In the present study, we evaluated the differential expression of seven stem-cell-associated markers in bronchoscopic biopsies of lung cancer and non-cancer patients by RT-PCR, and assessed their diagnostic value potential. Our data found Nanog mRNA had the highest specificity in lung cancer. We further confirmed the high diagnostic value of Nanog protein levels by IHC, Nanog was overexpressed in lung cancer tissues, but rarely expressed in non-malignant lung tissue. Taken together, these results demonstrate that Nanog mRNA is a potential diagnostic marker for lung cancer.
Nanog is a transcription factor that plays an important role in maintaining self-renewal of embryonic stem cells. Current studies have reported that the expression of Nanog was higher in multiple cancerous tissues than in their normal counterparts, including breast cancer , gastric adenocarcinomas , colorectal cancer , gliomas  and ovarian serous cystadenocarcinomas . In this study, we found the expression of Nanog mRNA in bronchoscopic biopsies of lung cancer patients was significantly higher compared to that in non-cancer patients. Although Nirasawa et al.  have also reported that the expression of Nanog mRNA was higher in surgically resected lung cancer tissues than in non-cancerous tissues, it is not known what cells express Nanog in non-cancerous lung tissues. Using IHC, we found Nanog was only expressed in metaplastic squamous bronchial epithelium cells in 2 out of 50 non-malignant lung tissues, and was negative in normal airway epithelia. Therefore, Nanog may be a good diagnostic marker for lung cancer.
In this study, our results showed that the mRNA levels of Bmi1, CD44 and CD133 were not significantly different between lung cancer and non-malignant lung tissues. Further analyzed by IHC, we observed that Bmi1, CD44 and CD133 were not only expressed in lung cancers, Bmi1 and CD44 were also abundantly expressed in lung interstitial cells, inflammatory cells and bronchial epithelium cells, and CD133 was diffusely expressed in some normal bronchial epithelium cells and bronchial smooth muscle cells, consistent with previous studies [11, 25, 26]. Hence, Bmi1, CD44 and CD133 are poor diagnostic markers for lung cancer.
Likewise, although the expression levels of Sox2 and Msi2 mRNA in lung cancer tissues were significantly higher as compared with non-malignant tissues, we found more than 80% of bronchoscopic biopsy specimens of non-cancer patients were positive for Sox2 and Msi2 mRNA, and all non-malignant tissues were positive for Sox2 and Msi2 protein expression, consistent with previous findings [10, 27, 28]. Therefore, Sox2 and Msi2 have poor diagnostic specificity in lung cancer.
It is still controversial whether lung cancer cells express OCT4. Some researchers believe that OCT4 is involved in tumorigenesis and metastasis of lung cancer, and therefore is a potential diagnostic marker and useful therapeutic target of lung cancer [9, 15, 29], while others did not detect OCT4 expression in lung cancer [10, 30]. Although we observed OCT4 mRNA expression in 85.7% of lung cancer and 38.8% of non-cancer bronchoscopic biopsy specimens, but OCT4 protein was nearly absent in 50 cases of lung cancer tissues. The reason for this discrepancy is unclear, but may be due to complex mechanism of post-transcriptional regulation, or potential presence of unknown OCT4 pseudogenes which cause false positive detection by RT-PCR. Therefore, the diagnostic value of OCT4 mRNA in bronchoscopic biopsy specimens requires further investigation.
In addition, we examined the correlation of seven stem cell markers expression in bronchoscopic biopsy specimens of lung cancer with patient clinical features. As we know, poorly differentiated cancers show stronger aggressive and metastatic ability . We found the positive expression rates of Nanog and Bmi1 mRNA was inversely correlated to differentiation of lung cancer, indicating these two markers may be useful to predict tumor progression and poor prognosis in lung cancer. Chiou et al.  reported that Nanog expression in surgically resected lung cancer tissues is an independent prognostic factors of poor prognosis for patients. Vrzalikova and colleagues  also believed that the expression of Bmi1 in surgically resected lung cancer tissues is a prognostic marker in lung cancer. However, surgical resection is not an option for all lung cancer patients, and therefore the use of these markers in bronchoscopic biopsies to predict prognosis would be a great clinical advantage.