Fax: (011) 81-88-880-2348
Sensitive detection of small cell lung carcinoma cells by reverse transcriptase–polymerase chain reaction for prepro–gastrin-releasing peptide mRNA
Article first published online: 30 APR 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 10, pages 2504–2511, 15 May 2003
How to Cite
Saito, T., Kobayashi, M., Harada, R., Uemura, Y. and Taguchi, H. (2003), Sensitive detection of small cell lung carcinoma cells by reverse transcriptase–polymerase chain reaction for prepro–gastrin-releasing peptide mRNA. Cancer, 97: 2504–2511. doi: 10.1002/cncr.11378
- Issue published online: 30 APR 2003
- Article first published online: 30 APR 2003
- Manuscript Accepted: 21 JAN 2003
- Manuscript Revised: 30 DEC 2002
- Manuscript Received: 14 OCT 2002
- prepro-gastrin releasing peptide;
- nested reverse transcriptase-polymerase chain reaction;
- small cell lung carcinoma;
- peripheral blood;
- pleural effusion
Gastrin-releasing peptide (GRP) is an autocrine growth factor in patients with small cell lung carcinoma (SCLC). The authors developed a reverse transcriptase-polymerase chain reaction (RT-PCR) assay for the detection of SCLC cells in the peripheral blood and the pleural effusion using preproGRP mRNA as a target.
The current study was conducted to determine the utility of preproGRP-specific nested RT-PCR on the peripheral blood, bone marrow, and pleural effusion samples from 32 patients with SCLC, 39 patients with non-small cell lung carcinoma (NSCLC), 28 patients with nonmalignant pulmonary disease, and 20 healthy volunteers. The internal primers were designed to amplify a 244-base pair PCR product, a sequence encompassing exon 1 and exon 2 by the nested RT-PCR assay.
Amplification of the preproGRP message was detected in SCLC cell lines (LU165, SBC1, SBC2, and SBC3) but not in other NSCLC cell lines (A549, ABC1, EBC1, and Oka-1). The SCLC cells (LU165) were detected in dilutions of tumor cells of up to 10−7 in hematopoietic cells from healthy donors. The preproGRP mRNA was detected in 16 of 32 (50%) blood samples, 2 of 11 (18%) marrow samples, and in all 6 (100%) pleural effusion samples. Blood samples gave positive results in 11 of 19 (58%) patients with extensive disease compared with 5 of 13 (38%) patients with limited disease. In contrast, only 1 blood sample (2.6%) from a patient with lung adenocarcinoma gave a positive result among patients with NSCLC. No other samples of blood, bone marrow, and pleural effusion from patients with NSCLC and none of the blood samples from patients with nonmalignant diseases and healthy volunteers were positive.
The current RT-PCR approach may be a sensitive and specific assay to detect SCLC cells in circulating blood as well as in pleural effusions from SCLC patients. Cancer 2003;97:2504–11. © 2003 American Cancer Society.
Small cell lung carcinoma (SCLC) represents 15–20% of all lung carcinomas. Compared with non-small cell lung carcinoma (NSCLC), SCLC is an aggressive disease with early and frequent development of widespread dissemination, despite a marked sensitivity to chemotherapy and radiotherapy. With conventional chemotherapy or chemoradiotherapy, 80–100% of patients with limited disease (LD) achieve an objective response to treatment and 15–20% of patients live more than 3 years.1 In contrast, 60–80% of patients with extensive disease (ED) achieve an objective response to treatment and less than 5% of patients live more than 3 years.2
As SCLC disseminates preferentially via the blood stream, sensitive techniques that detect circulating cells are needed. Recent advances in molecular technology, including polymerase chain reaction (PCR), enabled the detection of minimal residual tumor cells in the peripheral blood (PB) or bone marrow (BM) of patients with melanoma, prostate carcinoma, and neuroblastoma.3 CK 19, carcinoembryonic antigen, and MUCI were used to detect lung carcinoma cells in PB or BM samples, but they yielded inconsistent results.4–6
These assays were hampered by false-positive results due to either illegitimate transcription,7 coamplification of pseudogenes,8 or to detection of transcripts from dermal epithelial cells contaminating the blood samples in the course of venous puncture.4
Small cell lung carcinoma is a well known neuroendocrine tumor that produces a number of neuropeptides, including gastrin, vasopressin, corticotropic hormone, calcitonin, and gastrin-releasing peptide (GRP).9 Of these, GRP (bombesin-like peptide) is an autocrine growth factor in SCLC cell lines.10 It is a gut peptide hormone that was isolated from the porcine stomach. It is present in nerve fibers in nonantral stomach tissue, brain, and neuroendocrine cells in fetal lung and in pulmonary carcinoid and SCLC cells.11, 12 ProGRP, a precursor protein of GRP,13 is one of the reliable tumor markers for SCLC. It has been used in clinical practice in Japan for monitoring disease activity.14 A precursor of GRP, preproGRP is an appropriate target of reverse transcriptase (RT)-PCR because it is not expressed in the hematopoietic or endothelial cells in blood vessels, nor in the epithelial cells of the human skin due its unique mammalian tissue distribution. Therefore, the risk of contamination during needle puncture of the skin can be avoided. We developed a nested RT-PCR assay for preproGRP mRNA to detect SCLC cells in biologic fluids and investigated the utility of this method.
MATERIALS AND METHODS
Eight lung carcinoma cell lines were used: 4 SCLC cell lines (LU165, SBC1, SBC2, and SBC3), 2 lung adenocarcinoma cell lines (A549, ABC1), and 2 lung squamous cell carcinoma cell lines (EBC1, Oka-1). SBC1, SBC2, SBC3, ABC1, and EBC1 were provided by Dr. Hiroshi Ueoka (The Second Department of Internal Medicine, Okayama University Medical School, Okayama, Japan). LU165 was donated by Dr. Jiro Fujita (The First Department of Internal Medicine, Kagawa Medical School, Kagawa, Japan). Oka-1 was a poorly differentiated squamous cell carcinoma cell line.15 All cell lines were maintained under the standard condition (RPMI with 10%/fetal calf serum and penicillin/streptomycin).
The study population consisted of 32 patients with histologically proven SCLC, 39 patients with histologically proven NSCLC, 10 patients with chronic obstructive pulmonary disease (COPD), 8 patients with sarcoidosis, 10 patients with idiopathic pulmonary fibrosis, and 20 healthy volunteers. In addition, one patient with melanoma and three patients with congestive heart failure were recruited for the analysis of pleural effusion. All patients were admitted to the Koichi Medical School Hospital between January 1998 and December 2001.
Either 10 mL of PB, 2–3 mL of BM aspirates, or 20 mL of pleural effusion (PE) was obtained from patients at diagnosis. Thirty-two PB samples, 11 BM aspirates, and 6 PE samples were obtained from 32 SCLC patients. The median patient age was 73 years (range, 42–84 years). The female-to-male ratio was 4:28. Of the sample, 13 and 19 patients had LD and ED, respectively.
Thirty-nine PB, 2 BM, and 10 PE samples were obtained from 39 patients with NSCLC. The median age was 68 years (range, 39–84 years) and the female-to-male ratio was 4:35. Seventeen patients had adenocarcinoma, 17 had squamous cell carcinoma, and 5 had large cell carcinoma. Using the revised TNM classification,16 2 patients had Stage IA disease, 2 patients had Stage IB, 1 had Stage IIB, 14 had Stage IIIA, 13 had Stage IIIB, and 17 had Stage IV disease (Table 1).
|Disease type (No. of patients)|
|Squamous cell carcinoma||17|
|Large cell carcinoma||5|
|Age (median yrs)|
|SCLC (range)||73 (42–84)|
|NSCLC (range)||68 (39–84)|
|Disease stage (No. of patients)|
|IA + IB||4|
|IIA + IIB||1|
Twenty-eight PB samples from patients with nonmalignant disease (10 COPD patients, 8 sarcoidosis patients, and 10 idiopathic pulmonary fibrosis patients) and 20 PB samples from healthy volunteers (2 females, 18 males; median age, 37 years; range, 30–53 years) were tested in parallel. In addition, PE samples from a patient with melanoma and three patients with congestive heart failure also were tested. All samples were obtained after patients provided informed consent.
Samples of SCLC and NSCLC cell lines were centrifuged for 5 minutes at 450 × g and washed in phosphate-buffered saline (PBS) before RNA extraction. Mononuclear cells from PB and BM samples were isolated by a Ficoll-Hypaque density gradient centrifugation method (30 minutes at 2000 × g), washed in PBS, and stored-frozen at −80 °C until RNA extraction. Cell pellets obtained from centrifuged PE samples were also stored-frozen at −80 °C until RNA extraction. Total cellular RNA was extracted from the samples by a single-step method with guanidium thiocyanate and phenol/chloroform.17
Outer primer and internal primer pairs encompassing exon 1 and exon 2 of preproGRP18 (GenBank accession number K02054) were selected for nested RT-PCR (Fig. 1). The outer primers were 5′ TGC TGG CGC TGG TCC TCT GC 3′ (outer forward) and 5′ TGC TGC TAT CCT CTG AAT CC 3′ (outer reverse), yielding a 324-base pair (bp) transcript. The internal primers were 5′ GGA CCG TGC TGA CCA AGA TG 3′ (internal forward) and 5′ TCC CAC GAA GGC TGC TGA TT 3′ (internal reverse), yielding a 244-bp transcript. These primers were designed to contain the common sequence among three forms of the distinct preproGRP mRNA that are generated after alternative splicing.19, 20 The primers were selected using the oligo primer analysis software (Version 5.0; National Biosciences, Plymouth, MN.) according to the manufacturers instruction. The primer sequences for β-actin were adopted from a study by Kruger et al.21 The outer primers were 5′ GCG AGA AGA TGA CCC AGA TC 3′ (outer forward) and 5′ CCG ATC CAC ACG GAG TAC TT 3′ (outer reverse), yielding a 679-bp transcript. The internal primers were 5′ GGA CTT CGA GCA AGA TAT GG 3′ (internal forward) and 5′ GCA GTG ATC TCC TTC TGC ATC 3′ (internal reverse), yielding a 294-bp transcript.
RT-PCR and Analysis of Products
A 1-μg sample of RNA was subjected to RT with AMV reverse transcriptase (Takara Shuzou, Tokyo, Japan) and random nonamer in a total volume of 20 μL at 42 °C for 30 minutes. The nested PCR was performed as follows. In the first PCR assay, 5 of 20 μL of the cDNA preparation was subjected to a first preproGRP-specific amplification in a total volume of 25 μL using 1 U Taq DNA polymerase (Takara Shuzou), 1.5 mM MgCl2, 200 pM of each dNTP, and 10 pM of the outer primers. The PCR reactions were performed according to the following standard protocols.22 Five minutes of initial denaturation at 94 °C; 45 cycles: 94 °C for 1 minute, 68 °C for 1 minute, 72 °C for 1 minute; final elongation: 72 °C for 7 minutes. After the first amplification, an aliquot was diluted 1:10 and 5 μL was submitted to the second amplification with nested (internal) primers. The thermal cycling condition was the same as the first PCR assay except that the final volume was 50 μL with a correspondingly increased reaction mixture. To avoid contamination, positive displacement pipettes were used. Preparations of the amplification mix and analysis of the PCR products were performed in a laminar airflow hood. Amplification of β-actin sequences was carried out under similar conditions. Consequently, in case of failed amplification of β-actin transcripts, the sample analyzed was excluded from further analysis. Aliquots from amplified samples were analyzed by 2% agarose gel electrophoresis and visualized by ethidium bromide staining under ultraviolet (UV) light. All experiments were carried out twice.
After electrophoresis, the gels were transferred to Hybond N+ (Amersham Pharmacia Biotech, Little Chalfont, UK) by capillary blotting and fixed by means of a UV crosslinker. The membranes were hybridized to the digoxigenin-labeled product of nested RT-PCR at 42 °C overnight. The bound digoxigenin-labeled probes were immunostained with alkaline phosphatase-labeled murine antidigoxigenin antibody Fab fragment (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's instructions. Subsequently, nonradioisotopic chemiluminescence detection was performed using readymade CSPD (Roche Diagnostics).23 The preproGRP-specific digoxigenin-labeled PCR probe was made on the basis of a previous study.24 Briefly, digoxigenin was incorporated into the hybridization probe by means of PCR using a digoxigenin DNA-labeling mixture (Rosche Diagnostics) in place of dNTP and a 244-bp amplified nested PCR product as a template according to the manufacturer's protocol in some cases.
Sequence analysis of amplified fragments was performed in an automated sequencer (Applied Biosystems Japan, Tokyo, Japan). PCR products were sequenced by the dideoxynucleotide chain-termination method using fluorescent labels with dye terminator kits directly following standard methods.
Differences in the frequency of patients with preproGRP-expressing tumor cells within the different subgroups were compared using the Fisher exact test.
Lung Carcinoma Cell Lines
The primer pairs were tested for PCR amplification of cellular mRNA from lung carcinoma cell lines. Amplification of the SCLC cell lines (LU165, SBC1, SBC2, and SBC3) after reverse transcription produced PCR products of the expected size (244 bp) after the nested amplification. However, other NSCLC cell lines (A549, ABC1, EBC1, and Oka-1) did not produce PCR products (Fig. 2). To assess the quality of the RNA sample tested, the β-actin fragment was amplified in parallel. The specificity also was confirmed by Southern blotting.
To confirm specificity, we amplified cellular mRNA from the SBC2 cell line, an autopsy tumor tissue, and from cell pellets of PE from a patient with SCLC. PCR products (size, 244 bp) were amplified from all the samples after nested RT-PCR (Fig. 3). Direct sequencing of PCR products was performed using the BigDye terminator cycle sequencing ready reaction kit (Applied Biosystems Japan). Comparison with the sequence of preproGRP mRNA was carried out. The DNA sequences of amplified products and the published sequence of preproGRP18 were identical (data not shown).
The LU165 cell line was diluted by a factor of 10−1 to 10−7 in hematopoietic cells from healthy volunteer donors and an RNA sample was extracted and subjected to nested RT-PCR. The preproGRP message was detected in dilutions up to 10−7. The amplification products were visualized following ethidium bromide staining (Fig. 4). The specificity also was confirmed by Southern blotting.
Application to Clinical Samples
Typical examples of the results of the nested RT-PCR on samples from patients with SCLC and patients with NSCLC are shown in Figures 5 and 6, respectively. Nested RT-PCR was performed on 32 PB, 11 BM, and 6 PE samples from patients with SCLC. The preproGRP message was detected in 16 of 32 (50.0%) PB, 2 of 11 (18.1%) BM, and all 6 (100%) PE samples from patients with SCLC. In contrast, the preproGRP message was detected in 1 of 39 (2.6%) PB, 0 of 2 (0%) BM, and 0 of 10 (0%) PE samples from patients with NSCLC. Based on the histologic classification of patients with lung carcinoma, the preproGRP transcripts were detected in 1 of 17 (5.9%) patients, with adenocarcinoma, 0 of 17 (0%) patients with squamous cell carcinoma, and in 0 of 5 (0%) patients with large cell carcinoma of the lung. In addition, preproGRP transcripts were detected by RT-PCR assay in 0 of 10 (0%) patients with COPD, 0 of 8 (0%) patients with sarcoidosis, 0 of 10 (0%) patients with idiopathic pulmonary fibrosis, and 0 of 20 (0%) healthy donors (Table 2). The frequency of the positive rate for the preproGRP message among patients with SCLC was significantly greater than it was among patients with NSCLC (P < 0.01), patients with nonmalignant disease (P < 0.01), and healthy volunteers (P < 0.01).
|Patients with SCLC||16/32 (50.0)||2/11 (18.1)|
|Patients with NSCLC|
|Adenocarcinoma||1/17 (5.9)||0/2 (0)|
|Squamous cell carcinoma||0/17 (0)||——|
|Large cell carcinoma||0/5 (0)||——|
|Total||1/39 (2.6)||0/2 (0)|
|Patients with nonmalignant diseases|
|Idiopathic pulmonary fibrosis||0/10 (0)||——|
|Healthy donors||0/20 (0)||——|
Blood samples from 5 of 13 SCLC patients (38.5%) with LD were positive, compared with 11 of 19 SCLC patients (57.9%) with ED (Table 3).
|Histogic type and disease stage||PreproGRPRT-PCR-positive (%)|
|Small cell lung carcinoma|
|Limited disease||5/13 (38.5)|
|Extensive disease||11/19 (57.9)|
|Non-small cell lung carcinoma|
|I + II||0/5 (0)|
Only one PB sample from a patient with Stage IV adenocarcinoma of the lung was positive, but no other PB samples from patients with Stage I, II, or III NSCLC were positive.
The preproGRP RT-PCR assay detected SCLC cells in PE samples. Without Ficoll-Hypaque gradient procedures, preproGRP transcripts were detected successfully in all 6 cell pellets derived from PE samples from SCLC patients, 5 of whom were cytologically positive for SCLC. In contrast, preproGRP transcripts were not detected in any of the 10 PE samples from patients with NSCLC (5 adenocarcinoma, 4 squamous cell carcinoma, and 1 large cell carcinoma of the lung) nor in the 1 patient with melanoma. In addition, transcripts were not detected in any of the three PE samples from patients with congestive heart failure. The frequency of the positive rate of preproGRP transcripts in PE samples from SCLC patients was significantly greater than that of PE samples from NSCLC patients (P < 0.01).
We developed a sensitive and specific nested RT-PCR assay to detect SCLC cells in PB, BM, and PE samples by amplifying the preproGRP sequences.
Until recently, HuD25 and the GRP receptor26 were investigated solely to detect SCLC cells in PB samples. However, HuD transcripts also were amplified from the mononuclear cell fractions of all blood samples from healthy subjects, which indicated the lack of specificity of the assay. The GRP receptor detected SCLC cells in PB and BM samples. However, it also was expressed more frequently by NSCLC cells.27 The RT-PCR studies of Oceja-Garcia et al. were conducted on neuropeptides and their receptors in lung carcinoma cell lines.27 They demonstrated that the neuropeptides AVP and GRP, as well as the cholecystokinin receptor CCK-B, might be potential markers for the detection of SCLC cells in biopsy specimens or SCLC micrometastasis in blood samples.
Lacroix et al.28 demonstrated that amplification of preproGRP transcripts from clinical samples is a more sensitive and specific assay to detect disseminated or exfoliated lung carcinoma cells either in PB or sputum samples than RT-PCR assays for NCAM, PGP9.5, and the GRP receptor.
Consequently, we also focused on preproGRP as the target because proGRP, a processed form of preproGRP,29 is one of the most successful tumor markers of SCLC in Japan.14 ProGRP, which is produced by SCLC cells, functions as an autocrine growth factor and accelerates the continuous proliferation of these cells.10 In fact, proGRP is much more stable in the blood circulation than GRP itself. In addition, measurement of serum proGRP levels showed a wide range of values on blood samples from patients with SCLC which allowed us to assess the extent of progression of the disease.30
In the current study, SCLC cells (LU165) were detected in dilutions of tumor cells of up to 10−7 in hematopoietic cells from healthy donors. Amplification of the preproGRP message was detected in SCLC cell lines (LU165, SBC1, SBC2, and SBC3), but no amplification of the preproGRP message was detected in other NSCLC cell lines (A549, ABC1, EBC1, and Oka-1). Specificity was confirmed by Southern blotting using a digoxigenin-labeled PCR probe.
The preproGRP-based nested RT-PCR technique was applied to 32 PB, 11 BM, and 6 PE samples from 32 patients with SCLC. PreproGRP mRNA was detected in 16 of 32 (50%) PB, 2 of 11 (18%) BM and all 6 (100%) PE samples. Blood samples gave positive results in 11 of 19 (58%) patients with ED, compared with 5 of 13 (38%) patients with LD. These data suggest that cancer cells circulate in the PB of patients with SCLC, even in patients with LD. The positive rate of the RT-PCR assay of PB among patients with SCLC was similar when compared with an earlier study.28 In contrast, only 1 of the 39 PB samples (2.6%), from a patient with lung adenocarcinoma, was positive in patients with NSCLC. No other samples, including BM and PE from patients with NSCLC, were positive in this RT-PCR assay. Furthermore, none of the PB samples from patients with nonmalignant diseases and healthy volunteers were positive.
In the Lacroix et al. study,28 50% of PB samples from patients with SCLC and 27% of PB samples from patients with NSCLC were positive using a preproGRP-specific RT-PCR assay with primers encompassing exon 2 and exon 3, which were different assays from ours. Similarly, 50% of PB samples from SCLC patients in the current study were also positive. In contrast, only 2.6% of PB samples from NSCLC patients were positive. The reason for the discrepancy among patients with NSCLC between our RT-PCR assay and their assay (27% vs. 2.6%) is unknown. The selection of the target sequence of preproGRP mRNA might have influenced the positive rate among patients with NSCLC and contributed to the better specificity of our RT-PCR assay. In addition, our primers were designed carefully to contain the common sequence among three forms of the distinct preproGRP mRNA that are generated after alternative splicing. Yet another reason may be the production of proGRP by NSCLC cells with neuroendocrine function, because some NSCLC patients were positive for serum proGRP.12–14
The preproGRP RT-PCR assay detected SCLC cells in PE samples. PreproGRP transcripts were detected successfully in all 6 (100%) cell pellets derived from PE samples from SCLC patients, 5 (83%) of which were cytologically positive. In contrast, transcripts were not detected in any of the 10 PE samples from NSCLC patients (5 adenocarcinoma, 4 squamous cell carcinoma, and 1 large cell carcinoma of the lung), nor was it detected in the single patient with melanoma. In addition, transcripts were not detected in any of the three PE samples from patients with congestive heart failure. Further studies are necessary to confirm whether the RT-PCR assay is more sensitive than conventional cytologic examination for the detection of SCLC cells in PE samples.
In conclusion, the current RT-PCR assay for preproGRP mRNA may be a more sensitive and specific assay for detection of SCLC cells in the circulating blood and in PE samples from SCLC patients. In this context, real-time PCR for preproGRP message is promising for the detection of SCLC cells in biologic fluids.
- 2Randomized study of cyclophosphamide, doxorubicin, and vincristine versus etoposide and cisplatin versus alternation of these regimens in extensive small cell lung cancer: a phase III trial of Southeastern Cancer Study Group. J Clin Oncol. 1992; 10: 282–291., , , et al.
- 5Lung cancer metastatic cells detected in blood by reverse transcriptase-polymerase chain reaction and dot-blot analysis. J Clin Oncol. 1997; 15: 3383–3393., , , et al.
- 11Bombesin-like immunoreactivity in mammalian tissues. Biomed Res. 1978; 1: 767–774., , , , , .
- 12Non-amphibian bombesin-like peptides. In: BloomSR, PolakJM. Gut hormones, 2nd edition. Edinburgh: Churchill Livingstone, 1981: 407– 412..
- 13Growth factors and receptors in small cell lung cancer. In: KaneMA, BunnPJ. Biology of lung cancer. Lung biology in health and disease, vol. 122. New York: Marcel Dekker, 1998: 337–370..
- 17Molecular cloning: a laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 1989., , .
- 23Reverse transcriptase/polymerase chain reaction analysis of parathyroid hormone-related protein for the detection of tumor cell dissemination in the peripheral blood and bone marrow of patients with breast cancer. J Cancer Res Clin Oncol. 1997; 123: 514–521., , , et al.
- 30Serum levels of pro-gastrin-releasing peptide for follow up of patients with small cell lung cancer. Clin Cancer Res. 1996; 3: 123–127., , , et al.