Original Paper
Tumor Biol 1992;13:330-337
Department of Surgery, New England Deaconess Hospital and Division of Cellular and Molecular Biology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Mass., USA
Keyw ords Carcinoembryonic antigen Colorectal cancer mRNA
Differences in Messenger RIMA Expression of Carcinoembryonic Antigen in Surgical Specimens of Colorectal Carcinoma
Abstract
Carcinoembryonic antigen (CEA) is the most widely used tumor marker for colorectal cancer. Plasma CEA levels have been variably associated with prognosis. Since plasma CEA level is multifactorial, CEA gene expression in tumors may pro vide one precise mechanism to evaluate its functional role. This study evaluated CEA expression at the messenger RNA (mRNA) level in 22 human colorectal carcinomas and their adjacent normal mucosae by Northern blot hybridization using a 32P-labeled CEA probe (a loop-domain specific cDNA, LV7). Both tumor and normal mucosa displayed three mRNA species of 4.0, 3.6, and 3.0 kb in length. The expression of 3.6-kb mRNA which encodes for CEA was dominant and it was corre lated with another 4.0-kb CEA mRNA expression. The expres sion of 3.0-kb mRNA which encodes for nonspecific cross reacting antigen was weak and not detectable in 8 of 22 colon tumors and 12 of 22 normal colon mucosae. In only one tumor, a 4.5-kb mRNA (which might encode for a new family member of CEA) was expressed. A two- to fourfold higher expression of CFA mRNA (3.6 kb) was observed in 11 of 22 colorectal tumors (2 of 9 proximal colon tumors and 9 of 14 rectosigmoid tumors) when compared with morphologically normal adjacent mucosae. Preoperative plasma CEA levels and Dukes’ staging had no correlation with this CEA mRNA expression. CEA mRNA did not appear to correlate with metastasis because its expression in the primary colon cancers with métastasés (Dukes’ stage D tumor) was not always increased. These data also imply that factors other than mRNA expression in tumor might be important in regulating plasma CEA levels.
Received:
Januar>' 13,1992 Accepted: August 6.1992
Peter Thomas laboratory of Cancer Biology Department of Surgery New England Deaconess Hospital 50 Binney Street. Boston. MA 02115 (USA)
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Ken-ichi Mafuneab Kulvinder S. Saini a T.S. Ravikumar a Lan Bo Chen b Glenn D. Steele, Jr. a Peter Thomasa
Carcinoembryonic antigen (CEA) is the most widely used tumor marker for gastroin testinal malignancies, especially colorectal carcinoma. CEA was originally identified as a tumor-associated antigen of the digestive tract by Gold and Freedman [1], This discov ery was followed by a search for this antigen in patients’ sera. Thomson et al. [2] described a method to detect nanogram quantities of the antigen in the serum of patients who have tumors of the digestive tract. Several reports suggest that the preoperative plasma CEA level is helpful in predicting prognosis [3-5]. Postoperative monitoring of CEA is also used for diagnosis of recurrence [6-8], This highly glycosylated cell surface glyco protein (Mr = 180,000) is expressed at greatly increased levels in nearly all human colon car cinomas [9], CEA has been shown to be a member of a family of more than 10 crossreactive isoantigens which can be detected in a variety of normal and tumor tissue types [10, 11]. These closely related family mem bers, such as nonspecific cross-reacting anti gen (NCA) [12], are also produced both in colon and in a significant proportion of other prevalent tumors, such as breast and lung car cinomas [13, 14]. cDNAs for CEA [15-18], NCA [ 19, 20] and the related biliary glycopro teins [21] have been isolated and sequenced. These cDNAs provide a new approach to the question of CEA function. Recently, CEA was found to have the properties of an intercellu lar adhesion molecule. It has been suggested that the overproduction of CEA can be in volved in colon carcinogenesis either through effects on metastasis or on tissue architecture [22], Hostetter et al. [23] reported that CEA enhances the metastatic potential of human colorectal carcinoma. Thus, we postulated that the CEA expression in colorectal tumors at the messenger RNA (mRNA) level could be
a marker of biological behavior of tumors. It has been reported that both CEA and NCA mRNAs and proteins are overexpressed in colon tumor cells relative to their normal counterparts [13, 14, 24], However, the num ber of cases were very small and no correla tion was made with clinical information in these reports. In this investigation, we evalu ated CEA expression in colorectal carcinoma and its adjacent normal mucosa at the mRNA level and also correlated its expression with preoperative plasma CEA levels and clinicopathologic information.
Materials and Methods Total RNA Extraction from Tissues Twenty-two surgical specimen pairs of primary hu man carcinomas and adjacent normal colon tissue taken as far from the tumor margin as possible were obtained from the operating rooms at the New Eng land Deaconess Hospital. Necrotic and ulcerative parts of the tumors were removed and morphologically normal colonic epithelium was dissociated from mus cle and connective tissue immediately after resection. Tissues were immediately frozen in liquid nitrogen and kept at -7 0 °C until extraction of RNA. Samples of both tumor and normal colon were also fixed in 10% formol saline for subsequent histological exami nation. Total cellular RNA was extracted from specimens by a modification of the method of Chirgwin et al. [25], Each specimen was ground into fine particles in liquid nitrogen. Total RNA was extracted with guanid inium isothiocyanate and purified by ultracentrifuga tion through cesium chloride at 120,000 g for 20 h. The concentration of RNA in Tris/EDTA buffer was measured at a wavelength of 260 nm. Northern Blot Hybridization Northern blot hybridization was performed to compare the CEA mRNA expression in colon carci noma and adjacent normal colonic mucosa. Equal amounts (15 pg) of total cellular RNA extracted from 22 paired tumor and normal mucosa were loaded onto the lanes of 1.0% agarose/formaldehyde gel and electrophoresed overnight. RNA was transferred to nylon membranes (GreenScreenPlus, DuPont, Boston,
331
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Introduction
Clinicopathologic Information Age, sex, tumor location, the pathologic data relat ing to tumor size, histology, depth of invasion, lymph node metastasis, and outcome were obtained from hospital records of the patients. The study cohort for Northern blot analysis were surgical specimens from 22 patients with primary colo rectal cancer that included tumor and uninvolved nor mal mucosae. The mean age of the patients was 66 years (range: 48-82 years). The distribution of primary colorectal tumors was as follows: cecum 3, ascending colon 2, transverse colon 0, descending colon 3, sig moid colon 9, rectosigmoid colon 2, and rectum 3. His tologically. all 22 tumors were classified as moderately differentiated adenocarcinoma. The uninvolved mu cosae were histologically normal. The Dukes’ staging of primary tumors was: Dukes’ A 4, B 7, C 7, D 4. Preoperative plasma CEA levels were measured by an enzyme immunoassay in 18 of 22 patients. The plasma CEA level was elevated in 9 of 18 patients (50%). The normal range of plasma CEA level is less than 5 ng/ml.
332
Patterns
1
2
3
28 S • ★ ★ ▲
ft «• f i r
18SColorectal carcinoma Adjacent mucosa
13 10
8 12
1
Fig. 1. The patterns of CEA mRNA expression. LV7 cDNA usually hybridized with three mRNA spe cies: at 4.0 (*), 3.6 (★ ★ ), and 3.0 kb (A). Upper two mRNA (*, ★ ★ ) encode CEA, and their expression ratio was almost the same in all specimens. Another mRNA (a ) encodes NCA, but this mRNA expression was very faint and it was not detected in 20 of 44 speci mens. In only 1 tumor another higher band appeared at 4.5 kb (•). This mRNA might encode a new family member of CEA.
Results
Northern Blot Hybridization Northern blot analysis showed three pat terns of mRNA expression (fig. 1). This LV7 cDNA clone hybridized with three mRNA species of 4.0, 3.6, and 3.0 kb in length in both colorectal tumors and normal colonic muco sae. The expression of 3.6-kb mRNA which encodes for CEA was dominant among these three bands. This CEA mRNA expression was correlated with another 4.0-kb mRNA expres sion which also encodes CEA. The expression of 3.0-kb mRNA which encodes for NCA-55 was weak and was not detected in 20 of 44 samples (8 of 22 colon tumors and 12 of 22 normal colon mucosae). In only 1 tumor, a 4.5-kb mRNA which might encode for a new family member of CEA was expressed. This band was not seen in any of the other speci-
Mafune/Saini/Ravikumar/Chen/Steele/ Thomas
CEA mRNA Expression in Colorectal Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Mass., USA). A loop-domain specific CEA cDNA probe, LV7 [20, 21], was ,;!P-labelcd by the random primer method and used to hybridize with RNAs on the membranes. This probe recognizes all members of the CEA family. Specific activities of the DNA probes were about 2-5 X 10s cpm/gg of DNA. After over night hybridization at 42 °C, followed by washing at high stringency, 0.2 X SSPE, 0.25% SDS at 68 °C as described by Barnett et al. [20]. The membranes were exposed to X-ray films (Kodak X-OMAT AR) at - 70 °C. Hybridization signals on the films were quan titated using a LKB UltroScan XL Enhanced Laser Densitometer (LKB Produkter AB. Bromma, Swe den). The mRNA level in each of the blots was quanti tatively expressed as the integrated area under the curve. These membranes were used for further hybridiza tions after washing, using an actin probe to control for loading errors. The actin cDNA insert in pBR322 [26] was cloned into Pst I site of pBluescript vector (Stratagene, La Jolla, Calif., USA) and transformed in DH 5a Escherichia coli. Color-selected clones were used for large-scale preparation of plasmid DNA. The actin cDNA insert was digested with PST I and purified with GcncClcan Kit (BIO 101, La Jolla. Calif., USA). Puri fied actin cDNA was 32P-labeled by the random primer method.
C ase
1 T
2 N
T
4
3 N
T
N
T
5 N
T
N
•
mué
CEA- * * {HI
- - •
••
mm
Actin-
mens even on prolonged exposure of the blots. The level of CEA mRNA in colorectal car cinoma was 2- to 4-fold higher than in adja cent normal colonic epithelium in 11 of 22 patients (fig. 2). In the other 11 patients, CEA mRNA expression in tumor was less than that in adjacent normal mucosa. Correlation with Clinical Information With the limitation of small numbers in this analysis, there was a correlation between preoperative plasma CEA levels and Dukes’ staging (table 1). While patients with Dukes’ A and B tumors had normal plasma CEA lev els, those with Dukes’ C and D tumors had CEA levels over 5 ng/dl. However, there was no correlation between CEA mRNA expres sion and preoperative plasma CEA levels (ta ble 2). There was no correlation between CEA mRNA expression and Dukes’ staging (ta ble 3). CEA mRNA expression did not appear to correlate with metastasis, because this CEA mRNA expression was not higher in 2 of 4 primary colon cancers with métastasés (Dukes’ D primary tumors). In correlation studies of CEA mRNA expression with the location of primary colorectal cancers, 65% of rectosigmoid cancers demonstrated high tu-
Tablel. Preoperative plasma CEA levels and Dukes’ staging
Dukes’ stage
Plasma CEA level, ng/ml 0< 2.5 2.5 < 5 5 < 10
>10
total
A B C D Total
2 3 0 0 5
1 0 1 3 5
4 6 5 3 18
1 2 1 0 4
0 1 3 0 4
Table 2. CEA mRNA expression and preoperative plasma CEA levels
CEA
Plasma CEA level, ng/ml 0< 2.5 2.5< 5 5 < 10 > 1 0
T> N T< N Total
4 1 5
1 3 4
3 1 4
2 3 5
total 10 8 18
T > N indicates cases where mRNA expression is greater in the tumor than the normal mucosae while T < N indicates higher levels in the normal mucosae.
333
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Fig. 2. Northern blot analysis of CEA mRNA expression in surgical specimens of primary colorectal carcinoma (T) and adjacent nor mal colon mucosa (N) from the same patient. The CEA mRNA ex pression in T was lower than that in N in cases 1 and 2, and was higher than that in N in cases 3,4, and 5, in these five paired, representative samples out of 22. Actin mRNA expression (2.2 kb) was shown as control.
CEA mRNA
Dukes’ stage A
B
c
D
total
T> N T< N Total
3 1 4
4 3 7
2 5 7
2 2 4
11
11 22
Table 4. CEA mRNA expression and tumor
location CEA mRNA
T> N T< N Total
Tumor location proximal
recto sigmoid
total
2(25%) 6(75%) 8
9(64%) 5 (36%) 14
11 11 22
mor/normal ratios, only 25% of proximal co lon cancers had high tumor/normal ratios of CEA mRNA expression (table 4). A larger study will be required, however, to determine if this observation has any clinical signifi cance.
Discussion
With the delineation of the genes encoding the ‘CEA family’, the transcription of these genes in human colonic tissues and tumors has been reported. Cournoyer et al. [14] ob served that the levels of CEA-like mRNAs were approximately the same in colon carci nomas as in adjacent non-invaded mucosa from the same patients. However, Boucher et al. [24] reported a 6- to 10-fold more intense
334
major CEA mRNA band in colon tumors when compared with that in normal colonic mucosa from the same individual in 5 of 6 patients. Our results showed that the level of this CEA mRNA in colorectal carcinoma is 2to 4-fold higher than in adjacent normal co lonic epithelium only in half of the 22 patients in our series. This is a much lower frequency than was expected. Chi et al. [27] showed in a study of 32 colorectal cancers and normal mucosa that only 41 % had higher mRNA lev els for CEA in the tumor while 92% of the tumors expressed higher mRNAs for NCA than in the distant normal mucosa. It has been shown that CEA-related mate rial is present in normal colon mucosa [28— 30] and that purified CEA from normal colon is indistinguishable by immunological and physicochemical criteria from reference colon carcinoma CEA [29, 30]. An immunohistologic study showed that the expression of CEA epitopes 3d6 and COL-4 reacted with a large number of the mucosal biopsies, and that this expression is predominantly localized to the mucosal surfaces and is detectable throughout the colon [31]. Quantitatively, it was demon strated that the concentration of CEA-related material in normal colon mucosa was about 10-40 times lower than in primary large bowel carcinomas and 50-300 times lower than that in metastatic colon or rectum carci nomas [32, 33], This appears inconsistent with our results, which showed in 50% of the cases that the CEA mRNA expression in nor mal colon mucosa was greater than that in tumor. Furthermore, in cases where the CEA mRNA expression in tumor was higher, it was only 2- to 4-fold of that in normal mucosa. Taken together, these findings indicate a dis crepancy between the level of CEA mRNA and protein that suggest translational and/or post-translational events may be important. This is supported by the findings of Boucher et al. [24], who showed that there was no
Mafune/Saini/Ravikumar/Chcn/Steele/ Thomas
CEA mRNA Expression in Colorectal Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Table 3. CEA mRNA expression and Dukes'
staging
CEA levels with Dukes’ staging observed in this study and reported previously by others [4, 5] suggests that CEA might be an impor tant factor for tumor progression. However, our results also imply that factors other than mRNA expression might be important in reg ulating plasma CEA level; further molecular analysis of CEA expression and secretion may elucidate the discordant clinical data with respect to CEA as tumor marker and prognos ticator. Another question that arises is whether CEA plays a role in the mechanism of colorec tal cancer metastasis. It was shown using a nude mouse model that preinjection with CEA can enhance metastatic potential of hu man colorectal carcinoma [23]. As CEA can function as an intercellular adhesion mole cule, this provides a possible mechanism for CEA’s involvement in metastasis [22], NCA and other members of the immunoglobulin supergene family also have cell adhesion properties [36, 37]. Further studies will eluci date the relative roles of these adhesion mole cules and the role of other receptors for base ment membrane components (e.g. laminin re ceptors) in colorectal tumor metastasis. We will continue to increase the number of cases for this study and investigate CEA mRNA expression in metastatic tumors from colon. Such studies will enable us to evaluate the role of CEA as a factor in the development of metastasis.
Acknowledgments This study was supported by grants CA44583 and CA44704 from the National Cancer Institute, Bcthesda, Md. We thank Dr. Thomas R. Barnett at Molecular Diagnostics, West Haven, Conn., for kindly providing the CEA cDNA clone, LV7, Dr. Bruce Spiegelman at Dana-Farber Cancer Institute, Boston. Mass, for kindly providing the actin cDNA.
335
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
direct proportionality between mRNA levels and gene product expression and by the stud ies of Sato et al. [34] and Chi et al. [27], Dif ferences between mRNA and gene product levels may reflect regulation of CEA mRNA stability by exogenous and endogenous fac tors that affects ultimately the overall turn over of the CEA messenger molecules. It is possible that this regulatory mechanism plays a very important role in CEA secretion and CEA-related metastasis. An alternative expla nation for this discrepancy may be a purely anatomical one and this has been suggested by Kuroki et al. [35], who examined mRNA lev els and CEA secretion in three colorectal can cer-normal mucosa pairs. They found that even though the normal mucosa contained much smaller quantities of CEA than the adjacent tumor the expression of mRNA was comparable between the two. In normal colon tissue, CEA is expressed mainly on the lumi nal surface of the single layer of columnar epi thelial cells lining the upper parts of the crypts and secreted into the colon lumen. In colorec tal cancers or even in polyps, tumor cells no longer conform to the single-cell layer organi zation by bulging into the crypts and invading through the basement membrane in multicel lular arrays and CEA is found along the bor ders between cells [22], Thus, CEA may be secreted into the tumor mass or into lumens where it could fill and block blind ducts. High levels of CEA may then accumulate as a con sequence of its resistance to degradation. Thus, while CEA mRNA levels between nor mal and malignant mucosae are similar, the level of secreted protein in extracts of the tumor tissue is much higher because the CEA is trapped within the tissue. Other factors also regulate plasma CEA levels and include in creased CEA production by tumor invasion, organ-specific metastases and circulating tu mor cells, and factors affecting CEA clearance by the liver [11], The correlation of plasma
1 Gold P, Freedman SO: Specific carcinocmbryonic antigen of the hu man digestive system. J Exp Med 1965:122:467-481. 2 Thomson DMP, Krupcy J, Freed man SO. Gold P: The radioimmu noassay of circulating carcinoembryonic antigen of the digestive sys tem. Proc Natl Acad Sci USA 1969; 64:161-167. 3 Dhar P, Moore T, Zamchcck N, Kupchik HZ: Carcinoembryonic an tigen (CEA) in colonic cancer. Use in preoperative and postoperative diagnosis and prognosis. JAMA 1972:221:31-35. 4 Herrera MA, Chu TM, Holyoke EG: Carcinoembry onic antigen (CEA) as a prognostic and monitoring test in clinically complete resection o f colo rectal carcinoma. Ann Surg 1976: 183:5-9. 5 Goslin R, Steele G Jr, Macintyre J. Mayer R. Sugarbaker P. Cleghorn K. Wilson R, Zamcheck N: The use of preoperative plasma CEA levels for the stratification of patients after curative resection of colorectal can cers. Ann Surg 1980:192:747-751. 6 Sugarbaker PH, Zamchcck N, Moore FD: Assessment of serial car cinoembryonic antigen (CEA) as says in postoperative detection of recurrent colo-rectal cancer. Cancer 1976:38:2310-2315. 7 Welch JP. Donaldson GA: Detec tion and treatment of recurrent can cer of the colon and rectum. Am J Surg 1978:135:505-510. 8 Steele G Jr, Zamcheck N. Wilson R. Mayer R. Lokich J, Rau P, Maltz J: Results of CEA-initiatcd second look surgery for recurrent colorectal cancer. Am J Surg 1980; 139:544— 548. 9 Shuster J, Thomson DMP, Fuks A, Gold P: Immunologic approaches to the diagnosis of malignancy. Proc Exp Tumor Res 1980;25:89-139. 10 Shively J. Beatty J: CEA-related an tigens: Molecular biology and clini cal significance. CRC Crit Rev On col Hematol 1985;2:355-399.
11 Thomas P, Toth CA. Saini KS. Jes sup JM. Steele G Jr: The structure, metabolism and function of the car cinoembryonic antigen gene family. Biochim Biophys Acta 1990:1032: 177-189. 12 von Kleist S, Chavanel G, Burtin P: Identification of an antigen from normal human tissue that cross reacts with the carcinoembryonic antigen. Proc Natl Acad Sci USA 1972:69:2492-2494. 13 Zimmerman W, Weber B. OrtliebB. Rudert F, Schempp W, Fiebig HH, Shively JE, von Kleist S, Thompson JA: Chromosomal localization of the carcinoembryonic antigen gene family and differential expression in various tumors. Cancer Res 1988; 48:2443-2550. 14 Cournoyer D. Beauchemin N, Boucher D, Benchimol S. Fuks A. Stanners CP: Transcription of genes of the carcinoembryonic antigen family in malignant and nonmalignant human tissues. Cancer Res 1988;48:3153-3157. 15 Zimmerman W, Ortlieb R, Fried rich R, von Kleist S: Isolation and characterization of cDNA clones en coding the human carcinoem bryonic antigen reveal a highly con served repeating structure. Proc Natl Acad Sci USA 1987:84:29602964. 16 Oikawa S, Nakazato H. Kozaki G: Primary structure of human carci noembryonic antigen (CEA) de duced from cDNA sequence. Biochcm Biophys Res Commun 1987; 142:511-518. 17 Kamarck ME, Elting JJ. Hart JT, Goebel SJ, Rac PMH. Nothdurft MA, Nedwin JJ, Barnett TR: Carci noembryonic antigen family: Ex pression in a mouse L-cell transfectant and characterization of a partial cDNA in bacteriophage ygtl 1. Proc Natl Acad Sci USA 1987:84:53505354. 18 Beauchemin N, Benchimol S, Cour noyer D. Fuks A, Stanners CP: Isola tion and characterization of fulllength functional cDNA clones for human carcinoembryonic antigen. Mol Cell Biol 1987;7:3221-3230.
19 Tawaragi Y, Oikawa S, Matsuoka Y. Kosaki G, Nakazato H: Primary structure of nonspecific crossrcacting antigen (NCA), a member of the carcinoembryonic antigen (CEA) gene family, deduced from cDNA sequence. Biochem Biophys Res Commun 1988; 150:89—96. 20 Barnett T. Goebel SJ, Nothdurft MA, Elting JJ: Carcinoembryonic antigen family: Characterization of cDNAs coding for NCA and CEA and suggestion of nonrandom se quence variation in their conserved loop-domains. Genomics 1988:3: 59-66. 21 Barnett TR, Kretschmer A, Austin DA, Goebel SJ, Hart JT. Elting JJ. Kamarck ME: Carcinoembryonic antigens: Alternative splicing ac counts for the multiple mRNAs that code for novel members o f the carci noembryonic antigen family. J Cell Biol 1989;108:267-276. 22 Benchimol S, Fuks A, Jothy S. Beauchemin N. Shirota K, Stanners CP: Carcinoembryonic antigen, a hu man tumor marker, functions as in tercellular adhesion molecule. Cell 1989:51:327-334. 23 Hostetter RB. Campbell DE, Chi K. Kerchoff S, Cleary' KR, Ullrich S, Thomas P. Jessup JM: Carcinoem bryonic antigen enhances metastatic potential of human colorectal carci noma. Arch Surg 1990; 125:300— 304. 24 Boucher D, Stanners CP, Fuks A: Studies on the control of gene ex pression of the carcinoembryonic antigen family in human tissue. Cancer Res 1989;49:847-853. 25 Chirgwin J, Przybyla AE. MacDon ald RJ, Rutter WJ: Isolation of bio logically active ribonucleic acid from sources enriched in ribonu cléase. Biochemistry 1979:18:5294— 5299. 26 Spiegelman BM. Frank M, Gree H: Molecular cloning of mRNA from 3T3 adipocytes: Regulation of mRNA content for glycerophos phate dehydrogenase and other differentiation-dependent proteins during adipocyte development. J Biol Chem 1983:258:10083-10089.
336
Mafune/Saini/Ravikumar/Chen/Steele/ Thomas
CEA mRNA Expression in Colorectal Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
References
32 Khoo SK, Warner NL, Lie JT, MacKay IR: Carcinoembryonic an tigen activity of tissue extracts: A quantitative study of malignant and benign neoplasms, cirrhotic liver, normal adult and fetal organs. Int J Cancer 1973;11:681-687. 33 Zamcheck N, Doos WG, Prcdente R, Lurie BB, Gottlieb LS: Prognos tic factors in colon carcinoma: Cor relation of serum carcinoembryonic antigen level and tumor histopathology. Hum Pathol 1975;6:31-45. 34 Sato C, Miyaki M, Oikawa S, Nakasato H. Kosaki G: Differential ex pression of carcinoembryonic anti gen and nonspecific crossreacting antigen genes in human adenocarci nomas and normal colon mucosa. Jp n J Cancer Res 1988;79:433-437.
35 Kuroki M. Arakawa F, Yamamoto H, Shimura H, Ikchara Y, Matsuoka Y: Active production and mem brane anchoring of carcinoem bryonic antigen observed in normal colonic mucosa. Cancer Lett 1988: 43:151-157. 36 Oikawa S, Inuzuka C, Kuroki M, Matsuoka Y, Kosaki G, Nakazato H: Cell adhesion activity of non-spe cific cross-reacting antigen (NCA) and carcinoembryonic antigen (CEA) expressed on CHO cell sur face: Homophilic and hetcrophilic adhesion. Biochem Biophys Res Commun 1989:164:39-45. 37 Zhou H, Fuks A. Stanners C: Speci ficity of intercellular adhesion me diated by various members of the immunoglobulin supergene family. Cell Growth Differ 1990:1:209215.
337
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
27 Chi K, Jessup JM , Frazier ML: Pre dominant expression of mRNA cod ing for nonspecific cross-reacting antigen in colorectal carcinoma. Tu mor Biol 1991;12:298-308. 28 Burtin P, Martin E. Sabine MC. von Kleist S: Immunological study of polyps of the colon. J Natl Cancer Inst 1972;48:25-29. 29 Fritsche R, Mach JP: Isolation and characterization of carcinoembryonic antigen (CEA) extracted from normal human colon mucosa. Immunochemistry 1977; 14:119127. 30 Egan ML, Pritchard DG, Todd CW, Go VLW: Isolation and immuno chemical and chemical characteriza tion of carcinoembryonic antigen like substances in colon lavages of healthy individuals. Cancer Res 1977;37:2638-2643. 31 Wolf BC, Salem RR, Sears HF, Horst DA, Lavin PT, Hcrlyn M, Itzkowits SH, Schlom J, Steele GD Jr: The expression of colorectal carci noma-associated antigens in the normal colonic mucosa: An immu nohistochemical analysis of regional distribution. Am J Pathol 1989:135: 111-119.
Tumor Biol 1992;13:330-337
Department of Surgery, New England Deaconess Hospital and Division of Cellular and Molecular Biology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Mass., USA
Keyw ords Carcinoembryonic antigen Colorectal cancer mRNA
Differences in Messenger RIMA Expression of Carcinoembryonic Antigen in Surgical Specimens of Colorectal Carcinoma
Abstract
Carcinoembryonic antigen (CEA) is the most widely used tumor marker for colorectal cancer. Plasma CEA levels have been variably associated with prognosis. Since plasma CEA level is multifactorial, CEA gene expression in tumors may pro vide one precise mechanism to evaluate its functional role. This study evaluated CEA expression at the messenger RNA (mRNA) level in 22 human colorectal carcinomas and their adjacent normal mucosae by Northern blot hybridization using a 32P-labeled CEA probe (a loop-domain specific cDNA, LV7). Both tumor and normal mucosa displayed three mRNA species of 4.0, 3.6, and 3.0 kb in length. The expression of 3.6-kb mRNA which encodes for CEA was dominant and it was corre lated with another 4.0-kb CEA mRNA expression. The expres sion of 3.0-kb mRNA which encodes for nonspecific cross reacting antigen was weak and not detectable in 8 of 22 colon tumors and 12 of 22 normal colon mucosae. In only one tumor, a 4.5-kb mRNA (which might encode for a new family member of CEA) was expressed. A two- to fourfold higher expression of CFA mRNA (3.6 kb) was observed in 11 of 22 colorectal tumors (2 of 9 proximal colon tumors and 9 of 14 rectosigmoid tumors) when compared with morphologically normal adjacent mucosae. Preoperative plasma CEA levels and Dukes’ staging had no correlation with this CEA mRNA expression. CEA mRNA did not appear to correlate with metastasis because its expression in the primary colon cancers with métastasés (Dukes’ stage D tumor) was not always increased. These data also imply that factors other than mRNA expression in tumor might be important in regulating plasma CEA levels.
Received:
Januar>' 13,1992 Accepted: August 6.1992
Peter Thomas laboratory of Cancer Biology Department of Surgery New England Deaconess Hospital 50 Binney Street. Boston. MA 02115 (USA)
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Ken-ichi Mafuneab Kulvinder S. Saini a T.S. Ravikumar a Lan Bo Chen b Glenn D. Steele, Jr. a Peter Thomasa
Carcinoembryonic antigen (CEA) is the most widely used tumor marker for gastroin testinal malignancies, especially colorectal carcinoma. CEA was originally identified as a tumor-associated antigen of the digestive tract by Gold and Freedman [1], This discov ery was followed by a search for this antigen in patients’ sera. Thomson et al. [2] described a method to detect nanogram quantities of the antigen in the serum of patients who have tumors of the digestive tract. Several reports suggest that the preoperative plasma CEA level is helpful in predicting prognosis [3-5]. Postoperative monitoring of CEA is also used for diagnosis of recurrence [6-8], This highly glycosylated cell surface glyco protein (Mr = 180,000) is expressed at greatly increased levels in nearly all human colon car cinomas [9], CEA has been shown to be a member of a family of more than 10 crossreactive isoantigens which can be detected in a variety of normal and tumor tissue types [10, 11]. These closely related family mem bers, such as nonspecific cross-reacting anti gen (NCA) [12], are also produced both in colon and in a significant proportion of other prevalent tumors, such as breast and lung car cinomas [13, 14]. cDNAs for CEA [15-18], NCA [ 19, 20] and the related biliary glycopro teins [21] have been isolated and sequenced. These cDNAs provide a new approach to the question of CEA function. Recently, CEA was found to have the properties of an intercellu lar adhesion molecule. It has been suggested that the overproduction of CEA can be in volved in colon carcinogenesis either through effects on metastasis or on tissue architecture [22], Hostetter et al. [23] reported that CEA enhances the metastatic potential of human colorectal carcinoma. Thus, we postulated that the CEA expression in colorectal tumors at the messenger RNA (mRNA) level could be
a marker of biological behavior of tumors. It has been reported that both CEA and NCA mRNAs and proteins are overexpressed in colon tumor cells relative to their normal counterparts [13, 14, 24], However, the num ber of cases were very small and no correla tion was made with clinical information in these reports. In this investigation, we evalu ated CEA expression in colorectal carcinoma and its adjacent normal mucosa at the mRNA level and also correlated its expression with preoperative plasma CEA levels and clinicopathologic information.
Materials and Methods Total RNA Extraction from Tissues Twenty-two surgical specimen pairs of primary hu man carcinomas and adjacent normal colon tissue taken as far from the tumor margin as possible were obtained from the operating rooms at the New Eng land Deaconess Hospital. Necrotic and ulcerative parts of the tumors were removed and morphologically normal colonic epithelium was dissociated from mus cle and connective tissue immediately after resection. Tissues were immediately frozen in liquid nitrogen and kept at -7 0 °C until extraction of RNA. Samples of both tumor and normal colon were also fixed in 10% formol saline for subsequent histological exami nation. Total cellular RNA was extracted from specimens by a modification of the method of Chirgwin et al. [25], Each specimen was ground into fine particles in liquid nitrogen. Total RNA was extracted with guanid inium isothiocyanate and purified by ultracentrifuga tion through cesium chloride at 120,000 g for 20 h. The concentration of RNA in Tris/EDTA buffer was measured at a wavelength of 260 nm. Northern Blot Hybridization Northern blot hybridization was performed to compare the CEA mRNA expression in colon carci noma and adjacent normal colonic mucosa. Equal amounts (15 pg) of total cellular RNA extracted from 22 paired tumor and normal mucosa were loaded onto the lanes of 1.0% agarose/formaldehyde gel and electrophoresed overnight. RNA was transferred to nylon membranes (GreenScreenPlus, DuPont, Boston,
331
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Introduction
Clinicopathologic Information Age, sex, tumor location, the pathologic data relat ing to tumor size, histology, depth of invasion, lymph node metastasis, and outcome were obtained from hospital records of the patients. The study cohort for Northern blot analysis were surgical specimens from 22 patients with primary colo rectal cancer that included tumor and uninvolved nor mal mucosae. The mean age of the patients was 66 years (range: 48-82 years). The distribution of primary colorectal tumors was as follows: cecum 3, ascending colon 2, transverse colon 0, descending colon 3, sig moid colon 9, rectosigmoid colon 2, and rectum 3. His tologically. all 22 tumors were classified as moderately differentiated adenocarcinoma. The uninvolved mu cosae were histologically normal. The Dukes’ staging of primary tumors was: Dukes’ A 4, B 7, C 7, D 4. Preoperative plasma CEA levels were measured by an enzyme immunoassay in 18 of 22 patients. The plasma CEA level was elevated in 9 of 18 patients (50%). The normal range of plasma CEA level is less than 5 ng/ml.
332
Patterns
1
2
3
28 S • ★ ★ ▲
ft «• f i r
18SColorectal carcinoma Adjacent mucosa
13 10
8 12
1
Fig. 1. The patterns of CEA mRNA expression. LV7 cDNA usually hybridized with three mRNA spe cies: at 4.0 (*), 3.6 (★ ★ ), and 3.0 kb (A). Upper two mRNA (*, ★ ★ ) encode CEA, and their expression ratio was almost the same in all specimens. Another mRNA (a ) encodes NCA, but this mRNA expression was very faint and it was not detected in 20 of 44 speci mens. In only 1 tumor another higher band appeared at 4.5 kb (•). This mRNA might encode a new family member of CEA.
Results
Northern Blot Hybridization Northern blot analysis showed three pat terns of mRNA expression (fig. 1). This LV7 cDNA clone hybridized with three mRNA species of 4.0, 3.6, and 3.0 kb in length in both colorectal tumors and normal colonic muco sae. The expression of 3.6-kb mRNA which encodes for CEA was dominant among these three bands. This CEA mRNA expression was correlated with another 4.0-kb mRNA expres sion which also encodes CEA. The expression of 3.0-kb mRNA which encodes for NCA-55 was weak and was not detected in 20 of 44 samples (8 of 22 colon tumors and 12 of 22 normal colon mucosae). In only 1 tumor, a 4.5-kb mRNA which might encode for a new family member of CEA was expressed. This band was not seen in any of the other speci-
Mafune/Saini/Ravikumar/Chen/Steele/ Thomas
CEA mRNA Expression in Colorectal Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Mass., USA). A loop-domain specific CEA cDNA probe, LV7 [20, 21], was ,;!P-labelcd by the random primer method and used to hybridize with RNAs on the membranes. This probe recognizes all members of the CEA family. Specific activities of the DNA probes were about 2-5 X 10s cpm/gg of DNA. After over night hybridization at 42 °C, followed by washing at high stringency, 0.2 X SSPE, 0.25% SDS at 68 °C as described by Barnett et al. [20]. The membranes were exposed to X-ray films (Kodak X-OMAT AR) at - 70 °C. Hybridization signals on the films were quan titated using a LKB UltroScan XL Enhanced Laser Densitometer (LKB Produkter AB. Bromma, Swe den). The mRNA level in each of the blots was quanti tatively expressed as the integrated area under the curve. These membranes were used for further hybridiza tions after washing, using an actin probe to control for loading errors. The actin cDNA insert in pBR322 [26] was cloned into Pst I site of pBluescript vector (Stratagene, La Jolla, Calif., USA) and transformed in DH 5a Escherichia coli. Color-selected clones were used for large-scale preparation of plasmid DNA. The actin cDNA insert was digested with PST I and purified with GcncClcan Kit (BIO 101, La Jolla. Calif., USA). Puri fied actin cDNA was 32P-labeled by the random primer method.
C ase
1 T
2 N
T
4
3 N
T
N
T
5 N
T
N
•
mué
CEA- * * {HI
- - •
••
mm
Actin-
mens even on prolonged exposure of the blots. The level of CEA mRNA in colorectal car cinoma was 2- to 4-fold higher than in adja cent normal colonic epithelium in 11 of 22 patients (fig. 2). In the other 11 patients, CEA mRNA expression in tumor was less than that in adjacent normal mucosa. Correlation with Clinical Information With the limitation of small numbers in this analysis, there was a correlation between preoperative plasma CEA levels and Dukes’ staging (table 1). While patients with Dukes’ A and B tumors had normal plasma CEA lev els, those with Dukes’ C and D tumors had CEA levels over 5 ng/dl. However, there was no correlation between CEA mRNA expres sion and preoperative plasma CEA levels (ta ble 2). There was no correlation between CEA mRNA expression and Dukes’ staging (ta ble 3). CEA mRNA expression did not appear to correlate with metastasis, because this CEA mRNA expression was not higher in 2 of 4 primary colon cancers with métastasés (Dukes’ D primary tumors). In correlation studies of CEA mRNA expression with the location of primary colorectal cancers, 65% of rectosigmoid cancers demonstrated high tu-
Tablel. Preoperative plasma CEA levels and Dukes’ staging
Dukes’ stage
Plasma CEA level, ng/ml 0< 2.5 2.5 < 5 5 < 10
>10
total
A B C D Total
2 3 0 0 5
1 0 1 3 5
4 6 5 3 18
1 2 1 0 4
0 1 3 0 4
Table 2. CEA mRNA expression and preoperative plasma CEA levels
CEA
Plasma CEA level, ng/ml 0< 2.5 2.5< 5 5 < 10 > 1 0
T> N T< N Total
4 1 5
1 3 4
3 1 4
2 3 5
total 10 8 18
T > N indicates cases where mRNA expression is greater in the tumor than the normal mucosae while T < N indicates higher levels in the normal mucosae.
333
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Fig. 2. Northern blot analysis of CEA mRNA expression in surgical specimens of primary colorectal carcinoma (T) and adjacent nor mal colon mucosa (N) from the same patient. The CEA mRNA ex pression in T was lower than that in N in cases 1 and 2, and was higher than that in N in cases 3,4, and 5, in these five paired, representative samples out of 22. Actin mRNA expression (2.2 kb) was shown as control.
CEA mRNA
Dukes’ stage A
B
c
D
total
T> N T< N Total
3 1 4
4 3 7
2 5 7
2 2 4
11
11 22
Table 4. CEA mRNA expression and tumor
location CEA mRNA
T> N T< N Total
Tumor location proximal
recto sigmoid
total
2(25%) 6(75%) 8
9(64%) 5 (36%) 14
11 11 22
mor/normal ratios, only 25% of proximal co lon cancers had high tumor/normal ratios of CEA mRNA expression (table 4). A larger study will be required, however, to determine if this observation has any clinical signifi cance.
Discussion
With the delineation of the genes encoding the ‘CEA family’, the transcription of these genes in human colonic tissues and tumors has been reported. Cournoyer et al. [14] ob served that the levels of CEA-like mRNAs were approximately the same in colon carci nomas as in adjacent non-invaded mucosa from the same patients. However, Boucher et al. [24] reported a 6- to 10-fold more intense
334
major CEA mRNA band in colon tumors when compared with that in normal colonic mucosa from the same individual in 5 of 6 patients. Our results showed that the level of this CEA mRNA in colorectal carcinoma is 2to 4-fold higher than in adjacent normal co lonic epithelium only in half of the 22 patients in our series. This is a much lower frequency than was expected. Chi et al. [27] showed in a study of 32 colorectal cancers and normal mucosa that only 41 % had higher mRNA lev els for CEA in the tumor while 92% of the tumors expressed higher mRNAs for NCA than in the distant normal mucosa. It has been shown that CEA-related mate rial is present in normal colon mucosa [28— 30] and that purified CEA from normal colon is indistinguishable by immunological and physicochemical criteria from reference colon carcinoma CEA [29, 30]. An immunohistologic study showed that the expression of CEA epitopes 3d6 and COL-4 reacted with a large number of the mucosal biopsies, and that this expression is predominantly localized to the mucosal surfaces and is detectable throughout the colon [31]. Quantitatively, it was demon strated that the concentration of CEA-related material in normal colon mucosa was about 10-40 times lower than in primary large bowel carcinomas and 50-300 times lower than that in metastatic colon or rectum carci nomas [32, 33], This appears inconsistent with our results, which showed in 50% of the cases that the CEA mRNA expression in nor mal colon mucosa was greater than that in tumor. Furthermore, in cases where the CEA mRNA expression in tumor was higher, it was only 2- to 4-fold of that in normal mucosa. Taken together, these findings indicate a dis crepancy between the level of CEA mRNA and protein that suggest translational and/or post-translational events may be important. This is supported by the findings of Boucher et al. [24], who showed that there was no
Mafune/Saini/Ravikumar/Chcn/Steele/ Thomas
CEA mRNA Expression in Colorectal Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
Table 3. CEA mRNA expression and Dukes'
staging
CEA levels with Dukes’ staging observed in this study and reported previously by others [4, 5] suggests that CEA might be an impor tant factor for tumor progression. However, our results also imply that factors other than mRNA expression might be important in reg ulating plasma CEA level; further molecular analysis of CEA expression and secretion may elucidate the discordant clinical data with respect to CEA as tumor marker and prognos ticator. Another question that arises is whether CEA plays a role in the mechanism of colorec tal cancer metastasis. It was shown using a nude mouse model that preinjection with CEA can enhance metastatic potential of hu man colorectal carcinoma [23]. As CEA can function as an intercellular adhesion mole cule, this provides a possible mechanism for CEA’s involvement in metastasis [22], NCA and other members of the immunoglobulin supergene family also have cell adhesion properties [36, 37]. Further studies will eluci date the relative roles of these adhesion mole cules and the role of other receptors for base ment membrane components (e.g. laminin re ceptors) in colorectal tumor metastasis. We will continue to increase the number of cases for this study and investigate CEA mRNA expression in metastatic tumors from colon. Such studies will enable us to evaluate the role of CEA as a factor in the development of metastasis.
Acknowledgments This study was supported by grants CA44583 and CA44704 from the National Cancer Institute, Bcthesda, Md. We thank Dr. Thomas R. Barnett at Molecular Diagnostics, West Haven, Conn., for kindly providing the CEA cDNA clone, LV7, Dr. Bruce Spiegelman at Dana-Farber Cancer Institute, Boston. Mass, for kindly providing the actin cDNA.
335
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
direct proportionality between mRNA levels and gene product expression and by the stud ies of Sato et al. [34] and Chi et al. [27], Dif ferences between mRNA and gene product levels may reflect regulation of CEA mRNA stability by exogenous and endogenous fac tors that affects ultimately the overall turn over of the CEA messenger molecules. It is possible that this regulatory mechanism plays a very important role in CEA secretion and CEA-related metastasis. An alternative expla nation for this discrepancy may be a purely anatomical one and this has been suggested by Kuroki et al. [35], who examined mRNA lev els and CEA secretion in three colorectal can cer-normal mucosa pairs. They found that even though the normal mucosa contained much smaller quantities of CEA than the adjacent tumor the expression of mRNA was comparable between the two. In normal colon tissue, CEA is expressed mainly on the lumi nal surface of the single layer of columnar epi thelial cells lining the upper parts of the crypts and secreted into the colon lumen. In colorec tal cancers or even in polyps, tumor cells no longer conform to the single-cell layer organi zation by bulging into the crypts and invading through the basement membrane in multicel lular arrays and CEA is found along the bor ders between cells [22], Thus, CEA may be secreted into the tumor mass or into lumens where it could fill and block blind ducts. High levels of CEA may then accumulate as a con sequence of its resistance to degradation. Thus, while CEA mRNA levels between nor mal and malignant mucosae are similar, the level of secreted protein in extracts of the tumor tissue is much higher because the CEA is trapped within the tissue. Other factors also regulate plasma CEA levels and include in creased CEA production by tumor invasion, organ-specific metastases and circulating tu mor cells, and factors affecting CEA clearance by the liver [11], The correlation of plasma
1 Gold P, Freedman SO: Specific carcinocmbryonic antigen of the hu man digestive system. J Exp Med 1965:122:467-481. 2 Thomson DMP, Krupcy J, Freed man SO. Gold P: The radioimmu noassay of circulating carcinoembryonic antigen of the digestive sys tem. Proc Natl Acad Sci USA 1969; 64:161-167. 3 Dhar P, Moore T, Zamchcck N, Kupchik HZ: Carcinoembryonic an tigen (CEA) in colonic cancer. Use in preoperative and postoperative diagnosis and prognosis. JAMA 1972:221:31-35. 4 Herrera MA, Chu TM, Holyoke EG: Carcinoembry onic antigen (CEA) as a prognostic and monitoring test in clinically complete resection o f colo rectal carcinoma. Ann Surg 1976: 183:5-9. 5 Goslin R, Steele G Jr, Macintyre J. Mayer R. Sugarbaker P. Cleghorn K. Wilson R, Zamcheck N: The use of preoperative plasma CEA levels for the stratification of patients after curative resection of colorectal can cers. Ann Surg 1980:192:747-751. 6 Sugarbaker PH, Zamchcck N, Moore FD: Assessment of serial car cinoembryonic antigen (CEA) as says in postoperative detection of recurrent colo-rectal cancer. Cancer 1976:38:2310-2315. 7 Welch JP. Donaldson GA: Detec tion and treatment of recurrent can cer of the colon and rectum. Am J Surg 1978:135:505-510. 8 Steele G Jr, Zamcheck N. Wilson R. Mayer R. Lokich J, Rau P, Maltz J: Results of CEA-initiatcd second look surgery for recurrent colorectal cancer. Am J Surg 1980; 139:544— 548. 9 Shuster J, Thomson DMP, Fuks A, Gold P: Immunologic approaches to the diagnosis of malignancy. Proc Exp Tumor Res 1980;25:89-139. 10 Shively J. Beatty J: CEA-related an tigens: Molecular biology and clini cal significance. CRC Crit Rev On col Hematol 1985;2:355-399.
11 Thomas P, Toth CA. Saini KS. Jes sup JM. Steele G Jr: The structure, metabolism and function of the car cinoembryonic antigen gene family. Biochim Biophys Acta 1990:1032: 177-189. 12 von Kleist S, Chavanel G, Burtin P: Identification of an antigen from normal human tissue that cross reacts with the carcinoembryonic antigen. Proc Natl Acad Sci USA 1972:69:2492-2494. 13 Zimmerman W, Weber B. OrtliebB. Rudert F, Schempp W, Fiebig HH, Shively JE, von Kleist S, Thompson JA: Chromosomal localization of the carcinoembryonic antigen gene family and differential expression in various tumors. Cancer Res 1988; 48:2443-2550. 14 Cournoyer D. Beauchemin N, Boucher D, Benchimol S. Fuks A. Stanners CP: Transcription of genes of the carcinoembryonic antigen family in malignant and nonmalignant human tissues. Cancer Res 1988;48:3153-3157. 15 Zimmerman W, Ortlieb R, Fried rich R, von Kleist S: Isolation and characterization of cDNA clones en coding the human carcinoem bryonic antigen reveal a highly con served repeating structure. Proc Natl Acad Sci USA 1987:84:29602964. 16 Oikawa S, Nakazato H. Kozaki G: Primary structure of human carci noembryonic antigen (CEA) de duced from cDNA sequence. Biochcm Biophys Res Commun 1987; 142:511-518. 17 Kamarck ME, Elting JJ. Hart JT, Goebel SJ, Rac PMH. Nothdurft MA, Nedwin JJ, Barnett TR: Carci noembryonic antigen family: Ex pression in a mouse L-cell transfectant and characterization of a partial cDNA in bacteriophage ygtl 1. Proc Natl Acad Sci USA 1987:84:53505354. 18 Beauchemin N, Benchimol S, Cour noyer D. Fuks A, Stanners CP: Isola tion and characterization of fulllength functional cDNA clones for human carcinoembryonic antigen. Mol Cell Biol 1987;7:3221-3230.
19 Tawaragi Y, Oikawa S, Matsuoka Y. Kosaki G, Nakazato H: Primary structure of nonspecific crossrcacting antigen (NCA), a member of the carcinoembryonic antigen (CEA) gene family, deduced from cDNA sequence. Biochem Biophys Res Commun 1988; 150:89—96. 20 Barnett T. Goebel SJ, Nothdurft MA, Elting JJ: Carcinoembryonic antigen family: Characterization of cDNAs coding for NCA and CEA and suggestion of nonrandom se quence variation in their conserved loop-domains. Genomics 1988:3: 59-66. 21 Barnett TR, Kretschmer A, Austin DA, Goebel SJ, Hart JT. Elting JJ. Kamarck ME: Carcinoembryonic antigens: Alternative splicing ac counts for the multiple mRNAs that code for novel members o f the carci noembryonic antigen family. J Cell Biol 1989;108:267-276. 22 Benchimol S, Fuks A, Jothy S. Beauchemin N. Shirota K, Stanners CP: Carcinoembryonic antigen, a hu man tumor marker, functions as in tercellular adhesion molecule. Cell 1989:51:327-334. 23 Hostetter RB. Campbell DE, Chi K. Kerchoff S, Cleary' KR, Ullrich S, Thomas P. Jessup JM: Carcinoem bryonic antigen enhances metastatic potential of human colorectal carci noma. Arch Surg 1990; 125:300— 304. 24 Boucher D, Stanners CP, Fuks A: Studies on the control of gene ex pression of the carcinoembryonic antigen family in human tissue. Cancer Res 1989;49:847-853. 25 Chirgwin J, Przybyla AE. MacDon ald RJ, Rutter WJ: Isolation of bio logically active ribonucleic acid from sources enriched in ribonu cléase. Biochemistry 1979:18:5294— 5299. 26 Spiegelman BM. Frank M, Gree H: Molecular cloning of mRNA from 3T3 adipocytes: Regulation of mRNA content for glycerophos phate dehydrogenase and other differentiation-dependent proteins during adipocyte development. J Biol Chem 1983:258:10083-10089.
336
Mafune/Saini/Ravikumar/Chen/Steele/ Thomas
CEA mRNA Expression in Colorectal Cancer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
References
32 Khoo SK, Warner NL, Lie JT, MacKay IR: Carcinoembryonic an tigen activity of tissue extracts: A quantitative study of malignant and benign neoplasms, cirrhotic liver, normal adult and fetal organs. Int J Cancer 1973;11:681-687. 33 Zamcheck N, Doos WG, Prcdente R, Lurie BB, Gottlieb LS: Prognos tic factors in colon carcinoma: Cor relation of serum carcinoembryonic antigen level and tumor histopathology. Hum Pathol 1975;6:31-45. 34 Sato C, Miyaki M, Oikawa S, Nakasato H. Kosaki G: Differential ex pression of carcinoembryonic anti gen and nonspecific crossreacting antigen genes in human adenocarci nomas and normal colon mucosa. Jp n J Cancer Res 1988;79:433-437.
35 Kuroki M. Arakawa F, Yamamoto H, Shimura H, Ikchara Y, Matsuoka Y: Active production and mem brane anchoring of carcinoem bryonic antigen observed in normal colonic mucosa. Cancer Lett 1988: 43:151-157. 36 Oikawa S, Inuzuka C, Kuroki M, Matsuoka Y, Kosaki G, Nakazato H: Cell adhesion activity of non-spe cific cross-reacting antigen (NCA) and carcinoembryonic antigen (CEA) expressed on CHO cell sur face: Homophilic and hetcrophilic adhesion. Biochem Biophys Res Commun 1989:164:39-45. 37 Zhou H, Fuks A. Stanners C: Speci ficity of intercellular adhesion me diated by various members of the immunoglobulin supergene family. Cell Growth Differ 1990:1:209215.
337
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/25/2018 11:25:32 PM
27 Chi K, Jessup JM , Frazier ML: Pre dominant expression of mRNA cod ing for nonspecific cross-reacting antigen in colorectal carcinoma. Tu mor Biol 1991;12:298-308. 28 Burtin P, Martin E. Sabine MC. von Kleist S: Immunological study of polyps of the colon. J Natl Cancer Inst 1972;48:25-29. 29 Fritsche R, Mach JP: Isolation and characterization of carcinoembryonic antigen (CEA) extracted from normal human colon mucosa. Immunochemistry 1977; 14:119127. 30 Egan ML, Pritchard DG, Todd CW, Go VLW: Isolation and immuno chemical and chemical characteriza tion of carcinoembryonic antigen like substances in colon lavages of healthy individuals. Cancer Res 1977;37:2638-2643. 31 Wolf BC, Salem RR, Sears HF, Horst DA, Lavin PT, Hcrlyn M, Itzkowits SH, Schlom J, Steele GD Jr: The expression of colorectal carci noma-associated antigens in the normal colonic mucosa: An immu nohistochemical analysis of regional distribution. Am J Pathol 1989:135: 111-119.
