Clinical Cancer Research Versailles No Abst AACR Membership
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ho, S. B.
Right arrow Articles by Kratzke, R. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ho, S. B.
Right arrow Articles by Kratzke, R. A.
Clinical Cancer Research Vol. 10, 5777-5784, September 1, 2004
© 2004 American Association for Cancer Research

Molecular Oncology, Markers, Clinical Correlates

Quantification of Colorectal Cancer Micrometastases in Lymph Nodes by Nested and Real-Time Reverse Transcriptase-PCR Analysis for Carcinoembryonic Antigen

Samuel B. Ho1,2, Ann Hyslop1, Richard Albrecht1, Amanda Jacobson1, Michael Spencer1, David A. Rothenberger1,2, Gloria A. Niehans1, John D’Cunha1 and Robert A. Kratzke1,2

Departments of Medicine, Surgery, and Laboratory Medicine, 1 Veterans Affairs Medical Center, and 2 University of Minnesota and the Comprehensive Cancer Center, Minneapolis, Minnesota


   ABSTRACT
Top
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Purpose: Reverse-transcriptase PCR (RT-PCR) assays for carcinoembryonic antigen (CEA) have been described to identify lymph node micrometastases. These assays are not quantitative and can be confounded by false-positive results. The purpose of this study was to determine whether quantification of CEA in lymph nodes could more readily identify clinically relevant groups.

Experimental Design: Specimens included 400 lymph nodes from 64 patients undergoing colon resections. Specimens were tested by immunohistochemistry and by RT-PCR using nested primers for CEA. Specimens from 59 patients that were positive by nested RT-PCR were further quantified by detection of CEA mRNA fluorescence increase at a threshold PCR cycle.

Results: CEA was detected by nested RT-PCR analysis in 4 of 34 (12%) nodes of nonneoplastic disease, 2 of 13 (15%) nodes from T1N0 patients, 32 of 81 (40%) nodes of T2N0 patients, 49 of 109 (45%) nodes from T3N0 patients, and 92 of 163 (56%) nodes from T1–4N1–2 patients. The overall presence of any RT-PCR–detectable CEA in nodes did not differentiate patient groups. Immunohistochemistry was positive in nodes from 7% of T3N0 patients and 100% of T1–3N1–2 patients. CEA quantification revealed that 0 of 7 patients with nonneoplastic disease and 2 of 17 (12%) patients with stage I T1–2N0 cancers had one or more lymph nodes with ≥1.0 x 102 CEA transcripts per sample. In contrast, 4 of 13 (31%) patients with stage II T3N0 cancer and 10 of 22 (45%) stage III patients with known metastases had lymph nodes with ≥1.0 x 102 CEA transcripts.

Conclusions: These data suggest that quantification of CEA levels in lymph nodes may more accurately identify patients at risk for cancer recurrence than does routine nested RT-PCR or immunohistochemistry.


   INTRODUCTION
Top
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Colorectal cancer is the second leading fatal malignancy in the United States. The probability of cure is directly related to the stage of the cancer when it is removed by surgery. In stage I disease (Dukes’ stage A, B1, or T1–2N0M0), ~10 to 15% of patients will have a recurrence of the cancer. This recurrence rate increases to 25% in stage II disease (Dukes’ stage B2, or T3N0M0), and 65% in stage III (Dukes’ stage C, or T1–4N1–2M0; Ref. 1 ). Recurrence of cancer in patients without identifiable metastases has been thought to be due to residual cancer in an occult or microscopic stage. Detection of lymph node metastases in patients with colorectal cancer is clinically important because these patients have been shown to have improved survival if they are treated with postoperative adjuvant chemotherapy (2) .

Detection of carcinoembryonic antigen (CEA) and other tumor marker mRNA transcripts by use of reverse transcriptase-PCR (RT-PCR) techniques has been reported to identify micrometastatic disease in lymph nodes that appear histologically benign. This has been reported to identify patients at increased risk for cancer recurrence in some (3 , 4) , but not all studies (5) . One difficulty with this procedure is the frequently reported identification of CEA-positive transcripts in bone marrow, lymph nodes, and blood from patients with nonneoplastic disease (6, 7, 8, 9) . This may result from contamination that occurs during specimen collection. Alternatively, this may be due to very low levels of normal intestinal epithelial cells released into the blood or lymph circulation, or to low-level "illegitimate transcription" by nonepithelial cells (10 , 11) . Furthermore, standard RT-PCR assays are not quantitative, and the prognostic importance of low-level micrometastases may be quite different from the implications for patients with higher levels of PCR-detectable cells (6) .

This study was designed to determine the relationship between CEA detected in lymph nodes by immunohistochemistry and by a highly sensitive nested RT-PCR analysis with a quantitative measurement of CEA transcripts determined by real-time PCR (12) . We found that nested RT-PCR detected CEA transcripts in some lymph nodes from patients with benign or nonneoplastic colon disease and that these CEA transcripts were below the level of detection of a quantitative PCR technique. High levels of CEA transcripts were found most frequently in patients with more advanced histologic disease and in those with established nodal metastases. Immunohistochemistry rarely identified any patient with histology-negative disease. These data suggest that quantification of CEA levels in lymph nodes may more accurately identify patients at risk for cancer recurrence than either immunohistochemistry or routine nested PCR.


   MATERIALS AND METHODS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Specimen Collection.
Specimens of primary colorectal cancers and lymph nodes were collected at the time of surgery in a prospective fashion with methods designed to avoid contamination between specimens, as described previously (13) . Sterile instruments and gloves were used to dissect all lymph nodes out of the pericolic fat, with care given to avoid contact with the primary tumor and normal mucosa. All identifiable lymph nodes were bisected, and half were preserved at –70°C for RNA isolation and the other half placed in formalin for hematoxylin and eosin histochemistry and immunohistochemistry analyses. Gloves and scalpel blades were then changed, and a representative section of the primary tumor was harvested and frozen immediately at –70°C; the remainder of the tumor was placed in formalin. Cancers were staged after surgical and pathology analysis according to the TNM classification by the American Joint Committee on Cancer (14) . This consisted of patients with stage I (T1N0M0 and T2N0M0), stage II (T3N0M0 and T4N0M0), and stage III disease (TanyN1M0 and TanyN2M0, with N1 indicating metastasis in one to three regional lymph nodes and N2 indicating metastasis in four or more regional lymph nodes). Specimens were obtained with Institutional Review Board approval.

mRNA Isolation.
Total cell RNA was extracted from sample tissue by cell lysis with Trizol reagent (Life Technologies, Inc., Rockville, MD), as we have described previously (15) . Approximately 50 to 100 mg of frozen tissue were used for each assay. Extraction of total RNA was performed according to the manufacturer’s protocol, with an expected yield of 100 µg of RNA from each sample. The RNA was evaluated for quality by electrophoresis and ethidium bromide staining. All RNA samples were treated with DNase before RT-PCR processing to eliminate any potential genomic DNA. RNA was quantitated by spectrophotometry.

DNA Sequencing.
Representative nested PCR products were run on a 1% agarose gel and stained with ethidium bromide. The products were excised from the gel, isolated by use of Geneclean II (Bio101, Vista, CA), and sequenced by the Sanger dideoxy-mediated chain termination method (16) using Fmol DNA Cycle Sequencing System (Promega, Madison, WI).

CEA mRNA Standard Curves.
For these experiments the CEA-expressing carcinoma cell line HTB-174 and the CEA-negative fibroblast cell line NIH/3T3 (American Type Culture Collection, Rockville, MD) were cultured to near 80% confluence, washed with sterile PBS solution, trypsinized, centrifuged, washed, and quantitated with a hemocytometer. CEA-expressing cells (10-fold serial dilutions) were mixed with 107 NIH/3T3 cells (background cells), and total cellular RNA was extracted (15) . In addition, a standard source of CEA cDNA was made for standardization of assays. A human CEA pBluescript plasmid (587150 Z017C08.R1; ATCC no. 745558R) was obtained from the American Type Culture Collection. This was transformed into Escherichia coli DH5{alpha} competent cells (Invitrogen Life Technologies, Inc., Carlsbad, CA) according to the manufacturer’s protocol, and the E. coli were incubated overnight on agar plates selective for ampicillin resistance. Individual colonies were selected and grown in LB medium containing ampicillin. The plasmid DNA was isolated by use of the Wizard Plus Midiprep kit (Promega). We transcribed the RNA from the purified DNA, using the RNA polymerase as recommended by the manufacturer. RNA was quantified by spectrophotometry and stored at a concentration of 6 x 1010 copies/mL. Dilutions of this standard and of RNA purified from HTB-174 cells were used to prepare cDNA.

Preparation of cDNA.
Total cellular RNA was reverse transcribed with the Gene-Amp system (PE Corporation, Norwalk, CT) with random hexamer priming, as described previously (13) . Briefly, random hexamers (2.5 mmol/L) were annealed to 5 µg of total RNA for 10 minutes at 25°C. Reverse transcription (Multiscribe Reverse Transcriptase; PE Corporation) was performed for 30 minutes at 48°C, followed by enzyme inactivation at 95°C for 5 minutes.

Standard RT-PCR.
Standard RT-PCR for CEA was performed with a nested PCR approach, as described previously (15) . In brief, CEA primers that spanned a 160-bp fragment of the CEA open reading frame were synthesized and purified commercially (Genosys Biotechnology, Woodland, TX). Primer sequences used were as follows: primer A, 5'-TCTGGAACTTCTCCTGGTCTCTCAGCTGG-3'; primer B, 5'-TGTTAGCTGTTGCAAATGCTTTAAGGAAGAAGC-3'; and primer C, 5'-GGGCCACTGTCGGCATCATGATTGG-3'. Because the two reactions share one primer, this method is also called a "semi-nested" PCR reaction. A MJ Research PTC-200 thermocycler (Watertown, MA) was used for amplifications. A total of 200 ng of cDNA were amplified for 20 cycles of PCR at 95°C (1 minute) and 72°C (2 minutes), using primers A and B (200 nM), in a 25-µL final volume of 20 mmol/L Tris (pH 8.4); 50 mmol/L potassium chloride; 1.5 mmol/L magnesium chloride; 200 µmol/L each of dATP, dCTP, dGTP, and dTTP; and 0.05 units of TaqDNA polymerase (Life Technologies, Inc.). A final extension step was performed at 72°C for 10 minutes.

For second-round PCR, 1 µL of the first-round PCR product was subjected to 30 cycles of PCR at 95°C (1 minute), 69°C (1 minute), and 72°C (2 minutes), using primers B and C. Final extension was again performed at 72°C for 10 minutes. For analysis of PCR products, 10 µL of PCR product were analyzed by 2% agarose gel electrophoresis, with ethidium bromide staining used for PCR product analysis. In initial experiments, the 131-bp PCR product was eluted, and its identity was confirmed by sequence analysis. Analysis of serial dilutions of CEA-expressing HTB-174 cells indicated that the lower detection limit of this assay was 1,000 to 10,000 HTB-174 cells in 1 x 107 NIH3T3 cells, or 1 CEA-expressing cancer cell in 1,000 to 10,000 fibroblast cells.

Similarly, standard RT-PCR for actin was performed with a single-round PCR approach, as described previously (15) . This was performed on all samples initially to confirm the presence of viable mRNA. Previous studies have determined that the quantity of actin mRNA does not differ significantly among lymph node samples (15) .

Quantitative CEA RT-PCR.
For quantitative RT-PCR, a TaqMan probe and primer set for CEA was constructed with the PE Biosystems primer design software (PE Biosystems, Foster City, CA). The TaqMan probe was obtained commercially (PE Biosystems). Primer sequences spanned intron-exon junctions to avoid contaminating quantitative RT-PCR product with genomic DNA. To confirm the total gene specificity of sequences derived for primers and probes, we performed BLASTn (National Center for Biotechnology Information, Bethesda, MD) searches against dbEST and the nonredundant set of GenBank, European Molecular Biology Laboratory, and DNA Data Bank of Japan database sequences. In a 50-µL PCR reaction (final volume after cDNA was added), components included 1x TaqMan buffer; 200 mmol/L each of dATP, dCTP, dGTP, and dTTP, and dUTP; 5 mmol/L magnesium chloride, 1.25 units of AmpliTaq Gold (PE Biosystems); 0.5 units of AmpErase uracil N-glycosylase (PE Biosystems); 200 nmol/L 5' primer (5'-GAGGCTCCTGCTCACAGCC-3'), 200 nmol/L 3' primer (5'-TCAATAGTGAGCTTGGCAGTGG-3'), and 100 nmol/L TaqMan probe (5'-CACTTCTAACCTTCTGGAACCCGCCC-3'). We then added 200 ng of cDNA and subjected the mixture to the following cycling conditions in the PE Biosystems Prism 7900HT: 50°C for 2 minutes, 95°C for 10 minutes, and 45 cycles at 95°C for 30 seconds, at 68°C for 30 seconds, and at 72°C for 60 seconds. Each assay included a standard curve (described previously) and no-template control. Samples were processed in triplicate. Products were confirmed by 2% agarose gel electrophoresis with ethidium bromide staining.

In initial experiments, the expected 69-bp product was also eluted from the agarose gel, with its identity confirmed by sequence analysis. For data analysis, fluorescence emission from the reporter dye (6-carboxyfluorescein), a covalent modification to the 5' end of the TaqMan probe, exhibited peak fluorescence emission at 518 nm. A charge-coupled-device camera on the PE Biosystems Prism 7900HT instrument collected the laser-excited emission from each sample for real-time analysis of data. With the threshold cycle defined as the fractional cycle number at which the fluorescence increases with respect to background, the relative amounts of mRNA were calculated. Where appropriate, results are reported as mean ± SEM.

This assay is slightly more sensitive than the standard nested RT-PCR assay, with CEA quantifiable in samples containing as low as 1000 metastatic cells in 1 x 107 fibroblast cells (15) . This typically corresponded to a threshold cycle of less than 42. The coefficient of variation for triplicate reactions was less than 10%. The coefficient of variation between assays was also less than 10%.

For quantitative analysis of the number of CEA-expressing metastatic tumor cells per lymph node sample, standard curves were created from serially diluted CEA-expressing cells in a background of CEA-negative fibroblast cells as described previously (15) , and from serially diluted purified CEA transcripts. Regression analysis of standard curves from both the cell line and the purified CEA transcripts indicated a highly correlated linear relationship between starting quantity and threshold cycle (r = 0.99). Correlation of HTB-174 cells per sample and CEA transcripts per sample corresponded to the equation Y = 111.03 + 94.67X (r2 = 1.000) where Y is HTB-174 cells per sample and X is CEA transcript number per sample. For example, a CEA transcript number per sample of 1 x 102 is equal to ~9,578 HTB-174 cells per sample, which corresponds to a threshold cycle of 40. This is close to the lower limit of sensitivity for this assay. A CEA transcript number of 1 x 105 is equal to ~9.47 x 106 HTB-174 cells per sample. Quantities above this number are approximate and near the maximal quantification for this assay. A quantity ≥1 x 102 transcripts per specimen was considered a "high" level, because a previous study by Miyake et al. (17) found that CEA in lymph nodes quantified by real-time PCR was, on average, 102 transcripts per specimen and that the CEA level was always above this level in clinically significant, histologically positive nodes.

Immunohistochemistry.
The streptavidin-peroxidase technique was used as described previously (18) . Briefly, tissue sections were deparaffinized, rehydrated, incubated with fresh 3% hydrogen peroxide in methanol for 10 minutes, and then washed with PBS. Antigen retrieval was performed by high-pressure heat denaturation using a sodium citrate buffer in a pressure cooker at 15 psi for 5 to 10 minutes. Normal rabbit serum (5%) in 1% milk was applied for 20 minutes and removed by blotting. The sections were then incubated with a mouse monoclonal antibody against CEA (clone 11-7; Dako Corporation, Carpinteria, CA) for 90 minutes at room temperature, washed three times in buffer, and incubated with a biotinylated secondary antibody (rabbit antimouse IgG; 1:75 dilution in PBS) for 20 minutes. After washing, the sections were incubated with streptavidin-peroxidase conjugate (10 µg/mL) for 30 minutes, followed by repeated washing. The sections were incubated with diaminobenzidine in 0.03% hydrogen peroxide for 10 minutes, washed, counterstained with methyl green, rinsed in tap water, and mounted. Positive immunoreactivity was graded as any positive staining in cytologically malignant cells. Biotinylated rabbit antimouse IgG, streptavidin-peroxidase conjugate, and normal rabbit serum were obtained from Zymed Laboratories (San Francisco, CA). Diaminobenzidine was obtained from the Sigma Chemical Co. (St. Louis, MO). Negative controls included substituting similar dilutions of irrelevant antibodies for the primary antibody, which gave negative staining.

Statistical Analysis.
Variables are summarized using means and SEM as appropriate. Differences between groups were determined by use of contingency tables and {chi}2 analysis. All comparisons were two-tailed and a P value <0.05 was considered significant. The Prism v. 2.0 statistical software was used (GraphPad Software, Inc, San Diego, CA).


   RESULTS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Specimens from 64 patients having colon resections were analyzed, and the characteristics of these patients are listed in Table 1Citation . From these specimens, 400 lymph nodes were studied by RT-PCR analysis using nested primers for CEA sequences (mean, 6.2 nodes per case). Patients were classified as followed: colon cancer and histology-negative lymph nodes T1N0 (n = 4), T2N0 (n = 13), T3N0 (n = 16); colon cancer with histology-positive lymph nodes T1–4N1–2 (n = 24); and patients with nonneoplastic disease (n = 7). Specimens were processed by RT-PCR using nested primers for CEA. For 59 patients, the specimens that were positive by nested RT-PCR were further quantified by a real-time PCR assay (PRISM 7900HT; Perkin-Elmer), using a standard curve calculated from CEA-positive HTB-174 cancer cells and purified CEA transcripts.


View this table:
[in this window]
[in a new window]

  		Table 1 Patient characteristics

 
The percentages of lymph node samples that were positive by nested RT-PCR in each patient category are listed in Fig. 1Citation . CEA was detected by nested RT-PCR analysis in 4 of 34 (11.8%) nodes of nonneoplastic disease. The four CEA-positive nodes were from a specimen with colonic volvulus; nodes from patients with diverticular or hyperplastic disease were negative. In colon cancer specimens, CEA was detected by nested RT-PCR in 2 of 13 (15%) nodes from T1N0 patients, 32 of 81 (40%) nodes of T2N0 patients, 49 of 109 (45%) nodes from T3N0 patients, and 92 of 163 (56%) nodes from T1–4N1–2 patients. The frequency of any RT-PCR–detectable CEA in nodes was significantly higher in patients with histologically positive nodes and patients with T3N0 and T2N0 disease compared with nodes from patients with nonneoplastic disease (Fig. 1)Citation .



View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 1. Overall detection of CEA mRNA by nested RT-PCR in lymph node groups. *, P < 0.01 vs. patients with nonneoplastic disease (N); **, P = 0.07 vs. stage I (T1N0), P < 0.001 vs. patients with nonneoplastic disease; ***, P < 0.02 vs. patients with nonneoplastic disease and patients with stage I (T1N0) and stage 2 (T2N0) disease.

 
The overall presence of any RT-PCR–detectable CEA in nodes did not differentiate patients in the T1N0, T2N0, T3N0, and T1–4N1–2 groups. In T1N0 patients, 2 of 4 patients had at least one positive node; of T2N0 patients, 10 of 13 had at least one positive node; of T3N0 patients, 11 of 16 had at least one positive node; and of T1-4N1–2 patients, 19 of 24 had at least one positive node. Of the histology node-positive stage III patients in this study, RT-PCR identified CEA micrometastases in 6 of 11 patients who each had five or fewer nodes available for RT-PCR analysis, whereas CEA transcripts were detected in one or more of examined lymph nodes from 13 of 13 patients who each had six or more nodes available for analysis.

We determined the quantity of CEA in 125 lymph node specimens with a positive CEA signal determined by nested RT-PCR. As shown in Fig. 2Citation , CEA levels in nodes of patients with nonneoplastic disease were extremely low and not detectable by real-time PCR, or they represented false-positive results. CEA levels were quantifiable in only one positive node from a patient with a T1N0 cancer (CEA transcript number, 2.2 x 104) and in 8.7% of PCR-positive nodes from patients with T2N0 cancer (mean CEA transcript number, 1.2 x 104 ± 8400). In contrast, higher mean CEA transcript levels were found in patients with T3N0 (6.3 x 105 ± 5 x 105) and T1–4N1–2 disease (2.2 x 105 ± 7 x 104).



View larger version (22K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 2. Quantification of CEA transcript levels in specimens CEA positive by nested RT-PCR. The mean ± SEM for the CEA transcript level for each group is indicated in the bottom row. Specimens with undetectable CEA represent samples that were falsely positive by nested RT-PCR or samples with extremely low CEA levels detected by nested RT-PCR but not by quantitative PCR.

 
The majority of CEA levels in nodes from T1–4N1–2 patients were ≥102 transcripts (Fig. 3A)Citation . Using the quantity of ≥1.0 x 102 transcripts to designate high levels of CEA in nodes, we found that 0 of 7 patients with nonneoplastic disease, 1 of 4 (25%) patients with T1N0 cancers, and 3 of 13 (23%) patients with T2N0 cancers had any lymph nodes with high levels of CEA transcripts; whereas 4 of 13 (31%) patients with T3N0 cancer and 10 of 22 (45%) patients with known metastases had lymph nodes with high levels of CEA transcripts (Fig. 3B)Citation . Among the patients with T1–4N1–2 or stage III disease, more nodes from patients with N2 disease had high levels of quantified CEA (31 of 51; 61%) compared with nodes from patients with N1 disease (7 of 17; 41%).



View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 3. A, percentage of quantified nodes with 1 x 102 or more CEA transcripts per specimen. B, percentage of patients with high levels of CEA (≥1 x 102 transcripts per specimen) in one or more lymph nodes. N, patients with nonneoplastic disease.

 
Immunohistochemical analysis was also performed on serial sections of nodes examined by histochemistry with a monoclonal antibody against CEA. Good correlation of CEA-positive staining with lymph nodes previously identified as histologically positive was observed. In contrast, only one patient categorized as having histologically negative nodes (T3N0) had positive staining of a lymph node for CEA antigen (Table 2)Citation . This consisted of one single foci of tumor thrombus in the periphery of a lymph node (Fig. 4)Citation . This single patient also had one of four lymph nodes with detectable CEA mRNA at a level of 6.7 x 104 transcripts [see Table 3Citation , stage II (T3N0) patient 7].


View this table:
[in this window]
[in a new window]

  		Table 2 Immunohistochemical (IHC) staining for CEA in lymph nodes

 


View larger version (152K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 4. Lymph node micrometastases (arrows) identified by immunohistochemistry using a monoclonal antibody against CEA in a patient with negative lymph nodes (stage T3N0) by routine histochemistry (magnification, x100).

 

View this table:
[in this window]
[in a new window]

  		Table 3 Patient outcomes by stage and lymph node status

 
The clinical outcomes of patients with stage I and II node-negative colorectal cancer are listed in Table 3Citation . There was no correlation of cancer recurrence or overall mortality with the lymph node findings in this small group of patients. Only 1 patient of 31 evaluable patients had a clinical recurrence of colorectal cancer with liver metastases, and this patient was not identified by the presence of CEA in resected lymph nodes [Table 3Citation , stage I (T2N0), patient 3]; this patient also had long-standing ulcerative colitis.


   DISCUSSION
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The standard technique used to evaluate lymph nodes for the presence of metastases from solid tumors is light-microscopic assessment of paraffin-embedded tissue stained with hematoxylin and eosin. Since the 1980s, investigators have used various techniques in an attempt to improve the detection of metastatic cells in lymph nodes. Results from studies using immunohistochemistry to increase the prognostic accuracy of staging in node-negative colorectal carcinoma have been contradictory (19, 20, 21, 22, 23, 24) . More recently, sensitive RT-PCR techniques have been used as a method to detect cancer metastases in lymph nodes, bone marrow, and blood (25) . These techniques use RNA purified from the clinical specimen, which is then reverse-transcribed to cDNA, and then use PCR to amplify a specific tumor marker sequence. To maximize sensitivity, most of these PCR techniques involve the use of nested primers and two successive rounds of 25 to 40 cycles of amplification.

Several studies in relatively small numbers of patients have reported that nested RT-PCR methods frequently detect cancer micrometastases in lymph nodes of patients with colorectal cancer. Mori et al. (26) detected CEA mRNA in 47 of 87 histologically negative lymph nodes from 13 patients with colorectal cancer, whereas CEA was not detected in 15 lymph nodes from patients without colorectal cancer. The use of this RT-PCR technique increased the positive lymph node rate from 26% by histologic examination to 66% by RT-PCR. Leifers et al. (3) analyzed 192 lymph nodes from 26 patients with stage II colorectal cancer, using a CEA-specific nested RT-PCR. CEA transcripts were detected in one or more lymph nodes from 14 of 26 patients (54%). In this group of patients, the presence of CEA transcripts in lymph nodes was associated with a higher rate of cancer deaths. Bernini et al. (13) used a nested RT-PCR technique to detect MUC2 mucin transcripts in lymph nodes from patients with colorectal cancer. They found MUC2 transcripts in nodes from 1 of 16 (6%) patients with stage I colorectal cancer (T1–2N0) and in 11 of 27 (41%) patients with stage II colorectal cancer (T3–4N0). Similarly, Futamura et al. (27) found that 102 of 202 (50.5%) lymph nodes from 13 colorectal cancer patients with histologically negative nodes were positive for CEA and/or cytokeratin-20 using RT-PCR.

Studies using RT-PCR have also been criticized as being prone to false positives or illegitimate expression, and not all studies have demonstrated any prognostic significance (6) . For example, Zippelius et al. (7) found that 5 of 19 control patients without cancer had CEA transcripts detected by RT-PCR in the bone marrow. Similarly, Bostick et al. (8) found CEA detected by RT-PCR and Southern blot analysis in 3 of 3 lymph node specimens from patients without cancer and in 6 of 13 blood samples from normal donors. Manipulation of the annealing temperatures, concentrations of components (such as magnesium, deoxynucleotide triphosphates, and polymerase), and amplification cycle numbers can influence the detection limits of RT-PCR assays. In addition, the use of oligo(dT) primers instead of random hexamers in the reverse transcription steps may improve the specificity of a nested PCR technique, although the sensitivity may also change. Because standard RT-PCR provides a qualitative rather than a quantitative result, assay variations can considerably confound the results.

In this study we used a very sensitive nested RT-PCR technique to detect CEA mRNA and showed that CEA transcripts are increasingly detected in lymph nodes of primary cancers with increasing T stage. However, CEA transcripts were also found in nodes from some patients with nonneoplastic disease and in patients with stage T1N0 cancers, which are not expected to represent clinically significant micrometastases. We also found CEA transcripts in greater numbers of histologic node-negative patients than previously reported by others (3 , 13) . The relevance of these standard nested RT-PCR results was clarified by the use of quantitative RT-PCR. The quantitated CEA transcript level in nodes from patients with nonneoplastic disease and stage T1N0 cancers was low to undetectable, indicating that these nested-PCR–positive samples may represent illegitimate transcripts, "normal" epithelial cells, or RNA that escape into the lymph circulation, or they may represent low-level contamination from collection procedures. Conversely, patients with histologic metastases frequently have high levels of CEA transcripts measurable in lymph nodes. Interestingly, 31% of 14 patients with T3N0 cancers were found to have similarly "high" CEA transcript levels in resected lymph nodes. We speculate that these patients are at increased risk for cancer recurrence compared with comparable patients with low to nonquantifiable CEA transcript levels. No clinical cancer recurrence or metastases occurred in our small group of patients with T3N0 disease over a mean follow-up of 36 months, and five of these patients died from apparently unrelated causes. Further study of larger cohorts of patients with stage II colorectal cancer without comorbid diseases will be required to demonstrate the clinical utility of this assay for predicting cancer recurrence. Given the preliminary results in the present trial, over 100 patients with stage II colorectal cancer would need to be studied to have 80% power to detect a 20% difference in survival.

The CEA transcript levels found in lymph nodes of patients with colorectal cancer in this study were similar to the levels reported in a small number of nodes studied by Miyake et al. (17) . They used a variant of real-time PCR that included nonspecific DNA-binding fluorescent dyes and used a CEA-expressing cell line as a standard control. They found that the CEA transcript level in 16 histologically positive nodes were all greater than 1 x 102 transcripts and ranged from 1.3 x 103 to 5.7 x 106 transcripts; they found a somewhat lower level in 10 nodes that were histologically negative but RT-PCR positive (2.3 x 101 to 8.1 x 105 transcripts). A disadvantage of this study is the lack of specificity because of the requirement for an additional step of double-stranded DNA-based fluorescence melting curve analysis. The technique in the present study used a CEA-sequence–specific probe for the quantitative RT-PCR.

We did not find CEA by nested RT-PCR in all patients with histologically positive lymph nodes. This could be because we sampled many fewer nodes by RT-PCR than by routine histology. Of the histologic node-positive stage III patients in this study, RT-PCR identified CEA micrometastases in 6 of 11 patients who each had five or fewer nodes available for RT-PCR analysis, whereas CEA transcripts were detected in one or more of examined lymph nodes from 13 of 13 patients who each had six or more nodes available for analysis. Another limitation of this study is that nodes less than 5 mm in diameter were not sampled for RNA analysis because of difficulties in handling such small specimens. Previous studies have shown that small lymph nodes are just as likely as larger nodes to contain histologic metastases (28) . Furthermore, many histologic studies have shown that at least 12 nodes per resection is a minimum number to find or exclude metastases (29 , 30) . Therefore, sampling error can contribute to finding nodal metastases, and future studies using RT-PCR techniques should also attempt to sample at least 6 nodes and preferably up to 12 nodes per resection.

We found that use of immunohistochemistry to detect CEA in serial sections of lymph nodes identified micrometastases in one patient categorized as T3N0. Previous studies have shown that simply reexamining original or re-cut hematoxylin-and-eosin–stained nodes can identify additional metastases in 5 to 9% of patients with breast or colon cancer (24 , 31) . The additional use of immunohistochemical techniques with CEA or cytokeratin antibodies to examine nodes of colorectal patients in 15 reported series led to "upstaging" 0 to 68% of patients, with most studies (9 of 15) reporting that 25 to 39% of patients were upstaged. In 10 of 14 of these studies, the identification of metastases by immunohistochemical methods did not provide any prognostic significance for these patients (19, 20, 21, 22, 23, 24 , 32) . Recent practice guidelines have called for a minimum of 12 nodes to be examined histologically with each colorectal cancer resection (30 , 33) . One report has documented that during the 1990s pathologists increased the number of nodes that were routinely examined after a colon resection, which has led to upstaging of patients (29) . This change in routine practice would most likely lessen the impact of using immunohistochemical techniques on additional node sections and may explain the wide variation in the clinical utility of reported series using immunohistochemistry to identify micrometastases.

Noguchi et al. (34) also directly compared the sensitivity of RT-PCR and immunohistochemistry for the detection of micrometastases. Breast cancers and lymph nodes were tested for the presence of the MUC1 mucin protein by immunohistochemistry, and this was compared with detection of MUC1 mRNA by RT-PCR. They found 9 of 50 histologically negative lymph nodes from breast cancer patients to be positive for MUC1 protein by immunohistochemistry. In the remaining 41 lymph nodes, a total of 6 more nodes were found to be positive for MUC1 mRNA by RT-PCR. Immunohistochemical techniques are susceptible to lack of standards for node examinations, differences in staining technique and antibodies, and antibody cross-reactivity, which may contribute to the wide variation in results reported (19) . These data and the data in the present study highlight the increased sensitivity for detection of micrometastases provided by RT-PCR compared with immunohistochemical analysis.

In conclusion, this is the largest reported series to date demonstrating the utility of real-time quantitative PCR for the detection of CEA transcripts in lymph nodes of patients with colorectal cancer. This technique is more sensitive than immunohistochemical techniques for CEA detection. In addition, quantitative PCR is more specific than standard RT-PCR because no nodes from patients with nonneoplastic colon disease and few nodes from patients with stage I cancers had quantifiable CEA transcripts. Furthermore, nodes that contain at least 1 x 102 CEA transcripts were more likely in patients with T3N0 or stage II disease, who can be expected to have a greater chance of recurrence. These data suggest that categorization of lymph node micrometastases by the quantity of CEA transcripts is a more accurate way to identify clinically relevant patient groups. Further studies should be performed with greater numbers of patients to establish the clinical utility of this technique.


   FOOTNOTES
 
Grant support: NIH grant R21 CA81558-02, a grant from the Minnesota Medical Foundation, and the Research Service of the Minneapolis Veterans Affairs Medical Center.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: Samuel B. Ho, Gastroenterology Section 111-D, VA Medical Center, 1 Veterans Drive, Minneapolis, MN 55417. Phone: 612-725-2000, ext, 4109; Fax: 612-725-2248; E-mail: Samuel.Ho{at}med.va.gov

Received 10/30/03; revised 5/12/04; accepted 5/19/04.


   REFERENCES
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Bresalier RS. Malignant neoplasms of the large intestine Feldman M. Friedman L. S. Sleisenger M. H. eds. . Sleisenger and Fordtran’s gastrointestinal and liver disease, 2215-61, W. B. Saunders Philadelphia 2002.
  2. Moertel CG. Chemotherapy for colorectal cancer. N Engl J Med, 330: 1136-42, 1994.[Free Full Text]
  3. Liefers GJ, Cleton-Jansen AM, van de Velde CJ, et al Micrometastases and survival in stage II colorectal cancer. N Engl J Med, 339: 223-8, 1998.[Abstract/Free Full Text]
  4. Noura S, Yamamoto H, Ohnishi T, et al Comparative detection of lymph node micrometastases of stage II colorectal cancer by reverse transcriptase polymerase chain reaction and immunohistochemistry. J Clin Oncol, 20: 4232-41, 2002.[Abstract/Free Full Text]
  5. Clarke GA, Ryan E, Crowe JP, O’Keane JC, MacMathuna P. Tumour-derived mutated K-ras codon 12 expression in regional lymph nodes of stage II colorectal cancer patients is not associated with increased risk of cancer-related death. Int J Colorectal Dis, 16: 108-11, 2001.[CrossRef][Medline]
  6. Zippelius A, Pantel K. RT-PCR-based detection of occult disseminated tumor cells in peripheral blood and bone marrow of patients with solid tumors. An overview. Ann N Y Acad Sci, 906: 110-23, 2000.[CrossRef][Medline]
  7. Zippelius A, Kufer P, Honold G, et al Limitations of reverse-transcriptase polymerase chain reaction analyses for detection of micrometastatic epithelial cancer cells in bone marrow. J Clin Oncol, 15: 2701-8, 1997.[Abstract/Free Full Text]
  8. Bostick PJ, Chatterjee S, Chi D, et al Limitations of specific reverse-transcriptase polymerase chain reaction markers in the detection of metastases in the lymph nodes and blood of breast cancer patients. J Clin Oncol, 16: 2632-40, 1998.[Abstract]
  9. Ko Y, Klinz M, Totzke G, Gouni-Berthold I, Sachinidis A, Vetter H. Limitations of the reverse transcription-polymerase chain reaction method for the detection of carcinoembryonic antigen-positive tumor cells in peripheral blood. Clin Cancer Res, 4: 2141-6, 1998.[Abstract]
  10. Chelly J, Concordet JP, Kaplan JC, Kahn A. Illegitimate transcription: transcription of any gene in any cell type. Proc Natl Acad Sci USA, 86: 2617-21, 1989.[Abstract/Free Full Text]
  11. Jung R, Kruger W, Hosch S, et al Specificity of reverse transcriptase polymerase chain reaction assays designed for the detection of circulating cancer cells is influenced by cytokines in vivo and in vitro. Br J Cancer, 78: 1194-8, 1998.[Medline]
  12. Bustin SA. Absolute quantification of mRNA using real-time reverse transcription polymerase chain reaction assays. J Mol Endocrinology, 25: 169-93, 2000.[Abstract]
  13. Bernini A, Spencer M, Frizelle S, et al Evidence for colorectal cancer micrometastases using reverse transcriptase-polymerase chain reaction analysis of MUC2 in lymph nodes. Cancer Detect Prev, 24: 72-9, 2000.[Medline]
  14. American Joint Committee on Cancer. . Manual for staging of cancer, edition 5 J. B. Lippincott Philadelphia 1997.
  15. D’Cunha J, Corfits AL, Herndon JE, 2nd, et al Molecular staging of lung cancer: real-time polymerase chain reaction estimation of lymph node micrometastatic tumor cell burden in stage I non-small cell lung cancer—preliminary results of Cancer and Leukemia Group B Trial 9761. J Thorac Cardiovasc Surg, 123: 484-91, discussion, 491 2002.[Abstract/Free Full Text]
  16. Sanger F, Nicklen S, Coulson A. DNA sequencing with chain terminating inhibitors. Proc Natl Acad Sci USA, 74: 5463-7, 1977.[Abstract/Free Full Text]
  17. Miyake Y, Fujiwara Y, Ohue M, et al Quantification of micrometastases in lymph nodes of colorectal cancer using real-time fluorescence polymerase chain reaction. Int J Oncol, 16: 289-93, 2000.[Medline]
  18. Ho SB, Niehans GA, Lyftogt C, et al Heterogeneity of mucin gene expression in normal and neoplastic tissues. Cancer Res, 53: 641-51, 1993.[Abstract/Free Full Text]
  19. Calaluce R, Miedema BW, Yesus YW. Micrometastasis in colorectal carcinoma: a review. J Surg Oncol, 67: 194-202, 1998.[CrossRef][Medline]
  20. Isaka N, Nozue M, Doy M, Fukao K. Prognostic significance of perirectal lymph node micrometastases in Dukes’ B rectal carcinoma: an immunohistochemical study by CAM5.2. Clin Cancer Res, 5: 2065-8, 1999.[Abstract/Free Full Text]
  21. Oberg A, Stenling R, Tavelin B, Lindmark G. Are lymph node micrometastases of any clinical significance in Dukes Stages A and B colorectal cancer?. Dis Colon Rectum, 41: 1244-9, 1998.[CrossRef][Medline]
  22. Nakanishi Y, Ochiai A, Yamauchi Y, Moriya Y, Yoshimura K, Hirohashi S. Clinical implications of lymph node micrometastases in patients with colorectal cancers. A case control study. Oncology, 57: 276-80, 1999.[CrossRef][Medline]
  23. Clarke G, Ryan E, O’Keane JC, Crowe J, MacMathuna P. The detection of cytokeratins in lymph nodes of Duke’s B colorectal cancer subjects predicts a poor outcome. Eur J Gastroenterol Hepatol, 12: 549-52, 2000.[Medline]
  24. Tschmelitsch J, Klimstra DS, Cohen AM. Lymph node micrometastases do not predict relapse in stage II colon cancer. Ann Surg Oncol, 7: 601-8, 2000.[Abstract]
  25. Sidransky D. Nucleic acid-based methods for the detection of cancer. Science (Wash DC), 278: 1054-8, 1997.[Abstract/Free Full Text]
  26. Mori M, Mimori K, Inoue H, et al Detection of cancer micrometastases in lymph nodes by reverse transcriptase-polymerase chain reaction. Cancer Res, 55: 3417-20, 1995.[Abstract/Free Full Text]
  27. Futamura M, Takagi Y, Koumura H, Kida H, Tanemura H, Shimokawa K, Saji S. Spread of colorectal cancer micrometastases in regional lymph nodes by reverse transcriptase-polymerase chain reactions for carcinoembryonic antigen and cytokeratin 20. J Surg Oncol, 68: 34-40, 1998.[CrossRef][Medline]
  28. Herrera-Ornelas L, Justiniano J, Castillo N, Petrelli NJ, Stulc JP, Mittelman A. Metastases in small lymph nodes from colon cancer. Arch Surg, 122: 1253-6, 1987.[Abstract/Free Full Text]
  29. Goldstein NS, Sanford W, Coffey M, Layfield LJ. Lymph node recovery from colorectal resection specimens removed for adenocarcinoma. Am J Clin Pathol, 106: 209-16, 1996.[Medline]
  30. Association of Directors of Anatomic and Surgical Pathology. Recommendations for processing and reporting of lymph node specimens submitted for evaluation of metastatic disease. Am J Clin Pathol, 115: 799-801, 2001.[Free Full Text]
  31. Lilleng PK, Hartveit F. ‘Missed’ micrometastases—the extent of the problem. Acta Oncol, 39: 313-7, 2000.[CrossRef][Medline]
  32. Broll R, Schauer V, Schimmelpenning H, et al Prognostic relevance of occult tumor cells in lymph nodes of colorectal carcinomas: an immunohistochemical study. Dis Colon Rectum, 40: 1465-71, 1997.[CrossRef][Medline]
  33. Compton CC, Cancer Committee, College of American Pathologists. Updated protocol for the examination of specimens from patients with carcinomas of the colon and rectum, excluding carcinoid tumors, lymphomas, sarcomas, and tumors of the vermiform appendix. Arch Pathol Lab Med, 124: 1016-22, 2000.[Medline]
  34. Noguchi S, Aihara T, Nakamori S, et al The detection of breast carcinoma micrometastases in axillary lymph nodes by means of reverse transcriptase-polymerase chain reaction. Cancer (Phila), 74: 1595-600, 1994.



This article has been cited by other articles:


Home page
 Proc. Natl. Acad. Sci. USAHome page
K. S. Garman, C. R. Acharya, E. Edelman, M. Grade, J. Gaedcke, S. Sud, W. Barry, A. M. Diehl, D. Provenzale, G. S. Ginsburg, et al.
A genomic approach to colon cancer risk stratification yields biologic insights into therapeutic opportunities
PNAS, December 9, 2008; 105(49): 19432 - 19437.
[Abstract] [Full Text] [PDF]

Home page
 Clin. Cancer Res.Home page
T. Setoyama, S. Natsugoe, H. Okumura, M. Matsumoto, Y. Uchikado, S. Ishigami, T. Owaki, S. Takao, and T. Aikou
Carcinoembryonic antigen messenger RNA expression in blood predicts recurrence in esophageal cancer.
Clin. Cancer Res., October 15, 2006; 12(20): 5972 - 5977.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ho, S. B.
Right arrow Articles by Kratzke, R. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ho, S. B.
Right arrow Articles by Kratzke, R. A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online