Genetically engineered mice [B6.129P2-Lgr5tm1 (cre/ERT2) Cle/J] were donated to our laboratory by Professor Geng (Department of Biologic and Materials Sciences, University of Michigan School of Dentistry, Ann Arbor, MI, USA). This mouse model was produced by knock-in of the Lgr5-EGFP-IRES-creERT2 allele, which results in the loss of function of one Lgr5 allele and the expression of the CreERT2-EGFP fusion protein. Lgr5 mice do not exhibit any obvious abnormalities in terms of birth rate or any abnormal physiological properties. By contrast, the homozygous deletion of Lrg5 in mice resulted in neonatal lethality (12). To investigate the role of Lgr5 in the lungs, Lgr5^+/− mice were used in the present study. The mice were housed in an individual ventilated cage (IVC) environment at 24±2°C and 60±5% humidity with 12 h light/dark cycles. The mice were fed a bacteria-free diet. All mice were sacrificed under diethyl ether anesthesia. Guangdong College of Pharmacy Approved Animal Ethics Committee approved all murine experiments. To identify the genotype of the mice, DNA was extracted from a 0.5–1 cm portion of the tail. The DNA was amplified in polymerase chain reaction (PCR) amplification buffer (2X PCR mix; Promega, Beijing, China), and the PCR products were subsequently run on a 2% agarose gel. The PCR product of the homozygous mice was 174 bp, the heterozygous mice PCR product was 174 bp and the WT mice product was 298 bp. The primers used were as follows: Common (sequence no. 8060), 5′-CTGCTCTCTGCTCCCAGTCT-3′; WT reverse (sequence no. 8061), 5′-ATACCCCATCCCTTTTGAGC-3′; mutant reverse (oIMR9402) 5′-GAACTTCAGGGTCAGCTTGC-3′. The PCR system and PCR reaction conditions were in accordance with the protocols of the Jackson Laboratory (Bar Harbor, ME, USA).

Lgr5 expression was evaluated in a total of 42 primary non-small cell lung cancers (NSCLCs) and 28 matched normal adjacent lung tissue samples by immunohistochemistry (IHC). The 42 samples included 22 lung adenocarcinoma samples and 20 squamous cell carcinoma samples. In addition, a tissue microarray (TMA) that contained 80 cases of lung adenocarcinoma was also included. Therefore, the total number of lung adenocarcinoma cases was 102, whereas the total number of squamous cell carcinoma samples was 20. The clinical features of the patients with NSCLC are outlined in Table I. All specimens were collected from patients who were admitted to the Department of Thoracic Surgery, Sun Yat-Sen University Cancer Center, (Guangzhou, China) between October 9, 2009 and July 12, 2014 with patient consent and institutional review board approval. All patients signed informed consent for the use of their clinical specimens for medical research. The samples in this study were approved by the Committee for Ethical Review of Research at Sun Yat-Sen University.

The paraffin embedded tissues were sectioned at a thickness of 3–5 µm, and hematoxylin and eosin (H&E) staining was performed, as previously described (13). Briefly, IHC was performed as follows: Antigen retrieval with ethylenediaminetetraacetic acid solution (pH 9.0), incubation of the slides in 3% H₂O₂ at room temperature for 12 min, followed by washes and the blocking of non-specific proteins. The antibody was applied to the slides, which were subsequently incubated at 4°C overnight. The tissue sections were washed again and the secondary antibody was applied to the slides. Finally, diaminobenzidine was used to visualize the positive staining on the slides. The primary Lgr5 antibody was used at a dilution of 1:100 (cat. no. sc-135238; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). The anti-rabbit secondary antibody was conjugated to horseradish peroxidase and was used at a dilution of 1:500 (cat. no. GK500705). The diaminobenzidine chromogen system was purchased from Gene Tech Co., Ltd. (Shanghai, China). The slides were viewed on a BX51 Olympus microscope (Olympus Corporation, Shibuya, Japan).

All statistical analyses were performed using GraphPad Prism software version 5 (Graphpad Software, CA, USA). The χ² test or Fisher's exact test was used to analyze the association of Lgr5 expression with the clinicopathological parameters of the patients. The survival curves were plotted according to the Kaplan-Meier method to evaluate the Lgr5 expression levels with respect to the survival rate. The multivariate survival analyses were performed by a Cox proportional hazard model using the Wald test. P<0.05 was considered to indicate a statistically significant difference..

Lgr5 is well-characterized as a marker of normal stem cells in the intestine and as a marker of cancer stem cells. The Lgr5 homozygous knockout in mice resulted in neonatal lethality (12). The present study suggested that Lgr5 serves an important role in development. However, the role of Lgr5-positive cells in the lung remains unknown. To investigate whether cells in normal mouse lung express Lgr5, normal lung tissue was obtained from C57/BL WT mice and the expression of Lgr5 was determined by IHC using a specific antibody. To confirm that the antibody was specific, normal mouse intestine was stained. The results are shown in Fig. 1A. In the lung tissue, the results demonstrated that some Lgr5-positive cells were present in the bronchi and alveoli (Fig. 1B and C). The present study next determined whether the knockout of Lgr5 affected lung structure and found that the structure of the lung tissue of Lgr5 heterozygous knockout mice (Lgr5^+/− mice) differed from that of the WT mice. This finding indicated that the lung septa of Lgr5^+/− mice were thicker and that the lung structure was irregular compared with the WT mice (Fig. 1E).

Based on the above IHC and histology results, the present study hypothesized that Lgr5 may serve a function in the maintenance of the morphology of bronchi or alveoli in the normal lung. It must be considered that lung tumor initiating cells are partly derived from the bronchi and that Lgr5 is a well-known marker of intestinal cancer stem cells. Therefore, in lung cancer, the present study assumed that Lgr5 is associated with tumor formation or to the tumor, node, metastasis (TNM) stage. Next, 122 lung cancer cases, including 102 cases of adenocarcinoma and 20 cases of squamous carcinoma, were subjected to IHC. The results showed that Lgr5 is not expressed in squamous cell carcinoma (0/20; P<0.001; Table I; Fig. 2A), however, is expressed in adenocarcinoma (47/122). The representative image of the staining is presented in Fig. 2B and C. Notably, in certain cases, Lgr5-positive cells were present in areas of tumor necrosis (Fig. 2D), areas adjacent to the tumor (Fig. 2E), and the lymphatic spaces and vascular lumina (Fig. 2F).

The above data showed that Lgr5 is expressed in lung adenocarcinoma, but not in squamous carcinoma. Furthermore, the present study stained one tissue microarray that included 60 cases of lung adenocarcinoma to assess its expression by IHC using an Lgr5 antibody. These cases were classified into two groups as either ‘+’ or ‘-’, as shown in the representative image in Fig. 3A. The positive group was indicated by ‘+’ and the negative group was indicated by ‘-’.

A survival analysis was performed for all patients with lung adenocarcinoma (paraffin blocks, 22 cases; TMA, 80 cases) using the Kaplan-Meier method. The follow-up periods ranged between 4 and 157 months. To evaluate the association between Lgr5 expression in lung adenocarcinoma cells and the clinicopathological parameters, the data from the patients was summarized in Table I. Lgr5 expression was not associated with tumor size (P=0.075). In addition, no significant association was observed between Lgr5 expression and age, gender, tumor invasion or lymph node metastasis. However, Lgr5 expression was associated with TNM stage (P=0.026).

Additionally, the median survival was 30.5 months for patients in the Lgr5-negative group (n=55) and 12 months (n=47) for patients in the Lgr5-positive group, which indicated a significantly poorer survival rate of the Lgr5-positive group compared with the Lgr5-negative group (P=0.033). Next, univariable and multivariable analyses were performed to assess the effect of Lgr5 expression on survival. A Cox proportional hazards regression model was subsequently applied and the effect of Lgr5 on survival was estimated. The crude hazard ratio (HR) of the Lgr5-positive tumors compared with the Lgr5-negative tumors was 1.618 [95% confidence interval (CI), 1.06–2.48; P=0.027]. To estimate the independent pro-diagnostic effect of Lgr5 on survival, the present study adjusted for confounding factors. The present study involved 59 lung adenocarcinoma patients with lung cancer-associated mortalities and 5 variables that were included in the multivariate regression model. To determine whether Lgr5 expression was a single variable and not a potential confounding factor, a propensity score was applied. The five variables were clinicopathological factors, including age, tumor size, TNM stage, tumor invasion and metastasis. The adjusted HR of the Lgr5-positive group was 0.396 (95% CI, −0.28–1.06; P=0.15) compared with the Lgr5-negative group, which suggested that Lgr5 expression is not an independent risk factor for poorer survival (Table II).