Small cell lung carcinoma (SCLC) has the propensity to grow rapidly and metastasize extensively. Detection of micro-dissemination of SCLC may have clinical relevance. For its detection, tumor-specific gene expressions were examined in peripheral blood and bone marrow aspirate from patients with SCLC.
Expression of prepro-gastrin-releasing peptide (preproGRP), neuromedin B receptor (NMB-R) and gastrin-releasing peptide receptor (GRP-R) were examined by reverse transcriptase polymerase chain reaction (RT-PCR) in peripheral blood and bone marrow aspirate from 40 untreated patients with SCLC. Control samples consisted of peripheral blood samples from 5 patients with nonsmall cell lung cancer (NSCLC) and 20 healthy volunteers.
Positive rates of preproGRP, NMB-R, and GRP-R in bone marrow aspirate of patients with SCLC were 23% (9/40), 8% (3/40), and 10% (4/40), respectively. Those rates in peripheral blood were 11% (4/38), 5% (2/38), and 29% (11/38), respectively. Although GRP-R expression was detected in patients with NSCLC and in healthy volunteers, preproGRP and NMB-R expressions were not detected in patients with NSCLC and in healthy volunteers. All three gene expressions in bone marrow were more frequently observed in patients with bone marrow metastasis, accessed by biopsy, than in patients without. PreproGRP gene expression in bone marrow was also more frequent in patients with bone metastasis, accessed by bone scintigram, than in patients without, and was related to poorer survival.
Evaluation of serum tumor-specific antigens is clinically relevant for the diagnosis of cancer, monitoring of tumor response to therapy, detection of recurrence, and prediction of prognosis.1 Conversely, detection of tumor-specific gene expression with the technique of reverse transcriptase polymerase chain reaction (RT-PCR) can indicate the presence of viable tumor cells in the specimens. Because RT-PCR has the capability to detect very low levels of gene expression, even a small amount of viable cells undetectable by routine histologic examination can be detected with this technique. Such data regarding the presence of occult tumor cells in a given tissue may provide additional information for the choice of therapeutic modality.
For example, tumor-specific gene expressions, such as prostate-specific antigen, melanoma antigen recognized by T lymphocytes 1, tyrosinase and cytokeratin 19 have been detected in peripheral blood from some patients with prostate carcinoma, melanoma, and breast carcinoma, respectively.1–4 Although some of the reports showed a significant relationship between the evidence of occult dissemination and prognosis2 or disease progression/regression,5 a definite conclusion has never been established. The specificity of such examination in peripheral blood is sometimes one of the major problems.6–9
Conversely, small cell lung carcinoma (SCLC) is one of the most annoying types of solid tumors, possessing the propensity to grow and metastasize rapidly in most cases. It generally responds well to initial chemotherapy and radiotherapy. Limited disease (LD) stage SCLC is principally treated with a combination of chemotherapy and radiotherapy, while extensive disease (ED) stage is usually treated with chemotherapy alone. Approximately 20% of patients with LD stage SCLC show disease-free survival for three to five years or longer, while ED stage or recurrent SCLC is intractable and lethal in most cases.10 Detection of hematogenous occult dissemination may provide useful clinical information for the selection of treatment modality and predicting prognosis. An autocrine growth stimulatory loop governed by gastrin-releasing peptide (GRP) and other bombesin-like peptides, together with their receptors, participates in lung development11 and repair12 as well as promoting SCLC growth.13 PreproGRP is a direct product of the genes, and one of its cleaved products, proGRP, is currently being used as a serum tumor maker specific for SCLC.14, 15 Neuromedin B receptor (NMB-R) and gastrin–releasing peptide receptor (GRP-R) are subtypes of the receptors. Therefore, the current study examined preproGRP, NMB-R, and GRP-R in peripheral blood and bone marrow aspirates as tumor specific genes for SCLC, in an effort to improve stage diagnosis and prognostic prediction of the disease.
PATIENTS AND METHODS
The study population consisted of 40 patients with SCLC, 5 patients with nonsmall cell lung carcinoma (NSCLC), and 20 healthy volunteers. None of the patients had any prior chemotherapy or radiotherapy. The characteristics of the 40 patients with SCLC are summarized in Table 1. The five patients with NSCLC consisted of one patient with Stage IIB, one patient with Stage IIIB, and three patients with Stage IV disease. The ages of the 20 healthy volunteers ranged from 24 to 41 years, with a median 27.5 years. Written informed consent was obtained from every subject at enrollment in the study approved by our institutional review board. Staging procedures for SCLC included chest radiography, bronchoscopy, chest and abdominal computed tomography scan with contrast medium enhancement, brain magnetic resonance imaging with contrast medium enhancement, bone scintigram with 99mTc-methylene diphosphonate, and bone marrow aspiration and biopsy. Staging procedures for NSCLC were the same as for SCLC, except that bone marrow aspiration and biopsy were omitted. As to stage classification of SCLC, LD was defined as having all lesions within a hemithorax, with regional metastases to ipsilateral hilar, ipsilateral and/or contralateral mediastinal, supraclavicular lymphnodes, and ipsilateral pleural effusion. ED was defined as having lesions extending beyond a hemithorax.16 Stage classification for NSCLC was based on the International Union Against Cancer (UICC) TNM classification.
Table 1. Characteristics of Patients with Small Cell Lung Carcinoma (n = 40)
LD: limited disease; ED: extensive disease.
Accessed by bone marrow aspiration/biopsy.
Accessed by bone scintigram, bone radiogram, and bone magnetic resonance imaging.
Bone marrow aspiration and biopsy were performed as routine staging procedures for patients with SCLC. Briefly, after topical anesthesia with injection of xylocaine, the posterior iliac crest of either side was punctured and bone marrow aspirate was collected into a heparinized tube with an 11-gauge disposable Jamshidi bone marrow biopsy needle (Allegiance Healthcare Corporation, McGraw Park, IL), followed by biopsy at the same location with the same needle. A part of the aspirate and the biopsied specimen were used for cytologic and histologic examinations. The rest of the aspirate was used for RT-PCR. Prior to the aspiration and biopsy but on the same day, peripheral blood was collected into a heparinized tube immediately after needle puncture of a peripheral vein. Two mL of each sample were admixed with a 2 mL balanced solution consisting of 0.126 M NaCl, 0.01% anhydrous D-glucose, 5.0 × 10−6 M CaCl2, 9.8 × 10−5 M MgCl2, 5.4 × 10−4 M KCl, and 0.0145 M tris hydroxymethyl aminomethane, layered on Ficoll-Paque Plus (Amersham Biosciences, Piscataway, NJ) and centrifuged at 400 g to separate mononuclear cells, followed by washing the isolated cells twice with the balanced salt solution, according to the manufacturer's instructions. The mononuclear cells were lysed using an Isogen RNA isolation kit (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions. Extracted total RNA was suspended in 20 μL of distilled water. Eventually, bone marrow aspirates were examined in all 40 patients with SCLC, and peripheral blood of 38 of the 40 patients with SCLC, 5 patients with NSCLC, and 20 healthy volunteers was examined.
Complementary DNA was synthesized from 4 μL of the above prepared RNA solution by reaction with Moloney murine leukemia virus reverse transcriptase at 37 °C for one hour in a 20 μl reaction volume utilizing ProSTAR First-Strand RT-PCR Kits (Stratagene, La Jolla, CA) according to the manufacturer's instructions. Three genes, preproGRP, GRP-R, and NMB-R, supposedly specific for SCLC, were examined for their expression by PCR amplification of the synthesized cDNA. GRP-R and NMB-R were amplified by one-step PCR with the following primer sets: 5′-CTCCCCGTGAACGATGACTGG-3′ (forward, from 480 to 500 bp in the database sequence, GenBank #M73481)17 and 5′-ATCTTCATCAGGGCATGGGAG-3′ (reverse, from 848 to 868 bp), the same as previously reported18 to amplify the 0.4-kb fragment of GRP-R; 5′-CTTGCCCGCGGACAGTAAAC-3′ (forward, from 12 to 31 bp in the database sequence, GenBank # M73482)17 and 5′-CCCATGGCCTTCACACAGGT-3′ (reverse, from 617 to 636 bp) to amplify the 0.6-kb fragment of NMB-R. PreproGRP was amplified by two-step PCR, consisting of the first and nested PCR with the following primer sets: the outer primer set consisting of 5′-TGCTGGCGCTGGTCCTCT-3′ (forward, from 84 to 101 bp in the database sequence, GenBank # K02054)19 and 5′-GCTCGGGGCCTTGAATCTAA-3′ (reverse, from 746 to 765 bp) to amplify the first 0.7-kb fragment, and the inner primer set consisting of 5′-TGGGCGGTGGGGCACTTA-3′ (forward, from 185 to 202 bp) and 5′-TCAGCTGGGGGTTCCTTCCT-3′ (reverse, from 472 to 491 bp) to amplify the nested 0.3-kb fragment of preproGRP. All the primers were designed to span exon-intron boundaries so that amplification of possibly contaminated genomic DNA could be avoided.
The PCR mixture consisted of 1 μL of the synthesized cDNA, 1 × PCR buffer, 0.05 mM of each dNTPs mixture, 0.5 μM of a primer set, and 1 unit of Taq polymerase, in a total volume of 40 μL. For amplification of GRP-R and NMB-R, the reaction was repeated for 40 cycles; each cycle consisted of denaturing at 94 °C for 1 minute, annealing at 55 °C for 1 minute, and synthesis at 72 °C for 1 minute. For the first amplification of preproGRP, the reaction was repeated for 40 cycles; each cycle consisted of denaturing at 94 °C for 30 seconds; annealing at 56 °C for 30 seconds; and synthesis at 72 °C for 45 seconds. For the nested amplification of the gene, 1 μL of the first product in a total volume of 40 μL was re-amplified for 30 cycles, with each cycle the same as the first amplification. At the last cycle in every PCR for the three genes, the synthesis at 72 °C was extended to 5 minutes.
At every PCR, β-actin amplification with the primer set provided by the PCR kit was monitored as a positive control by visualization of the band after being electrophoresed in a gel followed by staining with ethidium bromide.
The PCR products (16 μL/lane) were electrophoresed on 3.0% agarose gel and blotted onto nylon membranes, Hybond N-plus (Amersham Biosciences). Oligonucleotide probes labeled with a fluorescein-dUTP labeling kit (Gene Images 3′-Oligolabelling module, Amersham Biosciences) at the 3′-ends were hybridized with the membranes, and bands were detected with an enhanced chemiluminescence detection kit, CDP-Star detection module (Amersham Biosciences), followed by autoradiography with Hyperfilm-MP (Amersham Biosciences). The sequences of the oligonucleatides for preproGRP, NMB-R and GRP-R were 5′-CAGAAACCACCAGCCACCTCAACCC-3′, 5′-TGGCGGCCGGGGACTTGCTGCTGCT-3′, and 5′-CTTCACACTCACGGCGCTCTCGGCA-3′, respectively.
Review of Bone Scintigram and Bone Marrow Pathology/Cytology
Bone scintigrams were reviewed for diagnosis of bone metastasis. When abnormal accumulations were found, each lesion was investigated by bone radiogram first, and then still–suspicious lesions were examined with bone MRI. All image analyses were performed by a radiologist (K.M.) who was blind to pathologic or RT-PCR results at the time of the image review. All pathologic and cytologic examinations of bone marrow were carried out by a pathologist (K.H.) who was also blind to image or RT-PCR results at the time of the examinations. Diagnosis of bone marrow metastasis required the existence of malignant cell cluster(s) in both or either the pathologic and cytologic specimens.
Frequencies of gene expression according to clinical factors were compared using the Fisher exact test. Univariate analysis with the Cox proportional hazards model assessed risk ratios according to gene expression and other clinical factors. Survival was analyzed using the Kaplan-Meier method and the log-rank test was used for comparisons. Differences with P values less than 0.05 were judged as statistically significant (two-sided). Survival time was calculated from the start of therapy.
Tumor Specific Gene Expression in Bone Marrow and Peripheral Blood
Frequencies of gene expression detected by RT-PCR are summarized in Table 2. Briefly, overall positive rates of preproGRP, NMB-R, and GRP-R in bone marrow aspirate of patients with SCLC were 23% (9/40), 8% (3/40) and 10% (4/40), and rates in peripheral blood were 11% (4/38), 5% (2/38) and 29% (11/38), respectively. The positive rates of gene expression are also compared in the table according to disease stage, existence of bone marrow, and bone metastasis. Expressions of preproGRP and NMB-R were not detected in peripheral blood samples from the 20 healthy volunteers and the 5 patients with NSCLC, while GRP-R expression was detected in 20% (4/20 and 1/5) of both groups. All three gene expressions were detected more frequently in patients with bone marrow metastasis than in those without, with statistical significance. The expression of preproGRP was also more frequent in patients with bone metastasis than in those without, again with statistical significance.
Table 2. Positive Rates of RT-PCR in Bone Marrow and Peripheral Blood Samples
Results of univariate analysis with the Cox proportional hazards model are shown in Table 3. Among the examined genes, only preproGRP expression in bone marrow contributed to poor prognosis in patients with SCLC with statistical significance. Likewise, the survival curve of patients with SCLC with preproGRP expression in bone marrow was significantly poorer than the one without preproGRP expression (Fig. 1).
Table 3. Univariate Analysis of Gene Expressions in Bone Marrow and Peripheral Blood for Correlation with Survival
The expressions of all three SCLC-related genes were observed in a certain population of patients with SCLC both in bone marrow and peripheral blood. Among the three gene expressions, the expression of GRP-R was evident in peripheral blood from patients with NSCLC and even from healthy volunteers, indicating a lack of its specificity for SCLC. The expressions of preproGRP and NMB-R in peripheral blood, however, seemed SCLC-specific because they were exclusively detected in SCLC patients. Although the current study contained no negative control for gene expression in bone marrow because of ethical considerations, the specificity of the three genes in peripheral blood gave rise to the speculation that preproGRP and NMB-R but not GRP-R expression in bone marrow might also be specific for SCLC.
The specificity of GRP-R expression in peripheral blood and bone marrow, however, has been divergent in previous publications. That is, Bessho et al.20 detected no gene expression of GRP-R in peripheral blood and bone marrow from 20 healthy volunteers, whereas Lacroix et al.21 detected it in 7 out of 10 peripheral blood samples from normal subjects. As RT-PCR results are sometimes largely dependent on technical aspects, direct comparison between different research results may not simply lead to definitive conclusions. In any event, similar approaches to detecting other tumor-specific gene expressions, including cytokeratin-19, cytokeratin-20,1, 2, 22 carcinoembryonic antigen,1 and HuD,23 all incur similar problems. It seems clear that such studies should be carried out with the careful selection of a negative control.
On the other hand, Bessho et al.20 detected NMB-R gene expression in 14 out of 44 peripheral blood samples and 2 out of 13 peripheral progenitor cell samples from patients with SCLC, while no NMB-R gene expression was detected in 20 peripheral blood and bone marrow samples from healthy volunteers. Lacroix et al.21 reported the expression rate of preproGRP to be 5 out of 5 SCLC cell lines, while no expression of the gene was detected among 67 peripheral blood samples from healthy volunteers. These studies seem concordant with ours, in which the expressions of preproGRP and NMB-R appeared to be specific for patients with SCLC.
With bone marrow samples, the expression rates of preproGRP, NMB-R, and GRP-R in patients with bone marrow metastasis, and of preproGRP in patients with bone metastasis, were significantly higher than in patients without bone marrow metastasis and without bone metastasis, respectively. In addition, patients with preproGRP expression in the bone marrow samples showed significantly poorer survival than those without this expression. As for the expression in peripheral blood, no significant differences in gene expression rate according to the disease extent and in survival time according to the existence of the gene expression were observed in the limited patient population currently presented. Although multivariate analysis was not performed because of the small patient number, the findings suggest that some tumor-specific gene expressions at least in bone marrow may be related to the disease extent of SCLC.
In addition to its potential as a prognostic factor, this kind of technique could be utilized for other purposes. The fact that some gene expressions of viable cancer cells were detected in peripheral blood or bone marrow samples suggests a potential use of this technique for the diagnosis or biologic investigation of cancers that could not be easily biopsied from the primary site. Further investigations with serial blood and/or bone marrow samples are warranted to see if expression rates are altered according to patients' response to treatment and to determine if minimal residual disease could be detected to predict the risk of relapse in patients who responded to the treatment. This approach, however, has obvious shortcomings for clinical application. Although the gene expressions are specific for some types of cancer, there would be no guarantee that these gene expressions are universal in all patients with the disease. Toi-Scott et al.24 detected gene expressions of GRP-R and NMB-R only in 17 and 11 out of 20 SCLC cell lines, respectively. Therefore, patients with SCLC expressing no such specific genes in nature cannot be investigated with this technique.
In conclusion, the current study revealed the presence of viable tumor cells in peripheral blood and bone marrow in a certain population of untreated patients with SCLC and showed that the RT-PCR detection of preproGRP and NMB-R was specific for SCLC. Because the patients with preproGRP expression in their bone marrow showed poorer prognosis than those without this expression, data obtained by this method might provide some additional information relating to disease extent or prognosis in patients with SCLC. As bone marrow examination is frequently performed as a staging procedure for SCLC, more studies elucidating its clinical relevance are warranted.
The authors thank their colleagues in charge of the patients. They also thank Reiko Kunii for technical assistance and Chieko Handa–Miyagi for secretarial assistance.