FGL1 in plasma extracellular vesicles is correlated with clinical stage of lung adenocarcinoma and anti-PD-L1 response

Abstract

Fibrinogen-like protein-1 (FGL1) is confirmed a major ligand of lymphocyte activation gene-3 which could inhibit antigen-mediated T-cell response and evade immune supervision. Although hepatocytes secrete large amounts of FGL1, its high expression also be detected in solid tumors such as lung cancer, leading to a poor efficacy of immune checkpoint inhibitors therapy. Here we reported that FGL1 was overexpressed in lung adenocarcinoma (LUAD) but not in lung squamous cell carcinoma. However, FGL1 in tissue and plasma can only distinguish LUAD patients from healthy donors and cannot correlate with clinical Tumor Node Metastasis (TNM) stage. Using lung cancer cell lines, we confirmed that FGL1 can be detected on extracellular vesicles (EVs) and we established a method using flow cytometry to detect FGL1 on the surface of EVs, which revealed that FGL1 could be secreted via EVs. Both animal model and clinical samples proved that plasma FGL1 in EVs would increase when the tumor was loaded. The level of FGL1 in plasma EVs was correlated with clinical TNM stage and tumor size, and a higher level indicated non-responsiveness to anti-programmed cell death ligand 1 (anti-PD-L1) immunotherapy. Its effect on tumor progression and immune evasion may be achieved by impairing the killing and proliferating capacities of CD8⁺ T cells. Our result demonstrates that FGL1 levels in plasma EVs, but not total plasma FGL1, could be a promising biomarker that plays an important role in predicting anti-PD-L1 immune therapy in LUAD and suggests a new strategy in LUAD immunotherapy.

Fibrinogen-like protein-1 (FGL1) was overexpressed in lung adenocarcinoma and could be secreted via extracellular vesicles (EVs). FGL1 on plasma EVs, not total plasma FGL1, is correlated with clinical TNM(Tumor Node Metastasis)staging and non-responsiveness to anti-programmed cell death ligand 1 therapy, which could be a new potential target for immunotherapy

Graphical Abstract

Graphical Abstract.

Introduction

Lymphocyte activation gene-3 (LAG3) is a type I transmembrane protein that exists mainly on the surface of activated T cells and negatively regulates the proliferation, activation, effector function, and homeostasis of both CD8⁺ and CD4⁺ T cells [1, 2]. Fibrinogen-like protein-1 (FGL1) is recently proved to be a novel ligand of LAG-3 independent from the canonical ligand, major histocompatibility complex II [3]. FGL1 may inhibit antigen-mediated T-cell response both in vitro and in vivo through binding with LAG3, which leads to T-cell depletion and subsequent dysfunction, as well as tumor cell escape from immune surveillance. Silencing the FGL1/LAG3 interaction by gene knockout or antibody blocking can preferentially stimulate the proliferation and activation of T cells in the tumor microenvironment, thus promoting tumor immunity and inhibiting tumor growth, which indicates that the FGL1/LAG3 pathway is an independent tumor immune escape mechanism [3]. The immune checkpoint inhibitors (ICIs) in recent years have been applied to clinical practice and made tremendous progress in targeting programmed death-1/programmed death ligand 1 (PD-1/PD-L1) and cytotoxic T lymphocyte antigen 4/B7 receptor (CTLA-4/B7) pathway [4]. However, the efficacy of immunotherapy is not satisfactory, only 20% of patients benefit from immune checkpoint blockade therapy in clinical studies [5], while the rest of the patients show primary or adaptive drug resistance to varying degrees [6–9]. Therefore, FGL1/LAG3 immune checkpoint pathway could attract more attention in further studies and provide promising applications in therapeutic strategies for malignant tumor treatment.

FGL1, which belongs to the fibrinogen family, is also known as liver fibrinogen-related gene-1 (LFIRE-1)/Hepassocin (HPS) or hepatocyte-derived fibrinogen-related protein-1 (HFREP-1) [10, 11]. As a proliferation and metabolism-related protein, FGL1 is mainly secreted by hepatocytes and involves in cell mitosis and metabolic process [12, 13]. Apart from its relatively high expression in the liver, FGL1 mRNA was shown to overexpress in human solid tumors, such as lung cancer, prostate cancer, melanoma, and colorectal cancer based on meta-analysis of the Oncomine databases [3]. FGL1 is largely secreted from tumor cells, and elevated plasma levels of FGL1 are involved in the resistance to ICIs and worse overall survival in non–small cell lung cancer (NSCLC) patients [3]. On the other hand, plasma FGL1 levels in NSCLC patients were significantly higher than healthy donors, but there was no difference among NSCLC patients with and without metastasis [3]. The reasons are still unclear, but it may be that FGL1 is present in plasma in a different way, or other factors are synergistically involved in the progression of NSCLC.

Extracellular vehicles (EVs) are small lipid membrane vesicles secreted into the extracellular space by almost all cell types. EVs from cancer cells play an important role in altering the tumor microenvironment and promoting tumor progression. Since EVs secreted by tumor cells often contain specific protein or nucleic acid components, they have the potential to become biomarkers [14–16]. Previous study reported that tumor-derived EVs carrying PD-L1 on their surface could suppress the function of CD8⁺ T cells and facilitate tumor growth in melanoma [17]. Clearance of PD-L1 in EVs has also been shown to inhibit tumor growth in prostate cancer, even in models resistant to the antiviral drug PD-L1 antibody [18]. Whether FGL1 could be secreted through EVs and play a role in tumor immune escape worth pondering. Here, we proved that FGL1 was overexpressed in lung adenocarcinoma (LUAD) and could be detected in EVs, and FGL1 level in plasma EVs was correlated with tumor stage of LUAD patients. EVs with overexpressed FGL1 inhibited CD8⁺ T-cell proliferation, effector function, and contributed to the resistance of anti-PD-L1 immunotherapy.

Material and methods

Clinical samples

Tissue microarray containing 141 cases of LUAD tissues was used for immunohistochemistry. Specimens for microarray were obtained from patients who underwent complete surgical resection at Shanghai Zhongshan Hospital, Fudan University (People’s Republic of China) in 2005, and the clinicopathologic characteristics of patients were listed in Supplementary Table S1. Tumor tissues and adjacent normal tissues for western blot (WB) were obtained during surgery immediately after specimens were resected. Peripheral blood specimens were randomly collected from lung LUAD patients receiving treatment at Zhongshan Hospital between 2019 and 2020, and plasma from 62 cases receiving surgery or chemotherapy was used for analysis correlation between FGL1 and clinical characters, Table 1 provides clinicopathologic characteristics of the patients enrolled in this study. Moreover, plasma from these patients and eight healthy donors was used to compare the difference in FGL1 expression. Plasma from 17 cases receiving anti-PD-L1 immunotherapy was used for the analysis of correlation between FGL1 level and immunotherapy resistance, the clinicopathologic characteristics of patients were listed in Supplementary Table S2. All patients provided informed consent for the use of their clinical information before surgery and the study was approved by the institutional review board (B2021-128).

Table 1.

Clinical characters of patients and their plasma FGL1 level and plasma EVs FGL1 level analysis

	No. of patients	Average of plasma FGL1 (ng/ml)	P-value	Average of FGL1 in plasma EVs (ng/µg)	P-value
Age
≤60	23	409.8	0.831	18.2	0.522
>60	39	398.4		20.9
Gender
Male	32	402.4	0.993	23.6	0.054
Female	30	402.8		15.9
Tumor size
≤3 cm	41	388.6	0.451	16.7	0.024
>3 cm	21	429.8		26.1
T stage
T1–2	51	406.8	0.728	17.8	0.021
T3–4	11	383.2		29.8
N stage
N0	40	399.2	0.860	17.0	0.047
N1–3	22	408.7		25.2
M stage
M0	42	379.5	0.195	16.1	0.005
M1	20	451.0		27.9
Group stage
I/II	37	388.9	0.521	15.4	0.006
III–IV	25	422.8		26.5
Grade
I/II	43	392.9	0.573	21.4	0.271
III	19	424.5		16.6
Smoking
No	44	381.9	0.210	18.1	0.171
Yes	18	453.1		24.2

The data of patients’ stage based on the 8th TNM staging system of lung cancer were used for T-stage, N-stage, and group-stage analysis.

Cell culture

Human bronchial epithelial cell lines HBE and BEAS-2B and human lung cancer cell lines H1299, A549, 95D, and H460 were purchased from the Chinese Academy of Sciences (Shanghai, China). These cell lines were correctly identified by STR. The cell lines in our study have been tested and found free of Mycoplasma. Cells were cultured in Dulbecco’s modified eagles medium or Roswell Park Memorial Institute (RPMI)-1640 medium with 10% fetal bovine serum (FBS) in 5% CO₂ at 37°C. Inhibitor of endoplasmic reticulum (Brefeldin A, BFA), Golgi complex (Monensin), and EVs (GW4869) was purchased from MedChemExpress (MCE, NJ, USA). The inhibitor was diluted in dimethyl sulfoxide (DMSO; Sigma, MO, USA), and the equal volume of DMSO was applied for control groups.

EVs extraction

For EVs purification from cell culture supernatants, cells were washed with phosphate buffered saline (PBS) at ~50–70% confluency and cultured for 24–48 h with 15% EVs-depleted FBS. EVs in FBS were depleted by overnight centrifugation at 100,000g for 14–16 h. Supernatants were collected and EVs were purified by a standard differential centrifugation protocol. Briefly, supernatants were centrifuged at 300g for 10 min, followed 2,000g for 10 min to remove cellular debris. Microvesicles were depleted by centrifuge at 10,000g for 30 min, then centrifuge at 100,000g for 70 min at 4°C. The EVs pellet was washed in PBS and centrifuged at 100,000g for an additional 70 min before collection at 4°C. For purification of plasma EVs, venous blood from patients or healthy donors was centrifuged at 2,000g for 10 min to obtain cell-free plasma. Then 1 ml of the obtained plasma was diluted with PBS and centrifuged by the above differential centrifugation protocol to obtain EVs.

Characterization of the purified EVs

For verification of purified EVs using electron microscopy, purified EVs suspended in PBS were dropped on formvar carbon-coated nickel grids. After staining with 2% uranyl acetate, grids were air-dried and visualized using a JEM-1011 transmission electron microscope. The size and concentration of EVs, purified from cell culture supernatants or patients’ plasma, were determined using nanoparticle tracking analysis (NTA) at VivaCellBiosceinces with ZetaView PMX 110 (Particle Metrix, Meerbusch, Germany).

EVs concentration assay

EVs were obtained from culture supernatant and plasma of all patients and nude mouse with tumor subcutaneous injected. The protein concentration of the EVs fraction was determined using a Pierce BCA protein assay kit (23225, Thermofisher), according to the manufacturer’s instructions.

Western blot

Samples were lysed in Radio Immunoprecipitation Assay (RIPA) lysis buffer system according to manufacturer’s instructions. Supernatant of cells cultured for 48 h was incubated with anti-FGL1 antibody and protein A/G (sc-2003, Santa Cruz) overnight, and then the immunoprecipitated protein A/G was centrifuged and resolved by SDS-polyacrylamide gel electrophoresis. Protein in cell lysates or immunoprecipitated protein A/G were separated in 10% SDS polyacrylamide gels and transferred to nitrocellulose membranes. Membranes were incubated with the following primary antibodies against FGL1 (16000-1-AP, Proteintech) and β-actin (20272, Abcam). EVs extracted from cell culture supernatant and plasma were treated with RIPA lysis buffer for 30 min and used for WB test following the same process. CD9 (223052, Abcam) and CD63 (216130, Abcam) were tested as EVs markers. All of the above primary antibodies were diluted using primary antibody dilution buffer(Beyotime, P0023A).

Enzyme linked immunosorbent assay (ELISA)

For detection of FGL1 protein in patients’ plasma and EVs, pre-coated ELISA plates (96-well) were purchased from Raybiotech (ELH-HFREP1-1, USA). Five microliters of plasma was diluted to 100 µl, 20 µg EVs were treated with the same volume of lysis buffer for 30 min at room temperature (RT) and added to 100 µl with diluent buffer. Then FGL1 levels in diluted plasma or EVs were measured using commercial ELISA kit according to the manufacturer’s protocol. The absorbance value was determined at 450 nm using Bio-Tek ELX800 microplate reader (Winooski, USA) within 30 min. Concentration standard curve was performed in each test to calculate the precise concentration of FGL1.

Immunohistochemistry and immunofluorescence

Paraffin-embedded tissue microarray was used to detect FGL1 protein level in tumor and normal lung tissues with anti-FGL1 antibody (16000-1-AP, Proteintech) for immunohistochemical staining. The protein level of FGL1 was scored according to standard protocols. The results were evaluated and scored by two pathologists individually, and in the case of disagreement in scoring, the results were reviewed and agreed upon jointly. Depending on the staining area, the score of protein expression was conducted: 0 (<5%), 1 (5–25%), 2 (26–50%), 3 (51–75%), and 4 (>75%). The staining intensity was scored as follows: 0 (no staining), 1 (weak staining), 2 (moderate staining), and 3 (strong staining). Composite expression score (CES) is calculated from intensity and area measurements for immunostaining (CES = 4 × (intensity score − 1) + area score), yielding a series of results ranging from 0 to 12. High expression was considered as a total score >8 and low expression with a total score ≤8.

H1299 cells and FGL1 overexpressed H1299 cells were used for immunofluorescence staining. The cells were incubated with rabbit anti-FGL1 antibody (16000-1-AP, Proteintech) and Alexa Fluor 647 donkey anti-rabbit IgG secondary antibody (ab150075, Abcam). In addition, EVs were stained with mouse CD63 antibody (ab1318, Abcam) and Alexa Fluor 488 donkey anti-mouse secondary antibody (A32766, Thermofisher). Stained cells were visualized under a Zeiss LSM710 confocal microscope (Carl Zeiss Meditec, Jena, Germany).

Flow cytometry analysis of EVs

The method for flow cytometry of EVs was described by Theodoraki et al. [19]. Briefly, 10 μg EVs purified from cell culture supernatants or plasma were first co-incubated with biotin-labeled anti-CD63 mAb (353018, Biolegend) adjusted to the concentration of 1 μg for 2 h at RT. Next, a 10 μl aliquot of Streptavidin magnetic beads (47503ES03, Yeasen) was added and the tubes were again incubated for 2 h at RT. Samples were washed 3× with PBS. The beads/anti-CD63 Ab/EVs complexes were then co-incubated with anti-human FGL1 (ab275091, Abcam) or with the isotype control Ab (ab172730, Abcam) for 1 h at RT and washed 3× with PBS. Alexa Fluor 647 donkey anti-rabbit IgG secondary antibody (ab150075, Abcam) was added and incubated for 1 h at RT. At last, the complexes were washed 3× and resuspended in 300 μl PBS for antigen detection by flow cytometry.

In vivo nude mice study

For establishing human lung cancer model in nude mice, H1299 cells or FGL1 overexpressed H1299 cells (5 × 10⁶ cells in 100 µl medium) were injected into flanks of 6-week-old female athymic nude mice. Mice were euthanized 14 days after cell inoculation, and the longest dimension of the tumors did not reach 2.0 cm. Immediately following euthanasia, blood samples were harvested by cardiac puncture, and EVs were purified and detected by WB and flow cytometry. PBS was injected into healthy nude mice with matched sex, age, and weight, and then EVs were purified as control.

Functional assays

Peripheral blood mononuclear cell (PBMC) was obtained from whole blood of healthy donors by centrifuge with human lymphocyte separation medium (KLSH1408, Dakewe Biotech). Briefly, 5 ml human lymphocyte separation medium was placed at the bottom of a 15-ml tube and 10 ml whole blood was added above the separation medium gently. After centrifuge at 800g for 20 min to stratify, PBMC was located in the white layer between the brown plasma and transparent separation medium. PBMC was absorbed and washed 3× with PBS, then resuspended in the RPMI-1640 medium with 10% fetal bovine serum (FBS) in 5% CO₂ at 37°C.

Human PBMC cells (5 × 10⁵ cells/well in a 48-well plate) were incubated with EVs derived from H1299 or FGL1 overexpressed H1299 for 48 h in the presence of anti-CD3/CD28 antibodies (317326/ 302934, Biolegend). The treated cells were collected, stained, and analyzed by flow cytometry. To detect the proliferation and killing function of CD8⁺ T cells, CD8 (344711, Biolegend), Ki-67 (151211, Biolegend), interferon γ (IFNγ) (502511, Biolegend), and CD107a (328607, Biolegend) were detected by flow cytometry.

Statistical analysis

Data were analyzed using GraphPad Prism software package (version 8; GraphPad Software Inc, La Jolla, CA) and were presented as the mean ± standard deviations. Differences of continuous variables between groups were analyzed using the Student’s unpaired t-test or One-way ANOVA analysis, categorical variables were compared with Chi-Square test. A P-value of <0.05 was considered statistically significant.

Results

Overexpressed FGL1 in LUAD was not correlated with clinicopathologic data

FGL1 was reported to overexpress in NSCLC and correlated with poor prognosis; however, it was not relevant to tumor metastasis [3]. To validate this result in larger scale sample, we compared mRNA data of FGL1 in The Cancer Genome Atlas (TCGA through the open-access web server “GEPIA” by Tang et al. [20]. FGL1 was significantly overexpressed among adenocarcinoma compared to paired adjacent normal lung, but not in lung squamous cell carcinoma (LUSC; Fig. 1A), so we focused on LUAD in the following study. The overexpression of FGL1 in LUAD was confirmed by WB in13 pairs resected tumor and adjacent normal tissues (Fig. 1B), and gray-scale statistics showed a significant difference (Fig. 1C). As a secretory protein, plasma FGL1 protein level could be detected by ELISA, so we collected plasma from 62 LUAD patients (Table 1). As expected, plasma FGL1 level of LUAD patients was higher than healthy donors (Fig. 1D). Moreover, FGL1 in the supernatant and lysis of normal bronchial epithelial cell line (BEAS-2B and HBE) and large cell lung cancer cell line (H460) was lower than that of LUAD cell line (H1299, A549, and 95D), which was consistent with the level of FGL1 in clinical samples (Fig. 1E). Neither the TCGA mRNA data (Supplementary Fig. S1A) nor our immunohistochemical data containing 141 LUAD tissue microarrays (Supplementary Fig. S2B; Supplementary Table S1) showed any correlation between FGL1 and the TNM staging system. Thus, FGL1 levels in tumors do not seem to vary with tumor progression. Meanwhile, total FGL1 in plasma also showed no correlation with tumor TNM stage (Table 1; Supplementary Fig. S1C).

Figure 1.

FGL1 is overexpressed in LUAD but not correlated with the clinical stage of tumor. (A) Comparison of FGL1 mRNA data in tumor and adjacent normal lung among LUAD (left) and LUSC (right) from TCGA. (B) WB analyzed the expression of FGL1 in 13 pairs of LUAD tumors and adjacent normal lung tissues. (C) FGL1 expression levels were quantified by densitometric analysis (Totallab 2.01), statistically analyzed from the above 13 pairs of tissues. *P < 0.05. (D) ELISA test of FGL1 expression in plasma from 8 healthy donors and 62 LUAD patients using two-tailed Student’s t-test. ***P < 0.001. (E) Immunoprecipitation (IP) assay for FGL1 in supernatants and lysates of normal bronchial epithelial cell lines (HBE and BEAS-2B), lung glandular epithelial cancer cell lines (H1299, A549, and 95D) and large cell lung cancer cell line (H460). Abbreviations: N, normal lung tissues; T, tumor tissues; HD, healthy donor; P, patient.

Detection of FGL1 on EVs

In contrast to exosome-bound PD-L1, soluble PD-L1 (sPD-L1) level in the plasma of neck squamous cell carcinomapatients was not proved to be associated with any clinicopathologic findings [19]. Cellular production and release of sPD-L1 from cells are likely to be independent of the complex packaging machinery responsible for the exosome biogenesis. To investigate the secretory mechanism of FGL1, H1299 cell was transfected with plasmid overexpressed FGL1 (Fig. 2A), and then we used BFA to block the protein endoplasmic reticulum (ER) trafficking, monensin to block the secretory transport pathway of Golgi, and GW4869 to inhibit the release of EVs. As shown in Fig. 2B, BFA completely blocked FGL1 secretion in the supernatant, monensin inhibited most secretion of FGL1, and GW4869 slightly reduced FGL1 level in the supernatant. It demonstrated that FGL1 protein was secreted mainly through the normal ER-Golgi pathway but may also be secreted through EVs in a relatively small amount. The schematic diagram is shown in Fig. 2C.

Figure 2.

FGL1 could be secreted through EVs and expressed on the surface of EVs. (A) WB analysis of FGL1 expression in H1299 cells transfected with vector or FGL1 overexpressed plasmid. (B) H1299 cells were transfected with FGL1 overexpressed plasmid and treated with Monesin, BFA, and GW4869, respectively, for 24 h, control group was treated with DMSO. IP and WB were performed to detect FGL1 expression in the supernatant. (C) Schematic diagram of FGL1 secretion through ER-Golgi and EVs pathway. (D) Characterization of circulating EVs purified from supernatants of H1299-FGL1 using NTA at VivaCellBiosceinces with ZetaView PMX 110 (Particle Metrix, Meerbusch, Germany). (E) TEM image of H1299-FGL1 derived EVs. The arrows indicate the EVs. (F) WB analysis of EVs derived from normal bronchial epithelial cell lines (HBE and BEAS-2B), lung glandular epithelial cancer cell lines (H1299, A549, and 95D), and large cell lung cancer cell line (H460), CD9 and CD63 were used as markers of EVs. (G) WB analysis of EVs derived from H1299 cells and H1299-FGL1 cells treated with different concentrations of GW4869. (H) Immunofluorescence of intracellular FGL1 (red) and EVs marker CD63 (green) in H1299 cells and H1299-FGL1 cells. Nuclei were stained with DAPI (blue). (I) Schematic diagram of flow cytometry analysis of EVs based on biotin-conjugated anti-human CD63 antibody and streptavidin magnetic beads. (J) FGL1 on the surface of EVs derived from H1299 and H1299-FGL1 cells was detected by flow cytometry, with RFI statistical analysis on the right. **P < 0.01. Abbreviation: RFI, relative fluorescence intensity.

Given that FGL1 might be secreted through EVs, EVs from the supernatant of FGL1 overexpressed H1299 cells (H1299-FGL1) were purified by differential centrifugation [21]. The quality was verified with nanoparticle tracking analysis and transmission electron microscopy (TEM; Fig. 2D and E). FGL1 in EVs was higher in human lung cancer cell lines compared to normal bronchial epithelial cell lines by WB (Fig. 2F). In FGL1-overexpressed H1299 cells, GW4869 inhibited EVs secretion and therefore reduced FGL1 expression in EVs (Fig. 2G). FGL1 got a relatively low expression in H1299 cells compared to FGL1 overexpressed H1299 cells in the same parameters, and immunofluorescence identified the co-localization of FGL1 and CD63 in cytoplasm (Fig. 2H). These results indicated that FGL1 could be loaded into EVs and secreted into the extracellular space. We also performed flowcytometry analysis of EVs based on biotin-conjugated anti-human CD63 antibody and streptavidin magnetic beads, as shown in Fig. 2I, to verify whether FGL1 could be detected on the surface of EVs like PD-L1 [17, 19]. FGL1 was indeed significantly upregulated on the surface of EVs originated from H1299-FGL1 cells (Fig. 2J).

The level of FGL1 on circulating EVs distinguishes patients from healthy donors

Whether tumor-secreted FGL1 could be detected in plasma EVs is important to determine whether FGL1 could act as a tumor biomarker and participant in immunosuppression. To verify whether tumor secretion of FGL1 via EVs can be detected in plasma, we established an animal model (Fig. 3A) in which H1299 and H1299-FGL1 cells were injected subcutaneously into immunodeficient nude mice and EVs were extracted from peripheral blood for flow cytometry. To confirm that the elevated FGL1 on plasma EVs was from the secretion of H1299 cells instead of the mice cells, we used a human-specific FGL1 antibody (ab275091, Abcam). Flow cytometry showed an increase of FGL1 on plasma EVs in the H1299-FGL1 group compared with the control group (Fig. 3B). In addition, FGL1 levels in EVs were significantly higher in LUAD patients than in healthy donors, as confirmed by WB (Fig. 3C and D) and flow cytometry (Fig. 3E and F).

Figure 3.

Tumor-secreted EVs containing FGL1 could be detected in plasma and identify LUAD patients and healthy donors. (A) Schematic diagram of the mouse model to verify the plasma EVs derived from subcutaneous tumors. (B) Flow cytometry analysis of EVs in plasma derived from nude mouse carrying H1299 and H1299-FGL1 tumors, with MFI and RFI statistical analysis on the right. ***P < 0.001. (C, D) WB analysis (C) of EVs derived from healthy donors (HD) and patients (P), with gray-scale statistical analysis on the right (D). **P < 0.01. Representative flow cytometry analysis (E) of EVs derived from healthy donors (HD) and patients (P), with statistical analysis on the right (F). **P < 0.01. Abbreviations: MFI, mean fluorescence intensity; RFI, relative fluorescence intensity.

The level of FGL1 in plasma EVs was correlated with LUAD clinical stage

To verify whether FGL1 in plasma EVs was correlated with cancer progression, plasma EVs were detected by WB in LUAD patients with different clinical status according to the 8th edition of the Lung Cancer Staging Grouping System [22], and advanced-stage patients (stages III–IV) tended to get higher FGL1 protein expression (Fig. 4A). In addition, the level of FGL1in plasma EVs positively correlated with tumor size (Fig. 4B). Flow cytometry analysis of EVs also showed higher FGL1 protein levels among stages III–IV patients compared to earlier stage patients (Fig. 4C), and patients with larger tumor size (2–4 cm, ≥4 cm) tended to have higher FGL1 levels than patients with tumors smaller than 2 cm (Fig. 4D). FGL1 protein levels in EVs from 62 LUAD patients were detected by ELISA and analyzed based on the TNM stage system. The clinical characteristics and FGL1 levels in EVs analysis of the patients are shown in Table 1. FGL1 levels in plasma EVs were not only correlated with primary tumor (T stage), but also with lymph node metastasis (N stage), distant metastasis (M stage), and grouping stage (I–IV; fig. 4E).

Figure 4.

FGL1 in plasma EVs is correlated with clinical stage and tumor size in LUAD. (A) WB analysis of plasma EVs in patients with different stages. (B) WB analysis of plasma EVs in patients with different tumor sizes. (C) Representative flow cytometry analysis of plasma EVs in patients with different stages, and statistical analysis on the right. ***P < 0.001. (D) Representative flow cytometry analysis of plasma EVs in patients with different tumor sizes, and statistical analysis on the right. *P < 0.05, ***P < 0.001. (E) FGL1 in plasma EVs by ELISA assay was correlated with the clinical TNM stage and group stage. *P < 0.05, **P < 0.01. Abbreviation: RFI, relative fluorescence intensity.

High FGL1 levels in plasma EVs are associated with poor outcomes to anti-PD-L1 therapy

High plasma FGL1 has been proved to be associated with poor outcomes of anti-PD-L1 therapy [3], as well as resistance to gefitinib inhibiting apoptosis in NSCLC [23]. Exosomal PD-L1 was proved to contribute to immunosuppression and could be a predictor for anti-PD-L1 therapy [17, 24, 25]. Our results of 17 cases of plasma FGL1 detected by ELISA also showed a trend with higher FGL1 levels in the progressive disease (PD)/stable disease (SD) group than the partial response (PR) group, but no statistical difference (Fig. 5A), which probably due to our small sample size. The clinical characteristics of the 17 patients are listed in Supplementary Table S2. The correlation between immunotherapy and levels of FGL1in EVs has never been studied, so we extracted plasma EVs from 17 patients to perform WB and flow cytometry assay. FGL1 in EVs inclined to be higher in the PD/SD group by WB (Fig. 5B), and showed a significant difference between groups (P < 0.01) by flow cytometry (Fig. 5C). The result indicated that total circulating FGL1 was less effective than FGL1 of EVs in distinguishing responders from non-responders. To determine the effect of FGL1 in EVs on CD8⁺ T cells, PBMC were isolated from human blood, and treated with EVs from H1299 or H1299-FGL1 for 48 h, then CD8⁺ T cells were labeled and detected with IFNγ, Ki-67, and CD107a antibodies. The results showed that CD8⁺ T cells in the H1299-FGL1 group have a decrease in cell killing and proliferation functions, as demonstrated by the reduced expression of IFNγ, Ki-67, and CD107a (Fig. 5D). Moreover, the effect of function was dose-dependent only in H1299-FGL1 EVs, which shown that only FGL1-containing EVs were key factors in the regulation of T-cell function. (Fig.5E; Supplementary Fig. S2).

Figure 5.

High level of FGL1 in plasma in EVs predicts a worse anti-PD-L1 therapeutic effect by diminishing CD8⁺ T-cell function. (A) No significant difference in plasma FGL1 between 17 patients in the PR and PD/ SD groups receiving anti-PDL1 immunotherapy by ELISA assay. (B) FGL1 in EVs detected by WB between the PR and PD/SD groups. (C) Flow cytometry and statistical analysis of FGL1 levels in plasma EVs between the PR and PD/SD groups. *P < 0.05. (D) Detection of CD8⁺ T-cell function by flow cytometry after human PBMC cells were treated with EVs (10 μg/ml) derived from H1299 or H1299-FGL1. (E) Statistical analysis of CD8⁺ T-cell function after human PBMC cells were treated with different concentrations of EVs derived from H1299 or H1299-FGL1 cell lines. Abbreviations: PR, partial response; PD, progressive disease; SD, stable disease; RFI, relative fluorescence intensity. *P < 0.05, **P < 0.01.

Discussion

Normally, FGL1 is secreted mainly by hepatocytes in the liver, as a product of hepatocyte regeneration and participates in hepatocyte mitosis and liver energy utilization (including lipid metabolism and blood glucose regulation) [13, 26, 27]. FGL1 is the main ligand of LAG3 found by Chen et al. [3]. Chen found that FGL1 can form a new immune escape pathway independently of PD-1/PD-L1, thus attracting the attention of researchers. The LAG3/FGL1 pathway plays a strong part in immune evasion during tumor progression, hence anti-FGL1 may help overcome cancer immunotherapeutic resistance as a promising novel checkpoint. FGL1 is thought to have a bright future in clinical applications, especially in immunotherapy of NSCLC [28, 29], which is attributed to its overexpression in NSCLC cells and it is closely associated with immune regulation, tumor neovascularization [30], EMT progression [31], resistance, and metastasis [32, 33]. So we explored whether FGL1 could be a potential biomarker for LUAD.

It is well known that some soluble secreted proteins, such as maspin, can be secreted via EVs. At the same time, they can also be secreted as free proteins in the vesicle-depleted conditioned media (VDCM) fraction [34]. Interleukin 1β (IL-1β) can be secreted via endolysosomal exocytosis and exosomes and many other non-classical pathways [35–37]. The secretion of FGL1 in EVs and the secretion of soluble FGL1 may proceed concomitantly, to different extents, through independent pathways. It remains unclear how cells coordinate the secretion of free FGL1 protein and FGL1-containing EVs.

EVs are one of the most important tools for intercellular communication. For example, PD-1 on CD8⁺ T cells interacting with PD-L1 on tumor-derived EVs can impair the function and induce the apoptosis of CD8⁺ T cells to promote tumor immune escape [25]. We confirmed that FGL1 could be detected on the surface of EVs, where it can perform its immune-repression function by binding its receptor, LAG3, on the T cells. For its immunosuppressive function, we found the levels of FGL1 in plasma EVs appeared to be a more sensitive and better biomarker to predict the unresponsiveness to anti-PD-L1 therapy than total plasma FGL1 levels in this study. It is well known that the PD-1/PD-L1 inhibitory pathway produces exhausted T cells. Anti-PD-1/PD-L1 antibodies can block this immune brake and release the antitumor activity of T cells. However, due to the existence of another inhibitory receptor, LAG3 on the T-cell surface, a newly characterized immune checkpoint is generated when LAG3 interacts with its ligand FGL1. Because of the high affinity between FGL1 and LAG3, FGL1/LAG3, and PD-1/PD-L1 can regulate T cells independently and blocking both checkpoints can produce synergistic antitumor effects. In our study, EVs carrying higher amounts of FGL1 significantly reduced the killing function and proliferation of CD8⁺ T cells, which may explain the poor prognosis and non-response to anti-PD-L1 therapy in patients with higher levels of FGL1 in EVs.

Conclusion

Our study proved that FGL1 was overexpressed in LUAD, but the FGL1 levels in tumors do not vary with tumor progression. Meanwhile, total FGL1 in plasma also showed no correlation with tumor TNM stage. FGL1 could be secreted by tumor cells via EVs, which can be detected in plasma. Furthermore, FGL1 levels in EVs can distinguish LUAD patients from healthy people, and the level is correlated with clinical TNM staging and non-responsiveness to anti-PD-L1 therapy. By impairing the killing and proliferating capacities of CD8⁺ T cells, high patients with higher FGL1 level in plasma EVs always have a poor prognosis and non-response to anti-PD-L1 therapy. Thus the level of FGL1 in plasma EVs may serve as a predictor of clinical outcome of anti-PD-L1 therapy and a new potential target for immunotherapy.

Glossary

Abbreviations

EVs

extracellular vesicles

FGL1

fibrinogen-like protein-1

ICIs

immune checkpoint inhibitors

IFNγ

interferon γ

LAG3

lymphocyte activation gene-3

LUAD

lung adenocarcinoma

LUSC

lung squamous cell carcinoma

MFI

mean fluorescence intensity

NSCLC

non–small cell lung cancer

PBMC

peripheral blood mononuclear cell

PD

progressive disease

PD-1

programmed cell death protein 1

PD-L1

programmed cell death ligand 1

PR

partial response

RFI

relative fluorescence intensity

SD

stable disease

Ethical approval

All animals received standard care, and animal procedures were approved by the Experimental Animal Ethics Committee, School of Basic Medical Sciences, Fudan University (Shanghai, China; approval number: 20190221-045). All patients provided informed consent for the use of their clinical information before surgery and the study was approved by The Institutional Review Committee of Zhongshan Hospital, Fudan University (Shanghai, China; B2021-128).

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Funding

This study was supported by the National Natural Science Foundation of China (81972617, 81772948, 82373371) and grant from Science and Technology Commission of Shanghai Municipal (20ZR1410800).

Data availability

The tissue-wise expression of FGL1 in LUAD and LUSC was analyzed by GEPIA (http://gepia.cancer-pku.cn/), which contains TCGA data (https://portal.gdc.cancer.gov/). The FGL1 mRNA data of LUAD patients analyzed in this study were downloaded from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6066282/bin/NIHMS978596-supplement-1.xlsx, which were integrated and obtained from the data portal of Genomic Data Commons (GDC, https://gdc-portal.nci.nih.gov/legacy-archive/). The materials can be applied to the corresponding author for use.

Author contributions

C.Y.Z., J.G., and D.G. designed the research; K.P.Z., Y.C.Z., and H.Y.W. performed the research; K.P.Z. and Y.C.Z. analyzed the data; K.P.Z., Y.C.Z., and C.Y.Z. wrote the paper. The authors read and approved the final manuscript.