Sage Journals: Discover world-class research

Abstract

BACKGROUND:

Circulating tumor cells (CTCs) is a promising biomarker for cancer prognosis and monitoring. Molecular characterizing of CTCs could provide beneficial information on the basis of CTCs counting.

OBJECTIVE:

To investigate the epithelial-mesenchymal transition (EMT) phenotypes and GALC mRNA expression of CTCs in non-small cell lung cancer (NSCLC) patients.

METHODS:

We analyzed the baseline number, EMT classification, and GALC expression of CTCs in 47 NSCLC patients using CanPatrol platform and RNA in situ hybridization technique.

RESULTS:

CTCs were detected in 91.5% patients ranging 0–47/5 mL blood. Increased CTCs were associated with advanced tumor stages (6/5 mL) compared with early stages (3.5/5 mL). Patients with effective treatment response presented lower CTCs (3.5/5 mL) than patients with insufficient response (7/5 mL). Epithelial, hybrid and mesenchymal CTCs were detected in 55.4%, 78.7% and 61.7% patients, respectively. Patients with distant metastasis and poor curative outcomes presented higher level of EMT-CTCs. GALC expression was positive in CTCs of 80.6% patients and closely correlated with tumor number and distant metastasis and treatment outcomes.

CONCLUSIONS:

EMT phenotypes and GALC expression of CTCs are correlated with cancer metastasis and therapeutic outcomes, suggesting them to be potential markers for the prognosis of NSCLC.

Keywords

CTCs detection cancer metastasis circulating biomarkers therapeutic monitoring and prognosis

1. Introduction

Lung cancer ranks the most common cancer worldwide with 1.82 million newly diagnosed cases a year according to the World Health Organization reports [1]. More than 80% of these cases are non-small cell lung cancers (NSCLC). It is also the most leading cause of cancer-related death in men because of its poor prognosis [2]. Physicians are facing great challenges to diagnose and manage the disease at early stage, as well as to predict and prevent the recurrence and metastasis.

Circulating tumor cells (CTCs) are tumors cells released into the peripheral blood from solid tumors or metastases. They are potential to disseminate to distant organs through the blood system, accounting for a crucial part of cancer metastasis. Recently, the breakthrough of liquid biopsy techniques opens a new era for physician decisions and clinical strategies in cancer management [3]. Compared with traditional tissue biopsy, CTCs test is advantageous for its invasiveness, conveniences and accessibility. Applications of CTCs test have been widely practiced in the clinical management of various solid tumors [4, 5]. In NSCLC, the increasing CTCs numbers has been proved correlated with organ metastasis and early relapse, suggesting that CTCs counting could be an independent predictive factor for the therapy-monitoring [6, 7] and prognosis [8, 9, 10] of NSCLC patients.

However, most of the mainstream CTCs detection methods depend on the recognition of epithelial markers on cancer cells, such as EpCAM and CKs, which would miss some CTCs subpopulation because of epithelial-mesenchymal transition (EMT). EMT plays a crucial role in the multi-step process of cancer metastasis including cell migration, invasion, and microenvironment modulation, as well as the generation and dissemination of CTCs [11]. EMT is characterized by the reduction of epithelial markers and acquisition of mesenchymal markers. Among CTCs isolated by cell size-based methods, 80%–100% CTCs were demonstrated to express mesenchymal protein (Vimentin) [12]. Co-expression of epithelial and mesenchymal markers has been proved to exist in over 80% CTCs from metastatic prostate cancer patients and over 75% CTCs from metastatic breast cancer patients [13]. Therefore, it is vital to combine the epithelial and mesenchymal markers for improving the sensitivity and accuracy of CTCs detection.

Galactocerebrosidase (GALC) is a lysosomal enzyme encoded by GALC gene. GALC catalyzes the hydrolysis of glycosphingolipids including galactosylsphingosine, lactosylceramide and galactocerebroside. GALC participates in the metabolism of sphingolipids and the aberration of GALC could cause severe metabolic disorders such as Krabbe disease [14, 15]. Recently, a few studies reported the dysfunction of GALC in tumors, because the metabolic products of glycosphingolipids such as bioactive lipids and ceramide are crucial structural and signal molecules in regulating cell adhesion, proliferation and apoptosis [16]. Aberrant expression of GALC has been observed in nasopharyngeal carcinoma [17] and glioblastoma [18]. Our previous research also revealed that epigenetic inactivation of GALC might play a role in tumorigenesis of lung and head and neck cancers [19]. Nevertheless, GALC expression in CTCs from NSCLC patients remains unclear.

In this study, we utilized the CanPatrol ${}^{\text{TM}}$ system (Surexam Biotech, Guangzhou, China) [20, 21] to detect CTCs in NSCLC patients and characterized them into three EMT phenotypes including epithelial CTCs, hybrid CTCs and mesenchymal CTCs. The expressions of GALC in individual CTCs were simultaneously investigated by an RNA in situ hybridization technique to explore the correlation of CTCs phenotypes and GALC expression with clinical pathological characteristics of NSCLC patients.

2. Materials and methods

2.1 Patients and samples

A total of 47 NSCLC patients were enrolled from The Third Affiliated Hospital of Southern Medical University (Guangzhou, China) during November 2015 to December 2016. Patients were diagnosed with NSCLC by histopathological examinations. The research protocol was approved by the Ethics Committee of The Third Affiliated Hospital of Southern Medical University. Written informed consents were provided by all the patients. 5 mL anticoagulated peripheral blood was collected for baseline CTCs detection and GALC analysis before any treatment.

2.2 CTCs isolation and EMT classification

CTCs isolation and identification were performed using the CanPatrol ${}^{\text{TM}}$ CTC analysis system (SurExam, Guangzhou, China) as previously described [20]. Br-iefly, blood samples were pretreated with a lysis buffer and centrifugated to remove the red blood cells. Then a filtration membrane (Millipore, billerica, MA, USA) was used to isolate CTCs from the precipitated karyocytes. Next the RNA in situ hybridization assay was performed to characterize and classify CTCs. Pooled capture probes targeting epithelial markers (CK8/18/19 and EpCAM), mesenchymal markers (Vimentin and Twist) and leukocyte markers (CD45) were labeled with different fluorescent dyes. Cell nucleus was stained with the 4’,6-diamidino-2-phenylindole (DAPI; Sigma, St. Louis, USA). Signals were detected with an automated imaging fluorescence microscope (Zeiss, Germany). The positive CTCs result of a sample was defined as CTCs counting $>$ 3/5 mL; otherwise the CTCs counting was negative ( $\leqslant$ 3/5 mL). The isolated CTCs were classified into three types including epithelial, hybrid and mesenchymal types according to the expression of the EMT markers.

Table 1
Capture probe sequences for GALC gene

Gene	Sequences (5’-3’)
GALC	TATCTGAGAACGATAGGGCT
	GGCACCAAAATTCGGCTTAA
	AGTGCATAATGCATGTGGGA
	CACTCGTATCCTCGGAAATA
	CTTAGCTTCTTTCATCAACC
	GTGTAATATTGGGATTCCTC
	CAGGGAATGACCATGGCAAC
	ATAAGGCCAGTCGAAACCTT

2.3 GALC expression analysis

The expression of GALC gene in CTCs was determined with specific capture probes in the RNA in situ hybridization assay. Sequences of GALC probes were synthesized by Invitrogen (Carlsbad, CA, USA) as listed in Table 1, labeled with fluorecent dye Alexa Fluor 647. GALC expression status was analyzed in patients with CTCs counting $\geqslant$ 3/5 mL blood. For the GALC ${}^{+}$ CTCs, the expressions of GALC were characterized into low and high level according to the levels of reference genes (TBP, TFRC and B2M). These genes exhibited different expression abundances (ranging low to high) in CTCs. The fluorescence intensities of them in CTCs were independently calculated in 100 clinical samples, and the corresponding cut-off value for the low and high expression level was defined as the 50 ${}^{\text{th}}$ percentile ( $P_{50}=$ 7.0) of the signal intensities. Thus the GALC expression was defined as low (signal $\leqslant$ 7.0) and high (signal $>$ 7.0) level according to the signal of Alexa Fluor 647.

2.4 Clinical pathological data collection

Except for basic information such as age and gender, some clinical pathological characteristics were recorded for analysis, including histology, differentiation, TNM stage, tumor size (maximum diameter detected by the computerized tomographic scan), tumor number, metastasis status and smoking history. Concentrations of serum tumor markers such as carcino-embryonic antigen (CEA), cytokeratin fragment antigen21-1 (CYFRA21-1) and neuron-specific enolase (NSE) were also detected simultaneously. Curative effect was determined as disease relief or progression according to the Response Evaluation Criteria in Solid Tumors (RECIST).

2.5 Statistical analysis

Figure 1.

EMT phenotypes and GALC expression of CTCs detected by the RNA in situ hybridization in NSCLC patients. (A) The epithelial, hybrid and mesenchymal CTCs with positive expression of epithelial markers (EpCAM and CK8/18/19), biphenotypic markers and mesenchymal markers (Vimentin and Twist). (B) Images of CTCs that had negative, low expression and high expression of GALC gene. (C) GALC expression in the epithelial, hybrid and mesenchymal CTCs. The arrows represent the fluoresence signals of GALC expression (Alexa Fluor 647). Scale bar $=$ 10 $\mu$ m.

Statistical analyses were conducted using SPSS 13.0 (SPSS Inc., Chicago, IL, USA). Data were expressed as medians for discontinuous variables and numbers (percentages) for categorical variables. Comparisons between groups were assessed by Mann-Whitney U test (for two groups) or Kruskal-Wallis test (for three groups). Chi-square test was used to evaluate the correlations between clinical pathological data and CTCs detection or GALC expression. All tests were two tailed, and a $P$ value $<$ 0.05 was considered statistically significant.

3. Results

3.1 Patient characteristics and CTCs counting

The basic characteristics of 47 NSCLC patients were summarized in Table 2, including histology, differentiation, TNM stage, tumor size, tumor number, metastasis status, smoking history and serum tumor markers CEA, CYFRA21-1 and NSE. Patients were treated with surgery or conventional chemotherapy according to pathological and clinical assessment. Assisted targeted therapy was applied if EGFR mutations occurred. Baseline CTCs were detected ( $\geqslant$ 1/5 mL) in 43 out of 47 patients (91.5%) with a median of 5 CTCs/5 mL blood, ranging from 1/5 mL to 47/5 mL. The proportions of patients with positive ( $>$ 3/5 mL) and negative ( $\leqslant$ 3/5 mL) CTCs counting were 68.1% and 31.9%, respectively.

Table 2
Characteristics of patients included in this study ( $n=$ 47)

Factor	Subgroups	n	%
Age	$<$ 60	27	57.4
	$\geqslant$ 60	20	42.6
Gender	Male	35	74.5
	Female	12	25.5
Histology ${}^{\rm a}$	ADC	33	70.2
	SCC	9	19.2
	Others	5	10.6
Differentiation	Poor	18	38.3
	Moderate	29	61.7
TNM stage	I $+$ II	20	42.6
	III $+$ IV	27	57.4
Tumor size (cm)	$\leqslant$ 3	25	53.2
	$>$ 3	22	46.8
Tumor number	1	30	63.8
	$\geqslant$ 2	17	36.2
Metastasis	Yes	16	34.0
	No	31	66.0
Smoking	Yes	35	74.5
	No	12	25.5
CEA (ng/mL)	$\leqslant$ 5	24	51.1
	$>$ 5	21	44.7
	Unknown	2	4.2
CYFRA21-1	$\leqslant$ 3.3	13	27.7
(ng/mL)	$>$ 3.3	24	51.1
	Unknown	10	21.2
NSE (U/mL)	$<$ 12.5	16	34.0
	$\geqslant$ 12.5	23	48.9
	Unknown	8	17.1
Therapy ${}^{\rm b}$	SR	9	19.1
	SR $+$ CHT	32	68.1
	CHT	6	12.8
CTCs count	$<$ 1	4	8.5
(/5 mL)	$\geqslant$ 1	43	91.5

${}^{\rm a}$ ADC: adenocarcinoma, SCC: squamous cell carcinoma. ${}^{\rm b}$ SR: surgical resection, CHT: chemotherapy.

3.2 EMT classification and GALC mRNA expression of CTCs

CTCs were defined as cell size over 8 $\mu$ m and CD45 ${}^{-}$ DAPI ${}^{+}$ cells. The isolated CTCs were identified and classified into three groups according to different fluorescent signals (Fig. 1A). They were epithelial CTCs (epithelial markers ${}^{+}$ ), hybrid CTCs (epithelial markers ${}^{+}$ and mesenchymal markers ${}^{+}$ ) and mesenchymal CTCs (mesenchymal markers ${}^{+}$ ). Epithelial, hybrid and mesenchymal CTCs were detected in 55.4% (26/47 cases), 78.7% (37/47 cases) and 61.7% (29/47 cases) patients, respectively. EMT-CTCs (hybrid and mesenchymal CTCs) were detected in 40 out of 47 patients (85.1%), with a median of 5 CTCs/5 mL blood (ranging from 1/5 mL to 43/5 mL).

We analyzed the expression of GALC gene in patients with CTCs counting $\geqslant$ 3/5 mL blood. The levels of GALC in CTCs were classified as negative, low expression and high expression (Fig. 1B) according to the signal intensity of Alexa Fluor 647. Among the 36 eligible patients, 80.6% patients (29/36 cases) demonstrated positive expression of GALC gene in CTCs. The rate of GALC expression in total CTCs was 48.1% (149/310 CTCs), with the GALC low expression- and high expression-CTCs accounting for 69.1% (103/149 CTCs) and 30.9% (46/149 CTCs). Besides, GALC gene expression was observed in three types of CTCs (Fig. 1C). The rates of GALC expression in epithelial, hybrid and mesenchymal CTCs were 48.6% (17/35 CTCs), 54.2% (96/177 CTCs) and 36.7% (36/98 CTCs), respectively.

3.3 CTCs counting and EMT-CTCs correlate with tumor stage and curative effect in NSCLC patients

Next we evaluated the correlations between CTCs counting or EMT status and clinical pathological characteristics, as shown in Table 3. The CTCs median of advanced stages (III $+$ IV) NSCLC patients was higher than that of early stages (I $+$ II) patients (6 CTCs/5 mL vs 3.5 CTCs/5 mL; $P<$ 0.05), indicating that higher CTCs number was associated with advanced tumor stages (Fig. 2A). Patients with effective response to treatment presented lower CTCs numbers compared with patients presenting insufficient response (3.5 CTCs/5 mL vs 7 CTCs/5 mL; $P<$ 0.05) (Fig. 2B). The assessment of curative effect was conducted 3 months after treatment according to the RECIST guideline. Disease relief or progression was then judged to reflect effective or ineffective response (Fig. 2C and D).

Figure 2.

Correlations between CTCs counting or EMT-CTCs and clinical pathological characteristics in NSCLC patients. (A–B) The CTCs number in patients with different TNM stages (A) and curative effects (B). (C–D) Computed Tomography images for disease relieved patients (C) or disease progressed patients (D) before therapy and after three months of therapy. (E–F) The rates of CTCs positive ( $>$ 3/5 mL) and negative ( $\leqslant$ 3/5 mL) result in patients with different metastatic (E) and smoking (F) statuses. (G–H) The rates of EMT-CTCs positive ( $>$ 3/5 mL) and negative ( $\leqslant$ 3/5 mL) result in patients with different metastatic statuses (G) and curative effects (F). ${}^{*}P<$ 0.05.

Furthermore, we assessed the positive rates of CTCs counting in different clinicopathological subgroups (Table 3). In consistent with CTCs numbers, higher CTCs positive ( $>$ 3/5 mL) rates were correlated with advanced tumor stages and insufficient therapeutic response. Noticeably, although the difference of CTCs median between metastasis subgroups or smoking subgroups was not statistically significant, CTCs positive rate was obviously higher in metastatic NSCLC patients (87.5%) than that of non-metastatic patients (65.0%) ( $P<$ 0.05; Fig. 2E), and was higher in smoking patients (77.1%) than that of non-smoking patients (41.7%) ( $P<$ 0.05; Fig. 2F). These results demonstrated that except for TNM stage and curative effect, CTCs counting also had a correlation with metastasis status and smoking history in NSCLC patients.

Table 3

Correlations between CTCs counting or EMT-CTCs and clinical data of NSCLC patients ( $n=$ 47)

Factor	Subgroups	CTCs	$P$	CTCs ${}^{\rm a}$	$P$	EMT-CTCs ${}^{\rm a}$	$P$
		median		$+/-$		$+/-$
Age	$<$ 60	6	0.300	19/8	0.696	18/9	0.250
	$\geqslant$ 60	4.5		13/7		10/10
Gender	Male	6	0.070	26/9	0.119	23/12	0.143
	Female	3		6/6		5/7
Histology ${}^{\rm b}$	ADC	5	0.753	22/11	0.366	20/13	0.784
	SCC	6		6/3		5/4
	Others	6		4/1		3/2
Differentiation	Poor	6.5	0.282	13/5	0.632	12/6	0.435
	Moderate	5		19/10		16/13
TNM stage	I $+$ II	3.5	0.013 ${}^{\ast}$	10/10	0.022 ${}^{\ast}$	9/11	0.080
	III $+$ IV	6		22/5		19/8
Tumor size (cm)	$\leqslant$ 3	4	0.054	15/10	0.205	13/12	0.259
	$>$ 3	6		17/5		15/7
Tumor number	1	4.5	0.213	18/12	0.114	16/14	0.247
	$\geqslant$ 2	6		14/3		12/5
Metastasis	Yes	6	0.075	14/2	0.040 ${}^{\ast}$	13/3	0.030 ${}^{\ast}$
	No	4		18/13		15/16
Smoking	Yes	6	0.053	27/8	0.023 ${}^{\ast}$	23/12	0.143
	No	2.5		5/7		5/7
CEA (ng/mL)	$\leqslant$ 5	6	0.278	17/7	0.526	15/9	0.884
	$>$ 5	4		13/8		12/9
	Unknown	5		2/0		1/1
CYFRA21-1	$\leqslant$ 3.3	3	0.276	6/7	0.225	6/7	0.478
(ng/mL)	$>$ 3.3	5		16/8		14/10
	Unknown	6		10/0		8/2
NSE (U/mL)	$<$ 12.5	4.5	0.315	10/6	0.918	9/7	0.981
	$\geqslant$ 12.5	5		14/9		13/10
	Unknown	6.5		8/0		6/2
Curative effect	Relief	3.5	0.001 ${}^{\ast}$	15/15	0.001 ${}^{\ast}$	12/18	0.001 ${}^{\ast}$
	Progression	7		17/0		16/1

${}^{\rm a}$ The positive/negative ( $+/-$ ) criteria of CTCs and EMT-CTCs are $>$ 3/5 mL and $\leqslant$ 3/5 mL. ${}^{\rm b}$ ADC: adenocarcinoma, SCC: squamous cell carcinoma; ${}^{*}P<$ 0.05.

EMT plays a vital role in metastasis and malignant progression of cancers. Therefore, we investigated the distribution of EMT phenotypes in CTCs from NSCLC patients. Correlations between EMT-CTCs (hybrid and mesenchymal CTCs) and clinical pathological data were evaluated. Positive EMT-CTCs ( $>$ 3 CTCs/5 mL) were associated with tumor metastasis and curative effect of NSCLC patients (Table 3). Metastatic patients had remarkable higher EMT-CTCs compared to non-metastatic patients, with a positive rate of 81.3% and 41.9% respectively ( $P<$ 0.05; Fig. 2G). EMT-CTCs positive rate was as high as 94.1% in patients with unfavorable therapeutic effect, while it was much lower (40.0%) in patients with effective response ( $P<$ 0.05; Fig. 2H).

No significant correlation was found between CTCs counting or EMT-CTCs and other clinical characteristics such as age, gender, histology, differentiation, tumor size, tumor number and serum biomarkers CEA, CYFRA21-1 and NSE.

3.4 GALC expression in CTCs correlates with metastasis and prognosis of NSCLC patients

CTCs number was $\geqslant$ 3/5 mL in 36 out of 47 NSCLC patients. We investigated the expression of GALC gene in CTCs from these 36 patients, as shown in Table 4. GALC expression in CTCs was closely correlated with distant metastasis and tumor number in NSCLC patients. GALC expression was detected in 100% of metastatic patients and 68.2% of non-metastatic patients (Fig. 3A; $P<$ 0.05). Similarly, patients with multiple tumors had higher GALC expressing rate than patients with single tumor (100% vs 68.2%) (Fig. 3B; $P<$ 0.05).

Table 4
Correlations between GALC expression and clinical pathological data in patients with CTCs $\geqslant$ 3/5 mL blood ( $n=$ 36)

Factor	n	GALC ${}^{+}$	$P$	GALC expression level
				Low	$P$	High	$P$
Age
$<$ 60/ $\geqslant$ 60	21/15	18/11	0.618	18/9	0.079	6/7	0.265
Gender
Male/Female	30/6	24/5	0.851	23/4	0.606	12/1	0.277
Histology ${}^{\rm a}$
ADC/SCC/Others	26/6/4	21/5/3	0.947	19/5/3	0.872	10/3/0	0.244
Differentiation
Poor/Moderate	14/22	11/18	0.810	11/16	0.693	5/8	0.968
TNM stage
I $+$ II/III $+$ IV	14/22	9/20	0.125	8/19	0.114	4/9	0.452
Tumor size (cm)
$\leqslant$ 3/ $>$ 3	20/16	15/14	0.346	13/14	0.121	8/5	0.587s
Tumor number
1/ $\geqslant$ 2	22/14	15/14	0.029 ${}^{\ast}$	14/13	0.048 ${}^{\ast}$	9/4	0.452
Metastasis
Yes/No	14/22	14/15	0.029 ${}^{\ast}$	12/15	0.236	5/8	0.968
Smoking
Yes/No	30/6	23/6	0.317	22/5	0.606	11/2	0.877
CEA (ng/mL)
$\leqslant$ 5/ $>$ 5/Unknown	18/16/2	15/13/1	0.526	15/11/1	0.435	7/6/0	0.548
CYFRA21-1 (ng/mL)
$\leqslant$ 3.3/ $>$ 3.3/Unknown	7/19/10	6/16/7	0.609	6/14/7	0.748	4/5/4	0.333
NSE (U/mL)
$<$ 12.5/ $\geqslant$ 12.5/Unknown	11/17/8	9/14/6	0.903	8/13/6	0.975	6/4/3	0.247
Curative effect
Relief/Progression	19/17	13/16	0.128	13/14	0.335	4/9	0.047 ${}^{\ast}$

${}^{\rm a}$ ADC: adenocarcinoma, SCC: squamous cell carcinoma; ${}^{*}P<$ 0.05.

Figure 3.

Correlations between the GALC expression status of CTCs and clinical pathological characteristics in NSCLC patients. (A–B) The detection rate of GALC ${}^{+}$ CTCs in patients with metastatic and non-metastatic tumor (A) and in patients with single and multiple tumors (B). (C–D) The detection rate (C) and the number (D) of CTCs that expressed high level GALC in patients with effective response (relief) and poor response (progression) to the therapy. ${}^{*}P<$ 0.05.

Moreover, the expression level of GALC in each CTC was defined as low or high level (Fig. 1B) for further analysis of their clinical relevance. High level of GALC expression was evidently associated with curative effect of NSCLC patients (Table 4; $P<$ 0.05). Both of the detection rate and number of CTCs that had high level GALC expression were higher in the patients with progression than that in patients with effective response to the therapy (Fig. 3C and D). There was no significant correlation between GALC expression of CTCs and other clinical characteristics such as age, gender, histology, differentiation, tumor size, TNM stage, smoking and serum biomarkers CEA, CYFRA21-1 and NSE (Table 4). These results indicated that GALC expression in CTCs might assist the diagnosis of tumor metastasis and the prediction of prognosis in NSCLC patients.

4. Discussion

CTCs test, as a liquid biopsy, has brought remarkable revolutions in the mechanism research and clinical management of metastatic cancers. The epithelial markers-based methods for CTCs detection might not be satisfactory due to the phenomenon of EMT during the dissemination of CTCs [22]. In the process of EMT, the loss of epithelial markers such as EpCAM and E-cadherin, along with the expression of mesenchymal markers such as Vimentin and N-cadherin, enables tumor cells to migrate and invade through the basilar membrane and enter the blood vessels [23]. The dynamic change of EMT could result in heterogeneous phenotypes of CTCs which may express epithelial markers, mesenchymal markers, or hybrid of two types of markers. Co-expression of epithelial and mesenchymal markers has been reported in CTCs from patients with breast, prostate and colorectal cancer [13, 24]. Missing of hybrid and mesenchymal CTCs may reduce the detection accuracy, which could cause the results bias and limit the clinical applications of CTCs analysis.

Here we utilized the CanPatrol ${}^{\text{TM}}$ system to analyze CTCs in NSCLC patients. It combined the tumor size-based filtration with the RNA in situ hybridization-based identification to isolate and characterize CTCs from the blood samples. While the sensitivity of EpCAM and/or CKs based CTCs detecting techniques (such as the CellSearch system) was reported to be 30.5%–42.6% [25, 26], we found CTCs were detectable in 91.5% NSCLC patients in this study. Epithelial, hybrid and mesenchymal CTCs were detected in 55.4%, 78.7% and 61.7% patients, respectively. EMT-CTCs (hybrid and mesenchymal CTCs) were positive in 85.1% patients with a median of 5 CTCs/5 mL blood. These results demonstrated that EMT-CTCs existed in most NSCLC patients and accounted for a large subpopulation of total CTCs. Lecharpentier et al. [27] also found that almost all the CTCs isolated by ISET (isolation by size of epithelial tumor cells) from lung cancer patients expressed EMT markers and some patients had only mesenchymal CTCs without epithelial CTCs. Besides, the detection rates of hybrid and mesenchymal CTCs were found higher than epithelial CTCs even if in early stage lung adenocarcinoma patients [28]. Therefore, the combination detection of epithelial CTCs and EMT-CTCs could contribute to improving the sensitivity and accuracy of CTCs test.

Previous studies have reported the significant value of CTCs counting in predicting treatment response and detecting early recurrence of NSCLC patients [6, 8]. In line with these results, we found that the total CTCs number and CTCs positive rate were correlated with cancer metastasis and curative effect of NSCLC patients. Therapeutic responses were favorable in patients with average CTCs counting of 3.5/5 mL but unsatisfactory in patients with average CTCs of 7/5 mL. Noticeably, some previous studies suggested that CTCs number did not correlate with the clinical stage of NSCLC [29, 30], which is contrary to our results. The variety of CTCs detection methods and diversity of targeted CTCs markers should account the most for this distinctness. Using CK19/TTF-1-based detecting method [29] and fluid/high definition imaging-based system [30] may achieve CTCs populations different from our study, which would result in diverse CTCs detection efficiencies and conflicting correlations with tumor stage. Therefore, CTCs test by the combined EMT markers might assist the tumor staging system at the diagnosis of lung cancer. Besides, the positive rates of CTCs counting were significantly higher in smoking patients compared with that in non-smoking patients. Inhalation of smoke, including exposure to passive smoking, can cause mutations of oncogenes and tumor suppressor genes such as TP53 and KRAS [31, 32]. These changes play a vital role in the drive of carcinogenesis and the prediction of prognosis in lung cancer patients.

In addition, EMT-CTCs might be more relevant to metastasis and progression of cancer than epithelial CTCs. Our results showed that the EMT-CTCs were remarkably associated with high risk of metastasis and poor curative response, suggesting them to be potential prognostic biomarkers. Studies of Cabel et al. [4] and Togo et al. [33] also demonstrated that the levels of hybrid and mesenchymal CTCs had negative correlations with the chemotherapeutic response, clinical outcomes, and overall survival. In an investigation of patient CTCs-derived eXplant (CDX) on immunocompromised mice, the researchers found that the absence of CellSearch detectable CTCs (EpCAM ${}^{+}$ ) did not hinder CDX generation [34], indicating the key roles of EMT-CTCs in NSCLC dissemination. EMT transcription factors such as Slug and Snail could activate multiple signaling pathways to protect CTCs from anoikis during the dissemination, thus promoting the formation of organ metastases [35]. In reverse, suppression of EMT could reduce the generation of CTCs and the appearance of distant metastasis [36]. These results support the vital roles of EMT in the generation and function of CTCs, and further promotion of cancer metastasis.

Furthermore, we analyzed the expression of GALC in CTCs from NSCLC patients for the first time. The downregulation or dysfunction of GALC could lead to the overexpression of glycosphingolipids in cancer cells, which could further result in changed cell surface antigenic characteristics, reduced cellular adhesion and strengthened cell motility [15]. In our previous investigation on the gene expression and DNA sequencing of lung cancer cell lines, we found that epigenetic inactivation of GALC might play a role in tumorigenesis of lung cancer [19]. In the present work, GALC expression in CTCs was proved to be associated with the tumor number and metastasis of NSCLC. Higher expression level of GALC in CTCs also indicated poor outcomes of therapy compared with lower GALC levels. These results suggested that GALC could be an assistant predictor of tumor progression and prognosis in NSCLC patients. The involved mechanism might be related to the impact of GALC gene on crucial transport factors. GALC promoter has been demonstrated to possess specific signal sequences susceptible to some transcription factors such as SP1, AP2 and YY1, which could consequently cause expression changes of key cancer modulating genes such as p53 [37, 38]. Nonetheless, further studies are in need to illuminate the function and mechanism of GALC in the development of NSCLC.

In conclusion, the present study simultaneously analyzed the EMT phenotypes and GALC mRNA expression of CTCs in NSCLC patients using an RNA in situ hybridization technique. The results demonstrated that the combination of EMT markers could detect CTCs in more than 90% NSCLC patients. Along with the total CTCs number, the level of EMT-CTCs and GALC expression in CTCs were closely correlated with the tumor stage, distant metastasis and curative effect of NSCLC, which indicated that they could be assistant markers for the monitoring and prognosis of these patients. The characterization of EMT phenotypes and GALC expression in CTCs should provide further insight on the role of CTCs in NSCLC metastasis, yet extra investigations on the molecular mechanisms and expanded sample size-based verification are needed in the future to promote the applications of CTCs analysis in cancer management and clinical practice.

Footnotes

Acknowledgments

This study is partially supported by the grant of National Natural Science Foundation of China (Grant No. 81301609).

Conflict of interest

The authors declare that they have no conflict of interest.

Abbreviations

References

World cancer report 2014. World health organization. http://publications.iarc.fr/Non-Series-Publications/World-Cancer-Reports/World-Cancer-Report-2014.

Chen

Zheng

Baade

P.D.

Zhang

Zeng

Bray

Jemal

X.Q.

and He

, Cancer statistics in China, 2015, CA: A Cancer Journal for Clinicians 66 (2016), 115–132.

Wang

Chang

and Sun

, Application of liquid biopsy in precision medicine: opportunities and challenges, Frontiers of Medicine 11 (2017), 522–527.

Cabel

Proudhon

Gortais

Loirat

Coussy

Pierga

J.Y.

and Bidard

F.C.

, Circulating tumor cells: clinical validity and utility, International Journal of Clinical Oncology 22 (2017), 421–430.

Chinen

L.T.D.

Abdallah

E.A.

Braun

A.C.

Flores

Corassa

Sanches

S.M.

and Fanelli

M.F.

, Circulating Tumor Cells as Cancer Biomarkers in the Clinic, Advances in Experimental Medicine and Biology 994 (2017), 1–41.

Zhou

Dong

Cui

and Tang

, The role of circulating tumor cells in evaluation of prognosis and treatment response in advanced non-small-cell lung cancer, Cancer Chemotherapy and Pharmacology 79 (2017), 825–833.

Punnoose

E.A.

Atwal

Liu

Raja

Fine

B.M.

Hughes

B.G.

Hicks

R.J.

Hampton

G.M.

Amler

L.C.

Pirzkall

and Lackner

M.R.

, Evaluation of circulating tumor cells and circulating tumor DNA in non-small cell lung cancer: association with clinical endpoints in a phase II clinical trial of pertuzumab and erlotinib, Clinical Cancer Research: An Official Journal of the American Association for Cancer Research 18 (2012), 2391–2401.

Milaki

Messaritakis

Koinis

Kotsakis

Apostolaki

Dermitzaki

E.K.

Perraki

Hatzidaki

and Georgoulias

, Prognostic value of chemotherapy-resistant CK19mRNA-positive circulating tumor cells in patients with advanced/metastatic non-small cell lung cancer, Cancer Chemotherapy and Pharmacology 80 (2017), 101–108.

Funaki

Sawabata

Nakagiri

Shintani

Inoue

Kadota

Minami

and Okumura

, Novel approach for detection of isolated tumor cells in pulmonary vein using negative selection method: morphological classification and clinical implications, European Journal of Cardio-Thoracic Surgery: Official Journal of the European Association for Cardio-thoracic Surgery 40 (2011), 322–327.

10.

Kim

M.Y.

Oskarsson

Acharyya

Nguyen

D.X.

Zhang

X.H.

Norton

and Massague

, Tumor self-seeding by circulating cancer cells, Cell 139 (2009), 1315–1326.

11.

Bardia

Wittner

B.S.

Stott

S.L.

Smas

M.E.

Ting

D.T.

Isakoff

S.J.

Ciciliano

J.C.

Wells

M.N.

Shah

A.M.

Concannon

K.F.

Donaldson

M.C.

Sequist

L.V.

Brachtel

Sgroi

Baselga

Ramaswamy

Toner

Haber

D.A.

and Maheswaran

, Circulating breast tumor cells exhibit dynamic changes in epithelial and mesenchymal composition, Science (New York, NY) 339 (2013), 580–584.

12.

Harouaka

R.A.

Zhou

M.D.

Yeh

Y.T.

Khan

W.J.

Das

Liu

Christ

C.C.

Dicker

D.T.

Baney

T.S.

Kaifi

J.T.

Belani

C.P.

Truica

C.I.

El-Deiry

W.S.

Allerton

J.P.

and Zheng

S.Y.

, Flexible micro spring array device for high-throughput enrichment of viable circulating tumor cells, Clinical Chemistry 60 (2014), 323–333.

13.

Armstrong

A.J.

Marengo

M.S.

Oltean

Kemeny

Bitting

R.L.

Turnbull

J.D.

Herold

C.I.

Marcom

P.K.

George

D.J.

and Garcia-Blanco

M.A.

, Circulating tumor cells from patients with advanced prostate and breast cancer display both epithelial and mesenchymal markers, Molecular Cancer Research: MCR 9 (2011), 997–1007.

14.

Hill

C.H.

Graham

S.C.

Read

R.J.

and Deane

J.E.

, Structural snapshots illustrate the catalytic cycle of beta-galactocerebrosidase, the defective enzyme in Krabbe disease, Proceedings of the National Academy of Sciences of the United States of America 110 (2013), 20479–20484.

15.

Spratley

S.J.

Hill

C.H.

Viuff

A.H.

Edgar

J.R.

Skjodt

and Deane

J.E.

, Molecular Mechanisms of Disease Pathogenesis Differ in Krabbe Disease Variants, Traffic (Copenhagen, Denmark) 17 (2016), 908–922.

16.

Yang

Sun

and Duerksen-Hughes

P.J.

, Ceramide and other sphingolipids in cellular responses, Cell Biochemistry and Biophysics 40 (2004), 323–350.

17.

Zhao

Guo

Wang

Xiao

Luo

Jing

Sun

Chen

and Zeng

, GALC gene is downregulated by promoter hypermethylation in Epstein-Barr virus-associated nasopharyngeal carcinoma, Oncology Reports 34 (2015), 1369–1378.

18.

Ling

G.Q.

Liu

Y.J.

Y.Q.

Chen

Jiang

X.D.

Jiang

C.L.

and Ye

, All-trans retinoic acid impairs the vasculogenic mimicry formation ability of U87 stem-like cells through promoting differentiation, Molecular Medicine Reports 12 (2015), 165–172.

19.

Peng

Chen

Shen

Deng

Liu

Xie

Gan

Huang

and Chen

, DNA promoter hypermethylation contributes to down-regulation of galactocerebrosidase gene in lung and head and neck cancers, International Journal of Clinical and Experimental Pathology 8 (2015), 11042–11050.

20.

Liu

Huang

Yang

Deng

Yang

and Xu

, Classification of circulating tumor cells by epithelial-mesenchymal transition markers, PloS One 10 (2015), e0123976.

21.

Liu

Y.K.

B.S.

Z.L.

and Lu

L.G.

, An improved strategy to detect the epithelial-mesenchymal transition process in circulating tumor cells in hepatocellular carcinoma patients, Hepatology International 10 (2016), 640–646.

22.

Alix-Panabieres

Mader

and Pantel

, Epithelial-mesenchymal plasticity in circulating tumor cells, Journal of Molecular Medicine (Berlin, Germany) 95 (2017), 133–142.

23.

Thiery

J.P.

, Epithelial-mesenchymal transitions in tumour progression, Nature Reviews Cancer 2 (2002), 442–454.

24.

Zhao

Cai

Cheng

Gao

Liu

Dong

Zheng

Zhang

and Yao

, Expression and clinical relevance of epithelial and mesenchymal markers in circulating tumor cells from colorectal cancer, Oncotarget 8 (2017), 9293–9302.

25.

Schulze

Gasch

Staufer

Nashan

Lohse

A.W.

Pantel

Riethdorf

and Wege

, Presence of EpCAM-positive circulating tumor cells as biomarker for systemic disease strongly correlates to survival in patients with hepatocellular carcinoma, International Journal of Cancer 133 (2013), 2165–2171.

26.

Guo

Yang

X.R.

Sun

Y.F.

Shen

M.N.

X.L.

Zhang

C.Y.

Zhou

Zhang

Zhou

and Fan

, Clinical significance of EpCAM mRNA-positive circulating tumor cells in hepatocellular carcinoma by an optimized negative enrichment and qRT-PCR-based platform, Clinical Cancer Research: An Official Journal of the American Association for Cancer Research 20 (2014), 4794–4805.

27.

Lecharpentier

Vielh

Perez-Moreno

Planchard

Soria

J.C.

and Farace

, Detection of circulating tumour cells with a hybrid (epithelial/mesenchymal) phenotype in patients with metastatic non-small cell lung cancer, British Journal of Cancer 105 (2011), 1338–1341.

28.

Jin

X.R.

Zhu

L.Y.

Qian

Feng

Y.G.

Zhou

J.H.

Wang

R.W.

Bai

Deng

Liang

and Tan

Q.Y.

, Circulating tumor cells in early stage lung adenocarcinoma: a case series report and literature review, Oncotarget 8 (2017), 23130–23141.

29.

Yoon

S.O.

Kim

Y.T.

Jung

K.C.

Jeon

Y.K.

Kim

B.H.

and Kim

C.W.

, TTF-1 mRNA-positive circulating tumor cells in the peripheral blood predict poor prognosis in surgically resected non-small cell lung cancer patients, Lung Cancer (Amsterdam, Netherlands) 71 (2011), 209–216.

30.

Wendel

Bazhenova

Boshuizen

Kolatkar

Honnatti

Cho

E.H.

Marrinucci

Sandhu

Perricone

Thistlethwaite

Bethel

Nieva

Heuvel

and Kuhn

, Fluid biopsy for circulating tumor cell identification in patients with early-and late-stage non-small cell lung cancer: a glimpse into lung cancer biology, Physical Biology 9 (2012), 016005.

31.

Kondo

Tsuzuki

Sasa

Sumitomo

Uyama

and Monden

, A dose-response relationship between the frequency of p53 mutations and tobacco consumption in lung cancer patients, Journal of Surgical Oncology 61 (1996), 20–26.

32.

Zhang

Jiang

Liang

Chen

and He

, Chromosome 15q25 (CHRNA3-CHRNB4) Variation Indirectly Impacts Lung Cancer Risk in Chinese Males, PloS One 11 (2016), e0149946.

33.

Togo

Katagiri

Namba

Tulafu

Nagahama

Kadoya

Takamochi

Suzuki

Sakurai

Mizuguchi

Urata

and Takahashi

, Sensitive detection of viable circulating tumor cells using a novel conditionally telomerase-selective replicating adenovirus in non-small cell lung cancer patients, Oncotarget 8 (2017), 34884–34895.

34.

Morrow

C.J.

Trapani

Metcalf

R.L.

Bertolini

Hodgkinson

C.L.

Khandelwal

Kelly

Galvin

Carter

Simpson

K.L.

Williamson

Wirth

Simms

Frankliln

Frese

K.K.

Rothwell

D.G.

Nonaka

Miller

C.J.

Brady

Blackhall

F.H.

and Dive

, Tumourigenic non-small-cell lung cancer mesenchymal circulating tumour cells: a clinical case study, Annals of Oncology: Official Journal of the European Society for Medical Oncology/ESMO 27 (2016), 1155–1160.

35.

Kajita

McClinic

K.N.

and Wade

P.A.

, Aberrant expression of the transcription factors snail and slug alters the response to genotoxic stress, Molecular and Cellular Biology 24 (2004), 7559–7566.

36.

Javaid

Zhang

Smolen

G.A.

Wittner

B.S.

Singh

Arora

K.S.

Madden

M.W.

Desai

Zubrowski

M.J.

Schott

B.J.

Ting

D.T.

Stott

S.L.

Toner

Maheswaran

Shioda

Ramaswamy

and Haber

D.A.

, MAPK7 Regulates EMT Features and Modulates the Generation of CTCs, Molecular Cancer Research: MCR 13 (2015), 934–943.

37.

Beier

U.H.

and Gorogh

, Implications of galactocerebrosidase and galactosylcerebroside metabolism in cancer cells, International Journal of Cancer 115 (2005), 6–10.

38.

Sui

Affar el

Shi

Brignone

Wall

N.R.

Yin

Donohoe

Luke

M.P.

Calvo

Grossman

S.R.

and Shi

, Yin Yang 1 is a negative regulator of p53, Cell 117 (2004), 859–872.

Epithelial-mesenchymal transition and GALC expression of circulating tumor cells indicate metastasis and poor prognosis in non-small cell lung cancer

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Patients and samples

2.2 CTCs isolation and EMT classification

Table 1 Capture probe sequences for GALC gene

2.4 Clinical pathological data collection

2.5 Statistical analysis

3.1 Patient characteristics and CTCs counting

Table 2 Characteristics of patients included in this study ( n = 47)

3.3 CTCs counting and EMT-CTCs correlate with tumor stage and curative effect in NSCLC patients

Table 4 Correlations between GALC expression and clinical pathological data in patients with CTCs ⩾ 3/5 mL blood ( n = 36)

Footnotes

Acknowledgments

Conflict of interest

Abbreviations

References

Table 1
Capture probe sequences for GALC gene

Table 2
Characteristics of patients included in this study ( $n=$ 47)

Table 4
Correlations between GALC expression and clinical pathological data in patients with CTCs $\geqslant$ 3/5 mL blood ( $n=$ 36)