sign in sign up
<<
Home , Carcinoembryonic antigen

ANTICANCER RESEARCH 26: 803-808 (2006)

Quantitative Estimation of CEA and CK20 Expression in Tumour Tissue of and its Liver Metastases with Reverse Transcription and Real-time PCR

MONIKA CERNA1,2*, LUBOS HOLUBEC JR1, MARTIN PESTA1,2, STANISLAV KORMUNDA1, ONDREJ TOPOLCAN1,2 and RADIM CERNY3

1Second Clinic of Internal Medicine, Faculty Hospital Pilsen, E.Benese 13, 305 99 Pilsen; 2Central Laboratory of Radioisotopes and 3Department of Biochemistry, Charles University, Faculty of Medicine, Karlovarska 48, 301 66 Pilsen, Czech Republic

Abstract. Background: Carcinoembryonic antigen (CEA) and remains surgery and the best prognostic method is tumour 20 (CK20) are established tumour markers and the staging based on histopathological examination. Multiple CEA pre-operative levels in serum have prognostic value. The biochemical markers have been suggested and tested for aim of this study was to verify the usefulness of the estimation of screening, aiding diagnosis, determining prognosis and CEA and CK20 gene expressions in tissues of colorectal cancers monitoring the treatment of the disease (1, 2). and their liver metastases. Patients and Methods: Two hundred The carcinoembryonic antigen (CEA or CEACAM 5) and twenty-two specimens of colorectal cancers, their liver holds a special position among all the tumour markers. CEA metastases, other liver tumours and control tissues were analysed is a cell-surface , normally expressed in foetal by reverse transcription combined with real-time PCR. Results: tissue but exhibiting only marginal expression in adults (3). It The expressions of CEA and CK20 were significantly higher in is a member of the immunoglobulin gene superfamily found tumours than in controls; there were differences between tumour on chromosome 19. The function and biological role of CEA types and no relationship to staging or clinical development was is not totally clear, but evidence that it serves as an found. CK20 expression was inversely dependent on grading. The intercellular adhesion molecule has been presented (4), while CEA expression in tumours did not correlate with the CEA levels its role as an intestine mucosa-protecting molecule has also in serum, but did correlate with serum tissue-specific polypeptide been documented (3). CEA has been known for almost 4 antigen (TPS). Conclusion: The measurement of CEA and decades and it is believed that, in normal colon, it is released CK20 gene expressions in tumours did not supply any new from the apical surface of polarised columnar epithelial cells prognostic information, but raised the question of the mechanism into the gut, whereas in dysregulated tumour tissue exfoliated releasing CEA into the blood. CEA is released into the blood and lymphatic vessels through intercellular spaces. In addition, CEA is overexpressed in Colorectal cancer (CRC) is not only the second most many tumours, mainly in CRC. Therefore, in many patients common cancer in the Czech Republic among both males the serum level of CEA rises and pre-operative CEA levels and females, but is also one of the most common cancers provide independent prognostic information and may also worldwide, and is accompanied by significant mortality. help with surgical management (2). Despite all the research efforts, the only curative treatment In contrast to the CEA levels in serum, the significance of the CEA levels in tumour tissue itself remains controversial. Two approaches have been exploited for the estimation of CEA expression in tumour tissue: (5, *Present address: Department of Clinical Genetics, Faculty 6) and determination of CEA in the cytosol or even in the Hospital Pilsen, E.Benese 13, 305 99 Pilsen, Czech Republic. pellet of tumour tissue (7-9). Both methods employ specific to quantify CEA expression. Estimations in Correspondence to: Dr. Radim Cerny, Charles University, Faculty of cytosols and pellet extracts exploit the standard techniques Medicine, Karlovarska 48, 301 66 Pilsen, Czech Republic. Tel: +420 for serum estimation, while immunohistochemistry requires 377593283, Fax: +420 377593149, e-mail: [email protected] the subjective grading of an experienced pathologist who Key Words: Colorectal cancer, liver metastases, CEA, CK20, evaluates the staining strength and the number of positive quantitative RT-PCR, mRNA, grading, tumour markers. cells. In some publications, a higher CEA expression in

0250-7005/2006 $2.00+.40 803 ANTICANCER RESEARCH 26: 803-808 (2006) tumour was connected with invasive growth (5), with a Gerhardt et al. (23), i.e. the CEA downstream primer was 5’- higher risk of local recurrence (6) or was correlated with TCTGGAACTTCTCCTGGTCTCTCAGCTGG-3’, the CEA grading (7). In other reports, no correlation with grading upstream primer was 5’-TGTAGCTGTTGCAAATGCTTTA AGGAAGAAGC-3’ forming a PCR fragment of 160 bp. The CK20 (10) or staging (11), or both (12), was observed. One report primers were designed in our laboratory; the downstream primer was revealed that the CEA content in the tissue pellet correlated 5’-CTGATGCAGATTCGGAGTAACA-3’, while the CK20 with the CEA level in the serum (13). An absence of upstream primer was 5’-TCTCTCTTCCAGGGTGCTTAAC-3’, correlation between CEA levels in the cytosol and serum was giving an amplicon length of 162 bp. All the primers were designed also described, together with the finding that higher CEA to produce PCR products extended across at least one exon-exon levels in the cytosol correlated with better prognosis (14). boundary, so possible RNA contamination with genomic DNA Here, we attempted to verify the usefulness of CEA would produce PCR products of considerably higher length than expected. All the quantitative estimations were done on a Rotor- expression measurement in tumour tissue by estimating the Gene 2000 apparatus (Corbett Research, Sydney, Australia) in a levels of specific CEA mRNA using reverse transcription total volume of 20 Ìl. In this volume, each sample contained 0.5 unit followed by real-time PCR. This technique has been used of Taq DNA-polymerase (Invitrogen, Life Technologies, KRD, repeatedly to detect micrometastases in lymph nodes (15- Prague, Czech Republic), 1x manufacturer’s buffer, 0.2 mM dNTP 17) or the circulating cells (18, 19) in the blood but, (each), 4.5 mM MgCl2, primers 0.6 ÌM each, and 0.5x Sybr-Green surprisingly, it has not yet been used for the estimation of (Molecular Probes Inc., BioTech, Prague, Czech Republic). DNA CEA gene expression in the tumour itself. Another standards for quantification were prepared by cloning individual PCR products into the pGEM-T vector (Promega Corporation, East epithelial cell and tumour marker, cytokeratin 20 (CK20) Port, Prague, Czech Republic). All the cloned standards were (16, 20, 21), was estimated in parallel. verified by DNA sequencing using the ALFexpress sequencer (Pharmacia Biotech, Uppsala, Sweden) and the Thermo Sequenase Patients and Methods fluorescent-labelled primer cycle sequencing with 7-deaza-dGTP (Amersham Pharmacia Biotech, Prague, Czech Republic). Plasmid Tissue samples. In this study, the following tissue samples were DNA was linearized by a suitable restriction endonuclease, included: 25 specimens of benign colon and liver tissues (from 15 quantified spectrophoto-metrically and diluted to concentrations of 3 8 patients, aged 19.2 - 70.3 years, mean 40.1), 12 specimens of colon 10 -10 molecule copies per Ìl. The results are presented as absolute, mucosae from healthy margins of operated CRC (only the cases with i.e., the number of mRNA copies per 150 ng of total RNA, and no recurrent disease and no generalization were included) from 12 relative, i.e., the number of copies of particular mRNA related to the patients (aged 50.8 - 77.0 years, mean 62.4), 4 specimens of healthy number of copies of GAPDH mRNA in the same RNA preparation. liver tissue (from 2 patients), 85 specimens of CRC tissue of type (from 67 patients, aged 28.4 - 87.3 years, mean Serum tumour markers. Tumour markers in patients’ serum samples 64.5), 75 specimens of liver metastases of CRC (from 42 patients, were estimated in the Laboratory for Immunoanalysis, Faculty aged 29.4 - 76.5 years, mean 59.1) and 21 specimens of malignant Hospital in Pilsen by routine techniques: CEA with the IRMA liver tumours other than metastases of CRC (21 patients, aged 39 - assay and thymidine kinase (TK) with the REA assay (both 73 years, mean 56.9). All the patients had undergone surgery in the Immunotech Prague, Czech Republic); tissue specific polypeptide Surgical Clinic of the Faculty Hospital in Pilsen, Czech Republic antigen (TPS) with the IRMA assay (IDL Biotech AB, Bromma, during the years 1998-2003. Tissue samples were taken during the Sweden); and tissue polypeptide antigen (TPA) with the IRMA surgery, frozen to –80ÆC and kept at this temperature until used. assay (DiaSorin, Sillwater, Minnesota, USA). Routine histology examinations of operated tumours were done in the Department of Pathology, Faculty Hospital in Pilsen. Statistical analysis. The SAS 8.02 software and STATISTICA 5.1 program were used. Standard statistical parameters such as mean, RNA extraction. Total RNA was isolated from 50-100 mg of tissue SE, range, median, interquartile range, minimum and maximum with the RNAgent Total RNA Isolation System (Promega were calculated. For comparing different groups, non-parametrical Corporation, East Port, Prague, Czech Republic). The amount of tests (Kruskal-Wallis, Wilcoxon) were used because of the non- isolated total RNA was assessed spectrophotometrically and the Gaussian distribution of the values. For correlations of different RNA was kept at –75ÆC. parameters within the group, the Spearman correlation coefficient was used. Reverse-transcription. Three Ìg of isolated total RNA were used for reverse transcription (RT), which was performed with Superscript Results II Reverse Transcriptase (Invitrogen, Life Technologies, KRD, Prague, Czech Republic), oligo d(T)21 was used as the primer. Non-cancerous control samples. The control group was comprised of 12 specimens of normal colon tissue (healthy Real-time polymerase chain reaction. The primers for glyceraldehydes- margins), 14 specimens of benign lesions or colon 3-phosphate dehydrogenase (GAPDH) mRNA estimation have been published previously (22): the downstream primer was 5’- inflammatory diseases, 11 specimens of benign liver lesions TCCCATCACCATCTTCCAG-3’, the upstream primer 5’- and 4 specimens of normal liver tissue. The obtained values CATCACGCCACAGTTTCC-3’, and the product length was 380 from healthy tissues were compared statistically in all bp. Primers for CEA mRNA estimation were those suggested by parameters with those of benign and inflammatory lesions,

804 Cerna et al: Quantitative Estimation of CEA and CK20 mRNAs in Tumour Tissue

Table I. Absolute and relative values of specific mRNA copy number. mRNA Colorectal cancer (n=85) Liver metastases (n=75) Other liver tumours (n=21)

IR Median p IR Median p IR p

GAPDH 4.2x104-6.3x105 1.4x105 0.0051 5.5x104-4.9x105 1.1x105 0.0272 7.6x104-5.6x105 0.0255 CEA 2.8x104-3.7x106 3.5x105 <0.0001 7.5x103-7.1x105 5.1x104 <0.0001 2.2x103-8.9x103 0.0012 CK20 0-7.8x105 4.5x103 0.0139 1.3x103-3.6x104 1.9x104 <0.0001 6.1x102-2.9x103 0.0060 CEA/GAPDH 0.10-37.26 3.35 <0.0001 0.01-12.1 0.85 <0.0001 0--0.02 0.0184 CK20/GAPDH 0-6.42 0.01 0.0225 0-3.87 0.17 0.0030 0--0.01 0.0477 p-value refers to statistical significance (Wilcoxon test) of the difference between the given group and the control group (results not shown). IR, interquartile range; GAPDH, glyceraldehydes-3-phosphate dehydrogenase; CEA, carcinoembryonic antigen; CK20, cytokeratin 20.

and no significant differences between these 2 groups were Table II. Significant differences in gene expression between individual found (p-values varied between 0.1572 and 0.9267). tumours (Wilcoxon two-sample test). Therefore, both groups were used as a single control group p in all statistical analyses. The combined control group represented 41 tissue samples from 29 patients with an age mRNA CRC CRC CRC liver range of 18–78 years (mean 51.7). versus versus metastases CRC liver other liver versus other liver metastases tumours tumours

Differences in gene expression between tumours and controls. CEA absolute 0.0089 <0.0001 <0.0001 The main purpose of this study was to verify the usefulness CEA/GAPDH 0.0003 0.0013 of quantitative mRNA estimation of CEA and CK20 in CK20 absolute 0.0050 tumour tissue. The obtained values are summarised in Table CK20/GAPDH 0.0380 0.0083 I. All the presented values show a significant difference CRC, colorectal cancer. between the tumour tissues and controls, as expected. Even if individual estimations gave good reproducibility in double or triple estimations (not shown), the overall distribution of the obtained values was extremely broad and did not show Staging and clinical follow-up. The TNM international the Gaussian parameters. Therefore, the interquartile range classification was applied for the staging procedure. Among and median, rather than the mean, characterized the the CRC patients, 10 were evaluated as stage I, 31 as stage particular group and the Wilcoxon and Kruskal-Wallis tests II, 24 as stage III, and 12 as stage IV. No significant were used to compare individual groups. The CEA was differences were found in the CEA and CK20 expressions expressed significantly in all CRC samples of the between the stage groups. On the other hand, the clinical 3- adenocarcinoma type, CK20 was, however, less specific and year follow-up after surgery revealed a very good correlation its expression was not detected in all specimens. with the expectation based on staging. Briefly, the disease- free interval (DFI) showed a median value of over 3 years for Differences and correlations in gene expression between stage I-III groups in the CRC group. In the liver metastases different tumours. Comparisons of gene expression levels group, the median values in groups I-IV were >3, 2.26, 1.95 between different types of tumours are summarised in Table and 1.07 years, respectively. The overall survival in all cancers II. A significant difference was found in absolute CEA gave a median value above 3 years in stage I-III groups and expression between CRC and their liver metastases. 2.1 years in the stage IV group. The expression levels were Additional differences were found between CRC and other also examined with respect to clinical development. All the liver tumours, most of which were between liver metastases cases were divided into 2 large subgroups for each expressed of CRC and other types of liver neoplasms. gene according to the mean value, i.e. one subgroup of those Gene expressions in individual tissue samples exhibited a patients with expression levels above the average and the number of correlations, the best being between the relative other subgroup with expression levels below the average. No values of CEA and CK20 in the group of other liver significant relationship was found between the levels of CEA tumours, albeit a good correlation between those values was or CK20 expression and positive or negative clinical also found for the CRC group (Table III). development in the 3-year period.

805 ANTICANCER RESEARCH 26: 803-808 (2006)

Table III. Correlations between the levels of different mRNAs within the Table IV. Correlation between the levels of specific mRNA and serum groups (Spearman test). markers (Spearman test).

Parameter 1 Parameter 2 Spearman p Parameter 1 Parameter 2 n Spearman p coefficient coefficient

CRC (n=85) Abs.CEA Abs.GAPDH 0.4972 <0.0001 CRC Abs.CEA serum CEA 40 –0.1376 0.3973 Abs.CEA Abs.CK20 0.6141 <0.0001 CEA/GAPDH serum CEA 40 –0.1799 0.2666 Abs.CEA CK20/GAPDH 0.5719 <0.0001 Abs.CEA serum TPS 10 0.7660 0.0098 CEA/GAPDH Abs.CK20 0.5627 <0.0001 CEA/GAPDH CK20/GAPDH 0.6164 <0.0001 Liver Abs.CEA serum CEA 27 0.1588 0.4290 metastases CEA/GAPDH serum CEA 27 0.0128 0.9494 Liver metastases Abs.CEA Abs.GAPDH 0.3041 0.0080 of CRC Abs.CEA serum TPS 23 0.5827 0.0035 of CRC Abs.CEA Abs.CK20 0.4910 <0.0001 CEA/GAPDH serum TPS 23 0.4364 0.0374 (n=75) Abs.CEA CK20/GAPDH 0.4804 <0.0001 CEA/GAPDH Abs.CK20 0.3373 0.0031 TPS, tissue-specific polypeptide antigen. CEA/GAPDH CK20/GAPDH 0.5299 <0.0001

Other liver CEA/GAPDH CK20/GAPDH 0.7033 0.0005 tumours (n=20) Table V. Significant differences in gene expression between grading groups (Wilcoxon two-sample test). Control tissue Abs.CEA Abs.GAPDH 0.3552 0.0227 (n=41) Cancer group Gene expressed Grading groups p compared

CRC abs.GAPDH G2 G4 0.0429 (n=76) G1+2 G4 0.0514 Correlations of gene expression and levels of tumour markers in serum. The results of standard tumour markers were CK20/GAPDH G1 G2 0.0387 available for some patients to the date of surgery and their serum levels were correlated with the mRNA expression in CRC + abs.CK20 G1 G2 0.0275 Liver G1 G4 0.0041 tissues. The available serum marker estimations were those metastases G2 G4 0.0240 of CEA, thymidine kinase (TK), tissue polypeptide antigen (n=156) G3 G4 0.0115 (TPA) and tissue-specific polypeptide antigen (TPS). The results are summarized in Table IV. No correlation between CK20/GAPDH G1 G4 0.0041 the CEA expression in cancer tissue and CEA levels in G2 G4 0.0240 G3 G4 0.0115 serum was found in the CRC group or even in the liver metastases group. In contrast, a good positive correlation between serum TPS and tissue CEA expression (absolute) in the CRC group and the liver metastases group was found. Other serum markers did not show any correlations with the Wallis test. The significant grading-dependent differences measured gene expression. were found in CK20 expression (p=0.0443 for relative values It should also be stated that the pre-operative serum in CRC, p=0.0095 for absolute values in the combined group levels of some markers correlated with clinical development and p=0.0275 for relative values in the combined group). during the 3-year follow-up, namely CEA with DFI The differences between any 2 grading groups were (p=0.0048) and overall survival (p=0.0005), CA19-9 with evaluated using the Wilcoxon two-sample test. Significant overall survival (p=0.0190) and TK with DFI (p=0.0082) differences (Table V) were observed in CK20 expression, and overall survival (p=0.0158), in the CRC group. where the higher expression was typical for more differentiated tumours (lower grading forms), but even Histological types, grading and gene expression. There were 2 GAPDH was more expressed in the lower grade groups. dominant histological types among the CRC and their liver metastases – adenocarcinoma and tubular adenocarcinoma. Discussion No significant differences in the expression of individual genes were found between these two types. RT-PCR remains the most sensitive method for the detection The examined samples were divided into groups according of specific mRNA and, in combination with fluorescence- to the histological grading (G1-G4) and the distribution of based quantification, it became an important tool in the gene expression parameters was evaluated using the Kruskal- determination of mRNA levels in tissue. The external

806 Cerna et al: Quantitative Estimation of CEA and CK20 mRNAs in Tumour Tissue standard system was used in each PCR run and the quality of The main conflict that remains is the absence of a correlation the PCR product was verified by melting curve analysis. Each between the CEA gene expression and the CEA levels in PCR reaction contained an equal amount of reverse serum. The details of CEA expression were studied earlier (3, 4, transcription mixture representing 150 ng of the original total 30) in normal colon and colonic and the existence of RNA. Therefore, our absolute values are related to the cell-bound CEA was demonstrated (4, 30), along with the free- standard amount of total RNA. For normalisation, we chose circulating form. The two forms can even be distinguished with the most popular house-keeping gene, GAPDH. During the a specific monoclonal (31). Circulating CEA is heavily analysis of our results, a significant difference in GAPDH glycated and its detection in serum might be dependent on the expression was observed between tumours and control tissues glycoprotein character of the particular determinants. (Table I). This is not unexpected since an increased glycolysis Unfortunately, details regarding the antigenic determinants rate in tumours was observed eighty years ago by Otto H. detected by the antibodies in commercial IRMA assay kits are Warburg (24) and, since then, many metabolic processes have not available (Immunotech Prague, private communication). In been shown to be more intensive in tumours. The problem of contrast, the CEA mRNA level in tissue is not connected with normalisation was recently approached in more systematic the glycation process and represents the instruction for both the studies, of which one group of authors recommended a panel cell-bound and circulating forms. It must be understood that of different standard genes (25), while the other authors (26) the elevation of any serum marker, in principle, may indicate demonstrated the use of any standard gene for normalisation its increased synthesis and/or its increased "leakage" from the as meaningless and recommended the use of absolute values tissue into the extracellular space. It seems that the "leakage" is (i.e., copy number per Ìg of total RNA). Therefore, we have more connected with the tumour character and prognosis, since presented both forms of results. the prognostic relationship with CEA synthesis in tissue was not The difference in gene expression between tumours and confirmed by the immunohistochemistry techniques (28). The controls does not require comment. The differences between mechanism responsible for the release of CEA into the the different cancer groups could be a result of the extracellular space seems to be of importance and remains to characteristic behaviour of the individual cancer groups. Liver be clarified. metastases of CRC differ in CEA and CK20 expression, despite their common origin. Liver metastases do not Acknowledgements represent average CRC tissue, but rather clones of exceptional cells that succeeded in the metastastation process. We would like to acknowledge Professor Vladislav Treska and The most significant expression was that of CEA, which many other surgeons of the Surgical Clinic of the Faculty Hospital in CRC was present in practically all tissue samples. This is in Pilsen, the Czech Republic, for supplying tumour tissue samples in contrast with the current belief that serum CEA is not and clinical data, Dr. Marie Ludvikova from the Department of Pathology for the histopathological data and Mrs. Jana Zehleova elevated in all primary CRC or even in the metastases for her excellent technical assistance. This work was supported by (Table IV). There was no correlation between CEA mRNA Research Aims No. CEZ:J13/98:111400003/206033/ and MSM levels in tissues and those of the CEA marker in serum. 0021620819 from the Ministry of Education of the Czech Republic. Similar findings were described after CEA estimation in the cytosol of non-small cell (14) References and, more recently, in another lung cancer study where no correlation was found between the levels of serum CEA and 1 Srivastava S, Verma M and Henson DE: Biomarkers for early CEA messengers in tissue (27). detection of colon cancer. Clin Cancer Res 7: 1118-1126, 2001. Our results do not prove a significant relationship 2 Duffy MJ, van Dalen A, Haglund C, Hansson L, Klapdor R, between cancer tissue CEA and CK20 expressions and Lamerz R, Nilsson O, Sturgeon C and Topolcan O: Clinical utility of biochemical markers in colorectal cancer: European staging, confirming results mostly reported for other Group on Tumour Markers (EGTM) guidelines. Eur J Cancer tumours and estimated by other means (11, 12, 28). Our 39: 718-727, 2003. results are also in contrast with other publications (5, 6, 14) 3 Hammarström S: The carcinoembryonic antigen (CEA) family: but, again, regarding different tumours and different structures, suggested functions and expression in normal and methods. On the other hand, we were able to show the malignant tissues. Sem Cancer Biol 9: 67-81, 1999. inverse dependence of CK20 (and CEA according to the 4 Benchimol S, Fuks A, Jothy S, Beauchemin N, Shirota K and Kruskal-Wallis test) expression on grading. It seems that Stanners CP: Carcinoembryonic antigen, a human , functions as an intercellular adhesion molecule. Cell CK20 (and CEA) is only expressed in tissues with a certain 57: 327-334, 1989. degree of differentiation. CK20 immunostaining is already 5 Lorenzi M, Vindigni C, Minacci C, Tripodi SA, Iroatulam A, used as a criterion in histological grading (29) in different Petrioli R and Francini G: Histopathological and prognostic tumours, but its detailed expression with respect to grading evaluation of immunohistochemical findings in colorectal has not yet been studied. cancer. Int J Biol Markers 12: 68-74, 1997.

807 ANTICANCER RESEARCH 26: 803-808 (2006)

6 Kim JC, Han MS, Lee HK, Kim WS, Park SK, Park KC, 19 Ito S, Nakanishi H, Hirai T, Kato T, Kodera Y, Feng Z, Kasai Bodmer WF, Rowan AJ and Kim OJ: Distribution of Y, Ito K, Akiyama S, Nakao A and Tatematsu M: Quantitative carcinoembryonic antigen and biological behaviour in colorectal detection of CEA expressing free tumor cells in the peripheral carcinoma. Dis Colon Rectum 42: 640-648, 1999. blood of colorectal cancer patients during surgery with real time 7 Quentmeier A, Moller P, Schwarz V, Abel U and Schlag P: RT-PCR on a LightCycler. Cancer Lett 183: 195-203, 2002. Carcinoembryonic antigen, CA 19.9, and CA 125 in normal and 20 Gunn J, McCall JL, Yun K and Wright PA: Detection of carcinomatous human colorectal tissue. Cancer 60: 2261-2266, micrometastases in colorectal cancer patients by K19 and K20 1987. reverse-transcription polymerase chain reaction. Lab Invest 75: 8 Cosimelli M, De Peppo F, Castelli M, Giannarelli D, Schinaia 611-616, 1996. G, Castaldo P, Buttini GL, Sciarretta F, Bigotti G, Di Filippo 21 Futamura M, Takagi Y, Koumura H, Kida H, Tanemura H, F et al: Multivariate analysis of a tissue CEA, TPA, and CA 19.9 Shimokawa K and Saji S: Spread of colorectal cancer quantitative study in colorectal cancer patients. A preliminary micrometastases in regional lymph nodes by reverse finding. Dis Colon Rectum 32: 389-397, 1989. transcriptase-polymerase chain reaction for carcinoembryonic 9 Di Carlo A, Mariano A, D’Alessandro V, Belli G, Romano G and antigen and cytokeratin 20. J Surg Oncol 68: 34-40, 1998. Macchia V: Evaluation of epidermal growth factor receptor, 22 Brown DC, Purushotham AD, Birnie GD and George WD: carcinoembryonic antigen and Lewis carbohydrate antigens in Detection of intraoperative tumor cell dissemination in patients human colorectal and liver neoplasias. Oncol Rep 8: 387-392, 2001. with by use of reverse transcription and 10 Miya T, Watanabe T, Adachi I, Itabashi M, Fukutomi T, polymerase chain reaction. Surgery 117: 95-101, 1995. Yamaguchi K, Abe K and Arakawa M: Analysis of cytosol CA 15- 23 Gerhard M, Juhl H, Kalthoff H, Schreiber HW, Wagener C and 3, carcinoembryonic antigen, estrogen and progesterone receptors Neumaier M: Specific detection of carcinoembryonic antigen- in breast cancer tissues. Jpn J Cancer Res 83: 171-177, 1992. expressing tumor cells in bone marrow aspirates by polymerase 11 Picardo AL, Torres AJ, Maestro M, Ortega D, Garcia-Asenjo chain reaction. J Clin Oncol 12: 725-729, 1994. JA, Muguerza JM, Hernando F, Diez M and Balibrea JL: 24 Warburg OH: Über den Stoffwechsel der Carcinomzelle. Quantitative analysis of carcinoembryonic antigen, squamous Naturwissenschaften 12: 1131-1137, 1924. cell carcinoma antigen, CA 125, and CA 50 cytosolic content in 25 Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, non-small cell lung cancer. Cancer 73: 2305-2311, 1994. De Paepe A and Speleman F: Accurate normalization of real- 12 Macchia V, Paduano D, Di Carlo A, Mariano A, Cecere C and time quantitative RT-PCR data by geometric averaging of Ferrante G: Serum and tissue levels of CEA, TPA, CA 125 and multiple internal control genes. Genome Biol 3: 1-34, 2002. CA 15.3 in patients with lung cancer. Int J Biol Markers 8: 215- 26 Tricarico C, Pinzani P, Bianchi S, Paglierani M, Distante V, 220, 1993. Pazzagli M, Bustin SA and Orlando C: Quantitative real-time 13 Molina R, Farrus B, Filella X, Jo J, Zanon G, Pahisa J, Latre reverse transcription polymerase chain reaction: normalization M, Munoz M and Ballesta AM: Carcinoembryonic antigen in to rRNA or single housekeeping genes is inappropriate for tissue and serum from breast cancer patients – relationship with human tissue biopsies. Anal Biochem 309: 293-300, 2002. steroid receptors and clinical applications in the prognosis and 27 Nakamura H, Saji H, Idiris A, Kawasaki N, Hosaka M, Ogata early diagnosis of relapse. Anticancer Res 19: 2557-2562, 1999. A, Saijo T and Kato H: Comparison of immunohistochemistry 14 Vegh I, Sotelo T, Estenoz J, Fontanellas A, Navarro S, Millan I and real-time reverse transcription-polymerase chain reaction and Enriquez de Salamanca R: Tumor cytosol carcinoembryonic to detect expression of carcinoembryonic antigen in lung antigen as prognostic parameter in non-small cell lung cancer. cancer. Oncol Rep 10: 1231-1235, 2003. Tumori 88: 142-146, 2002. 28 Jantscheff P, Terracciano L, Lowy A, Glatz-Krieger K, Grunert 15 Miyake Y, Fujiwara Y, Ohue M, Yamamoto H, Sugita Y, F, Micheel B, Brummer J, Laffer U, Metzger U, Herrmann R Tomita N, Sekimoto M, Shiozaki H and Monden M: and Rochlitz C: Expression of CEACAM6 in resectable Quantification of micrometastases in lymph nodes of colorectal colorectal cancer: a factor of independent prognostic cancer using real-time fluorescence polymerase chain reaction. significance. J Clin Oncol 21: 3638-3646, 2003. Int J Oncol 16: 289-293, 2000. 29 Weaver EJ, McCue PA, Bagley DH, Kovatich AJ and Bibbo M: 16 Bustin SA, Siddiqi S, Ahmed S, Hands R and Dorudi S: Expression of cytokeratin 20 and CD44 protein in upper urinary Quantification of cytokeratin 20, carcinoembryonic antigen and tract transitional cell carcinoma: cytologic-histologic guanylyl cyclase C mRNA levels in lymph nodes may not correlation. Anal Quant Cytol Histol 23: 339-344, 2001. predict treatment failure in colorectal cancer patients. Int J 30 Jothy S, Yuan SY and Shirota K: Transcription of Cancer 108: 412-417, 2004. carcinoembryonic antigen in normal colon and colon 17 Öberg ANV, Lindmark GE, Israelsson ACE, Hammarström SG carcinoma. In situ hybridization study and implication for a new and Hammarström MLKC: Detection of occult tumour cells in in vivo functional model. Am J Pathol 143: 250-257, 1993. lymph nodes of colorectal cancer patients using real-time 31 Stewart LM, Young S, Watson G, Mather SJ, Bates PA, Band quantitative RT-PCR for CEA and CK20 mRNAs. Int J Cancer HA, Wilkinson RW, Ross EL and Snary D: Humanisation and 111: 101-110, 2004. characterisation of PR1A3, a specific for 18 Guller U, Zajac P, Schnider A, Bosch B, Vorburger S, Zuber cell-bound carcinoembryonic antigen. Cancer Immunol M, Spagnoli GC, Oertli D, Maurer R, Metzger U, Harder F, Immunother 47: 299-306, 1999. Heberer M and Marti WR: Disseminated single tumor cells as detected by real-time quantitative polymerase chain reaction Received June 22, 2005 represent a prognostic factor in patients undergoing surgery for Revised November 9, 2005 colorectal cancer. Ann Surg 236: 768-775, 2002. Accepted November 22, 2005

808