Introduction

Liver cancer is a life-threatening illness and one of the fastest-growing cancer types in the globe. Hepatocellular carcinoma (HCC) is the most common form of liver cancer and constitutes more than 90% of the primary malignant tumor of the liver [1]. Nowadays, HCC is the fifth most common cause of cancer worldwide, and becomes the second leading cause of cancer death after lung cancer in men. Over the years, it has been widely accepted that various risk factors involve the initiation and progression of HCC including viral hepatitis (hepatitis B and hepatitis C), alcoholic liver disease, and non-alcoholic liver steatohepatitis/non-alcoholic fatty liver disease. The estimated 5-year survival rate for HCC is less than 20%, owing to a large proportion of patients being diagnosed at advanced stages when curative options are not feasible [2]. Large randomized, controlled trials (RCTs) found that regular surveillance reduced HCC-related mortality and was beneficial to improve overall survival [3]. Blood test-based biomarkers including alpha fetal protein (AFP) together with prothrombin induced by vitamin K absence-II (PIVKA-II), osteopontin (OPN), annexin A2, Golgi protein-73 and squamous cell carcinoma antigen (SCCA) are used in clinical practice, whereas the value of these indicators for HCC exhibits low specificity in clinical diagnosis and prognostic values [4]. Thus, there is urgent clinical need to identify novel HCC biomarkers for improving HCC diagnosis and monitoring treatment outcome.

HCC occurs in 80–90% of patients with cirrhosis, while ascites is one of the major complications of liver cirrhosis and indicates advanced disease and poor prognosis. Ascites is regarded as pathological fluid accumulation within the abdominal cavity. The malignant ascites in HCC is characterized by high volume, persistence and recurrent occurrence. Besides, ascites may also be related to large tumor burden and vascular invasion of HCC, and severe ascites could be life-threatening [5]. Therefore, molecular analysis of malignant ascites from HCC patients may provide valuable information for differential diagnosis, clinical surveillance, medical screening, and intervention. Though there is a close link between ascites and prognosis, the diagnostic efficacy of ascitic fluids in HCC patients remains ambiguous and needs to be further elucidated.

The hallmark of tumor progression is significantly elevated inflammatory cytokines and chemokines accompanied by infiltration of neutrophils in the ascites. It has been reported that tumor microenvironment (TME) can regulate neutrophil recruitment and polarization. Tumor-associated neutrophils (TANs) correlate with increased tumor growth, invasiveness and metastasis in vast solid tumors [6, 7]. Notably, TANs could exert dual effects on tumor progression based on their activation status, and can be further divided into subgroups through cell surface molecule. Growing evidence supports that TANs contribute to the pathogenesis of HCC, since neutrophils serve as key mediators of the immunosuppressive environment to promote tumor growth, metastatic capacity as well as extracellular matrix remodeling [8]. Also, a cohort study showed that neutrophils in patients with HCC were the only cell type related to patient outcome, emphasizing the importance of TANs in HCC [9]. Recently, several studies have illustrated that neutrophil extracellular traps (NETs) could trigger tumorous inflammatory responses and promote tumor metastasis, and treatments targeting NETs and its related signaling pathways display therapeutic potential to reduce tumor metastasis [10,11,12]. These finding implicated that targeting NETs rather than neutrophils themselves could be an effective approach against tumor progression [10]. It has been reported that elevated NETs was found in peripheral blood and tumor tissues in patients with HCC [13], while the expression level of NETs in ascites is unknown.

In this study, we aimed to collect clinical samples to investigate NETs level in ascites, and explored the diagnostic efficiency of ascitic NETs together with current biochemical parameters to illustrate the clinical significance of NETs in patients with HCC. Moreover, the formation of NETs is triggered by which components in TME and which subpopulation of TANs are responsible for NETs release, as well as the function and clinical significance of NETs in ascites would be further elucidated. Taken together, our investigation revealed the critical value of NETs in ascitic fluid samples, thus providing a scientific basis for developing ascitic NETs-based diagnosis and treatments in patients with HCC.

Materials and methods

Human patients and controls

HCC patients with malignant ascites (n = 35) and age- and sex- matched patients with benign ascites (n = 35) were enrolled from the Second Affiliated Hospital of Chongqing Medical University from September 2022 to March 2024. Clinical data of patients were recorded including counts of white blood cells (WBC), neutrophils, lymphocytes and C-reactive protein (CRP), as well as ascitic glucose, total protein, albumin, LDH and ADA levels. Informed written consents were obtained from all recruited subjects. All samples were obtained and approved by the Institutional Ethics Committee of the Second Affiliated Hospital of Chongqing Medical University in accordance with the Declaration of Helsinki.

Routine laboratory tests for ascites and blood

Biochemical indicators including lactate, glucose, total protein, albumin, lactate dehydrogenase (LDH) and adenosine deaminase (ADA) in ascites and serum were detected by AU 5800 analyzer (Beckman Coulter Inc., CA, USA) with the test kits (Medicalsystem Biotechnology, Ningbo, China).

Quantification of NETs concentrations

Blood and ascites samples from each patient were obtained, and then serum and supernatant of ascites were isolated by centrifugation at 3000 rpm for 5 min at 4 °C and were stored in aliquots at − 80 °C. For detection of NETs concentrations, calprotectin, double strain DNA (dsDNA), Citrullinated Histone H3 (CitH3), myeloperoxidase (MPO) and MPO-DNA were performed using ELISA kits (elabscience, Wuhan, China) and all samples were run in duplicate according to the manufacture’s instruction.

Cell culture and reagents

Fresh peripheral blood obtained from healthy volunteers for the purification of primary human neutrophils. PBS-diluted fresh human blood were carefully layered to surface of solution A. After centrifugation for 25 min at 1800 rpm, neutrophil fractions were obtained and then red blood cell lysis buffer were added to cells. After lysis of residual erythrocytes, cells were washed with cell washing buffer for 3 times. When performed correctly, this method has been shown to yield samples of > 85% neutrophils with > 95% viability (P9040, Solarbio life sciences, Beijing, China). After isolation, neutrophils were maintained in RPMI 1640 medium (Gibco Invitrogen Corp., Carlsbad, CA, USA) containing 10% FBS (Thermo Fisher, Waltham, MA USA). For in vitro study, neutrophils were stimulated with different concentrations of lactate (MedChemExpress, NJ, USA) for 2 h at 37 °C and 5% CO2.

Identification and transcriptional profile of lactate target genes using RNA-sequencing

After optimization, fresh neutrophils were treated with lactate (20 mM) for 2 h in vitro. Neutrophils were collected in TRIzolTM Reagent and then sent to APEXBIO Ltd., Shanghai, China, for library preparation and transcriptomic sequencing with next-generation sequencing. Samples were sequenced on an Illumina NovaseqTM 6000 (LC Sciences, USA). Differential gene expression analysis of two experiments with three biological replicates was performed using DESeq2 and genes with P < 0.05 were regarded as differentially expressed. Cluster analysis of the DEGs was applied to describe the expression patterns of the DEGs under different treatments.

Flow cytometric analysis of CD66b, CD11b and PD-L1

For analyzing the expressions of CD66b, CD11b and PD-L1, neutrophils were stained with FITC anti-human CD66b antibody (1:100, BioLegend), PerCP/Cyanine5.5 anti-human CD11b antibody (1:100, BioLegend) and APC anti-human CD274 (B7-H1, PD-L1) antibody (1:100, BioLegend) for 30 min at room temperature. Then cells were washed with PBS for 3 times and cells were acquired by flow cytometer (Celula Sparrow, China) and data were analyzed by FlowJo (v10) software.

Flow cytometric analysis of intracellular MPO levels

To analyze intercellular MPO levels, neutrophils were stimulated with lactate (5 mM, 10 mM, 20 mM) for 2 h. Cells were collected and fixed with fixation buffer for 30 min, and then incubated with permeabilization buffer for another 30 min at room temperature. Next, cells were stained with FITC labeled anti-human MPO (1:100, BioLegend, Inc.,San Diego, CA, USA). After last wash, cells were acquired by flow cytometry and data were analyzed by FlowJo (v10) software.

RNA extraction and QT-PCR

Total RNA from cell pellets of ascites specimens were extracted using QIAzol reagent (Qiagen Inc., Valencia, USA) and reverse transcribed into complementary DNA using Prime-Script RT Master Mix (Takara Bio Inc., Shiga, Japan). Quantitative real-time PCR reactions were conducted with the SYBR Primix Ex TaqT (Takara Bio Inc.). The housekeeping gene GAPDH was considered as a reference gene. All primer sequences are shown in Table 1.

Table 1 Primer sequences
Full size table

Statistics

Data were analyzed using GraphPad PRSIM software version 8.0 with Mann–Whitney to determine the significance of NETs in HCC patients with malignant ascites. Spearman’s correlation coefficient was used to evaluate the correlation between NETs-related markers and LDH. Cell-based data was analyzed by Student’s t-test or one-way ANOVA. To analyze the diagnostic value of biomarkers, ROC curves were generated and the AUC were determined by using SPSS 17.0. A significance level of p < 0.05 was considered significantly different.

Results

Elevated tumor-associated neutrophils in ascites from patients with HCC

Recently, the role of neutrophils in cancer development has become an area of great interests. In this set, ascites specimens were collected to identify whether neutrophils were involved in pathogenesis in HCC. The basic demographic and characteristics of HCC patients with malignant ascites and patients with benign ascites caused by other chronic liver diseases enrolled in this study are summarized in Table 2. Increased white blood cells and neutrophils were observed in HCC patients with malignant ascites (Table 2). Consistently, clinical routine tests for ascites revealed that elevated polymorphonuclear leukocyte (PMNL) proportion (%) was found in malignant ascitic fluids (Table 3). Then, we observed tumor cells accompanied by abundant TANs in malignant ascites from HCC patients through cytological analysis of ascitic fluids when compared with benign ascites (Fig. 1A). Next, we detected the proportion of CD11b+CD66b+ TANs in blood and ascites using flow cytometry. Our data showed that the infiltration of CD11b+CD66b+ TANs was significantly increased in both periphery blood and ascites in HCC patients with malignant ascites than chronic liver diseases patients with benign ascites (Fig. 1B, C), indicating TANs in ascitic fluid exerted an essential role in HCC pathogenesis.

Table 2 Baseline characteristics of HCC patients with malignant ascites and chronic liver diseases patients with benign ascites
Full size table
Table 3 Differential white cell counts of HCC patients with malignant ascites and patients with benign ascites
Full size table
Fig. 1
figure 1

Increased tumor-associated neutrophils in HCC patients with malignant ascites. A Cytological analysis of malignant ascitic fluids from HCC patients and benign ascites from chronic liver diseases patients. B Flow cytometric analysis of TANs proportion in periphery blood. C Flow cytometric analysis of TANs proportion in ascites. **P < 0.001, ***P < 0.0001

Full size image

Accumulation of pro-tumor PD-L1+ TANs in HCC ascites

Recently, the plasticity and diversity of TANs have been described in several tumors, since TANs are able to polarize into pro-tumor phenotype or anti-tumor phenotype in TME. It has been reported that PD-L1-expressing neutrophils are considered pro-tumor as they can suppress cytotoxic T cells which in turn results in promoting disease progression [14]. Therefore, we analyzed the percentage of CD11b+CD66b+PD-L1+ TANs in peripheral blood and ascites from the same patients. The proportion of blood PD-L1+ TANs showed no statistical difference between HCC and control group (Fig. 2B, C), while PD-L1+ TANs was significantly elevated in malignant ascites when compared with benign ascites (Fig. 2D, E), implicating TANs acquired the polarization potential into pro-tumor phenotype in ascitic fluids in patients with HCC. Then, representative markers of pro-tumor TANs were further identified, the results revealed that pro-tumor TANs-associated gene expression including Arg-1, CCL2, CCL5 and TGF-β (Fig. 2F–I) were dramatically upregulated in HCC ascites. However, we found no significant differences of CXCR2 in these two groups (Fig. 2J). Thus, our data revealed that TANs polarized into pro-tumor phenotype in ascitic fluids in HCC patients.

Fig. 2
figure 2

Elevated PD-L1+ TANs in HCC patients with malignant ascites. A Gating strategy of PD-L1+ TANs in ascites. B Percentage of PD-L1+ TANs in periphery blood and cell surface expression of PD-L1 (C) were analyzed by flow cytometry. D Proportions of PD-L1+ TANs in ascites and PD-L1 levels (E) were analyzed by flow cytometry. Cell pellets of ascites were obtained for PCR analysis, the mRNA expression level of pro-tumor-related genes including Arg-1 (F), CCL2 (G), CCL5 (H), TGF-β (I) and CXCR2 (J) were measured. *P < 0.05, **P < 0.001, ns not significant

Full size image

Lactate was involved in TANs polarization in ascites

It has been reported that lactate dehydrogenase (LDH) serves as a predictor of clinical outcomes in HCC [15]. Consistently, significantly higher levels of LDH were detected in both serum (Fig. 3A) and ascitic fluids (Fig. 3B) in HCC patients compared with control individuals. Then, serum and ascites lactate were examined by lactic acid detection kits. The results showed that significantly increased lactate were detected in serum (Fig. 3C) and ascites (Fig. 3D) in HCC.

Fig. 3
figure 3

PD-L1 expression on neutrophils mediated by lactate. LDH levels were detected in serum (A) and ascitic fluids (B), and lactate concentrations were measured in serum (C) and ascitic fluids (D). Fresh neutrophils were isolated from peripheral blood, and then stimulated with 20 mM lacate for 2 h. Total RNA was extracted from neutrophils and transcriptome analysis was performed to identify lactate target genes. Volcano plot comparison of total differentially expressed genes (E). GSEA analysis enrichment in VEGF signaling pathway (F), PD-L1 expression (G) and NETs formation (H). Heat map diagrams of PD-L1 expression and NETs formation-related genes (I). RNAseq results were confirmed using real-time qPCR and the GAPDH housekeeping gene was used as the reference gene (J-O). Fresh neutrophils were isolated from peripheral blood, and then stimulated with 5 mM, 10 mM and 20 mM lacate for 2 h. Gating strategy of PD-L1 expression on neutrophils as shown in (P). Percentage of PD-L1+ neutrophils treated by lactate (Q) and mean fluorescence intensity of PD-L1 on cell surface (R-S) were analyzed by flow cytometry. *P < 0.05, **P < 0.01, ***P < 0.001. Abbreviation: C, control group; L, Lactate-treated group

Full size image

To elucidate the effect of lactate on neutrophils, fresh peripheral neutrophils were isolated and then stimulated with lactate in vitro. Then, RNAsequencing (RNAseq) was applied to identify lactate target genes in neutrophils at the transcriptional level. Volcano plot revealed that lactate engagement on neutrophils resulted in distinct transcriptomic states with 212 upregulated and 236 downregulated genes. Next, gene set enrichment analysis (GSEA) was performed to analyze the gene expression profiles (Fig. 3E). The results showed that differentially expressed genes (DEGs) were mainly enriched in VEGF signaling pathway (Fig. 3F), PD-L1 expression and PD-1 checkpoint pathway in cancer (Fig. 3G) together with neutrophil extracellular traps formation (Fig. 3H). Furthermore, cluster analysis of DEGs implicated that PD-L1 expression and NETs formation-related genes were changed after lactate treatment. Real-time quantitative PCR (qPCR) was used to validate these target genes in three biological replicates and the results were coincident with the RNAseq findings (Fig. 3I). Lactate significantly upregulated CD274 (also known as PD-L1), STAT1, protein tyrosine phosphatase 6 (PNPT6), advanced glycation end-product (AGER) expression and downregulated BATF2 and SAMD4B expression in neutrophils (Fig. 3J–O). To further confirm these RNAseq data and investigate the potential role of lactate on neutrophils in TME, neutrophils were stimulated with lactate in vitro. Consistently, we observed that lactate promoted the percentage of CD66b+PD-L1+ neutrophils (Fig. 3P, Q) as well as mean fluorescence intensity (MFI) (Fig. 3R, S) of PD-L1 in a dose-dependent manner using flow cytometry. These results illustrated that lactate played a key role in polarizing TANs with pro-tumor phenotype in TME.

Lactate facilitated NETs generation of PD-L1+ TANs

RNAseq data revealed that lactate was associated with NETs formation. In addition, several studies have reported that PD-L1 might participate in NETs formation. To evaluate whether lactate involved in NETs formation, peripheral neutrophils were treated with lactate for 2 h, and then stained with NETs-related marker MPO through intracellular staining for flow cytometric analysis. Our data revealed that lactate promoted the percentage of PD-L1+ MPO+ TANs (Fig. 4A, B) and induced MPO production from PD-L1+ TANs (Fig. 4C, D), indicating that lactate could promote NETs release. Then, cell supernatant was collected to measure MPO-DNA by ELISA, we found that lactate accelerated NETs generation in a dose-dependent manner (Fig. 4E). Taken together, our results confirmed that lactate was able to induce NETs production from PD-L1+ neutrophils, implicating high concentration of lactate in malignant ascites regulated polarization and NETs formation of TANs in HCC.

Fig. 4
figure 4

Enhanced NETs release by PD-L1+ neutrophils stimulated with lactate. Fresh neutrophils were isolated from peripheral blood, and then stimulated with 5 mM, 10 mM and 20 mM lacate for 2 h. A Gating strategy of MPO in PD-L1+ neutrophils. B Percentage of MPO+PD-L1+ neutrophil stimulated with lactate and C, D MPO expression level in PD-L1+ neutrophils were analyzed by flow cytometry. E Concentration of MPO-DNA were detected using ELISA kits. **P < 0.01, ***P < 0.001

Full size image

Enhanced level of NETs in HCC ascites rather than periphery

We analyzed NETs-related markers including calprotectin, dsDNA, CitH3, MPO and MPO-DNA in both ascites and blood specimens from HCC patients and control individuals. Interestingly, there was no statistical difference in serum NETs concentrations between these two groups (Fig. 5A–E). However, dramatically elevated calprotectin, dsDNA, CitH3, MPO and MPO-DNA were examined in ascitic fluid from HCC patients when compared to those in benign ascites (Fig. 5F–J). These results illustrated that ascitic NETs level rather than that in the circulation had greater sensitivity to distinguish HCC patients with malignant from benign ascites, suggesting that ascitic NETs could serve as biomarker candidates in HCC patients with malignant ascites from individuals with benign ascites.

Fig. 5
figure 5

NETs-related markers levels in serum and ascites. Serum and ascites samples from the same patients were collected. Serum calprotectin (A), dsDNA (B), CitH3 (C), MPO (D) and MPO-DNA (E) and ascitic calprotectin (F), dsDNA (G), CitH3 (H), MPO (I) and MPO-DNA (J) were measured by ELISA kits. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. not significant

Full size image

Clinical relevance of NETs in HCC patients with malignant ascites

Biochemical analysis of ascitic fluid samples is widely carried out for clinical routine tests. Glucose, total protein (TP), albumin, lactate dehydrogenase (LDH) and adenosine deaminase (ADA) are useful indicators for clinical utility. To clarify the precise effects of NETs in HCC ascites, we analyzed the correlation between NETs-related markers and biochemical indicators in ascites. Our results showed that ascitic calprotectin, dsDNA, CitH3, MPO and MPO-DNA levels positively correlated with TP, albumin, LDH and ADA, whereas glucose negatively correlated calprotectin, dsDNA, CitH3, MPO and MPO-DNA in ascitic fluids, respectively (Fig. 6 and Fig. S1–S5). Of note, our data showed that a significantly positively correlation between LDH and NETs-related markers including calprotectin, dsDNA, CitH3, MPO and MPO-DNA in patients with malignant ascites (Fig. 6A–E). In all, our results indicated an essential role of NETs in HCC progression.

Fig. 6
figure 6

Relationship of asctic LDH with ascitic NETs levels. Correlation between LDH with calprotectin (A), dsDNA (B), CitH3 (C), MPO (D) and MPO-DNA (E) in HCC patients with malignant ascites

Full size image

Diagnostic value of ascitic NETs for predicting HCC

To clarify the diagnostic efficiency of ascitic NETs in discriminating HCC patients with malignant ascites from patients with benign ascites, ROC analysis was used for generating area under curve (AUC) of NETs-related markers. As shown in Fig. 7, the AUC of MPO-DNA (0.7878, p < 0.0001) was higher than AUC for ascitic LDH (0.7404, p = 0.0005), calprotectin (0.7176, p = 0.0017), dsDNA (0.6690, p = 0.0150), MPO (0.6812, p = 0.0091), CitH3 (0.7249, p = 0.0012), indicating NETs-related MPO-DNA served as a better biomarker than LDH. Then, the combination of NETs-related markers and ascitic LDH were analyzed, the combination of LDH with calprotectin, dsDNA, MPO, CitH3 and MPO-DNA generated AUC values of 0.725, 0.723, 0.718, 0.763 and 0.813, respectively. The AUC values of combined LDH with CitH3 and combined LDH with MPO-DNA were larger than that of LDH alone. Moreover, ROC curves of LDH combined CitH3 and MPO-DNA showed highest value of AUC (0.841, p < 0.0001) (Supplementary Table ). Collectively, these results suggested that combination of NETs-related markers including MPO-DNA and CitH3 with traditional ascitic biochemical index could improve diagnostic efficacy in malignant ascites diagnosis.

Fig. 7
figure 7

Differentiating power of ascitic NETs for diagnosis of malignant ascites. ROC of NETs-related calprotectin (A), dsDNA (B), CitH3 (C), MPO (D) and MPO-DNA (E) and NETs markers in combination with LDH (F) for differentiating malignant ascites from benign ascites

Full size image

Discussion

HCC is usually diagnosed at an advanced stage and related to high mortality. Currently, clinical trials involving combined locoregional approaches and systemic therapies, together with immune checkpoint inhibitors (ICIs) and other anti-tumor agents have been carried out to exhibit compelling clinical efficacy, which laid a solid foundation in the immunotherapy of HCC [16, 17] Of note, ICIs are associated with various immune-related adverse events (irAEs) including hearing loss and neurotoxicity [18,19,20]. Therefore, research on immune-related responses and biomarkers are further required. Malignant ascites represents a significant clinical challenge in the management of advanced HCC [14]. A growing number of studies have identified that neutrophils play a central role in pathogenesis of HCC, such as tumorigenesis, local tumor progression and metastasis. Accumulation of TANs is a biomarker of poor prognosis in HCC [21]. Here, we observed significantly increased CD11b+CD66b+ TANs in malignant ascites from HCC patients compared with that in benign ascites. With the deep investigations of TANs, neutrophils are found to display considerable phenotypic heterogeneity, their phenotypes are highly dynamics as they are influenced by tissues they reside, maturation status and components in TME [22]. However, the subpopulation of TANs in HCC has not been determined. In the current study, we observed enhanced infiltration of CD11b+CD66b+PD-L1+ TANs in ascitic fluid instead of peripheral blood in HCC patients, as well as increased expression of pro-tumor TANs-associated genes such as Arg-1, CCL2, CCL5 and TGF-β. PD-L1+ TANs exhibited pro-tumor phenotype and have been studied in several types of tumors. For example, gastric tumor-derived GM-CSF induces neutrophil PD-L1 expression via JAK and STAT3 pathways to suppress T cell immunity [23]. In pancreatic ductal adenocarcinoma, increased levels of PD-L1 on neutrophils are associated with metabolic reprogramming and followed by tumor metastasis [24]. He and colleagues reported that PD-L1+ TANs were dominantly accumulated in peritumoral tissues than intratumoral tissues [25]. Moreover, PD-L1+ TANs negatively correlated with overall survival of HCC patients [26]. Our data described that neutrophils polarized into pro-tumor phenotype in HCC ascites which are a crucial step to suppress immune responses of T cells and promote tumor progression.

To investigate the components in ascites involved in TAN polarization, biochemical parameters used commonly in ascitic tests were analyzed (Table 2). Ascitic LDH was significantly elevated in HCC patients compared to that in control group. The role of LDH have been widely studied in heterogeneous populations of HCC patients, serum LDH has been tested as a prognostic biomarker to predict disease outcome [27]. In addition, ascitic LDH could be helpful in distinguishing between malignant and benign conditions [28]. LDH is an important enzyme in the anaerobic metabolic process; its function is to catalyze the conversion of pyruvate to lactate [29]. Thus, we examined lactate concentrations in ascites, and the results revealed that ascitic lactate level was higher in HCC patients. High lactate levels can be derived from HCC tumor cells as the reprogramming of glucose metabolism in tumor cells. Various research groups have described that lactate was associated with PD-L1 expression on neutrophils [30]. Therefore, we isolated neutrophils from peripheral blood and followed by stimulated with lactate for RNAseq analysis. It was notable that lactate induced a significant elevated CD274, STAT1, PTPN6 and AGER expression and dramatically decreased SAMD4B and BATF2 expression. It has been reported that STAT1 was positively correlated with PD-L1 expression in several tumors and regulated NETs formation to aggravate inflammation [31, 32]. Studies revealed that higher expression of PTPN6 was associated with poor prognosis in cancer since PTPN6 could mediate proliferation, invasion and migration of cancer cells [33]. AGER has been reported to function as a tumor promoter to facilitate tumor growth and metastasis [34]. Additionally, AGER is closely related to NETs signaling immune pathway [35]. Recently, SAMD4B was recognized as a tumor suppressor in HCC cohort, low expression of SAMD4B could induce PD-L1 expression, therefore evoking immune escape of tumor cells [36]. In non-small cell lung cancer, BATF2 was negatively correlated with PD-L1 expression in patients to exert anti-tumor responses [37]. Above all, lactate might induce these PD-L1 expression-related genes to mediate TAN polarization in TME. Consistently, our results validated that lactate induced PD-L1 expression on neutrophils in a dose-dependent manner by flow cytometry. Collectively, these results suggested that ascitic lactate might be responsible for TAN polarization to exacerbate tumor progression in HCC. As shown by GSEA analysis, lactate was able to accelerate NETs formation. Awatshi and colleagues found that lactate induced elastase production and DNA release, indicating that lactate could be related to NETs generation [38]. In this study, we observed significantly elevated MPO production in PD-L1+ neutrophils stimulated with lactate in a dose-dependent manner, and higher concentrations of MPO-DNA were measured in the cell supernatant. These data indicated that ascitic lactate was capable of inducing TAN polarization on neutrophils as well as NETs release from PD-L1+ neutrophils.

Nowadays, NETs has become a “hot spot” in HCC research. Emerging evidence showed that NETs displayed an important role in tumor growth and metastasis in HCC via mediating inflammatory responses and metabolic reprogram [39]. More importantly, we previously found NETs level were significantly elevated in liver tissues and serum of HBV-related HCC patients, and NETs might serve as a biomarker for predicting HBV-related HCC progression [13]. Since differentiating malignant ascites from benign ascites has always been a clinical difficult, we therefore explored the clinical significance of ascitic NETs in HCC patients. It has been reported that one of NETs markers calprotectin in ascites could predict early occurrence of HCC. In this study, ascites and serum from HCC patients were collected and NETs-related markers including calprotectin, dsDNA, CitH3, MPO and MPO-DNA were then detected. Interestingly, we found no statistical differences in circulating NETs levels between HCC patients and individuals suffered from other chronic liver diseases. However, significantly higher ascitic calprotectin, dsDNA, CitH3, MPO and MPO-DNA levels were measured in malignant ascites compared to benign ascites, implicating that ascitic NETs had greater potential in discriminating malignant ascites from benign ascites than that in serum. Furthermore, ROC analysis was carried out to identify clinical diagnostic efficiency of NETs. We found that AUC value of MPO-DNA was highest among classic ascitic marker LDH and other NETs-associated markers, and combination of LDH with CitH3 or MPO-DNA improved diagnostic efficacy. Notably, combined LDH, CitH3 and MPO-DNA exhibited the largest values of AUC, these results indicated that NETs might be a candidate biomarker, and NETs combined with ascitic LDH effectively increased diagnostic efficacy in discriminating malignant from benign ascites. However, there are some limitations of this study, First of all, sample size in our study is limited and all specimens were recruited from one hospital, large cohort of HCC patients with malignant ascites and sex- and age- matched patients with benign ascites should be introduced and our results need to be verified in multicenter study. Secondly, dynamic change of serum and ascitic NETs level should be measured to describe the fluctuation of NETs concentrations in HCC progression. Last but not least, HCC-related disease scoring systems should be included to analyze the correlation of ascitic NETs with tumor grades, stages and recurrence for further evaluating prognostic values of NETs.

In conclusion, we observed accumulation of pro-tumor phenotype PD-L1+ TANs, as well as significantly elevated LDH and lactate in malignant ascites. Moreover, in vitro study was conducted to confirm that lactate was able to induce PD-L1 expression and NETs release. Next, dramatically increased NETs-associated parameters were examined in malignant ascites and NETs was proved to be a diagnostic biomarker for malignant ascites. To our knowledge, this is the first study to explore the clinical values of NETs in distinguishing HCC patients with malignant ascites from chronic liver diseases patients with benign ascites. This study provides a new perspective for differential diagnosis between malignant and benign ascites.