Differential Expression of Immune-Regulatory Protein C5AR1, NLRP3 and CLEC4A on Peripheral Blood Mononuclear Cells in Early-Stage Non-Small Cell Lung Cancer Patients

Background: Previous studies from our group reported paracrine signal-induced peripheral mononuclear cell (PBMC) gene upregulation in various cancer types through epigenetic regulation. We speculated protein expression on circulating T- lymphocytes which might represent T-lymphocyte tracking before inltrating into tumor microenvironment. The possibility to use protein expression on circulating T-lymphocytes as a biomarker to discriminate early stage lung cancer has been explored. Methods: PBMC gene expression was explored, using 4 independent gene expression microarray datasets (GSE12771, GSE13255, GSE20189 and GSE3934). We selected 3 candidate proteins, C5AR1, NLRP3 and CLEC4A, based on their signicant protein expression in tumor-inltrating lymphocytes but not in normal lymphoid tissue. A validation study using automated ow cytometry was conducted in 121 participants including 44 treatment-naïve early stage non-small cell lung cancer patients (NSCLC), 19 non-malignant pulmonary diseases and 62 healthy individuals. Results: The ratio of C5AR1, NLRP3 and CLEC4A specic antibody staining to CD3 positive was signicantly higher in early stage NSCLC compared to healthy control. Median ratio of C5AR1, NLRP3 and CLEC4A expression in early stage NSCLC were 0.65 [range 0.27-0.96; 95% CI: 0.55-0.70], 0.83 [range 0.27-0.99; 95% CI: 0.73-0.86] and 0.75 [range 0.21-0.98; 95% CI: 0.65-0.81], respectively. While, median ratio of C5AR1, NLRP3 and CLEC4A expression in healthy control were 0.21 [range 0.05-0.81; 95% CI: 0.23-0.42, p-value <0.001], 0.32 [range 0.04-0.94; 95% CI: 0.28-0.46, p-value <0.001] and 0.22


Background
Lung cancer is the most common cancer and the leading cause of cancer-related death worldwide with around 2.1 million patients and an estimated 1.8 million deaths in 2018 (1). The majority of lung cancer patients are diagnosed with advanced stage disease in which curative treatment is not suitable. Despite the improvement in therapy, the overall 5-year survival rate for all stages is approximately 18% (2). Thus, there is a major effort for discovery of screening tools for early-stage disease. The National Lung Screening Trial (NLST) using low-dose helical computed tomography (LDCT) demonstrated a 20% reduction in lung cancer mortality among high-risk individuals but included a high rate of false-positive results which often leads to subsequent invasive procedures (3).
Expression of immune-related molecules on peripheral blood mononuclear cell (PBMC) is considered to play a role in antitumor activity. The changes of gene expression pro ling in PBMC seem to represent the host's response to the cancer cells via paracrine signal (4). Several studies have shown that the characterization of gene expression in PBMC may be useful as a diagnostic or for early detection (5)(6)(7). However, many technical issues limit gene expression on PBMC in investigational research. Therefore, protein expression usually correlates to mRNA expression (8). Protein expression is less complex and has a more physiologically relevant environment.
In this study, we postulate that the PBMC from early-stage non-small cell lung cancer (NSCLC) patients would have a different immune-related protein expression compared to healthy control. Discriminative expression of non-malignant pulmonary disease could enhance the potential possibility of these bloodbased biomarkers. Therefore, speci c protein expression on PBMC from three separate groups of earlystage NSCLC patients, non-malignant pulmonary disease and healthy control were conducted.

Candidate protein discovery
To discover the potentially different protein expression in PBMC of early-stage NSCLC patients vs. healthy control, 4 gene expression microarray datasets provided by the Gene Expression Omnibus database (available at https://www.ncbi.nlm.nih.gov/geo/) were retrieved, including GSE12771(9), GSE13255 (5), GSE20189 (6) and GSE39345 (10). Demographic characteristic of patients from 4 gene expression datasets is shown in Table S1. Candidate up-and down-regulated gene expressions from these four independent microarray experiments were identi ed by using CU-DREAM (Connection Up-and Down-Regulation Expression Analysis of Microarrays) (11). 1,885 signi cant gene expressions with p-values and odd ratios > 1 which indicated strong association between the two independents studies were retrieved (Table S2). Overlapping candidate gene expressions from each gene expression microarray dataset is illustrated ( Figure S1). Overlapping signi cant up-regulation genes of at least 3 datasets were retrieved for biological functions using the PANTHER (Protein ANalysis THrough Evolutionary Relationships) classi cation system (available at http://www.pantherdb.org) (12) ( Figure S2). Seventyve genes with immune system processes were identi ed and then mapped with pathology-based protein expression pro ling in the Human Protein Atlas (available at https://www.proteinatlas.org) (13). Three signi cant up-regulated genes which had the presence of antibody-speci c, immunohistochemistry-based protein expression on tumor-in ltrating lymphocytes (TILs) in tumor specimen but not in normal lymphoid tissue were selected, including CLEC4A, C5AR1, NLRP3 ( Figure S3). Details of selected gene ontology in homo sapiens based on the PANTHER are summarized in Table S3.

Study Population
Forty-four patients with early-stage non-small cell lung cancer (NSCLC), 20 patients with non-malignant pulmonary disease and 57 healthy volunteers who received treatment at The King Chulalongkorn Memorial Hospital, Bangkok, Thailand were recruited in a prospective manner. Inclusion criterion were ≥18 year-old with either resectable NSCLC staging system based on pathologic result of the seventh edition of the American Joint Committee on Cancer (AJCC) Cancer Staging System (14) or non-malignant pulmonary disease were enrolled. Healthy volunteers must have had no history of malignant disease with normal chest radiography within 6 months before or after enrollment. The exclusion criterion were prior systemic therapies, including cytotoxic chemotherapy, small molecule targeted therapy, immunotherapy, history of another malignancy within 5 years before enrollment and receipt of corticosteroids or any immunosuppressive agents. All study participants gave written informed consent. Demographic characteristics were obtained from individual patients. This study was approved by the Ethics Committee of the Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand (IRB 211/61).

Blood Collection and Preparation
Whole blood samples were collected in EDTA containers from all participants. Peripheral mononuclear cell (PBMC) were isolated using Ficoll-Paque™ PLUS (GE Healthcare Life Sciences, Canada) as previously described (15). Brie y, blood samples were centrifuged at 1600 × g at 16 °C for 10 minutes. After removal plasma, the specimen was diluted (#4082, Cell Signaling Technology) at the concentration of 1 µg/ml was dropped on the slides. Slides were incubated at 4 °C for 15 minutes in a dark place and washed slides three times with PBS for 10 minutes. Slides were tapped off on tissue paper to eliminate the excess water, then one drop of PRoLong® Gold Antifade Reagent (#9071, Cell Signaling Technology) was added and covered with glass coverslips, all edges of slides were coated with clear ngernail polisher to prevent cells from drying.
Slides were examined under a uorescent microscope ( Figure S4).

Flow cytometry analysis
To shorten the time-consuming process, we adapted the protocol of protein expression assessment to automated ow cytometry. A monoclonal antibody to the CD3 antigen was used to identify target lymphocyte populations. PBMCs were seeded in 96-well V-bottom plate at 2 × 10 6 cells per well. The plate was centrifuged at 500 g for 5 minutes at 4 °C and was washed with 50 µl of cold PBS. Three primary antibodies were used in this study including C5AR1 (Cat. No. HPA014520, Sigma), CLEC4A (Cat. No.

Statistical Analysis
Statistical parameters considered for characterization protein expression on the PBMC is as follows; 1) ratio of speci c-antibody positive to CD3 positive, 2) average uorescence intensity of speci c-antibody staining with CD3 positive 3) average adjusted uorescence intensity calculated by average uorescence intensity × ratio of speci c-antibody positive to CD3 positive. Mann-Whitney U test was used to assess differences among non-parametric continuous ratios and intensity of uorescence staining positive between groups. Receiver operative characteristic (ROC) curve analysis was used to analyze sensitivity and speci city of particular gene expressions as a potential diagnostic marker for early-stage NSCLC. The two-sided test p-values < 0.01 were considered signi cant. All the statistical analyses were performed with the use of R program, version 3.6.3.

Demographic characteristics of study participants
We enrolled at total of 121 study participants, 44 treatment-naïve early-stage NSCLC patients, 20 nonmalignant pulmonary diseases and 57 healthy individuals. The median age of early-stage NSCLC patients was 63.5 years [range 34-83]. Seventy-two percent of patients were female. All patient had ECOG performance of 0-1. A majority (97.7%) of histology was adenocarcinoma. Pathological stage I disease accounted for 68.2%. All patients received curative attempt operation such as lobectomy and bilobectomy. The median age for the non-malignant pulmonary diseases and healthy controls were 48.5 years [range 30-76] and 58 years [range 30-86], respectively. Ninety-ve percent of non-malignant pulmonary disease patients presented with pulmonary nodule mimic lung cancer and had operative procedure. A majority (75%) of resected lung tissue histology was infection and in ammation. Details of patient demographics and disease characteristics at study enrollment are provided in Table 1.

Protein expression on PBMC in early-stage lung cancer patients and healthy controls
The potential of protein expression as a biomarker for early lung cancer detection was rst explored comparing to healthy control. The ratio of speci c antibody staining to CD3 positive was calculated from the number of speci c antibody staining to CD3 positive. Median ratio of C5AR1, NLRP3 and CLEC4A expression in early-stage NSCLC patients were 0. In addition to the number of positive staining, ow cytometry also enables the determining of the uorescence intensity on the cell surface of each speci c antibody-stained cell.  Table 2). Both ratio and average adjusted intensity of C5AR1, NLRP3 and CLEC4A were signi cant higher in early-stage NSCLC patients compared to healthy controls. This suggests that the number of speci c protein expression of C5AR1, NLRP3 and CLEC4A on immune cell surface changes in the presence of cancer. The comparison of ratio and average adjusted uorescence intensities of speci c protein expression on CD3 positive lymphocytes between early NSCLC patients and healthy controls are shown in Fig. 1. *Pass QC accounting percentage of specimen that had more than 10,000 CD3 + in each speci c antibody #Ratio calculated by percentage positive cells of speci c antibody to CD3 Protein expression on PBMC as potential candidate to discriminate early-stage non-small cell lung cancer patients from healthy controls To assess the potential to discriminate early-stage NSCLC patients from healthy controls, optimal cut-off ratio of each speci c antibody staining positive with CD3 positive was conducted. The C5AR1 expression ratio cut-off value at 0.28 could distinguish early-stage NSCLC patients from healthy controls at 100% sensitivity and 50.9% speci city (0.77 area under the ROC curve, accuracy rate 71.4%). The NLRP3 expression ratio cut-off value at 0.58 could distinguish early-stage NSCLC patients from healthy controls at 87.5% sensitivity and 64.7% speci city (0.83 area under the ROC curve, accuracy rate 74%). The CLEC4A expression ratio cut-off value at 0.34 could distinguish early-stage NSCLC patients from healthy controls at 97.4% sensitivity and 57.1% speci city (0.80 area under the ROC curve, accuracy rate 75%) (Fig. 1). Speci c protein expression ratio to CD3 positive staining showed better discrimination power than average adjusted protein expression.
Protein expression on PBMC in early-stage lung cancer and non-malignant pulmonary disease

Discussion
Previous studies from our group reported paracrine signal-induced PBMC gene upregulation in various cancer types through epigenetic regulation (4,16,17). On the basis of paracrine signal-induced PBMC gene upregulation, we explored the possibility to use protein expression on circulating T-lymphocytes as a biomarker to discriminate early-stage NSCLC and healthy individuals by automated ow cytometry analysis. We speculated that protein expression on circulating T-lymphocytes might represent Tlymphocyte tra cking before in ltrating into tumor microenvironment. In the rst-stage in this report, we chose proteins which related to immune function in either innate and/or adaptive immune processes. We found that ratios of all 3 chosen markers; C5AR1, CLEC4A, NLRP3 protein expression had been signi cantly increased in early-stage NSCLC patients compared to healthy controls. The discriminative sensitivity ranged 87.5-100% and speci city ranged 50.9-64.7%. Thus, measuring these proteins may allow us to distinguish early-stage NSCLC patients from healthy controls. Our proof-of-principle ndings strengthen the hypothesis that malignancies generate distinctive protein expression ngerprints on circulating T-lymphocytes.
In more detail of the 3 chosen markers, CLEC4A (C-type lectin domain family 4 member a) or dendritic immunoreceptor (DCIR) has been reported as an immune suppressor of dendritic cells which play a crucial role in the adaptive immune response. CLEC4A expresses on various immune cells depending on stage of maturation (18). Current evidence shows a role of CLEC4A down-regulation via small hairpin RNA inhibited tumor progression in animal models (19). NLRP3 (NOD-, LRR-and pyrin domain-containing 3) in ammasome complex plays a role in the innate immune signaling. Aberrant NLRP3 involves in various in ammatory conditions, infection immune responses and including cancer control (20). C5AR1 (complement component 5a receptor 1) is a G protein-coupled receptor for C5a. It functions as a complement receptor and modulates in ammatory responses via chemokine and cytokine signaling pathway. C5AR1 signaling contributes to the promotion of tumor growth by suppressing the adaptive immune response against tumor antigens and recruits myeloid-derived suppressor cells (MDSCs) into tumor (21). Considering function and mechanistic-relation of these 3 protein expressions, it might be the reason why we could not use them to discriminate non-malignant pulmonary disease with a majority of in ammation and infection from early-stage non-small cell lung cancer.
The technique measures uorescence signals through a laser source. Gating techniques allows the simultaneously measurement of the cell surface protein expression on a single cell level (22). The detection of a speci c antigen by ow cytometry is a rapid, easy and a semi-quantitative assay.
Therefore, ow cytometry with immuno uorescence-labeled cells is a promising tool for the pro ling of cell surface protein signaling from peripheral blood samples in clinical practice. It may provide valuable information for assessment and cancer monitoring during treatment. Further elucidation in cancerspeci c PBMC protein expression might provide information for lung cancer screening.

Conclusions
Our proof-of-principle ndings strengthen the hypothesis that malignancies generate distinctive protein expression ngerprints on circulating T-lymphocytes. Represents ratio expression of C5AR1, NLRP3 and CLEC4A between healthy control and early-stage NSCLC patient and area under the curve (above). Cut-off value at 0.28, 0.58 and 0.34 for C5AR1, NLRP3 and CLEC4A ratio expression could represent 87.5-100% sensitivity and 50-64.7% speci city. Average adjusted expression (calculated by average intensity × ratio expression) of C5AR1, NLRP3 and CLEC4A between healthy control and early-stage NSCLC patient and area under the curve are also represented (below).