Lunx Is a Superior Molecular Marker for Detection of Non-Small Lung Cell Cancer in Peripheral Blood

Abstract

The clinical management of non-small cell lung cancer (NSCLC) would benefit greatly by a test that was able to detect small amounts of NSCLC in the peripheral blood. In this report, we used a novel strategy to enrich tumor cells from the peripheral blood of 24 stage I to IV NSCLC patients and determined expression levels for six cancer-associated genes (lunx, muc1, KS1/4, CEA, CK19, and PSE). Using thresholds established at three standard deviations above the mean observed in 15 normal controls, we observed that lunx (10 of 24, 42%), muc1 (5 of 24, 21%), and CK19 (5 of 24, 21%) were overexpressed in 14 of 24 (58%) peripheral blood samples obtained from NSCLC patients. Patients who overexpressed either KS1/4 (n = 2) or PSE (n = 1) also overexpressed either lunx or muc1. Of patients with presumed curable and resectable stage I to II disease (n = 7), at least one marker was overexpressed in three (43%) patients. In advanced stage III to IV patients (n = 17), at least one marker was overexpressed in 11 patients (65%). These results provide evidence that circulating tumor cells can be detected in NSCLC patients by a high throughput molecular technique. Further studies are needed to determine the clinical relevance of gene overexpression.

Non-small cell lung carcinoma (NSCLC) is the most common cause of cancer-related death among men and women in the United States. Standard therapies for patients with NSCLC include surgery, chemotherapy, and radiation therapy, and the stage of disease dictates the choice of therapy. The current staging system for lung cancer uses the American Joint Committee on Cancer (AJCC) TNM system, and its goal is to classify patients into groups based on the extent of disease. This system relies heavily on the pathological evaluation of the primary tumor (T), regional nodes (N), and distant metastases (M). Patients with early stage disease (no evidence of disease cancer beyond the lung, stage I or II), are considered candidates for surgery for cure with an operation. In contrast, patients with evidence of disease in mediastinal lymph nodes or distant organs (stage III or IV) are not considered candidates for surgery. Thus, surgery is considered primary therapy for patients with early stage disease, while non-surgical therapy is recommended for patients with advanced stage disease. 1 Reliable detection of metastatic disease is therefore critical for appropriate staging and treatment of patients with NSCLC.

Non-small cell lung cancer metastasizes through both lymphatic and hematogenous routes. The current staging evaluation addresses these possibilities by defining stage based on the presence or absence of disease in lymph nodes (hilar and mediastinal) and distant organs (principally bone, brain, adrenal glands, and liver). 2 Following a thorough evaluation, approximately 35% of patients will be considered to have early stage disease and will be candidates for curative lung resection. Unfortunately, 50% of these patients will develop metastases within 5 years of surgery and die from their disease. Further, even in patients with the very earliest stage of disease (pathologically confirmed stage I), the 5-year survival is only 75%. 3 These statistics indicate that conventional staging techniques lack the sensitivity necessary to identify patients who are unlikely to benefit from surgery. 4 Indeed, recent studies of sensitive assays for detection of rare disseminated lung cancer cells have shown that detection of occult metastases by immunocytology or immunohistochemistry in lymph nodes or bone marrow aspirates is a reliable predictor of poor prognosis. 5^, 6^, 7 However, the detection of circulating tumor cells in peripheral blood has remained a challenge.

Reverse transcriptase polymerase chain reaction (RT-PCR) has been recognized as the method with the highest sensitivity for detection of micrometastatic disease, allowing identification of one cancer cell in 10⁶ to 10⁷ normal cells. 8^, 9^, 10^, 11^, 12^, 13^, 14^, 15^, 16 Detection of metastatic cancer cells by RT-PCR is based on the fact that cancer cells continue to express genes (or markers) that are specific to the tissue from which they originate, but are not expressed in tissue compartments that frequently harbor metastatic foci, such as lymph nodes and bone marrow. 17^, 18 Although RT-PCR-based tumor cell detection assays often yield higher sensitivity than conventional immunohistochemistry, 4^, 17^, 19^, 20 such assays for lung cancer have been limited by the availability of molecular markers. 21 However, two genes have recently shown promise by virtue of the fact that they are overexpressed in metastatic mediastinal lymph nodes of NSCLC patients: 22 lunx, 21 (also known as palate, lung, and nasal epithelium carcinoma-associated (Plunc) gene 23^, 24^, 25 ) and the epithelial carcinoma-associated gene KS1/4. 26

In addition to aspects of NSCLC marker identification, advances have also been made in tumor cell enrichment. A porous barrier density gradient (PBDG) centrifugation system has recently been developed that results in reliable recovery of tumor cells from peripheral blood with a concomitant >500-fold depletion of white blood cells. 27 This new tumor cell enrichment system can be combined with real-time RT-PCR, a high throughput technology that allows for sensitive detection and quantitation of gene expression. Real-time RT-PCR measures the on-line fluorescence of products as they are amplified from one cycle to the next and is quickly becoming recognized as the technology of choice for the precise measurement of gene expression levels. 8^, 28 In this report, we combined real-time RT-PCR with PBDG centrifugation and show that of the five NSCLC-associated markers examined, the one with the highest sensitivity was lunx.

Materials and Methods

Oligonucleotides

All primers were designed using Primer Express software (ABI, Foster City, CA), spanned at least one intron, and failed to amplify negative control cDNA in which reverse transcriptase enzyme was omitted. Sequences of the internal control β₂-microglobulin primers were: 5` GCCGTGTGAACCATGTGA (forward) and 5` CCAAATGCGGCATCTTCA (reverse). Other sequences (previously described 22 ) were: lunx, CCCTGGAAGCCTGCAAATT (F) GAACCAACTCAGGCAGGACTTT (R); KS1/4, CGCAGCTCAGGAAGAATGTG (F), TGAAGTACACTGGCATTGACGA(R); CK19, CATGAAAGCTGCCTTGGAAGA (F), TGATTCTGCCGCTCACTATCAG (R); CEA, GGGCCACTGTCGCATCATGATTGG (F), TGTAGCTGTTGCAAATGCTTTAAGAAAGAAGC (R); PSE, AGTGCTCAAGGACATCGAGACG (F), AGCCACTTCTGCACATTGCTG (R). Synthetic lunx fragment for gene copy determination: CCCTGGAAGCCTGCAAATTUCUCUGCUUGAUGGACUUGGCCCCCUCCCCAUUCAAGGUCUUCUGGACAGCCUCACAGGGAUCUUGAAUAAAGTCCTGCCTGAGTTGGTTC.

Peripheral Blood Specimens

This study was approved by the Medical University of South Carolina (MUSC) Institutional Review Board, and informed consent was obtained from all patients enrolled. Peripheral blood specimens (10 to 20 ml) were collected using K₃ EDTA tubes (Vacutainer; Becton Dickinson, Franklin Lakes, NJ) and immediately placed in ice. Samples were then processed using a porous barrier density gradient centrifugation media (OncoQuick, Hexal Gentech, Holzkirchen, Germany) per manufacturer’s instructions. Briefly, pre-cooled 50-ml centrifugation tubes containing 15 ml of separation medium below a porous barrier were filled with peripheral blood and centrifuged at 1600 × g for 20 minutes. The entire volume of the upper compartment was then collected and washed for 10 minutes at 200 × g. Cells were pelleted and evaluated as described below.

RNA Isolation and Gene-Specific cDNA Synthesis

Total cellular RNA was isolated from pelleted cells using a guanidinum thiocyanate-phenol-chloroform solution (RNA STAT-60; TEL-TEST, Friendswood, TX). Briefly, pelleted cells recovered from peripheral blood specimens were resuspended in 1 ml of RNA STAT-60. Total RNA was isolated as per manufacturer’s instructions with the exception that 1 μl of a 50 mg/ml solution of glycogen (Sigma, St. Louis, MO) was added to the aqueous phase before addition of isopropanol. Final RNA pellet was dissolved in 50 μl of 1X RNA secure buffer (Ambion, Austin, TX). RNA was quantified by UV absorbance at 260 nm. Complementary DNA (cDNA) was made from 5 μg of total RNA using 200 U of M-MLV reverse transcriptase (Promega, Madison, WI) and the following gene-specific primers (70 ng each): CCAAATGCGGCAT (β₂-microglobulin), TGAAGTACACTGG (KS1/4), GAACCAACTCAGGC (lunx), GCCACCATTACCT (muc1), TGATTCTGCCGC (CK19), GTTCCCATCAATCAG (CEA), and AGCCACTTCTGC (PSE). Final reaction volume was 20 μl.

Real-Time RT-PCR

Real-time RT-PCR was performed on a PE Biosystems Gene Amp 5700 Sequence Detection System (Foster City, CA). All reaction components were purchased from PE Biosystems. Standard reaction volume was 10 μl and contained 1X SYBR Green PCR buffer, 3.5 mmol/L MgCl₂, 0.2 mmol/L each of dATP, dCTP, dGTP, and 0.4 mmol/L of dUTP, 0.25U AmpliTaq Gold, 0.1U AmpErase UNG enzyme, 0.7 μl cDNA template, and 0.25 mmol/L of forward and reverse primer. Initial step of RT-PCR was 2 minutes at 50°C for AmpErase UNG activation, followed by a 10-minute hold at 95°C. Cycles (n = 40 first round) consisted of a 15-second melt at 95°C, followed by a 1 minute annealing/extension at 60°C. The final step was a 60°C incubation for 1 minute. All reactions were performed in triplicate and a negative control lacking cDNA was included. For a blood sample to be considered evaluable, we set a cutoff value for the β₂-microglobulin internal control gene at ≤25 (corresponding to approximately 2 × 10⁴ gene copies).

Results

Reliable Detection of 20 Lunx Gene Copies by Real-Time PCR

To determine whether a single round of real-time RT-PCR could be used for the sensitive detection of NSCLC, we first performed studies on a synthetic fragment encoding a portion of lunx, a gene previously shown to be expressed in metastatic lymph nodes of NSCLC patients. 21^, 22 C_t values for various fragment dilutions were obtained and plotted as a function of initial fragment copy number. Figure 1 demonstrates a strong linear relationship between the C_t value and the log of fragment copy number (R² = 0.9981). Reliable fluorescent signals were obtained for reactions containing as few as 20 gene copies (Figure 1 , log value = 1.4). In contrast, reliable fluorescent signals were not obtained for samples that contained only two gene copies (data not shown), regardless of the fluorescent threshold setting used for real-time measurements. These data provide evidence that a single round (as opposed to two) of real-time PCR reliably amplifies ≥20 gene copies, a result amenable to detection of circulating tumor cells in peripheral blood.

Figure 1.

Reliable detection of 20 gene copies in a single round of PCR. Real-time RT-PCR reactions were performed in triplicate as described in Materials and Methods using the lunx primer pair and the lunx synthetic sequence listed in Materials and Methods. Gene copy number was determined by UV absorbance measurements at 260 nm. The line through the data points was obtained by linear regression analysis using Microsoft Excel software.

Detection of Lunx Gene Expression in Peripheral Blood of NSCLC Patients

To assess the ability of real-time RT-PCR to detect circulating tumor cells in the peripheral blood of NSCLC patients, samples from 15 healthy volunteers and 24 patients with stage I-IV NSCLC were obtained. Tumor cells were first enriched from peripheral blood by a newly developed porous barrier density gradient (PBDG) centrifugation system. 27 The depletion of mononuclear cells in the enriched cell fraction after PBDG centrifugation is approximately 300- to >500-fold. 27^, 29 Mean tumor cell recovery rates for PBDG are comparable to that achieved by Ficoll purification. 27^, 29 Previous studies in the breast cancer setting have shown that the upper limit of detection using real-time RT-PCR is one cancer cell among 5 × 10⁸ peripheral blood cells. 29

Using a single round of real-time PCR (40 cycles), we determined expression levels for five genes associated with NSCLC: lunx, KS1/4, muc1, CK19,and CEA, 22 as well as one gene (PSE) associated with prostate 30 and breast cancer. 31^, 32 Mean expression levels of the cancer-associated genes were normalized to β₂-microglobulin using Q-gene software. 33 We observed that in the normal control peripheral blood samples, expression of the lunx gene was not detectable (Figure 2) . For other genes, expression was detected in a limited number of patients: muc1 and CK19 (four samples), CEA and PSE (three samples), and KS1/4 (two samples) (Figure 2) . Based on data obtained from the normal control population, we set threshold values for marker positivity at three standard deviations beyond the mean normalized expression values of each respective gene (Figure 2 , horizontal lines). Assuming a normal distribution of the control peripheral blood samples, three standard deviations correspond to a test-specificity level of 99.9%. In the control patient group, no gene was overexpressed above threshold levels.

Figure 2.

Multimarker real-time RT-PCR analysis of NSCLC in peripheral blood. Real-time PCR analyses of peripheral blood specimens from 15 healthy volunteers (open triangles), and 24 NSCLC patients (open diamonds) were performed as described in the text using primer pairs for the indicated genes. Threshold levels of marker positivity for each gene were calculated as described in the text and are depicted by the horizontal line on the left side of each data set. Expression levels of each gene were calculated with Q-gene software 33 and are expressed as the ratio of the target gene relative to β₂-microglobulin.

In the peripheral blood samples derived from NSCLC patients (n = 24), we observed that 14 of 24 (58%) overexpressed at least one marker gene (Figure 2 , Table 1 ). The gene most highly overexpressed was lunx (10 of 24 samples (42%)). muc1 and CK19 were each overexpressed in 5 of 20 (21%) of patients, three of whom overexpressed both markers. Overexpression of KS1/4 and PSE was observed in two patients and one patient, respectively, all of who overexpressed either lunx or muc1 (Table 1) . Of patients with presumed curable and resectable stage I to II disease (n = 7), lunx was overexpressed in two (29%) blood samples.

Table 1.

Detection of Gene Overexpression in NSCLC Patients

Patient information	Real-time RT-PCR results*
Patient no.	Stage	Age	LUNX	MUC1	CK19	KS1/4	PSE	CEA
1	IA	78	—	—	—	—	—	—
2	IA	67	1	—	1	—	—	—
3	IB	57	—	—	—	—	—	—
4	IB	52	—	—	—	—	—	—
5	IB	75	1	—	—	—	—	—
6	IIB	75	—	—	—	—	—	—
7	IIB	41	—	1	1	—	—	—
8	III	55	—	—	—	—	—	—
9	III	67	—	—	—	—	—	—
10	IIIA	64	—	—	—	—	—	—
11	IIIA	71	—	—	—	—	—	—
12	IIIB	59	1	—	—	—	—	—
13	IIIB	63	1	—	—	—	—	—
14	IIIB	54	1	1	1	—	—	—
15	IIIB	54	1	—	—	—	—	—
16	IIIB	67	1	—	—	—	—	—
17	IV	66	—	—	—	—	—	—
18	IV	62	—	1	—	—	—	—
19	IV	62	—	—	—	—	—	—
20	IV	74	1	—	—	—	—	—
21	IV	46	1	—	—	—	—	—
22	IV	77	—	1	—	—	1	—
23	IV	67	—	1	1	1	—	—
24	IV	52	1	—	1	1	—	—
Total	10	5	5	2	1	0

No overexpression of the respective gene is indicated by “—”; overexpression is indicated by “1”.

Discussion

The ability to detect nucleic acid fragments by PCR is directly proportional to gene copy number, fragment amplification efficiency, and detection threshold, and inversely proportional to the formation of primer dimers. Due to their extremely low concentration (and hence, gene copy numbers), the molecular detection of cancer cells in peripheral blood has proven challenging compared to other tissues such as lymph node. In an effort to increase the senstivity of NSCLC detection in peripheral blood, reseachers have adopted nested PCR strategies for various genes such as CK19, preproGRP, syndcan 1, collagen 1A2, and CEA. 34^, 35^, 36^, 37^, 38^, 39 In this study, we provide evidence that the use of real-time PCR and SYBR Green I chemistry allows for reproducible detection of 20 copies of an aritificial lunx sequence by PCR cycle number 36 (Figure 1) . These results agree well with those of Karsai et al, 40 who used SYBR Green I chemistry to detect 10 to 20 copies of UBQ-5 RNA or dsDNA, and Leutenegger et al, 41 who used a real-time TaqMan PCR assay to detect as few as 50 copies of SIV RNA/ml. These results provide evidence that a single round of real-time RT-PCR is sufficient for detection of genes present in low abundance.

Using a single round of real-time PCR, we analyzed the peripheral blood of NSCLC patients for expression of five genes associated with NSCLC:lunx, KS1/4, muc1, CK19,and CEA. 22 With respect to normal control samples, overexpression of at least one gene was observed in 14 of 24 (58%) NSCLC patients. Of stage I to II patients (n = 7), 3 (43%) were positive for at least one marker, while 11 of 17 (65%) stage III to IV patients were marker positive. Ten NSCLC blood samples were positive for lunx, providing evidence that this marker was the most sensitive for detection of circulating NSCLC cells.

In a previous study, we used the same markers described in this paper and observed that KS1/4 had the highest sensitivity for detection of NSCLC in mediastinal lymph nodes. 22 However, in the present study, we observed that KS1/4 was overexpressed in the peripheral blood of only 2 of 24 (8%) patients. The low level of overexpression of this gene in peripheral blood may be due to several factors, including small patient sample size, lack of appropriate growth factors in peripheral blood, and/or increased background gene expression. In support of the later possibility, Zhong and colleagues 42 found by RT-PCR that KS1/4 was expressed in 40% (16 of 40) of normal peripheral blood samples and 100% of normal bone marrow samples (n = 8). DeGraffe et al, 43 using quantitative nested RT-PCR, found a consistently low level of KS1/4 mRNA (4 × 10 ⁻⁴ copies/cell) in peripheral blood mononuclear cells from normal donors while several breast cancer cell lines expressed 20 to 100 copies/cell. It also should be noted that KS1/4 has been extensively studied as a molecular marker of various cancers. However, this fact is not obvious since the KS1/4 marker and/or the antibody which recognizes its gene product is known by various names (TROP1, 44 AUA1, 45 HeGP314, 46 CO17–1A, 47 EpCAM, 48 MK-1, 49 M4S1, 50 EGP40, 51 EGP2, 52 TACSTD1, 50 KSA, 48 and GA733–2 53 ).

The results described in this paper provide evidence that lunx was the most sensitive marker for detection of circulating NSCLC cells. Examination of the predicted amino acid sequence of plunc, the mouse homologue of lunx, indicates that it is a secreted protein and related to proteins expressed in high abundance in saliva. 23 During mouse embryogenesis, plunc/lunx expression is limited in a temporal manner to the dorso-lateral epithelium of the developing palatal shelves. In the adult (mouse), expression is confined to discrete epithelial bands on the exposed surfaces of the nasal columella, turbinates, and common nasal passage. 23 Recent mass spectroscopy studies indicate that the human lunx gene product is also expressed in normal adult nasal lavage fluid. Further, its expression might be up-regulated in response to certain airway irritants such as cigarette smoke and dimethylbenzylamine. 54^, 55 Although lunx mRNA has been detected in metastatic lymph nodes of NSCLC patients, 21^, 22 this is the first study to demonstrate expression of lunx in peripheral blood of NSCLC patients. Additional clinical studies with follow-up data are required to determine whether detection of lunx in early stage II NSCLC patients correlates with decreased survival.