B cells compared to those from HCs (Fig. 3C). Moreover, the poly(I: C) treatment of BEAS-2B cells significantly increased HPNs recruitment, which can be inhibited by SB225002 (Fig. 3C). Additionally, the poly(I: C)-induced mRNA expression of TNF-α, IL-1β, IL-6, IL-8, and IL-10 in HPNs can be inhibited by pre-treated with SB225002 (Fig. 3D), and the same outcomes were validated by detecting the concentrations of the above cytokines in the supernatants of HPNs (Fig. 3E). Fig. 3Impact of CXCR2 on HPNs recruitment and function in response to poly(I: C) stimulation. A A549 cells (5 × 105) or BEAS-2B cells (5 × 105) were exposed to poly(I: C) (20 µg/ml) for 20 h. The mRNA expression levels of CXCL1 (left), CXCL2 (middle) and CXCL8 (right) were quantified using RT-qPCR. B Chemotactic experiment schematic diagram: HPNs from COVID-19 patients and HC were placed in the upper compartment with 3 μm pore inserts; the lower chamber contained conditioned medium from BEAS-2B cells, supplemented with either poly(I: C), the CXCR2 antagonist SB225002, or left unstimulated as a control. C (Left) Representative images showing migration of HPNs from HC and COVID-19 patients, which were pre-treated with SB225002 (10 µM) and/or poly(I: C) (20 µg/ml), toward conditioned medium of BEAS-2B cells. (Right) Numbers of migrated HPNs from HC or COVID-19 patients, pre-treated by SB225002 (10 µM) or poly(I: C) (20 µg/ml), toward conditioned medium of BEAS-2B. All images were photographed at the time point of 4 h of Transwell assay. Five random 200X fields were selected and numbers of cells of each field were counted to calculate the average migrated number. D HPNs were pretreated with/without SB225002 (10 µM) for 1 h, and then stimulated by poly(I: C) (20 µg/ml) for 4 h. mRNA levels of TNF-α, IL-1β, IL-6, IL-8 and IL-10 were analyzed using RT- qPCR. E ELISA was used to measure concentrations of TNF-α, IL-1β, IL-6, IL-8 and IL-10 in HPNs supernatant, which were pre-treated by SB225002 (10 µM) and poly(I: C) (20 µg/ml). F (Left) Typical flow cytometry results of the apoptotic percentage of BEAS-2B cells co-cultured with supernatant of HPNs with different treatment. (Right) Quantification of the apoptotic percentage of BEAS-2B cells co-cultured with supernatant of HPNs with different treatment. All samples were biologically independent and three or more independent experiments were performed. G (Left) Typical flow cytometry results of ROS expression on HPNs, pre-treated with/without SB225002 (10 µM) and/or poly(I: C) (20 µg/ml). (Right) ΔMFI value of ROS in pre-treated HPNs. All data were shown as the mean ± SD. For (A), Student paired t-test for two groups was used (normal distribution). For (C, D, E), Kruskal-Wallis test for multiple groups was used (non-normal distribution). For (F, G), ANOVA for multiple groups was used (normal distribution). NS (not significant), *P < 0.05, **P < 0.01, and ***P < 0.001 Impact of CXCR2 on HPNs recruitment and function in response to poly(I: C) stimulation. A A549 cells (5 × 105) or BEAS-2B cells (5 × 105) were exposed to poly(I: C) (20 µg/ml) for 20 h. The mRNA expression levels of CXCL1 (left), CXCL2 (middle) and CXCL8 (right) were quantified using RT-qPCR. B Chemotactic experiment schematic diagram: HPNs from COVID-19 patients and HC were placed in the upper compartment with 3 μm pore inserts; the lower chamber contained conditioned medium from BEAS-2B cells, supplemented with either poly(I: C), the CXCR2 antagonist SB225002, or left unstimulated as a control. C (Left) Representative images showing migration of HPNs from HC and COVID-19 patients, which were pre-treated with SB225002 (10 µM) and/or poly(I: C) (20 µg/ml), toward conditioned medium of BEAS-2B cells. (Right) Numbers of migrated HPNs from HC or COVID-19 patients, pre-treated by SB225002 (10 µM) or poly(I: C) (20 µg/ml), toward conditioned medium of BEAS-2B. All images were photographed at the time point of 4 h of Transwell assay. Five random 200X fields were selected and numbers of cells of each field were counted to calculate the average migrated number. D HPNs were pretreated with/without SB225002 (10 µM) for 1 h, and then stimulated by poly(I: C) (20 µg/ml) for 4 h. mRNA levels of TNF-α, IL-1β, IL-6, IL-8 and IL-10 were analyzed using RT- qPCR. E ELISA was used to measure concentrations of TNF-α, IL-1β, IL-6, IL-8 and IL-10 in HPNs supernatant, which were pre-treated by SB225002 (10 µM) and poly(I: C) (20 µg/ml). F (Left) Typical flow cytometry results of the apoptotic percentage of BEAS-2B cells co-cultured with supernatant of HPNs with different treatment. (Right) Quantification of the apoptotic percentage of BEAS-2B cells co-cultured with supernatant of HPNs with different treatment. All samples were biologically independent and three or more independent experiments were performed. G (Left) Typical flow cytometry results of ROS expression on HPNs, pre-treated with/without SB225002 (10 µM) and/or poly(I: C) (20 µg/ml). (Right) ΔMFI value of ROS in pre-treated HPNs. All data were shown as the mean ± SD. For (A), Student paired t-test for two groups was used (normal distribution). For (C, D, E), Kruskal-Wallis test for multiple groups was used (non-normal distribution). For (F, G), ANOVA for multiple groups was used (normal distribution). NS (not significant), *P < 0.05, **P < 0.01, and ***P < 0.001 To better understand the effects of HPNs on epithelial cells, BEAS-2B cells were cultured in conditioned medium from HPNs under different treatments, and apoptosis was analyzed. The results demonstrated that the conditioned medium of poly(I: C)-stimulated HPNs significantly promoted apoptosis in BEAS-2B cells, whereas pretreatment with the SB225002 (CXCR2 pathway inhibitor) reversed the promotion of apoptosis induced by poly(I: C)- HPNs (Fig. 3F). In addition to chemotaxis, the role of CXCR2 in the regulation of HPNs function has also been observed to be related to ROS production [15], a core marker of the inflammatory response and oxidative stress [35]. Hence, we also detected the ROS levels produced by HPNs. Poly(I: C) significantly promoted ROS production of HPNs, whereas SB225002 treatment reversed the stimulatory effect of poly(I: C) (Fig. 3G). Moreover, stimulation with serum from COVID-19 patients significantly upregulated ROS production in HPNs from both HC and COVID-19 patients (Fig. S3). These results collectively indicating that the CXCR2 signaling pathway plays a pivotal role not only in directing the chemotaxis of HPNs toward infected respiratory epithelium but also in potentiating their pro-inflammatory effector functions, thereby contributing to the pathogenic neutrophil-mediated inflammation in COVID-19. Poly(I: C) promotes CXCR2 expression on HPNs through RIG-I-MAPK signaling RIG-I plays a crucial role in antiviral response [22–24]. We analyzed whether there was a difference in RIG-I expression between HC and patients. The results showed that RIG-I protein levels were significantly upregulated in the HPNs of COVID-19 patients compared to those in HC (Fig. 4A), confirming the potential regulatory roles of RIG-I in antiviral immune responses in COVID-19. Therefore, we explored the regulatory relationship between RIG-I and CXCR2 expression in HPNs. HPNs from HCs were collected, pre-treated with/without the RIG-I inhibitor BX795, and then stimulated with/without poly(I: C). Our findings demonstrated that poly(I: C) stimulation significantly upregulated RIG-I expression and increased CXCR2 mRNA expression in HPNs, whereas BX795 inhibited this increase (Fig. 4B-D). To further evaluate the underlying mechanisms leading to CXCR2 overexpression, HPNs were isolated from HC and COVID-19 patients for RNA-sequencing (RNA-seq) analysis. KEGG analysis revealed significant differences in the MAPK pathway between HC and COVID-19 patients (Fig. 4E). We then examined changes in the MAPK signaling pathway in HPNs treated by BX795 and/or poly(I: C) in vitro. As shown in Fig. 4F, poly(I: C) stimulation significantly upregulated the phosphorylation of ERK 1/2 and MEK in HPNs, whereas the poly(I: C)-induced phosphorylation of ERK 1/2 and MEK was suppressed by BX795 (Fig. 4F and G). Fig. 4Poly(I: C) promotes CXCR2 expression on HPNs through RIG-I-MAPK signaling. A Western blot result of RIG-I expression in HPNs from HCs and COVID-19 patients. H: Healthy controls; P: COVID-19 patients. B Representative images of RIG-I expression in HPNs which were pre-treated with BX795 (10 µM) for 1 h and then stimulated with/without poly(I: C) (20 µg/ml) for 4 h. C The density of RIG-I in HPNs pre-treated with BX795 (10 µM) and/or poly(I: C) (20 µg/ml). D mRNA expression of CXCR2 in HPNs pre-treated with/without BX795 (10 µM) for 1 h and then stimulated with/without poly(I: C) (20 µg/ml) for 4 h. E Bubble plot showed the KEGG functional annotations in terms of Environmental Information Processing signatures enriched in HPNs of COVID-19 patients. F Represented Western blot images of phosphorylated and total ERK1/2 and MEK in HPNs which were pre-treated with BX795 (10 µM) for 1 h and stimulated with/without poly(I: C) (20 µg/ml) for 4 h. G Ratio of density of phosphorylated/total ERK1/2 and MEK in HPNs pre-treated with BX795 (10 µM) and/or poly(I: C) (20 µg/ml). H (Left) Typical flow cytometry results showed CXCR2 expression on HPNs pre-treated with PD98059 (10 µM) and then stimulated by poly(I: C) (20 µg/ml). (Right) ΔMFI value of CXCR2 in HPNs pre-treated with PD98059 (10 µM) and/or poly(I: C) (20 µg/ml). I The mRNA expression levels of CXCR2 in HPNs treated with PD98059 (10 µM) and/or poly(I: C) (20 µg/ml). J The concentration of IL-8 in the supernatant of HPNs pretreated with/without BX795(10 µM) or PD98059 (10 µM) for 1 h, followed by the treatment of poly(I: C) (20 µg/ml) for 4 h. All samples were biologically independent and three or more independent experiments with similar results were performed. All data were shown as the mean ± SD. For (C, D, H, J), ANOVA for multiple groups was used (normal distribution). For (G, I), Kruskal-Wallis test for multiple groups was used (non-normal distribution).*P < 0.05 and **P < 0.01 Poly(I: C) promotes CXCR2 expression on HPNs through RIG-I-MAPK signaling. A Western blot result of RIG-I expression in HPNs from HCs and COVID-19 patients. H: Healthy controls; P: COVID-19 patients. B Representative images of RIG-I expression in HPNs which were pre-treated with BX795 (10 µM) for 1 h and then stimulated with/without poly(I: C) (20 µg/ml) for 4 h. C The density of RIG-I in HPNs pre-treated with BX795 (10 µM) and/or poly(I: C) (20 µg/ml). D mRNA expression of CXCR2 in HPNs pre-treated with/without BX795 (10 µM) for 1 h and then stimulated with/without poly(I: C) (20 µg/ml) for 4 h. E Bubble plot showed the KEGG functional annotations in terms of Environmental Information Processing signatures enriched in HPNs of COVID-19 patients. F Represented Western blot images of phosphorylated and total ERK1/2 and MEK in HPNs which were pre-treated with BX795 (10 µM) for 1 h and stimulated with/without poly(I: C) (20 µg/ml) for 4 h. G Ratio of density of phosphorylated/total ERK1/2 and MEK in HPNs pre-treated with BX795 (10 µM) and/or poly(I: C) (20 µg/ml). H (Left) Typical flow cytometry results showed CXCR2 expression on HPNs pre-treated with PD98059 (10 µM) and then stimulated by poly(I: C) (20 µg/ml). (Right) ΔMFI value of CXCR2 in HPNs pre-treated with PD98059 (10 µM) and/or poly(I: C) (20 µg/ml). I The mRNA expression levels of CXCR2 in HPNs treated with PD98059 (10 µM) and/or poly(I: C) (20 µg/ml). J The concentration of IL-8 in the supernatant of HPNs pretreated with/without BX795(10 µM) or PD98059 (10 µM) for 1 h, followed by the treatment of poly(I: C) (20 µg/ml) for 4 h. All samples were biologically independent and three or more independent experiments with similar results were performed. All data were shown as the mean ± SD. For (C, D, H, J), ANOVA for multiple groups was used (normal distribution). For (G, I), Kruskal-Wallis test for multiple groups was used (non-normal distribution).*P < 0.05 and **P < 0.01 To better understand whether RIG-I affects CXCR2 expression through the MAPK signaling pathway, HPNs were pre-treated with an MEK inhibitor PD98059, which can effectively reverse the poly(I: C)-induced increase in CXCR2 expression in HPNs (Fig. 4H). This was further confirmed by RT-qPCR and was consistent with the flow cytometry results (Fig. 4I). Additionally, BX795 or PD98059 also dramatically inhibited poly(I: C)-induced IL-8 expression in the supernatant of neutrophils (Fig. 4J). These results indicated that viral mimic poly(I: C) promotes CXCR2 expression on HPNs through regulation of RIG-I-MAPK signaling. RIG-I-specific agonist 3p-hpRNA directly induces CXCR2 expression in HPNs Given that poly(I: C) can also activate TLR3/MDA5, which share downstream signaling components with RIG-I, we employed 3p-hpRNA, a well-established RIG-I-specific agonist [36, 37], to further establish the specific role of RIG-I in regulating CXCR2 expression and to exclude potential contributions from TLR3/MDA5. Stimulation of HPNs with 3p-hpRNA significantly upregulated CXCR2 expression (Fig. 5A). Pretreatment with the RIG-I pathway inhibitor BX795 markedly reversed 3p-hpRNA-induced CXCR2 upregulation (Fig. 5A). Moreover, 3p-hpRNA stimulation significantly increased IL-8 production in HPNs, which was also abrogated by BX795 pretreatment (Fig. 5B). These results are consistent with our poly(I: C)-based findings and directly demonstrate that RIG-I activation is sufficient to drive CXCR2 expression and pro-inflammatory cytokine production in HPNs. Notably, TLR7 expression was markedly higher than TLR3 in HPNs, and 3p-hpRNA stimulation did not significantly affect TLR7 expression but slightly increased TLR3 mRNA levels (Fig. 5C), suggesting that TLR3 is minimally expressed in HPNs and unlikely to contribute substantially to the observed CXCR2 upregulation. Fig. 5RIG-I-specific agonist 3p-hpRNA directly induces CXCR2 expression and IL-8 production in HPNs. A Representative flow cytometry histograms (left) and quantification of CXCR2 MFI (right) in HPNs stimulated with 3p-hpRNA (1 µg/ml, 20 h) with or without BX795 (10 µM) pretreatment. B IL-8 levels in the supernatant of HPNs stimulated with 3p-hpRNA (1 µg/ml, 20 h) with or without BX795 (10 µM) pretreatment, measured by ELISA. C mRNA expression levels of TLR3 and TLR7 in HPNs. Data are presented as mean ± SD from three independent experiments. ANOVA for multiple groups was used (normal distribution). *P < 0.05, **P < 0.01; NS, not significant RIG-I-specific agonist 3p-hpRNA directly induces CXCR2 expression and IL-8 production in HPNs. A Representative flow cytometry histograms (left) and quantification of CXCR2 MFI (right) in HPNs stimulated with 3p-hpRNA (1 µg/ml, 20 h) with or without BX795 (10 µM) pretreatment. B IL-8 levels in the supernatant of HPNs stimulated with 3p-hpRNA (1 µg/ml, 20 h) with or without BX795 (10 µM) pretreatment, measured by ELISA. C mRNA expression levels of TLR3 and TLR7 in HPNs. Data are presented as mean ± SD from three independent experiments. ANOVA for multiple groups was used (normal distribution). *P < 0.05, **P < 0.01; NS, not significant Inhibiting CXCR2 or RIG-I expression ameliorated poly(I: C)-induced acute lung injury in vivo To test the roles of CXCR2 and RIG-I in the inflammatory response in the lungs in vivo, a poly(I: C)-induced acute lung injury mouse model was established by intranasal administration of poly(I: C) (10 mg/kg/day) [30, 38], then mice were treated with SB225002, BX795, or anti-Ly6G antibodies (a neutrophil depleting agent) (Fig. 6A). Compared with control mice, poly(I: C)-challenged mice exhibited significant pulmonary edema and congestion (Fig. 6B). SB225002, BX795, and anti-Ly6G significantly diminished poly(I: C)-induced lung injury (Fig. 6B). H&E staining revealed marked alveolar septal thickening, airway wall thickening, and extensive perivascular inflammatory cell infiltration (predominantly mononuclear cells and neutrophils), as well as congestion of alveolar septal capillaries in poly(I: C)-challenged mice (Fig. 6C). SB225002, BX795, and anti-Ly6G significantly diminished poly(I: C)-induced pathological change (Fig. 6C). Moreover, SB225002, BX795, and anti-Ly6G significantly decreased the W/D ratio in lung tissues (Fig. 6D). Additionally, poly(I: C) stimulation led to protein accumulation in the bronchoalveolar lavage fluid (BALF), and interventions with SB225002, BX795, and anti-Ly6G decreased poly(I: C)-promoted BALF total protein levels (Fig. 6E), which was consistent with the histological evaluations of lung injury. Fig. 6Inhibition of CXCR2 or RIG-I expression ameliorate poly(I: C)-induced acute lung injury in vivo. A Schematic diagram of the establishment of poly(I: C)-induced acute lung injury model. Some elements were created with BioRender.com. B Lung injury score in acute lung injury models with different treatment (n = 4 per group). C Representative images of H&E -stained lung sections after poly(I: C) (10 mg/kg/day) stimulation. Black arrows indicate inflammatory cell infiltration. (Scale bar for 100X, 100 μm; Scale bar for 200X, 50 μm.) (D) The lung W/D ratio in acute lung injury models with different treatment (n = 4 per group). E Total protein concentrations of BALF. F Total cell numbers in peripheral blood. G Total cell numbers in BALF. H Representative images of immunohistochemical staining of MPO in acute lung injury model with different treatment (n = 4 per group). Black arrows indicate MPO+ cells. (Scale bar for 100X, 100 μm; Scale bar for 200X, 50 μm.) (I) The quantitative analysis of the percentage of MPO+ cells in acute lung injury models with different treatment (n = 4 per group). J Concentration of IL-6 (left) and IL-8 (right) in BALF. Data were presented as mean ± SD. For (B, D, J), Kruskal-Wallis test for multiple groups was used (non-normal distribution). For (I), ANOVA for multiple groups was used (normal distribution). NS (not significant), *P < 0.05, **P < 0.01, and ***P < 0.001 Inhibition of CXCR2 or RIG-I expression ameliorate poly(I: C)-induced acute lung injury in vivo. A Schematic diagram of the establishment of poly(I: C)-induced acute lung injury model. Some elements were created with BioRender.com. B Lung injury score in acute lung injury models with different treatment (n = 4 per group). C Representative images of H&E -stained lung sections after poly(I: C) (10 mg/kg/day) stimulation. Black arrows indicate inflammatory cell infiltration. (Scale bar for 100X, 100 μm; Scale bar for 200X, 50 μm.) (D) The lung W/D ratio in acute lung injury models with different treatment (n = 4 per group). E Total protein concentrations of BALF. F Total cell numbers in peripheral blood. G Total cell numbers in BALF. H Representative images of immunohistochemical staining of MPO in acute lung injury model with different treatment (n = 4 per group). Black arrows indicate MPO+ cells. (Scale bar for 100X, 100 μm; Scale bar for 200X, 50 μm.) (I) The quantitative analysis of the percentage of MPO+ cells in acute lung injury models with different treatment (n = 4 per group). J Concentration of IL-6 (left) and IL-8 (right) in BALF. Data were presented as mean ± SD. For (B, D, J), Kruskal-Wallis test for multiple groups was used (non-normal distribution). For (I), ANOVA for multiple groups was used (normal distribution). NS (not significant), *P < 0.05, **P < 0.01, and ***P < 0.001 During the lung injury assessment, it was noted that poly(I: C) caused significant aggregation of inflammatory cells in the alveoli [39]. Further analysis indicated that poly(I: C) induced inflammatory cells aggregation in the peripheral blood and infiltration in the BALF, which was decreased by SB225002, BX795 and anti-Ly6G. (Fig. 6F and G). To evaluate the accumulation and activation of HPNs upon poly(I: C)-induced lung injury, we stained mouse lung sections with myeloperoxidase (MPO). The results implied that the number of MPO+ HPNs significantly increased in poly(I: C)-induced mice lung tissues, whereas SB225002, BX795, and anti-Ly6G markedly reduced the number of MPO-positive cells (Fig. 6H and I). Moreover, SB225002, BX795, and anti-Ly6G significantly decreased poly(I: C)-induced production of the pro-inflammatory cytokines (IL-6 and IL-8) in the BALF of mice (Fig. 6J). These in vivo experiments further emphasized the regulatory roles of RIG-I and CXCR2 in mimic poly(I: C) induced lung injury. Validation of the CXCR2-dependent axis using TLR7/8 agonists Although SARS-CoV-2 is an ssRNA virus, its replication cycle generates dsRNA intermediates that serve as potent ligands for RIG-I, and the synthetic dsRNA analogue poly(I: C) has been widely used to model these immunogenic intermediates [40, 41]. In our study, poly(I: C) was employed as the primary mimic to investigate the RIG-I-CXCR2 axis. To further validate that the observed mechanism is also relevant to the ssRNA genome of SARS-CoV-2, we performed additional experiments using the TLR7/8 agonist ssRNA40 and R848 as alternative viral mimics that directly engages ssRNA sensing pathways. In vitro, as shown in Fig. 7A, stimulation of A549 and BEAS-2B cells with ssRNA40 (2 µg/ml for 24 h) significantly upregulated CXCL1, CXCL2, and CXCL8 mRNA expression (Fig. 7A). Conditioned medium from ssRNA40-treated BEAS-2B cells promoted HPN chemotaxis, an effect that was abrogated by the CXCR2 antagonist SB225002 (Fig. 7B). In vivo, we established an acute lung injury model using intranasal administration of R848 (2.5 µg/g body weight). R848 challenge induced significant lung inflammation, characterized by inflammatory cell infiltration, alveolar septal thickening, and increased lung injury scores (Fig. 7C). Treatment with SB225002 significantly attenuated R848-induced lung injury (Fig. 7C). Immunohistochemical staining for MPO revealed that R848 promoted substantial neutrophil infiltration into lung tissues, which was markedly reduced by SB225002 (Fig. 7D). Taken together, these results demonstrate that both dsRNA intermediates (mimicked by poly(I: C)) and ssRNA viral genomes (mimicked by TLR7/8 agonists) activate the CXCR2-dependent inflammatory axis, supporting the translational relevance of our findings to SARS-CoV-2 infection. Fig. 7TLR7/8 agonists validate the CXCR2-dependent axis in response to ssRNA viral genome mimic. A A549 and BEAS-2B cells were stimulated with ssRNA40 (2 µg/ml) for 24 h. The mRNA expression levels of CXCL1, CXCL2, and CXCL8 were quantified by RT-qPCR. B (Left) Representative images of migrated HPNs toward conditioned medium from BEAS-2B cells treated with ssRNA40 (2 µg/ml) in the presence or absence of SB225002 (10 µM). Five random 200X fields were selected and numbers of cells of each field were counted to calculate the average migrated number. Scale bar, 200 μm. (Right) Quantification of migrated HPNs per 200× field. C Representative images of H&E -stained lung sections after poly(I: C) (10 mg/kg/day) stimulation (n = 5 per group). (Scale bar for 100X, 100 μm; Scale bar for 200X, 50 μm.) (D) Representative images of immunohistochemical staining of MPO in acute lung injury model with different treatment (n = 5 per group). (Scale bar for 100X, 100 μm; Scale bar for 200X, 50 μm.) Data are presented as mean ± SD. For (A, B), Mann-Whitney U test for two-group comparisons was used (non-normal distribution). For (C, D), Kruskal-Wallis test for multiple groups was used (non-normal distribution). *P < 0.05, **P < 0.01 TLR7/8 agonists validate the CXCR2-dependent axis in response to ssRNA viral genome mimic. A A549 and BEAS-2B cells were stimulated with ssRNA40 (2 µg/ml) for 24 h. The mRNA expression levels of CXCL1, CXCL2, and CXCL8 were quantified by RT-qPCR. B (Left) Representative images of migrated HPNs toward conditioned medium from BEAS-2B cells treated with ssRNA40 (2 µg/ml) in the presence or absence of SB225002 (10 µM). Five random 200X fields were selected and numbers of cells of each field were counted to calculate the average migrated number. Scale bar, 200 μm. (Right) Quantification of migrated HPNs per 200× field. C Representative images of H&E -stained lung sections after poly(I: C) (10 mg/kg/day) stimulation (n = 5 per group). (Scale bar for 100X, 100 μm; Scale bar for 200X, 50 μm.) (D) Representative images of immunohistochemical staining of MPO in acute lung injury model with different treatment (n = 5 per group). (Scale bar for 100X, 100 μm; Scale bar for 200X, 50 μm.) Data are presented as mean ± SD. For (A, B), Mann-Whitney U test for two-group comparisons was used (non-normal distribution). For (C, D), Kruskal-Wallis test for multiple groups was used (non-normal distribution). *P < 0.05, **P < 0.01 Discussion Dysregulated immune responses during respiratory viral infections can lead to adverse clinical outcomes; therefore, controlling excessive immune reactions and mitigating lung damage is crucial. Beyond the well-documented increases in number of HPNs in severe COVID-19 patients [42, 43], more recent evidence pointed to significant differences in CXCR2 expression on HPNs among patients with varying disease severities [44]. However, the specific mechanisms underlying these effects remain unclear. In this study, we investigated the CXCR2 expression patterns on HPNs and the molecular mechanisms regulating CXCR2 expression in the context of viral pneumonia, particularly in COVID-19. The results indicated that CXCR2 plays a crucial role in HPN-mediated chemotaxis, inflammatory response, and damage to bronchial and alveolar epithelial cells. RIG-I-mediated MAPK pathway contributes to CXCR2 upregulation in HPNs. Our findings suggest that CXCR2 is a key therapeutic target for controlling inflammatory responses in viral pneumonia. This study provides new insight into the role of RIG-I in the regulation of CXCR2 expression in HPNs during the COVID-19 pandemic. It highlights potential therapeutic targets for managing excessive immune responses and reducing lung injury in affected patients. Interestingly, we observed that on admission, the CXCR2 expression on HPNs in patients with severe COVID-19 was significantly elevated than that in mild cases, which was consistent with previous findings [44]. More importantly, longitudinal analysis revealed that CXCR2 expression on HPNs progressively declined in severe patients during hospitalization, a trend potentially associated with inflammatory control, as supported by the correlation between high initial CXCR2 MFI and poor prognosis. In contrast, CXCR2 expression on HPNs exhibited the opposite trend in mild patients during hospitalization. These dynamic patterns underscored CXCR2 as a potential marker of disease severity and resolution. Several studies have previously reported the phenomenon of overexpression of CXCR2 in HPNs in severe COVID-19 patients [15, 44], but none have explored the specific mechanisms responsible for the upregulation of its expression. Although CXCR2 levels are known to increase during HPNs maturation [45], and can be downregulated via ligand-induced internalization [46], neither fully explains our findings. Previous studies revealed that HPNs in patients with moderate-to-severe COVID-19 were mostly immature, with some appearing as precursor cells [15, 44]. Concurrently, we also observed significantly elevated levels of CXCR2 ligands such as CXCL8 in severe COVID-19 patients. In this study, we proposed a novel and plausible mechanism for CXCR2 overexpression in severe COVID-19: SARS-CoV-2 infection may directly activate the RIG-I pathway in HPNs, leading to transcriptional upregulation of CXCR2 and thereby fueling excessive neutrophil recruitment and inflammation. Dysfunction of RIG-I is reported related to the antiviral immune responses in pneumonia [47]. In our study, RIG-I signaling was confirmed to be activated in HPNs of COVID-19 patients and poly(I: C)-stimulated HPNs, accompanied by an increase in CXCR2 expression. Importantly, the inhibition of the RIG-I significantly attenuated this virus-induced increase in CXCR2, establishing a causal link between RIG-I activation and CXCR2 expression in HPNs. Following, we sought to delineate the underlying mechanism between RIG-I activation and CXCR2 expression. To illustrate this point, The HPNs were isolated from HC and COVID-19 patients for RNA-seq analysis, and the bioinformatic analysis predicted a central mediated mechanism-related to MAPK signaling pathway. Previous studies have confirmed that RIG-I signaling is responsible for the activation of the downstream targets NF-κB and p38 MAPK in influenza and that the inhibition of RIG-I signaling leads to a reduction in pro-inflammatory cytokine production and prevents lung immune damage [48, 49]. The MAPK signaling pathway, including ERK, p38, and JNK, is a major pathway that promotes CXCL8 transcription, regulates HPNs migration, and promotes the expression of CXCR2 in human granulocytes [50, 51]. Based on the bioinformatical prediction and the described information, we hypothesized that MAPK pathways serve as the critical intermediary bridging RIG-I to CXCR2 expression in HPNs. As expected, in the poly(I: C)-stimulated HPNs, RIG-I inhibition significantly blocked the MAPK signaling, and MAPK blocking further inhibited expression of CXCR2. Based on these results, we deduce that the RIG-I/MAPK axis is a crucial signaling cascade responsible for upregulating CXCR2 expression in HPNs during viral infection. Moreover, in poly(I: C)-induced lung injury animal model, inhibition of this signaling cascade significantly attenuated lung tissue damage, reduced the infiltration of neutrophils and other inflammatory cells into the lungs, and downregulated the expression of key pro-inflammatory cytokines (e.g., IL-8, IL-6). This evidence substantially strengthens our proposed mechanism that the RIG-I–MAPK–CXCR2 regulatory cascade is critical during inflammatory response of viral infection. Previous studies have established that CXCR2 activation leads to downstream MAPK signaling, particularly ERK1/2 phosphorylation, in neutrophils [52, 53]. In contrast, the present study reveals a distinct regulatory axis wherein RIG-I activation promotes CXCR2 expression via the MAPK pathway, representing an upstream regulatory mechanism rather than a downstream effector function, which has also been reported previously [54, 55]. Taken together, these findings suggest a potential bidirectional regulatory relationship between CXCR2 signaling and the MAPK pathway. Importantly, to our knowledge, this is the first demonstration that RIG-I controls CXCR2 expression at the transcriptional level in HPNs, providing a novel mechanistic link between viral sensing and neutrophil-mediated inflammation. Because of its critical role in the inflammatory process, CXCR2 has been investigated as a pivotal therapeutic target for controlling inflammation in several studies. For instance, in a mouse model of lipopolysaccharide-induced neuroinflammation, the CXCR2 antagonist SB225002 reduces the recruitment of HPNs to the brain [56]. Similarly, in a mouse model of influenza-induced lung injury, the combination of a CXCR2 antagonist with antiviral treatment significantly inhibited HPNs infiltration into the lungs compared with antiviral treatment alone [56]. These results are coincidence with our finding that SB225002 treatment decreased the HPN migration to epithelial cells and inflammatory factors expression in our chemotaxis experiment, further confirming that CXCR2 inhibitors may serve as potential therapeutic targets for the prevention of HPNs-mediated inflammatory diseases, thereby attenuating tissue damage caused by inflammation [57]. Beyond its role in HPN chemotaxis, previous research has demonstrated that CXCR2 deficiency (in CXCR2−/− mice) leads to phagocytic activity reduction in HPNs, along with decreased levels of ROS, F-actin, and α-tubulin [58], indicating that the CXCR2 pathway plays a role in regulating HPNs function. Our in vitro experiments confirmed that the CXCR2 pathway regulates various HPNs functions, including cytokine secretion and ROS production. Collectively, this evidence supports the role of CXCR2 in modulating HPNs activity, making it a promising target for therapeutic intervention. Collectively, these findings offer a promising strategy to mitigate excessive neutrophilic inflammation and prevent the progression to acute respiratory distress syndrome in severe viral pneumonia. Despite the insights provided, this study has several limitations. First, although detail information has been provided, the cohort from a single-center source is still limited, encouraging us to conduct larger, multi-center longitudinal studies to validate the generalizability of our findings. Second, to address the concern that poly(I: C) is a dsRNA mimic, we performed additional validation experiments using the TLR7/8 agonists ssRNA40 (in vitro) and R848 (in vivo), which directly mimic the ssRNA genome of SARS-CoV-2. Our results demonstrated that these agonists also activate the CXCR2-dependent inflammatory axis, confirming that our findings are relevant to both dsRNA intermediates and ssRNA genomes. Nevertheless, these viral mimics may not fully recapitulate the complexity of live SARS-CoV-2 infection, including viral replication, cytopathic effects, and adaptive immune responses. Future studies using live SARS-CoV-2 under appropriate biosafety conditions are warranted to further strengthen the clinical relevance of our findings. Finally, from a translational perspective, although inhibiting the RIG-I/MAPK/CXCR2 axis showed efficacy in our models, feasibility of systemic inhibition of these pathways, particularly RIG-I and MAPK, remains obscure. Conclusion This study examines the dynamical changes in CXCR2 expression in the HPNs of COVID-19 patients and uncover its relationship with inflammation. CXCR2 overexpression in HPNs is found in severe patients, which is emphasized associated with a deteriorated inflammatory status and poorer prognosis. Mechanically, RIG-I, a protein that is highly expressed in the HPNs of COVID-19 patients, is proposed to upregulate CXCR2 expression by activating the MAPK pathway. Targeting this mechanism could control the immune responses and reduce lung injury in viral pneumonia, thereby providing a critical basis for future translational research. Supplementary information Supplementary Material 1. Supplementary Fig. S1. Supplementary Fig. S2. Supplementary Fig. S3. Supplementary Material 1. Supplementary Fig. S1. Supplementary Fig. S2. Supplementary Fig. S3. Supplementary Material 2. Supplementary Table S1. Supplementary Table S2. Supplementary Material 2. Supplementary Table S1. Supplementary Table S2. Supplementary Material 3 Supplementary Material 3