Abstract Macrophages play a crucial role in malignant pleural effusion (MPE), a frequent complication of advanced cancer. While C1q^+ macrophages have been identified as a pro-tumoral cluster, direct evidence supporting the role of C1q-mediated macrophages remains to be elucidated. This study employed global and macrophage-specific knockout mice to investigate the role of C1q in MPE. The data demonstrated that C1q deficiency in macrophages suppressed MPE and prolonged mouse survival. scRNA-seq analysis of the C1qa^-/- mouse MPE model revealed that C1q deficiency significantly decreased the proportion of M2 macrophages in MPE. In vitro experiments suggested that C1q expression was gradually upregulated during M2 polarization, which was C1q-dependent, as was antigen presentation. Deficiency of C1q in macrophages rescued the exhausted status of CD8^+ T cells and enhanced the immune activity of CD8^+ T cells and NK cells in both MPE and pleural tumors. Cell-to-cell interaction analysis demonstrated that C1q deficiency attenuated the immunoinhibitory effects of macrophages on NK cells by downregulating the CCR2-CCL2 signaling axis. Metabolomic analysis revealed significantly elevated hippuric acid levels in C1q-deficient mouse MPE. Treatment with either hippuric acid or a CCR2 antagonist inhibited MPE and tumor growth, with an even more pronounced effect observed when both treatments were combined. Keywords: malignant pleural effusion, C1q, macrophages, metabolomics, NK cells Introduction Malignant pleural effusion (MPE) is a frequent complication of advanced cancer [37]^1, accounting for 15% to 35% of exudative pleural effusions [38]^2. The overall survival (OS) of MPE patients ranges from 3 to 12 months [39]^3^, [40]^4. The primary mechanism of MPE formation is the increase in vascular permeability, usually associated with the lymphatic drainage obstruction at the pleura and/or mediastinum [41]^2. Furthermore, the inflammatory signaling network and immune cell regulation in the pleural cavity are key participants in the formation of MPE [42]^1^, [43]^5^, [44]^6. In addition to tumor cells, MPE contains large numbers of immune cells including lymphocytes, macrophages, and granulocytes [45]^7^-[46]^10, all of which function in the occurrence and development of MPE; however, the exact mechanism remains incompletely elucidated. Tumor-associated macrophages play vital roles in the tumor environment. It is currently believed that a pro-tumoral differentiated monocyte population exhibiting an M2-like signatures is thought nowadays to be one of the key factors contributing to tumor promotion [47]^11. Our research has revealed a significantly higher presence of CD206^+ macrophages (M2 macrophages) in MPE compared to non-MPE, suggesting that CD206^+ macrophages are a promising candidate for the lung cancer-originated MPE diagnosis [48]^12. Furthermore, macrophages with M2 signatures are associated with even worse prognosis in lung adenocarcinoma (LUAD) patients [49]^13. In recent years, an increasing number of signature genes, such as TREM2, SSP1, APOE, and C1Q, have been identified to classify macrophages into novel subsets beyond the traditional M1 or M2 categories [50]^14^-[51]^16. Previous research has established the presence of complement proteins can be detected in MPE [52]^17. C1q, the initial recognition molecule of the classical complement pathway [53]^18, is an 18-subunit protein consisting of C1QA, C1QB, and C1QC in equal molar ratios [54]^19. Beyond its role in the immune tolerance maintenance and the apoptotic cell clearance [55]^20, C1q directly influences cell differentiation, adhesion, migration, and proliferation in various cell types, participating in numerous immune-related physiological and pathological processes [56]^21^-[57]^23. In the context of cancer, studies have associated C1q with the progression of diverse malignancies, including lung cancer, mesothelioma, breast cancer, renal clear cell carcinoma, melanoma, and colon cancer [58]^16^, [59]^22^-[60]^24. Recent advancements in single-cell mRNA sequencing (scRNA-seq) have unveiled C1q acted as a distinctive marker for identifying a novel macrophage subset. scRNA-seq analyses have demonstrated that the C1q^+ macrophages and other immune cells interactions contribute to immunosuppression, poor prognosis in cancer patients, and suboptimal responses to immunotherapies [61]^16^, [62]^25. However, the precise role of C1q as a pro-tumor or anti-tumor factor remains controversial and may vary depending on the specific tumor microenvironment. Recent investigations have demonstrated that the C1q levels in pleural effusion can differentiate tuberculous pleural effusion (TPE) from non-TPE[63]^26^, [64]^27, exhibiting higher diagnostic accuracy than tumor necrosis factor (TNF)-α and IL-6. However, relevant reports regarding the role of C1q in MPE needs further investigations. In contrast to most other complement molecules, which are primarily synthesized by the liver, C1q is mainly secreted by cells derived from myeloid precursor cells, including macrophages [65]^19 and immature dendritic cells (DCs) [66]^19^, [67]^28. Recently, C1q^+ macrophages were found to lead to immunosuppression in a fatty acid metabolic reprogramming-dependent manner in MPE [68]^29. The researchers utilized MHC-II/HLA-DR^+CX3CR1^+ and MHC-II/HLA-DR^-CX3CR1^- to define and isolate C1q^+ and C1q^- macrophages, respectively. Subsequently, these "C1q^+" and "C1q^-" macrophages were employed in their further investigation. However, the direct evidence to verify the internal role of C1q and the corresponding mechanism in MPE using C1q-deficient mice remains to be elucidated. This study discovered that C1q functioned in MPE independently of the classical complement pathway, participating in antigen presentation and macrophage polarization. Data from scRNA-seq and flow cytometry of cells from C1q-deficient mice revealed that C1q^+ macrophages promote MPE by establishing an immunologically inhibitory tumor environment. This process was mediated by the regulation of natural killer (NK) cells through CCR2 signaling and the presence of hippuric acid in pleural effusion. Materials and Methods Study populations and sample processing The research received ethical approval from the committees of Beijing Chao-Yang Hospital, Capital Medical University (2021-ke-9). Informed consent was obtained in writing from participants involved in the study. Pleural effusion patients were recruited at Beijing Chao-Yang Hospital. Patients who had undergone any invasive procedures involving the pleura or received chest trauma within three months prior to hospitalization; and who had undergone any anti-tuberculosis, antitumor, glucocorticoids, or other non-steroidal anti-inflammatory treatment were excluded. The diagnosis of MPE was made if malignant cells were detected in the pleural fluid (PF) and/or through pleural biopsy specimens. Diagnostic thoracentesis was performed to collect PF from each participant. Concurrently, blood samples from the periphery were also gathered, together with PF, they were promptly brought to researchers under refrigeration at 4° C and subjected to centrifugation at 400 g for 10 minutes under the same temperature. Nucleated cells were separated using a Ficoll-Paque gradient technique (Pharmacia, Uppsala, Sweden), the cell pellets were assessed within one hour, while the supernatant was preserved at -80 °C for future applications. Cell lines MC38, the cell line derived from mouse colon adenocarcinoma, was gifted by Dr. G.T. Stathopoulos who worked for University of Patras, Rio Patras, Greece. LLC, the murine Lewis lung carcinoma cell line, BEAS-2B, the human bronchial epithelial derived cell line, and A549, the human non-small cell lung carcinoma (NSCLC) derived cell line, all of which were sourced from the American Type Culture Collection, Manassos, VA. Cells were maintained in Dulbecco's modified Eagle's medium (DMEM) that was enriched with 10% heat-inactivated fetal bovine serum (FBS), 100 U/ml Penicillin, 100 mg/ml Streptomycin, 4 mM L-glutamine, and 4.5 g/L HEPES in 5% carbon dioxide at 37 ° C. Mice The mouse related protocols in the current study have adhered to the guidelines approved by the Institutional Animal Care and Use Committee of the Capital Medical University (AEEI-2021-290, AEEI-2023-237). Wild-type (WT) C57BL/6 mice were purchased from Beijing Vital River Laboratory Animal Technology (Beijing, China). C1qa^fl/flLyz2-Cre (conditional knockout (CKO) of C1qa in macrophages) mice were bred by crossing C1qa^fl/fl and Lyz2-Cre transgenic mice. C1qa^fl/fl, Lyz2-Cre, OT-I mice (CD8^+ T cell transgenic mice expressing a TCR recognizing the K^b associated epitope SIINFEKL which is derived from OVA) and OT-II mice (CD4^+ T cell transgenic mice expressing a TCR recognizing the I-A^b associated epitope ISQAVHAAHAEINEAGR which is derived from OVA) were purchased from Cyagen Biotech Co., Ltd (Suzhou, China). The C1qa globally knockout (KO) mice (C1qa^-/-) were kind gifts from Chao Xiong (Fuwai Hospital, China). Mice were kept in an air-conditioned room with the temperature 23-25° C and the humidity 40-70% under a 12-h dark-light cycle. All mice were used at 6-8 weeks of age. Murine models and sample processing Murine MPE models were established by intrapleural injection of 1.5×10^5 LLC, or MC38 cells in each mouse as previously described [69]^10^, [70]^30^-[71]^32. Mice were euthanized by CO[2] asphyxiation fourteen days post LLC cell injection or, ten days post MC38 cell injection and MPE, pleural tumors and blood samples were obtained for subsequent experiments. The pleural tumors were cut into small pieces with ophthalmic scissors to incubate with digestion solution (130 ug/mL DNase I, 130 U/mL Collagenase Type IV, in Medium/RPMI 1640) for 1 hour at 37℃. The digested samples were screened with an 80-100 μm filter, the passed-through samples and MPE cell pellet were subjected to Ficoll-Paque gradient centrifugation to obtain nucleated immune cells. To verify the effects of hippuric acid and RS504393 on MPE, hippuric acid and RS504393 were freshly dissolved in dimethyl sulfoxide (DMSO), diluted into a solution with DMSO (10%), PEG400 (40%), Tween-80 (5%), and saline (45%) in sequence. Then mice were treated with hippuric acid (0.14 mg/kg/day) [72]^33 and/or RS504393 (6 mg/kg/3 days) [73]^34 (purchased from MedchemExpress) by intraperitoneal injection after the MPE model was established. Flow cytometry The antibodies (Abs) for flow cytometry analysis including anti-human C1q mAb was purchased from Quidel Corporation; anti-human CD3, CD15, CD19, CD45, CD14 mAbs, anti-mouse CD45, CD3, CD8, LAG-3, TIM-3 and PD-1 mAbs were purchased from eBioscience; anti-human HLA-DR, CD86, CD206, anti-mouse CD45, CD3, CD8, NK1.1, F4/80, CD11b, Ly6C, CX3CR1, MHC II, CD86, CD206, and Goat anti-Rat IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 488 were purchased from Invitrogen. For staining of human C1q, cells were incubated with the C1q mAb at 4°C for 30 minutes, then the Alexa Fluor 488 goat anti-rat IgG were added and incubated at 4°C for 30 min. For staining of cell surface receptors including human CD3, CD15, CD19, CD45, CD14, HLA-DR, CD86, CD206, mouse CD45, CD3, CD8, F4/80, CD11b, Ly6C, CX3CR1, MHC II, CD86, LAG-3, TIM-3 and PD-1. The corresponding fluorescent Abs were incubated with the cells at 4°C for 15 minutes. For intracellular staining of transcription factors or cytokines, cells were suspended and fixed with fixation/permeabilization solution (BD Biosciences) for 25 minutes, and then anti-CD206, IFN-γ, IL-17, FOXP3, TNF-α, GZMB, and Perforin mAbs were added and incubated for 30 minutes, and then washed by permeabilization buffer (BD Biosciences). Finally, cells were resuspended in PBS and subjected to flow cytometry (BD Biosciences) for further analysis. Survival analysis Mice exhibiting MPE were closely observed, humanely euthanized, and their cases were documented as part of the Kaplan-Meier survival analysis at the point they showed severe illness and were nearing death. The comparison of overall survival was performed by a pairwise log-rank test (GraphPad Software). Antigen presentation by macrophage in vitro Immune complexes (ICs) consisting of IgG and OVA were formed by mixing OVA that was linked to Alexa Fluor 647 (Invitrogen) with polyclonal rabbit anti-OVA IgG (Sigma) for 30 minutes at a temperature of 37 °C. After that, macrophages isolated from the MPE of C1qa^fl/fl and C1qa^fl/flLyz2-Cre mice were exposed to the OVA ICs for an hour. Following a washing step, these macrophages were then incubated with OT-I or OT-II T cells (pre-labeled with CFSE). The levels of proliferation in CD8^+ T cells and CD4^+ T cells were assessed three days later using flow cytometry. Immunoturbidimetric assay Immunoturbidimetric method was used to determine levels of C3 and C4 by commercial kits (3V biotech, Weifang, China), and values were expressed as mg/dL. Enzyme-Linked Immunosorbent Assay (ELISA) Hippuric acid ELISA Kit (SPS-21964, SAIPEISEN BIOLOGY, Shanghai) was used to determine hippuric acid level in MPE according to the manufacturer's instructions. Luminex assay The levels of IFN-γ, TNF-α, CCL2, CCL3, CCL4, CCL7 and CCL8 in mouse MPE were determined using Luminex assay with Mouse Cytokine & Chemokine Panel 1A (36 plex) (Invitrogen), the concentration of GZMB was determined with Mouse ProcartaPlex Simplex Kit (EPX010-26074-901). Luminex 200 was used to detect and calculate the concentrations. Preparation of conditional medium When the LLC, A549, and BEAS-2B cells were cultured to the logarithmic growth phase, 5×10^6 cells were inoculated into a 75cm^2 culture flask. After twenty-four hours of culture, the medium was changed into serum-free medium. Forty-eight hours later, the cell supernatant was centrifuged and filtered by a 0.22 μm sterile filter, and then stored at -80° C for later cell culture. RNA extraction, reverse transcription, and quantitative RT-PCR TRIzol (Thermo Fisher Scientific) were used to extract the total RNA. The RNA with the amount of 1000 ng/20 μL/test underwent processing for cDNA synthesis utilizing the Reverse Transcription Kit (Takara) utilizing random primers. To quantify the mRNA expression levels, one microliter of the synthesized cDNA was analyzed using SYBR Green I Master (Roche, Basel. Switzerland), and the calculations of the results were performed applying the 2^-ΔΔCT methodology, with GAPDH serving as the control. Each sample underwent testing in triplicate. Murine bone marrow-derived macrophage (BMDM) isolation and polarization Bone marrow (BM) was flushed from femurs and tibia of WT, C1qa^fl/flLyz2-Cre or C1qa^fl/fl mice on C57BL/6 background. BM cells were cultured in DMEM containing 10% heat-inactivated FBS, 100 U/ml Penicillin, 100 mg/ml Streptomycin, 4 mM L-glutamine, 4.5 g/L HEPES and 20 ng/mL macrophage colony-stimulating factor (M-CSF) in 5% CO[2] at 37°C. The medium was replenished every three days to support the maturation of M0 macrophages. On the seventh day, the M1 phenotype was induced by LPS (100 ng/mL, Sigma), and IFN-γ (40 ng/mL, Peprotech), while recombinant murine IL-4 (10 ng/mL, Peprotech) was utilized to promote the M2 phenotype. To investigate the expression of markers related to antigen presentation and the polarization of macrophages, cells were treated with hippuric acid (MedchemExpress) for periods of 3 and 10 days, respectively. Immunofluorescence Tumors located in the pleura were preserved using 4% paraformaldehyde and subsequently incorporated into paraffin. Thin sections measuring 5 μm were then produced from the paraffin blocks. These sections underwent incubation with monoclonal antibodies for immunofluorescence, specifically targeted against CD31 (Servicebio, Wuhan). Imaging was performed utilizing a Leica TCS SP5 microscope. Cell transfection assay The C1qa, C1qb, and C1qc-specific siRNAs (siC1qa, siC1qb, siC1qc, we used the mixture of three with equal molar ratio), negative control and 5-Fam (all from RiboBio) were introduced into murine BMDM using Lipofectamine RNAiMAX (Life Technologies, Carlsbad, CA). Apoptosis assay Murine BMDMs before and after transfection were suspended in DMEM containing 10% heat-inactivated FBS and cultured at 2×10^6 cells/mL. Twenty-four hours later, cells were collected and the apoptotic fractions were detected and analyzed using an FITC Annexin V Apoptosis Detection Kit (BD Biosciences) adhering to the manufacturer's instruction. Phagocytosis assay Mouse macrophages isolated from C1qa^fl/flLyz2-Cre or C1qa^fl/fl mice were suspended in DMEM containing 10% heat-inactivated FBS, cultured at 2×10^6 cells/ml, and collected 24, 48 and 72 hours later. Phagocytosis capacity was detected using a pHrodo™ Green E. coli BioParticles™ Conjugate (Invitrogen) adhering to the manufacturer's instruction. Single-cell RNA sequencing analysis The cells with nuclei extracted from murine MPE and pleural tumors were derived following the methodologies specified for mouse models and sample analysis, subsequently subjected to trypan blue staining (Sigma, Shanghai, China) and assessed under a microscope for cell viability. Cells exhibiting a viability rate of 80% or more were selected for further processing. The resultant single-cell suspensions were adjusted to 1 × 10^5 cells/mL in concentration in PBS and introduced into microfluidic platforms, while single-cell RNA sequencing libraries were generated in accordance with the Singleron GEXSCOPE® protocol utilizing the GEXSCOPE® Single-Cell RNA Library Kit as well as the Singleron Matrix® Automated single-cell processing system (Singleron Biotechnologies). The resulting libraries were adjusted to 4 ng/µL in concentration and then combined for the sequencing process. The pooled libraries underwent sequencing on the Illumina novaseq6000 platform with 150 bp paired-end reads. The resulting reads were aligned to the reference genome GRCh38.93. Outcomes from the Cell Ranger analysis provided UMI counts associated with each gene across all cells for each sample employing all mapped reads. Data analysis objects were constructed using the R package Seurat (v4.0.0). Initially, genes expressed in fewer than 200 cells (about 0.1% of the total) were filtered out. Subsequently, cells were excluded if they demonstrated expression of fewer than 200 genes, more than 6000 genes, or over 10% of mitochondrial genes. The DoubletFinder (v2.0.1) was then applied to identify and remove doublets from subsequent analysis. Finally, the graph-based clusters were annotated with known biological cell types utilizing canonical marker genes (v.3.12). by referring to the results of SingleR (v.3.12) as well. The FindAllMarkers and FindMarkers functions were used to determine the differentially expressed genes between a certain cell subgroup and all other clusters or a specific cluster with the following signatures (i) genes expressed in at least 30% of the cells and (ii) log2(fold change) > 1. Interactions were calculated based on ligand and receptor expression on certain two types of cells. CellChatDB ([74]http://www.cellchat.org/) served as references for ligand-receptor