Identification and Exploration of Serine Peptidase Inhibitor Kazal Type 1 (SPINK1) as a Potential Biomarker Correlated with the Progression of Non-Small Cell Lung Cancer

Background Lung cancer is one of the most common malignancies worldwide and is the leading cause of cancer-related death. Approximately 85% of lung cancer patients represent a group of histological subtypes collectively known as non-small cell lung cancer (NSCLC). Methods To explore the molecular mechanisms underlying tumorigenesis and progression of NSCLC, mRNA expression profiles were downloaded from GEO database (GSE19804, GSE18842, GSE27262, and GSE43458) and differentially expressed genes (DEGs) in NSCLC tissues were analyzed by GEO2R. DEGs in NSCLC tissues were further analysed in TCGA (The Cancer Genome Atlas), GTEx (The Genotype-Tissue Expression Project) and IST (In Silico Transcriptomics) online databases. Serum of NSCLC patients and normal controls were collected and serum concentration of SPINK1 were analysed by ELISA. mRNA and protein expression levels of SPINK1 were analysed by qRT-PCR and western assays. siRNA targeting SPINK1 and normal controls were used for the silence of SPINK1 and GAPDH. CCK-8 assays were employed for cell proliferation detection. Flow cytometric analysis and western blot assays were conducted to assess cell cycle distribution and apoptosis. Western blot assays were performed for the evaluation of cell autophagy and signaling pathways. Results Among these DEGs, SPINK1 was distinctively up-regulated in NSCLC tissues, which was further validated in TCGA, GTEx and IST databases. Furthermore, serum SPINK1 concentration notably increased in NSCLC patients compared with normal controls. Besides, both mRNA and protein expression levels of SPINK1 significantly increased in human NSCLC cell lines A549 and H1299 compared with normal human bronchial epithelial cell line HBE. Silence of SPINK1 significantly inhibited proliferation of NSCLC cell lines, and exogenous addition of rhSPINK1 partially rescued proliferation suppressed by endogenous knockdown of SPINK1. Mechanistically, silence of SPINK1 could inhibit MEK/ERK signaling pathway, while rhSPINK1 could activate MEK/ERK signaling pathway and then promote proliferation of NSCLC in vitro experiments, we demonstrated that SPINK1 played a role in cell cycle distribution of NSCLC cell lines. In eukaryotes, DNA replication is limited in S-phase, and chromosome segregation occurs at M-phase. Two gap phases, known as G1 and G2, separate S-phase and M-phase. Multiple cell cycle checkpoints participate in the regulation of cell cycle progression such as cyclin-dependent kinases (CDKs), cyclins and CDKs inhibitors ( CKI ). CDKs act as engine to drive cell cycle progression (44). Previous reports suggested that during the G1 phase, CDK4/cyclin

for 30 minutes with TBST and subsequently incubated with the corresponding secondary antibody (#656120, HRP goat anti rabbit, 1:10000, Invitrogen; #A32723, HRP goat anti mouse, 1:10000, Invitrogen) for 1 h at 37℃. Membranes were then washed another three times for 30 minutes with TBST. Protein bands were visualized using Super Signal electrochemiluminescence (#34580, Thermo Scientific) and quantitated with image Pro Plus 6.0, and the data were normalized to GAPDH.
Statistical analysis. Data represent three independent experiments and are presented as mean ± SD (n = 3). GraphPad Prism software (version 5.01) was applied for statistical analysis. Unpaired t tests were used for data with a normal distribution. Differences with p < 0.05 were considered statistically collected to detect the SPINK1 concentration by ELISA. As shown in Fig.2d, serum SPINK1 in NSCLC patients was significantly higher than normal controls (1092.9pg/mL vs 644.5pg/mL, p<0.01).

SPINK1 promotes proliferation of NSCLC cells
To investigate the effect of SPINK1 on the proliferation of NSCLC cell lines, siRNAs targeting SPINK1 were transfected into NSCLC cell lines. qRT-PCR (Fig.3a, S Fig.1a) and western blot (Fig.3b, S Fig.1b) results confirmed successful knockdown of SPINK1 in A549 and H1299 cell lines respectively. CCK8 assay and growth curve were applied to detect the vitality of cells and results indicated that silence of SPINK1 significantly inhibited cell proliferation of A549 and H1299 (Fig.3c, Fig.3d, S Fig.1c). Given that SPINK1 is a secreted soluble protein, we speculated that exogenous addition of rhSPINK1 could stimulate the proliferation of H460 with relatively low SPINK1 expression. CCK8 assay results confirmed that rhSPINK1(1ng/mL) promoted the proliferation of H460 (Fig.3e). Moreover, addition of exogenous rhSPINK1(5ng/mL) could rescue cell vitality of A549 and H1299 after transfection with si-SPINK1-1 (Fig.3f, S Fig.1d). Apart from that, silencing SPINK1 in A549 and H1299 caused cell cycle arrest, with strikingly increased proportion of cells in G 1 -phase and decreased proportion of cells in Sphase ( Fig. 4a, S Fig. 2a). Consistently, silencing SPINK1 decreased the expression of cell cycle checkpoint protein CDK4 in A549 (Fig.4b) and H1299 (S Fig.2b). Above results validated that SPINK1 promotes proliferation of NSCLC cell lines.

Silence of SPINK1 activates autophagy and apoptosis of NSCLC cells
Both autophagy and apoptosis are vital process during progression of development and oncogenesis.
As shown in Fig. 5a, silence of SPINK1 in A549 activated autophagy characterized by increased expression of Beclin-1 and ATG7, decreased P62 and increased ratio of LC3 II/LC3I. Consistently, knockdown of SPINK1 in H1299 induced excessive autophagy (S Fig. 3a). Meanwhile, silence of SPINK1 increased expression of pro-apoptosis protein BAX and decreased expression of anti-apoptosis protein BCL-2 (Fig.5b). Consistently, flow cytometry analysis indicated that silence of SPINK1 in A549 and H1299 resulted in cell apoptosis (Fig.5c, S Fig.3b, S Fig.3c).

SPINK1 accelerates the proliferation of NSCLC via activation of MEK/ERK signaling pathway
It has been universally accepted that SPINK1 acting as a pro-survival factor accelerated the progression of multiple cancers by interacting with EGFR ( 25, 28, 29, 38, 53). In order to clarify signaling pathways downstream of EGFR mediating the carcinogenic effect of SPINK1 in NSCLC, the phosphorylation levels of MEK1/2, ERK, PI3K and AKT were assessed by western blot. Results showed that silence of SPINK1 in A549 and H1299 cells decreased the phosphorylation levels of MEK1/2, ERK, PI3K and AKT, and MEK/ERK signaling pathway represented more significant inhibition (Fig. 6a, S Fig.   4a). In addition, WB results demonstrated that rhSPINK1(5ng/mL) could increase the phosphorylation levels of p-ERK, which was inhibited by U0126 (10μM, the specific inhibitor of MEK) (Fig. 6b, S Fig. 4b).

Discussion
In our present study, mRNA expression profiles were downloaded from GEO, TCGA, GTEx and IST databases and analyzed by integrated bioinformatic analysis. 144 DEGs were identified in NSCLC tissues compared with normal lung tissues. Among these DEGs, SPINK1 attracted our attention, for there are few reports about the effect and mechanism of SPINK1 on the progression of NSCLC up to date.
Numerous studies have shown that SPINK1 plays multiple biological roles under various physiological and pathological conditions. As a serpin typically produced by pancreatic acinar cells, SPINK1 functions to defense against trypsinogen activation in the acini and the pancreatic ducts (34). Serving as an acute phase reactant, SPINK1 is increasingly secreted into blood in a state of stress such as major surgery and serious inflammatory and septic complications (16). Ken-ichi Yamamura et al.
reported that SPINK3, the homologous gene of SPINK1 in mouse, may participate in the differentiation and development of mouse embryo (35). In addition, SPINK1 secreted by cancer cells could coimmmunoprecipitate with serine protease granzyme A (GzmA), a cytolytic granule secreted by natural killer cells and cytotoxic T lymphocytes, thereby suppressing GzmA-mediated apoptosis and establishing a tolerance of cancer cells to the immune surveillance system (32). More importantly, SPINK1, as a growth factor, is aberrantly expressed in multiple human cancers such as ovarian cancer (33), prostatic cancer (36), colorectal cancer (37), hepatocellular carcinoma (38) and pulmonary adenocarcinoma (39). Based on big data analysis, we detected a significant increase of SPINK1 expression in NSCLC tissues. Consistently, a recent study reported that tissue SPINK1 is overexpressed in NSCLC patients (40). Additionally, SPINK1 could be detected in serum and act as independent prognostic factor in prostatic cancer (41), hepatocellular carcinoma (42), bladder transitional cell carcinoma (21), colorectal cancer (26) and renal cancer (43). So we speculated that serum SPINK1 might serve as a potential and promising biomarker in NSCLC. In our study, we founded that concentration of SPINK1 in serum of NSCLC patients was higher than normal controls, and concentration of SPINK1 in pleural effusions was even higher than in serum, which may shed new insight into early body fluid diagnosis of NSCLC. In the following study, we will further explore relationships between serum SPINK1 level and NSCLC stage, metastasis and overall survival of patients.
In our study, we founded that the expression of SPINK1 increased in A549 and H1299 cells compared with HBE, which was consistent with the report that well-differentiated adenocarcinoma tissues express stronger SPINK1 (39). In in vitro experiments, we demonstrated that SPINK1 played a role in cell cycle distribution of NSCLC cell lines. In eukaryotes, DNA replication is limited in S-phase, and chromosome segregation occurs at M-phase. Two gap phases, known as G1 and G2, separate S-phase and M-phase. Multiple cell cycle checkpoints participate in the regulation of cell cycle progression such as cyclin-dependent kinases (CDKs), cyclins and CDKs inhibitors ( CKI ). CDKs act as engine to drive cell cycle progression (44). Previous reports suggested that during the G1 phase, CDK4/cyclin D and CDK2/cyclin E complexes could phosphorylate Rb, inducing the activation of E2F and the transcription of E2F responsive genes, which are required for G1/S transition (45,46). In our studies, we indicated that silence of SPINK1 in A549 and H1299 cells resulted in cell cycle arrest, with increased percentage of cells in G 1 -Phase and decreased percentage of cells in S-Phase. In addition, western blot assays showed decreased expression of CDK4 in SPINK1 silencing groups. Consistent Autophagy is a process characterized by engulfing cytoplasmic proteins, complexes or organelles into the autophagosome. Accumulating evidence confirmed that there is a crosstalk between autophagy and apoptosis. In basal state, Beclin-1 binds to BCL-2, an anti-apoptotic protein, and then autophagy and apoptosis are inhibited. A short period (4 h) of nutrient starvation could activate c-Jun N-terminal protein kinase 1 (JNK1), following that BCL-2 was phosphorylated by p-JNK1 and then dissociated with Beclin-1, thereby activating autophagy. During the short period of nutrient starvation, phosphorylated BCL2 binds with pro-apoptotic protein BAX and exerts an anti-apoptotic effect (47,48). However, BCL-2 is hyperphosphorylated and the BCL-2/Bax complex is disrupted after 16 hours of nutrient starvation, followed by the activation of caspase 3 and the initiation of apoptosis (49). YAMAMURA et al. reported autophagic cell death of pancreatic acinar cells in SPINK3 deficient mice, with increased ratio of LC3 II/LC3 I and increased formation of autophagosome (50). In our study, NSCLC cells were transfected with siRNAs and cultured with incomplete culture medium containing 2% FBS for 48 h.
Consistent with these reports, we showed that silence of SPINK1 activated autophagy and apoptosis.
To some extent when autophagy is no longer able to matain cell survival, apoptosis will be initiated, with decreased expression of anti-apoptosis protein BCL-2 and increased expression of pro-apoptosis protein BAX. In our models, we speculated that silence of SPINK1 conferred sensitivity to starvation and induced excessive autophagy and apoptosis.
At present, most reports indicated that SPINK1 phosphorylates EGFR and thereby acting the downstream signaling pathways to exert oncogenic effect. MEK/ERK (25, 29 38, 53) p38 (28) and PI3K/AKT (51) signaling pathways were reported as executants downstream of EGFR. In our studies, we validated that silence of SPINK1 decreased the phosphorylation levels of MEK1/2, ERK, PI3K and AKT, while producing no effect on phosphorylation levels of p38. Notably, compared with PI3K/AKT signaling pathway, more significant decrease of MEK/ERK phosphorylation levels was observed after treatment with rhSPINK1. Further, MEK/ERK signaling pathway and cell proliferation of NSCLC cells were suppressed by higher doses of MEK inhibitor, U0126, indicating that SPINK1 could accelerate progression of NSCLC partially through MEK/ERK signaling pathways.
In summary, our study revealed that silence of SPINK1 in NSCLC cell lines could inhibit cell proliferation, induce cell cycle arrest, and activate cell autophagy and apoptosis. SPINK1 promotes progression of NSCLC partially through MEK/ERK signaling pathway. Our study presented evidence in support of SPINK1 as a potential bio-target in the progression of NSCLC and provides new insight into the underlying mechanism.