Original Paper Oncology 1992;49:209-214

3 Department of Medical Oncology. J. Graham Brown Cancer Center, University of Louisville, Ky., USA b Department of Surgery, College of Medicine. Seoul National University and Seoul National University Hospital, Seoul, Korea c Institute of Medical Pharmacology. University of Pisa, Italy d Biostatistics and Data Management Section, and * NCI-Navy Medical Oncology Branch, National Cancer Institute and Naval Hospital, Bethesda, Md., USA

KeyWords Colon cancer Cell lines Gene expression Clinicobiological correlation

Pattern of Growth Factor, ProtoOncogene and Carcinoembryonic Antigen Gene Expression in Human Colorectal Carcinoma Cell Lines

Abstract The aim of the present study was to examine whether the expression of growth factor genes, proto-oncogenes and carcinoembryonic antigen (CEA) gene in human colorectal cancer cell lines was related to their clinicobiological beha­ vior. A significant variability among cell lines was detected for both insulin-like growth factor II and transforming growth factor (3 gene message. Detectable levels of c-myc, Her-2, c-myb, K-rrw and EGF receptor mRN A were found in mostcell lines, whereasonly 1/11 and 2/11 cell lines were positiveforN-»nrand c-sis message, respectively. N -/m r expression was limited to a cell line orig­ inated from a tumor with neuroendocrine features, while high levels of K-ras message were found only in a cell line derived from a radioresistant tumor. CEA mRN A levels correlated well with the concentration of antigen in each cell line. On the basis of these results, our findings demonstrated that human colorec­ tal cancer cell lines show heterogeneous expression of growth factor and CEA genes and proto-oncogenes; however, with the exception of K-ras, N -myc and CEA, other correlations between gene expression and the clinicobiological characteristics of these cell lines could not be demonstrated.

Introduction Colorectal cancer is the third most common neoplastic disease in the United States with respect to incidence and is second in mortality, after lung cancer [16]. In spite of ad­ vances made in diagnosis and therapy, the clinical out­ come of patients is still disappointing. Today, carcinoem­ bryonic antigen (CEA) remains the principal serum tumor marker to monitor response to therapy and tumor recur­ rence; it is part of a family of closely related glycoproteins whose biological functions remain unknown [11 ]. With the

advent of molecular biology techniques, the identification of genes associated with the biological characteristics of a particular malignancy may indeed result in a new genera­ tion of tumor markers [7], Cytogenetic analysis of the human cancer cell lines, object of the present study, has revealed that three of them contain homogeneous staining regions, whereas five have greater than 50% of their meta­ phases containing double minute chromosomes [10], a characteristic of cells having amplified gene sequences. The aim of this study was to characterize the pattern of proto-oncogene, growth factor and CEA gene expression

Romano Danesi, MD Institute o f Medical Pharmacology University of Pisa ~ Via Roma 55 I 56126 Pisa (Italy)

< 1992 S.K arger AG. Basel 0030 2414/92/ 0493-0209 S 2.75/0

Downloaded by: University of Connecticut 132.174.250.220 - 6/12/2018 8:42:30 AM

Renato V. La Roccaa Jae-Gahb Park b Romano Danesic Mario Del Taccac Seth M. Steinbergd Adi F. Gazdarc

in human colorectal cancer cell lines and to identify genet­ ically discrete subgroups, which in turn could correlate with their clinicobiological behavior.

Materials and Methods Cell Lines and Nucleic Acid Analysis The human colorectal cancer cell lines NCI-H548, NCI-H630. NCI-H684. NCI-11958, NCI-H498S, NCI-H508. NCI-H7I6. SNUC l, SNU-C2, SNU-C4, and SNU-C5 [10] were cultivated in RPMI 1640 plus 10% heat-inactivated fetal bovine serum (Gibco, Grand Island, N.Y., USA) and incubated at 37 C in an atmosphere of 5% CCb. 95% air and 100% humidity. High molecular weight genomic DN A was isolated from cell lines [8]; 20 pg were digested with specified restriction enzymes (BRL, Rockville, Md., USA), run on 0.8% agarose gel inTris-borate buffer and transferred to nitrocellulose filters. Total cellular RNA was ex­ tracted from cells using the guanidinum isothiocyanate/cesium chloride method [8]: 30 gg were loaded on 1.1% agarose-formaldehyde gel and transferred to nylon membranes. Filters were from Schleicher & Schuell (Keene, N.H., USA). cDNA probes were radiolabellcd with [u-':P]dCTP (3,000 Ci/ mmol, 10 mCi/ml, NEN-Dupont, Wilmington, Del., USA) by nick translation to a specific activity of « 10* cpm/gg of DNA. The fol­ lowing DNA probes were used: c-myc: pmyc3pst [ 18]; N-myc: pNB 1 [13]; L-myc: 1.8-kb fragment cloned into the SmaI-£coRI sites of pJB327 [9]: c-myb: 2.2-kb Hae III cDNA fragment of a human c-myb cDNA plasmid (courtesy of Dr. J. Fredric Mushinski, NCI, Bethesda, Md., USA); epidermal grow'th factor receptor (EGP'-R): 2.4-kb cDNA fragment containing the human EGF-R cloned into the Cla\ site of pBR322 [21]: K-ras: 0.618-kb fragment and v-sis: 0.983-kb fragment both from Oncor (Gaithersburg, Md., USA); Her-2: cDNA fragment pHER2-436-3 cloned into the £coRI site ofpUC12 [3]; insulin-like growth factor II (IGF-II): 0.78-kb cDNA clone of rat IGF-II inserted into the Pst I site of pUC12 [20]; gastrin-releasing peptide (GRP): 0.9-kb insert cloned into pBR322 (courtesy of Dr. J. Battey, NCI-Navy, Bethesda. Md., USA); transforming growth fac­ tor P (TGF-P): 1.05-kb fragment cloned in the £coRI site of pSP64 [5];CEA: pCEA 1clone (courtesy of Dr. W. Zimmerman) [22]. Prehy­ bridization, hybridization and stringent washing were performed fol­

Results Growth factor gene expression showed significant va­ riability among cell lines (table 2). IGF-II transcripts (6.0, 4.0, 2.2 and 1.7 kb) were detected in two well-differen­ tiated cell lines, NCI-H548 and NCI-H684 (fig. 1, top). TGF-P mRNA was found in eight cell lines, with marked variability in the degree of expression, ranging from high levels (NCI-H716) to trace levels (SNU-C2 and NCIH630, fig. 1, middle). Detectable c-sis message (4.2 kb) was found in two poorly differentiated cell lines derived from xenografts in nude mouse, SNU-C2 (fig. 2, bottom) and SNU-C5. GRP mRNA could not be detected. Detectable levels of c-myc, EGF-R and K-ra.v message were found in all cell lines (table 2) without variations in transcript size. The degree of c-myc expression was rela­ tively uniform among cell lines, whereas EGF-R expres­ sion varied in intensity. K-rasmRNA levels were uniform among cell lines with the exception of NC1-H630, a cell line derived from a radioresistant tumor, which showed a marked gene amplification (fig. 2). Detectable levels of the

Patient age, sex, race or blood group Site or size of primary tumor Site of cell line culture Dukes' stage at diagnosis or at the time of cell line Lymph node involvement Prior radiotherapy or chemotherapy Tumor histology Differentiation of primary tumor or cell line Presence of double minute chromosomes a

210

Data Analysis and Statistics The signal intensity of messages was compared to the colorectal cell line NCI-H548 which was chosen since it expresses detectable levels of all genes studied except for L-myc, c-sis and GRP. The signal intensity was examined for association with the parameters listed in table 1 by using various data analysis techniques. When appropriate, the Spearman rank correlation coefficient was calculated, and y; tests of association were performed to assess the strength of the relation­ ships among the data.

Cloning efficiency in agar or in media Cell doubling time Plating efficiency Pattern of growth (monolayer or cell suspension) Tumorigenicity in nude mouse Chemosensitivity' CEA protein (cell/supernatant) CA19-9 protein (cell/supernatant) TAG-72 protein (cell/supernatant)

Drugs: 5-fluorouracil, doxorubicin, ch-DDP, melphalan, etoposide, BCNU.

La Rocca/Park/Danesi/Del Tacca/ Steinberg/Gazdar

Gene Expression in Colon Cancer Cell Lines

Downloaded by: University of Connecticut 132.174.250.220 - 6/12/2018 8:42:30 AM

Table 1 . Parameters examined for cor­ relation to growth factor and CEA gene and proto-oncogene expression

lowing the specifications of the filter manufacturer. Sample blots were rehybridized with P-actin probe to document equivalent amounts of intact RNA per lane. Autoradiograms of nucleic acid fil­ ters were evaluated by densitometry.

Fig-1 . Growth factor gene expression in human colorectal cancer cell lines. Total RNA (30 pg) was hybridized with [’’P]-labelled IGFII (top), TGF-p (middle) and v-sis(bottom)cDNA probes. Fig. 2. Amplification of K-ra.v in a human colorectal tumor cell line. High molecular weight genomic DNA (20 pg) was digested with £eoRI. Blot was hybridized with [ 100fold as intense as those for the other cell lines.

Table 2. Growth factor geneand proto-oncogene expression in colorectal cancer cell lines

TGF-p

Differentiation

Cell line

c-myc

N-myc

L-mvc

c-myb

EGF-R K-ras

Her-2

c-sis

IGF-II

GRP

Well

NCI-H548 NCI-H630 NCI-H684 NCI-H958

+ + + +

+ 0 0 0

0 0 0 0

+ + + +

-F

+ /0 ++

+ + ++ + +

+ + + +

0 0 0 0

+ 0 + 0

0 0 0 0

0 +

Mucinous

NCI-H498S

+

0

0

+

+

+

ND

0

0

0

+

Moderate

NCI-H508 SNU-C1

+ +

0 0

0 0

+ + + +

+ /0 + /0

+ +

++ +

0 0

0 0

0 0

1 0

Poor

NCI-H716 SNU-C2 SNU-C4 SNU-C5

+ + + •

0 0 0 0

0 0 0 0

-F/0 0 + + /0

+ ++

+

0

0

+

+

+ +

+ +

0 +

0 0 0 0

0 0 0 0

+ + + /0 + ND

+/o

+/o ++

+

+

+/o

211

Downloaded by: University of Connecticut 132.174.250.220 - 6/12/2018 8:42:30 AM

0 = No detectable signal; -1-/0 = signal intensity that is detectable but less than of NCI-H548; + = signal intensity equivalent to that of NCI-H548 (in the case of c-sis, + denotes detectable signal only; for L-myc and GRP. known positive controls were used); + + = at least twofold increase in signal intensity as compared to that of NCI-H548; -F + + = 100-fold or greater increase in signal intensity as compared to that of NCI-H548; ND = not determined.

Fig. 3. CEA gene expression in human colorectal cancer cell lines and correlation with previously reported CEA protein levels derived from cell extracts. Total RNA was hybridized with [s’P]-labelled pCEA I probe. In the lower panel is shown the CEA antigen concentration/106 cells from each of the eleven colorectal cell lines as well as an ad­ ditional cell line (NCI-H747).

Discussion The results of the present study demonstrated that growth factor and CEA genes and proto-oncogenes show heterogeneous expression among the human colorectal cancer cell lines studied. A correlation between theclinicobiological characteristics of cell lines and the amplification of a particular gene was found in cells derived from tumors characterized by radioresistance (K -ras), neuroendocrine features (N-mvc) and elevated production of CEA antigen (CEA gene). It has been shown that EGF-R levels were increased in moderately to well-difTerentiated human colon tumor cell lines as compared with poorly differentiated ones [1].

212

EGF-R transcripts levels, however, do not appear in our hands to correlate well with either the degree of differentia­ tion or with the CEA antigen production by the cell line; this would imply that EGF-R is posttranscriptionally re­ gulated. A marked variability was observed in the degree of expression of IGF-II, TGF-P and c-sis, which encodes for the B chain of platelet-derived growth factor [19], IGFII message was present in two well-differentiated cell lines originated from tumors of the sigmoid colon. Marked variability in the degree of expression ofTGF-P among the cell lines was found, with two of them (NCI-H684 and SNU-C1) lacking gene mRNA. TGF-P elicits a differen­ tiation-like effect as well as inhibition of proliferation on a variety of epithelial cells [17]. However, we were unable to find a correlation between TGF-P message and the degree of differentiation, since NCI-H716, a poorly differentiated cell line, showed the highest level ofTG F-P expression. c-sis mRNA was found in two poorly differentiated cell lines. c-myc message was found in each cell line; a correlation between c-myc expression of a primary colon tumor and its more distal location along the gastrointestinal tract [12] or a well-differentiated phenotype [15] has been reported. We have found that c-myc expression was relatively uniform

La Rocca/Park Danesi/Dcl Tacca/ Steinberg/Gazdar

Gene Expression in Colon Cancer Cell Lines

Downloaded by: University of Connecticut 132.174.250.220 - 6/12/2018 8:42:30 AM

4.6-kb Her-2 message could be seen in 9 colon cancer cell lines. Expression of c-myb was found in ten cell lines, whereas N-myc expression was limited to NCI-H548, which originated from a tumor with neuroendocrine fea­ tures. L-myc expression was not detected (table 2). CEA mRNA was demonstrated in five cell lines; a correlation between a detectable CEA message and a concentration of CEA antigen greater than 500 ng/106cells is noted (fig. 3).

among the cell lines regardless of the source (primary ver­ sus metastasis), the histologic grading or location of the primary tumor. As noted, K-rai expression was uniform in the eleven colorectal cell lines tested with the exception of NCI-H630, which demonstrated gene amplification and markedly increased gene expression; this line was also re­ markable for the presence of homogeneous staining re­ gions. NC1-H630 originated from a tumor that had been previously irradiated and subsequently progressed, and it was found to be the most radioresistant of four colorectal cancer cell lines [2], ras oncogenes have in fact been asso­ ciated with an increased resistance to ionizing radiations [14]. N-mvc expression has previously been associated with cells of neuroendocrine origin [6] and was detected in NCIH548 only, derived from a well-differentiated neoplasm of the sigmoid colon with carcinoid features; these features, however, were not present in the derived cell line, and no evidence for chromogranin A mRNA could be observed. Thus, N -mycexpression in a colon carcinoma cell line may indicate the presence of neuroendocrine features in the pri­ mary tumor specimen. NCI-H716, derived from a poorly differentiated carcinoma of the caecum, did contain cyto­ plasmic granules with dense core, elevated DOPA decar­ boxylase (DDC) activity as well as chromogranin A gene message. It lacked N-/mrmRN A but did show c-myc gene amplification, elevated levels of TGF-(3 mRNA and ab­ sence of detectable Her-2 message. Subcutaneous xeno­ grafts of NCI-H716 readily invaded surrounding muscle tissue, unlike the xenografts from other colorectal cell lines [10]. Also, this cell line had the highest cloning efficiency of

the eleven cell lines tested. It is worth noting that NCIH716 was derived from a patient with a history of ulcera­ tive colitis. In our analysis c-myb message was found in ten cell lines, but no significant correlation was found with re­ spect to its expression and the differentiation or doubling time of each cell line. pCEAl gene message correlated well with the amount of CEA present in the cell line extract: it could be found only in cell lines whose CEA concentration was equal to or greater than 500 ng/106cells; this implies that CEA is not posttranscriptionally regulated. Although there was heterogeneous expression of pCEAl mRNA. in none of the poorly differentiated lines could the pCEA I transcript be found; thus, as in the case of CEA antigen, the presence of detectable pCE A 1transcript correlates with a more dif­ ferentiated cell phenotype [4]. Finally, with the exception of K-ras, N -myc and CEA gene, whose levels of expression were associated with tu­ mors characterized by radioresistance, neuroendocrine features, or elevated CEA production, respectively, the biological characteristics of human colorectal cancer cell lines evaluated in the present study are not correlated with their pattern of gene expression.

Acknowledgments R.D. gratefully acknowledges the financial support from the Italian Association forCancer Research and European Organization for Research and Treatment of Cancer-U.S. National Cancer In­ stitute Exchange Training Program.

References 4 Denk H.TappeincrG, Eckerstorer R, Holzner JH: Carcinoembryonic antigen (CEA) in gas­ trointestinal and extragastrointestinal tumors and its relationship to tumor-cell differentia­ tion. Int. J Cancer 1972;10:262-272. 5 Dcrynck R, Jarrct JA. Chen EY. Eaton DH. Bell J R, Assoian R K, Roberts A B. Sporn M B. Gocddel DV: Human transforming growth l'actor-p complementary DNA sequence and expression in normal and transformed cells. Nature 1985;316:701-705. 6 Kohl NE, Gee CE, Alt FW: Activated expres­ sion of the N-myc gene in human neuroblasto­ mas and related tumors. Science I984;226: 1335-1337.

7 Israel MA. Hclman LJ, Miser J: Patterns of proto-oncogene expression: A novel approach to the development of tumor markers; in De Vita VT. Heilman S, Rosenberg SA (eds): Im­ portant Advances in Oncology. Philadelphia. Lippincott, 1987, pp 87 104. 8 Maniatis T. Fritsch EF. Sambrook J: Mo­ lecular Cloning: A Laboratory Manual, ed 2: New York, Cold Spring Harbor. 1989. 9 Nau MM. Brooks BJ, Battcy J. Sausvillc E. Gazdar A F, Kirsch IR, McBride OW, Bertness V, Hollis G F, Minna JD: L-mrr, a new mye-rclatcd gene amplified and expressed in human small cell lung cancer. Nature 1985; 318:69-73.

213

Downloaded by: University of Connecticut 132.174.250.220 - 6/12/2018 8:42:30 AM

1 Bradley SJ. Garfinklc G. Walker E, Salem R. Chen LB. Steele G Jr: Increased expression of the epidermal growth factor receptor on hu­ man colon carcinoma cells. Arch Surg 1986; 121:1242-1247. 2 Carmichael J. Park JG. DeGraff WG. Gamson J. Gazdar A F, Mitchell J B: Radiation sen­ sitivity and study of glutathione and related enzymes in human colorectal cancer cell lines. Eur J Cancer Clin Oncol 1989;24:1219-1224. 3 Coussens L. Yang-Feng TL, Liao YC, Chen E, Gray A, McGrath J, Seeburg PH. Libcrmann TA, Schlessinger J, Francke U, Levin­ son A. Ullrich A: Tyrosine kinase receptor with extensive homology to EGF receptor shares chromosomal location with neu oncogene. Science 1985;230:1132-1 139.

15 Sikora K, Chan S, Evan G, Gabra H. Mark­ ham N, Stewart J. Watson J: c-myc oncogene expression in colorectal cancer. Cancer 1987;59:1289-1295. 16 Silverberg E. LuberaJ: Cancer statistics. 1986. CA 1986:36:9 25. 17 Sporn MB. Roberts AB. Wakefield LM, Assoian RK: Transforming growth factor p: Biological function and chemical structure. Science 1986;233:532-534. 18 Vennstrom B. Moscovici C, Goodman HM, Bishop JM: Molecular cloning of the avian myelocytomatosis virus genome and recovery of infectious virus by transfection of chicken cells. J Virol 1981;39:625-631. 19 Waterfield MD.ScraceGT. Whittle N.Stroobant P, Johnsson A. Wasteson A, Westermark B. Heldin CH. Huang JS, Deuel TF: Platelet derived growth factors structurally related to the putative transforming protein p28"\ Na­ ture 1983;304:35-39.

20 W hitficldHJ.BruniCB,FrunzioR,TerrelJE, Nissley SP. Rcchler M M: Isolation of a cDNA clone encoding rat insulin-like growth factorII precursor. Nature 1984;312:277-280. 21 Xu YH, Richert N. ItoS. MerlinoGT, Pastan I: Characterization of epidermal growth factor receptor gene expression in malignant and nor­ mal human cell lines. Proc Natl Acad Sci USA 1984:81:7308-7312. 22 Zimmerman W, Ortlieb B. Friedrich R, von Klcist S: Isolation and characterization of cDNA clones encoding the human carcinocmbryonic antigen reveal a highly conserved re­ peating structure. Proc Natl Acad Sci USA 1987;84:2960-2964.

214

La Rocca/Park/Danesi/Del Tacca/ Steinberg/Gazdar

Gene Expression in Colon Cancer Cell Lines

Downloaded by: University of Connecticut 132.174.250.220 - 6/12/2018 8:42:30 AM

10 Park JG, Oie HK. Sugarbaker PH. Henslee JG. Chen TR, Johnson BE, Gazdar A: Characteristics of cell lines established from human colorectal carcinoma. Cancer Res 1987:47:6710-6718. 11 Rogers GT: Carcinoembryonic antigens and related glycoproteins: Molecular aspects and specificity. Biochem Biophys Acta 1983:695: 227-243. 12 Rothberg PG: The role of the oncogene c-myc in sporadic large bowel cancer and familial Polyposis coli. Semin Surg Oncol 1987;3: 152 158. 13 Schwab M. Alitalo K, Klempnaucr K H, Varmus HE. Bishop JM. Gilbert F. Brodcur G, Goldstein M. Trent J: Amplified DNA with limited homology to myc cellular oncogene is shared by human neuroblastoma cell lines and a neuroblastoma tumor. Nature 1983:305: 245-248. 14 Sklar M D: The rus oncogene increases the in­ trinsic resistance of NIH-3T3 cells to ionizing radiations. Science 1988:239:645-647.

Pattern of growth factor, proto-oncogene and carcinoembryonic antigen gene expression in human colorectal carcinoma cell lines.

The aim of the present study was to examine whether the expression of growth factor genes, proto-oncogenes and carcinoembryonic antigen (CEA) gene in ...
970KB Sizes 0 Downloads 0 Views