Circulating tumor cells in early stage lung adenocarcinoma: a case series report and literature review

Purpose The study aimed to monitor circulating tumor cells (CTCs) in early stage lung adenocarcinoma patients. Results CTCs were characterized and classified to epithelial (E-) CTCs, mesenchymal (M-) CTCs and epithelial- mesenchymal (E&M-) CTCs, as per epithelial-mesenchymal transition(EMT) biomarkers. CTCs could not be found in healthy controls. However, in cohort A, CTCs were found in 17 (17/18) cases. Detection rate of E-CTCs was lower (5/18) compared with M-CTC (10/18) or E&M-CTC (14/18). Highly abundant M-CTCs were prone to being in the tumors > 2 cm. In cohorts A and B, CTCs count increased significantly in all patients with tumor progression (7/7). Higher CTCs level or change range could be found postoperatively in the patients with tumor progression, as compared with patients with disease free survival (P < 0.01). Additionally, CTCs detected by CanPatrolTM could be validated by CytoploRare or Pep@MNPs. Materials and Methods We included four cohorts of patients and 20 healthy controls. In cohort A, CTCs were detected by a newly established approach, i.e., CanPatrolTM, prior to anesthesia and monitored after operation longitudinally. In cohort B, CTCs were not assessed prior to operation, but were longitudinally detected after operation. For validation, we detected FOLR(+)-CTCs by using CytoploRare and EPCAM(+)-CTCs by using Pep@MNPs prior to operation, in cohorts C and D, respectively. Conclusion CTCs can be detected in early stage lung adenocarcinoma, even in adenocarcinoma in situ, and CTCs detection can effectively monitor tumor progression. The distinguishing of biomarkers of highly invasive and aggressive CTCs warrants further robust study.

Circulating tumor cells (CTCs), which shed from the primary tumor into the vasculature or lymphatics, can be regarded as a new prognostic factors of metastatic process
In stage 1A and 1B lung cancer, the detection of CTCs has found to be a sensitive biomarker to predict the prognosis [12,13]. In this study, we identified CTCs in early-stage lung adenocarcinoma patients with CanPatrol TM (Surexam Biotech, Guangzhou, China) [14,15], and explored the subtypes of CTCs as per EMT markers, showing that aberrant activation of EMT could be involved in lung cancer dissemination. Our results showed for the first time that CTCs can be detected in the case with adenocarcinoma in situ (AIS) of lung. Additionally, a longitudinal study was performed to assess the clinical implications of continuous monitoring of CTCs. Furthermore, we validated CanPatrol TM by using other two ways to detect Folate receptor (FOLR) (+)-CTCs [16,17] and EPCAM (+)-CTCs [18], respectively. Finally, we reviewed the published studies regarding the detection of CTCs in a variety of AIS and early stage lung cancer.

RESULTS
CTCs could not be found in 20 healthy controls, but were detectable in 17 (17/18) cases in cohort A, as shown in Table 1. As shown in Figure 1A, RNA-in situ hybridization clearly identified EMT markers in an epithelial (E-) CTC (red dots), a mesenchymal (M-) CTC (green dots) and an epithelial-mesenchymal (E&M-) CTC(red and green dots), respectively. Therefore, these CTCs were classified as E-CTCs, M-CTCs and E&M-CTCs, respectively. In cohort A, E-CTCs, M-CTCs and E&M-CTCs were detected in 5,10 and 14 cases, respectively. There were no statistically significant correlations between the number of CTCs and tumor size, age and gender. Figure 1B demonstrated that highly abundant CTCs (total number of E-CTCs, M-CTCs and E&M-CTCs) were prone to being in the tumors in right side, e.g., case A2(CTCs count = 17), case A10(CTCs count = 14) and case A18(CTCs count = 18), although there was no statistical significance difference between right and left side.
As shown in Figure 1C, we divided cohort A into three groups as per tumor size, i.e., < 2 cm, 2-3 cm and > 3 cm, respectively. Intriguingly, E-CTCs were prone to being in the tumors < 3 cm ( Figure 1C). Furthermore, highly abundant M-CTCs were prone to being in the tumors > 2 cm ( Figure 1C), suggesting these CTCs that have undergone EMT potentially render high risk of metastasis in these cases.
We performed longitudinal studies in 14 cases of cohort A after operation as shown in Figure 2. Figure 2A showed total CTCs in cases A1, A2,A10,A14,A15 and A18 decreased significantly, ranging from one month to one year after operation, respectively. CTCs count seemed to be stable in case A6 [5 (before operation) to 5 (one year after operation)] and in case A16 [0 (before operation) to 0 (one year after operation)]. In case A13, two CTCs prior to operation slightly increased to three CTCs in one year after operation. Follow-up studies demonstrated disease free survival in the abovementioned nine cases.
However, as compared with before operation, cases A5, A11 and A12 had significant increase of CTCs after operation (Figure 2A), while the values of tumor markers, e.g., CEA, CA125 and CA199 were still normal (data not shown). Careful follow up will be performed among these cases in spite of the current absence of definitive evidence regarding tumor progression.
Cases A3 and A8 respectively had a recurrence in six months and one year after the operation, proved by the increased CTCs ( Figure 2B), computed tomography ( Figure 2B) and biopsy.
Longitudinal CTCs monitoring was performed in case B1 as shown in Figure 2C, suggesting CTCs count can effectively monitor tumor activity. CTCs count significantly increased from 0 (three months after operation) to 6 (six months after operation). Docetaxel & cisplatin were given to the patient, and CTCs count gradually decreased to 2 after two cycles of chemotherapy and 1 after four cycles of chemotherapy. However, CTCs count increased to 16 and bone metastasis was proved, six months after finished chemotherapy. Therefore, Pemetrexed & cisplatin plus Pamidronate Disodium were given to the patient, and CTCs count decreased to 8 after two cycles of chemotherapy.

DISCUSSION
CTCs can be found in early stage of tumors, even in a variety of AIS, e.g., breast cancer in situ, melanoma in situ and bladder cancer in situ (Online Table 2). In our study, CTCs were detectable before or after operation in cases A6 and B16 with lung AIS. Interestingly, CTCs were found in 5 COPD patients without clinically detectable lung cancer [19], and all of them developed into cancer in 1 to 4 years, suggesting CTCs may predict the progression of lung cancer in COPD cases [19].
CTCs had been reported to be detectable in early stage NSCLC (Online Table 3) . However, another study demonstrated CTCs test had insufficient capability of discrimination between lung cancer and nonmalignant diseases, although CTC counts were significantly higher in lung cancer patients than in nonmalignant patients [20]. Indeed, both EMT-CTCs and FOLR(+)-CTCs were detectable in a case with tuberculoma as shown in Table 1.
The diagnostic value of CTCs detection could be enhanced when combined with tumor markers, e.g., CEA, Ki-67, CA125, CA199, Cyfra21-1, and SCCA [21]. In our study, both CTCs count and tumor markers remarkably increased after operation in cases B2-B4, prior to appearance of radiographic evidence of tumor progression. Furthermore, more advanced technologies for CTCs enrichment and detection with high specificity and sensitivity warrant further study.
Oncotarget 23136 www.impactjournals.com/oncotarget motility receptor(HMMR), and human telomerase catalytic subunit(hTERT) could be detected in 74.0% of 254 lung cancer patients and present a reliable prognosis value. In addition, Folate receptor-positive CTCs showed a sensitivity of 72.5%-73.2% and a specificity of 84.1%-88.7% in the diagnosis of NSCLC, especially a sensitivity of 67.2% in stage I disease [16,17]. In our study, we tried three methods to detect CTCs in a small series of cases, and found CanPatrol TM can be validated by CytoploRare or Pep@MNPs in lung cancer patients. However, the sensitivity and specificity of CTCs detected by these methods warrant further study.
After surgical removal of tumor mass, some earlystage NSCLC patients will relapse or develop metastases finally. It emphasizes the importance of risk assessment of tumor progression in these patients, and CTCs are considered as the predictive biomarker for guiding cancer treatment strategies [25]. CTCs count was proved to predict the radiation therapy response of NSCLC patients [26] . With regards to the role as a prognostic and predictive factor during chemotherapy, the value of CTCs detection is still inconclusive [27][28][29] . In NSCLC patients, numerous studies concluded that preoperative or postoperative high CTCs count suggested the shorter lifetime of patients than those with low count [11,12,30] and EGFR+ CTCs could predict the high risk of early recurrence after operation [31]. Interestingly, two studies [32,33] used CellSearch ® system to detect CTCs before and after surgery, and found no statistical correlation between EpCAM (+) CTCs count and survival or recurrence, demonstrating EpCAM (-) CTCs or M-CTCs also should be evaluated to monitor tumor progression. In our longitudinal study, CTCs decreased significantly after operation, probably due to the reduced tumor burden, and CTCs significantly increased in all the cases (6/6) of cohorts A and B with tumor relapse or progression, demonstrating CTCs detection can effectively monitor tumor activity.
Theoretically and empirically, CTCs can be erased by anoikis, i.e., a form of programmed cell death that occurs in anchorage-dependent cells when they detach from the surrounding extracellular matrix [34]. As shown in Figure 2A, the number of CTCs in cases A2, A10, A14, A18 before operation was relatively high but decreased remarkably after operation, and the cases were stable after operation. We postulated these CTCs in the above-mentioned cases had been eliminated and formed no tumorous lesion. However, in cases A3 and A8, some critical signaling pathways in CTCs had been activated to help the tumor cells escape from anoikis, leading to tumor recurrence. Indeed, our pilot study had found highly expressed BCAR1, i.e., one of the Crk-associated substrate (cas) protein family members [35] in CTCs in cases A3 and A8(data not shown). Furthermore, our previous studies indicated serum BCAR1 levels were signifi cantly higher in lung cancer compared with the control group, gradually increasing with the progression of tumor staging, and decreasing following malignant lesion removal [36]. In a cohort of 151 Chinese patients with NSCLC, elevated BCAR1 protein expression levels in tumor tissues were shown to predict a poor prognosis [37,38]. In addition, BCAR1 was found to be required for TGF-beta1-mediated EMT in lung cancer, and BCAR1knockdown caused cell migration inhibition and arrest of cell growth and the cell cycle in lung cancer cells [35,37]. Importantly, BCAR1 degradation had been proved to lead to anoikis of cancer cells in vitro [39]. Whether these BCAR1(+)-CTCs are more invasive and aggressive than BCAR1(-)-CTCs warrant further robust study.

CONCLUSION
Collectively, CTCs could not be found in 20 healthy controls, but were preoperatively found in 94% (17/18) cases with lung adenocarcinoma in early stage. E-CTCs,

Clinical-demographical characteristics
From January 2013 to August 2016, four cohorts of patients (From Daping hospital: cohorts A, C and cases B1-B4 and cases B6-B12 in cohort B; From Peking union medical college hospital: cohort D and case B5 and cases B13-B19 in cohort B) and 20 healthy controls (nonsmokers) were enrolled, respectively as shown in Table 1. Informed consent was written by each patient or control who participated this research.
Cohorts A and B included 18 and 19 patients with stage I lung adenocarcinoma, respectively (Table 1). In cohort A, CTCs had been detected prior to anesthesia in all 18 cases. Among them, one patient died of pneumonia and three patients denied re-evaluation of CTCs after operation, hence, CTCs were monitored after operation in 14 cases. In cohort B, CTCs were not assessed prior to operation, but were longitudinally detected after operation.
All the cases, except Case A10 who had CTguided percutaneous lung biopsy prior to operation, did not undergo invasive diagnostic procedure. All the cases underwent video-assisted thoracoscopic surgery(VATS). i.e., lobectomy and lymphadenectomy. The diagnosis was confirmed pathologically in all the cases. The clinical and demographical characteristics were shown in Table 1.

Detection of EMT-CTCs by using CanPatrol TM
Ten ml of blood was collected from the healthy controls and cases, and transferred into sample preservative tubes (Surexam Biotech, Guangzhou, China) containing ammonium chloride-based lysing buffer by a tailored connection device (Surexam Biotech, Guangzhou, China) and incubated at room temperature for 30 min.
CanPatrol TM was used to detect EMT-CTCs, which is a newly established technology to detect CTCs, containing the following steps [14,15]: (1) To remove erythrocytes by red blood cell lysis and deplete CD45+ leukocytes in 10 ml blood sample using a magnetic bead separation method; (2) To enrich CTCs by 8-μm-diameterpore calibrated membrane filters; and (3) To identify and characterize CTCs by using RNA-in situ hybridization (ISH), based on the branched DNA (bDNA) signal amplification technology, to detect EMT markers, e.g., cytokeratins(CK) 8, 18 and 19, epithelial cell adhesion molecule (EpCAM), vimentin and twist.
The details of classification of EMT-CTCs by using CanPatrol TM was depicted in the recently published protocol [40]. Finally, the EMT-CTCs were clustered into three subtypes, as per the EMT markers, i.e., epithelial (E-) CTCs, mesenchymal (M-) CTCs and epithelialmesenchymal (E&M-) CTCs.
Firstly, erythrocytes were removed by red blood cell lysis and CD45+ leukocytes were depleted by using a magnetic bead separation method. Then it was labeled with a conjugate of a tumor-specific ligand folic acid and a synthesized oligonucleotide [16,17]. Thereafter, the CTCs were collected for quantitative PCR analysis. Before amplification, the conjugate first annealed and extended on the reverse transcriptase primer. After immunofluorescence staining the enriched CTCs, FOLR(+)-CTCs were defined as cells expressing folate ligands and cytokeratin and 4′,6-diamidino-2-phenylindole-stained nucleus.
In this study, we used an arbitrarily defined CTC unit, which was defined as the number of CTCs detected in 3 ml of blood. FOLR value > 8.7 FU/3 ml was defined as positive [16,17] .

Isolation of EpCAM(+)-CTCs by using Pep @ MNPs
EpCAM(+)-CTCs were isolated by using the EpCAM recognition peptide functionalized iron oxide magnetic nanoparticles (MNPs) (Pep@MNPs) [18]. 10 μLof the obtained anti-EpCAM@MNPs was added to 1 mL human blood in a 1.5 mL centrifuge tube. After incubation of the mixed suspension in a shaker at 37°C for 30 min (the optimal conditions), the captured cells were gently washed with PBS at least 5 times under a high magnetic field (116 mT). In order to confirm the cell type of the captured cells from blood samples, the commonly used three-color immunocytochemistry method was applied. The captured cell samples were incubated with FITC-labeled anti-CD45 and APC-labeled anti-CK for 30 min respectively and followed by PBS washing at least 3 times. The CTCs confirmation was performed using a fluorescence microscope.