KLRK1 as a prognostic biomarker for lung adenocarcinoma cancer

Lung cancer is one of the most common malignancy worldwide and causes estimated 1.6 million deaths each year. Cancer immunosurveillance has been found to play an important role in lung cancer and may be related with its prognosis. KLRK1, encoding NKG2D, is a homodimeric lectin-like receptor. However, there has not been one research of KLRK1 as a biomarker in lung cancer. Data including patients` clinical characteristics and RNAseq information of KLRK1 from TCGA were downloaded. A total of 1019 patients with lung cancer were included in this study, among which 407 patients were female and 611 patients were male. Evaluations of mRNA expression, diagnostic value by ROC (receiver operating characteristic) curves and prognostic value by survival curve, Cox model and subgroup analysis were performed. The level of KLRK1 expression in lung adenocarcinoma cancer tissues and normal lung tissues was detected by qRT-PCR. The CCK-8 assay investigated the proliferation rate and the wound healing assay assessed the migratory ability in vitro. The expression of KLRK1 in tumor was lower than that in normal tissue. KLRK1 expression was associated with gender, histologic grade, stage, T classification and vital status. Patients with high KLRK1 expression presented an improved overall survival (P = 0.0036) and relapse free survival (P = 0.0031). KLRK1 was found to have significant prognostic value in lung adenocarcinoma (P = 0.015), stage I/II (P = 0.03), older patients (P = 0.0052), and male (P = 0.0047) by subgroup overall survival analysis, and in lung adenocarcinoma (P = 0.0094), stage I/II (P = 0.0076), older patients (P = 0.0072), and male (P = 0.0033) by subgroup relapse free survival analysis. Lung adenocarcinoma cancer patients with high KLRK1 expression presented an improved overall survival (P = 0.015) and relapse free survival (P = 0.0094). In vitro studies indicated that KLRK1 inhibited tumor cell proliferation and migration. KLRK1 was an independent prognostic factor and high KLRK1 expression indicated a better overall and relapse free survival. KLRK1 may be a prognostic biomarker for lung adenocarcinoma cancer.

www.nature.com/scientificreports/ in R was used and Cox model was established, followed by evaluation of subgroups including different genders, different ages, and different stages.

Methods
Data mining. To collect original data from patients suffering from lung cancer, data mining was carried out.
Specifically, we download data including patients' clinical characteristics and RNAseq information of KLRK1 from TCGA (The Cancer Genome Atlas) database by UCSC Xena. Given that all the data were open to public, no ethical approval was needed. A total of 1019 patients with lung cancer were included in this study, among which 407 patients were female and 611 patients were male. The RNA expression of KLRK1 was shown in boxplots as center line represents median, top line and bottom line of the box represents upper and lower quartiles and vertical lines represents 95% confidence intervals (95% CI).
To study the diagnostic value of KLRK1 in patients with lung cancer, ROC (Receiver operating characteristic) curves was plotted by pROC package 14 . The calculated area under curves (AUC) indicated the diagnostic value. Furthermore, according to the identified threshold level of KLRK1, the patients were grouped into high expression group and low expression.
To study the prognostic value of KLRK1 in patients with lung cancer, the survival package in R was used and Cox model was established 15 . Evaluation of subgroups was performed as well.
Wound healing assay. Wound-healing assays were performed as previously described 16 . The migration of cells toward the wound was photographed under a Nikon fluorescence microscope.
Immunohistochemical staining in HPA database and qRT-PCR. The Human Protein Atlas (HPA, http:// www. prote inatl as. org/) online database was explored to validate the KLRK1 protein expression in lung cancer by immunohistochemical (IHC) staining by CAB021896 antibody 17 . Ten pairs of lung adenocarcinoma cancer tissues and normal lung tissues were obtained from primary adenocarcinoma cancer patients at the Affiliated Hospital of Weifang Medical University. According to the manufacturer's instructions, total RNA was extracted using TRIzol reagent, cDNA was synthesized and the qRT-PCR was performed and calculated by means of 2−ΔΔCt methods. The related primer of KLRK1 were displayed as following: F: 5ʹ-TGG ATT CGT GGT CGG AGG TCTC-3ʹ, R: 5ʹ-GGA CAT CTT TGC TTT TGC CAT CGT G-3ʹ.
Cell culture and cell transfection. A549 cell lines were purchased from American Tissue Culture Collection, and cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37 °C with 5% CO 2 .
The KLRK1 sequence was amplified and inserted into pCMV vector (Beyotime). The transfection of KLRK1 over-expression or control plasmid was performed using Lipofectamine 3000 (Invitrogen).
Cell proliferation assay. The cells were treated with plasmids and cultured for 24 h. Then, 10 μL of CCK-8 reagent was added and cultured for 20 min. A microplate reader was used to measure the absorbance at 490 nm. The cell viability was calculated relative to the untreated control.
To evaluate the cell viability, co-staining of calcein AM and PI was performed. The cells were seeded and incubated for 16 h. After different treatments for 24 h, the cells were stained and observed using a Nikon fluorescence microscope.
Wound healing assay. Wound-healing assays were performed as previously described 16 . The migration of cells toward the wound was photographed under a Nikon fluorescence microscope.
Statistical analysis. R version 3.5.2 packages (https:// www.R-proje ct. org) was used for bioinformatics analysis 18 . Data were presented using the ggplot2 package in R 19 . The Wilcoxon rank-sum test was used for comparison between two groups, and the Kruskal-Wallis test was used for comparison among three or more groups. For assessment of associations between KLRK1 expression and clinical parameters, the chi-squared test was used, and corrected by Fisher's exact test. Data from in vitro and in vivo experiments were analyzed by the Student's t-test (unpaired, two-tailed). P < 0.05 was statistical significance.
Ethical approval. This study was approved by the ethics committee of the Affiliated Hospital of Weifang Medical University and conducted in strict accordance with the National Institutes of Health guidelines.

Results
Characteristics of patients with lung cancer. Clinical characteristics of the patients with lung cancer, including age, gender, histological type, stage, T classification, N classification, M classification, radiation therapy, residual tumor, vital status, sample type, KLRK1 expression, were shown in Table 1. The percentage of two histological types, lung adenocarcinoma (50.74%) and lung squamous cell carcinoma (49.26%), was close. About half of the patients were in stage I (51.13%). As for T, N, M classification, T2 (56.04%), N0 (63.98%), and M0 (74.39%) were highest among each classification. Most patients (99.80%) were primary tumor. Low KLRK1 expression in tumor. As shown in Fig. 1 www.nature.com/scientificreports/ histological type, sample type, age, gender, radiation therapy, was further evaluated. KLRK1 expression in lung squamous cell carcinoma was lower than that in lung adenocarcinoma (P = 2.2e−05). KLRK1 expression in male was lower than that in female (P = 3e−04). Other clinical features did not show statistical differences (P > 0.05). Association of KLRK1 expression with stage, T classification, N classification, M classification, residual tumor and vital status was evaluated (Fig. 2). The KLRK1 expression got decreased progressively with higher stages (P = 0.0015) and T classification (P = 2.6e−06). No significances were observed in N classification (P = 0.200), M classification (P = 0.091) and residual tumor (P = 0.790). The KLRK1 expression in living patients was a little higher than that in deceased patients (P = 0.011).
Diagnostic value of KLRK1 for lung cancer. As shown in Fig. 2, the ROC analysis was first performed in all lung cancer patients, indicating a modest diagnostic value with AUC of 0.789. Moreover, different stages of lung cancer were analyzed. It was suggested that the diagnostic value was increasing with stage getting higher from stage I (AUC = 0.766) to stage II (AUC = 0.797) to stage III (AUC = 0.835) finally to stage IV (AUC = 0.853) (Fig. 3).

Correlation of KLRK1 expression with clinical features.
According to the threshold value ascertained by ROC analysis, patients were divided to two subgroups, high (n = 701) and low (n = 318) KLRK1 expression group. The relationship between the clinical features and KLRK1 expression in patients with lung cancer was shown in Table 2. KLRK1 expression was associated with gender (P = 0.005), histologic grade (P = 0.001), stage (P = 0.001), T classification (P < 0.001) and vital status (P = 0.001). No correlation was found between KLRK1 expression and age (P = 0.431), N classification (P = 0.218), M classification (P = 0.383), radiation therapy (P = 0.399), residual tumor (P = 0.336), and sample type (P = 0.850).

IHC and qRT-PCR result.
Comparing the IHC experimental pictures obtained in the HPA database, it can be seen that the expression of KLRK1 in lung adenocarcinoma cancer tissues was significantly lower than that in normal lung tissues. As shown in Fig. 4, we utilized qRT-PCR to validate the KLRK1 expression in lung adenocarcinoma cancer tissues and found the KLRK1 expression was down-regulated in the lung adenocarcinoma cancer (N = 10) compared with normal lung tissues (N = 10; P < 0.001).
Overall survival and relapse free survival of KLRK1 for lung cancer. Given that KLRK1 was correlated with survival, the prognostic value of KLRK1 was further studied. As shown in Fig. 5, lung cancer patients KLRK1 is an independent risk factor for lung cancer's overall survival. The univariate Cox model of overall and relapse free survival in patients with lung cancer was established (Tables 3, 4, Table S1), and the multivariate analysis of overall and relapse free survival in patients with lung cancer was further performed (Tables 5, 6, Table S2). As shown in Table 3, no obvious differences were observed in age (P = 0.867), gender (P = 0.201), histological type (P = 0.282), M classification (P = 0.220), and radiation therapy (P = 0.983). T classification (P < 0.001), N classification (P < 0.001), residual tumor (P = 0.016), stage (P < 0.001) and KLRK1 expression (P = 0.004) showed significant differences for lung cancer. As shown in Table 4, T classification (P < 0.001), N classification (P < 0.001), M classification (P = 0.029), residual tumor (P = 0.037), stage (P < 0.001) and KLRK1 expression (P = 0.016) showed significant differences for lung adenocarcinoma cancer. As shown in Table 5, T classification (HR 1.18; 95% CI 1.02-1.36; P = 0.025) and stage (HR 1.34; 95% CI 1.14-1.57; P < 0.001) were risk factors for lung cancer.

KLRK1 inhibits lung cancer cell proliferation and migration in vitro.
We further studied the effect of KLRK1 on cells by overexpressing KLRK1 in lung cancer A549 cell lines. The CCK-8 assay showed that KLRK1 upregulation significantly inhibited cell proliferation (P < 0.05; Fig. 9A). Compared with control group, KLRK1 increased the percentage of dead cells by live/dead staining (Fig. 9B). KLRK1 decreased the migration distance of cancer cell (P < 0.05; Fig. 9C,D).
Recently, research about biomarkers by data mining is popular 20,21,23,24,27,30,[36][37][38][39][40] . Although some protein blood biomarkers have already been put into clinical practice, the ability of their diagnosis and prognosis is limited and he exploration of biomarkers for lung cancer is ongoing 13 . CEA (carcinoembryonic antigen) and CTC (Circulating tumor cells) are used in some lung cancers 41,42 . Also, there are studies reporting some novel biomarkers. Jiang et al. found thymidine kinase 1 combined with CEA, CYFRA21-1 and NSE improved its diagnostic value for lung cancer 43 45 . In this study, we first found KLRK1 as a prognostic biomarker for lung cancer.
Researches of biomarkers for lung cancer are not only limited in prognosis but also in diagnosis. Wu et al. found UCK2 (Uridine-cytidine kinase 2) as a potential diagnostic and prognostic biomarker for lung cancer, and identified UCK2 highly expressed in stage IA lung cancer with AUC > 0.9 46 34 . Moreover, blockade drugs for EGFR and PD-1 can enhance the effective of NKG2D, encoded by KLRK1, lead to cancer cell recognition and killing by NK effector cells 50,51 . From our results, high KLRK1 expression is associated with a better overall and relapse free survival, which may attributes to the immunosurveillance of KLRK1 therefore suppressing the proliferation and metastasis of lung cancer 4 . Of note, our results showed the significances of KLRK1 in both overall survival and relapse free survival in lung cancer.
Our research first suggested the diagnostic and prognostic value of KLRK1 for lung cancer. However, the major limitation is that this study analyzed the data from a single database by data mining. Further verifications in different areas and populations are required. Besides, in vivo function experiments and exploration of its molecular mechanism would further illuminate the role of KLRK1 in lung cancer.

Conclusions
In conclusion, KLRK1 was lower expressed in lung cancer in comparison with normal lung tissue. KLRK1 expression was associated with gender, histologic grade, stage, T classification and vital status. KLRK1 had a diagnostic value for lung adenocarcinoma cancer. KLRK1 was an independent prognostic factor and high KLRK1 expression indicated a better overall and relapse free survival. KLRK1 may be a novel biomarker for lung adenocarcinoma cancer.