influence of the organ microenvironment on the ex- pression of CEA in human KM colon carcinoma cell lines with different metastatic potentials. Based on.
American Jouernal of Pathology, Vol. 149, No. 4, October 1996 Copyright © American Society for Investigative Pathology
Regulation of Carcinoembryonic Antigen Expression in Human Colon Carcinoma Cells by the Organ Microenvironment
Yasuhiko Kitadai,* Robert Radinsky,* Corazon D. Bucana,* Yutaka Takahashi,* Keping Xie,* Eiichi Tahara,t and Isaiah J. Fidler* From the Department of Cell Biology,* The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030 and the Department of Pathology,t Hiroshima University School of Medicine, Hiroshima, Japan
The expression of carcinoembryonic antigen (CEA) is thought to be involved in homotypic adhesion and has been associated with the progression of human colon carcinomas (HCC) to the metastatic state. Three cell lines established from surgical specimens of Dukes' stage D (KM20) or Dukes' stage B (KM12C, KM12SM) with high and low preoperative CEA serum levels, respectively, were studied subsequent to growth in culture, in the subcutis (ectopic) or cecal waU (orthotopic) of nude mice. In all cell lines, CEA expression was higher in cecal tumors than in subcutaneous lesions. The degree of differentiation and CEA expression by HCCgrowing in the cecal waU of nude mice correlated with the pathological diagnosis and preoperative CEA level of the original patients. To better understand the regulation of CEA expression, the HCC ceUs were grown in culture as sparse and confluent monolayers or as multiceUl spheroids. The CEA expression level increased in aU three ceU lines growing as confluent monolayers and was highest in multiceU spheroids. Treatment of sparse monolayer cultures of KM12SM ceUs with mitomycin-C inhibited ceU division and was associated with higher production of CEA protein. Moreover, conditioned media from confluent monolayer cultures or from spheroids up-regulated CEA production in sparse monolayer ceUs. These data show that CEA expression in HCC ceUs may be regulated by ceUl density and by factors from the organ environment. (Am J Pathol 1996, 149:1157-1166)
Carcinoembryonic antigen (CEA), first described in 1965 by Gold and Freedman,1 is the most widely used clinical tumor marker for several neoplastic diseases, especially colorectal carcinomas.23 CEA serves as a prognostic marker because, in most cases, an elevated preoperative serum CEA level is associated with a poor prognosis.2'3 Recent studies have demonstrated that CEA is a member of an immunoglobulin supergene family functioning as an intercellular adhesion molecule.4 8 Cells expressing CEA have been shown to aggregate under in vitro conditions, suggesting that CEA may promote the adhesion of tumor cells to each other (homotypic) or to host cells (heterotypic).8 Hence, tumor cells in the circulating aggregates have an increased capacity to arrest in capillary beds, which gives the cells an increased chance to produce metastases.9 Although the expression of CEA by human colon carcinoma (HCC) cells has been directly correlated with their metastatic potential,10'11 its regulation is poorly understood.12 Some recent evidence indicates that the correlation may be causal. We have recently demonstrated that the organ microenvironment can profoundly influence the biological behavior of tumor cells, including growth,13 invasion,14,15 angiogenesis, 16 17 resistance to chemotherapy,18 and differentiation.19 Previous reports have demonstrated a positive relationship between the degree of differentiation of colorectal cancer and CEA production.10 Several agents that alter the degree of cellular differentiation have also been shown to increase the level of CEA expression in these cell types.20'21 To examine a potential mechanism for organ-specific metastasis produced by HCC, we studied the Supported in part by the University Cancer Foundation, Cancer Center Support Core grant CA 16672, and grants R29-CA 67952 (RR) and R35-CA 42107 (IJF) from the National Cancer Institute, National Institutes of Health, and by the Josef Steiner Foundation. Accepted for publication June 12, 1996. Address reprint requests to Dr. Isaiah J. Fidler, Department of Cell Biology, Box 173, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77030.
1157
1158
Kitadai et al
A/P October 1996, Vol. 149, No. 4
influence of the organ microenvironment on the expression of CEA in human KM colon carcinoma cell lines with different metastatic potentials. Based on the relation between tumor size and CEA expression, we also examined the effect of cell density on the production of CEA. The results demonstrate that both growth in orthotopic organs and homotypic cell density in vitro induce cellular differentiation and CEA expression in the HCC cell lines.
Materials and Methods Human Colon Carcinoma Cell Lines The heterogeneous, poorly metastatic KM12C cell line was originally isolated from a primary colon carcinoma classified as Dukes' stage B2.2' The highly metastatic KM12SM cell line was isolated from a liver metastasis produced by parental KM12C cells growing in the cecal wall of a nude mouse.22 The highly metastatic KM20 cells were established in culture from a primary colon carcinoma classified as Dukes' stage D.23
Animals and Production of Tumors Male athymic BALB/c nude mice were obtained from the Animal Production Area of the National Cancer Institute-Frederick Cancer Research Facility (Frederick, MD). The mice were maintained under specific pathogen-free conditions in facilities approved by the American Association for Accreditation of Laboratory Animal Care and in accordance with current regulations and standards of the United States Department of Agriculture, Department of Health and Human Services, and the National Institutes of Health. The mice were used according to institutional guidelines when they were 8 weeks old. To produce tumors, the cells growing in culture were harvested by a brief treatment with 0.25% trypsin and 0.02% EDTA. A single-cell suspension of 1 x 106 cells with a viability of >95% were implanted into the subcutis or cecal wall of nude mice.22 After 3 to 4 weeks, the mice were killed, and tumors exceeding 6 mm in diameter were resected for the study.
Histology Tumors in nude mice were fixed with 10% buffered formalin, paraffin embedded, processed routinely, and stained with hematoxylin and eosin. Alcian blue staining was used to detect mucin-producing cells. Assignment of the degree of differentiation was
based on how gland-like the tumor was, and the volume and number of mucin-producing cells.
In Vitro Culture Conditions All tumor cell lines were maintained as monolayer cultures in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS), sodium pyruvate, nonessential amino acids, L-glutamine, and twofold vitamin solution (GIBCO, Grand Island, NY). The cell cultures were maintained on plastic and were incubated in 5% C02/95% air at 37°C. The cultures were free of Mycoplasma and the following pathogenic murine viruses: reovirus type 3, pneumonia virus, K virus, Theiler's encephalitis virus, Sendai virus, minute virus, mouse adenovirus, mouse hepatitis virus, lymphocytic choriomeningitis virus, ectromelia virus, and lactate dehydrogenase virus (assayed by M. A. Bioproducts, Walkersville, MD). The cultures were maintained for no longer than 12 weeks after recovery from frozen stocks.
Spheroid Cultures Tumor cells (2 x 105) were seeded onto 1% agarose-coated 24-well plates. Because cells were prevented from attaching to the plastic surface, large multicell spheroids were formed.24 After 3 to 5 days, the spheroids were quickly frozen in liquid nitrogen for mRNA extraction or processed for immunohistochemistry.
Northern Blot Analysis For Northern blot analysis, we used in vivo growing tumors and HCC cells growing as sparse (90% confluent) monolayers, or as multicell spheroids. Polyadenylated mRNA was extracted from the cultured cells or tumor tissue using the FastTrack mRNA isolation kit (Invitrogen Co., San Diego, CA). The mRNA was electrophoresed on a 1 % denaturing formaldehyde/agarose gel, electrotransferred at 0.6 A to a GeneScreen nylon membrane (DuPont Co., Boston, MA), and ultraviolet light cross-linked with 120,000 ,uJ/cm2 using a UV Stratalinker 1800 (Stratagene, La Jolla, CA). Hybridizations were performed as described previously.25 Nylon filters were washed at 650C with 30 mmol/L NaCI, 3 mmol/L sodium citrate (pH 7.2), and 0.1% sodium dodecyl sulfate (w/v).
cDNA Probes The cDNA probes used in this analysis were a 1.3-kb Pstl cDNA fragment corresponding to rat GAPDH26
Regulation of CEA Expression in HCC Cells 1159 AJP October 1996, Vol. 149, No. 4
and a 0.8-kb Pstl cDNA fragment of human CEA corresponding to the open reading frame in CEA that contains two closely related repeats of 178 amino acids each. pCEA1 contains repeat a and part of repeat b.7 Each cDNA fragment was purified by agarose gel electrophoresis, recovered using GeneClean (BIO 101, Inc., La Jolla, CA), and radiolabeled with the random primer technique using [a-32P]deoxyribonucleotide triphosphates.27
diamino benzidine (Research Genetics, Huntsville, AL). The sections were then washed three times with distilled water, counterstained with Mayer's hematoxylin (Biogenex Laboratories, San Ramon, CA), washed once with distilled water and once with PBS, and rinsed again with distilled water. The slides were mounted with Universal mount (Research Genetics) and examined in a brightfield microscope. A positive reaction was indicated by a reddish brown precipitate in the cytoplasm.
Densitometric Quantitation The cDNA probe detected three bands. The 3.0-kb band is the NCA transcript, and the 3.6-kb and 4.0-kb bands are CEA-specific mRNA transcripts. Steady-state mRNA expression (3.6 kb and 4.0 kb) was quantitated by densitometry of autoradiograms using the Image Quant software program (Molecular Dynamics, Sunnyvale, CA). Each sample measurement was calculated as the ratio of the average area of the specific mRNA transcripts to the average area of the 1.3-kb GAPDH mRNA transcript in the linear range of the film.
Immunohistochemistry HCC cells were cultured in chamber slides (Nunk, Naperville, IL) under sparse (90% confluent) conditions. The cells were fixed with cold acetone (-20°C) for 10 minutes at room temperature and rinsed with phosphate-buffered saline (PBS). Frozen sections (8 ,tm) of spheroid cultures were also fixed with cold acetone. (Samples were fixed and stored in PBS at 4°C if the procedure could not be finished on the same day.) Tissue sections (4 ,um) of formalin-fixed, paraffinembedded specimens were deparaffinized in xylene, dehydrated in alcohol, and transferred to PBS. The slides were rinsed twice with PBS, and endogenous peroxidase was blocked using 3% hydrogen peroxide in PBS for 12 minutes. The samples were washed three times with PBS and incubated for 20 minutes at room temperature with a protein-blocking solution consisting of PBS (pH 7.5) containing 5% normal horse serum. Excess blocking solution was drained and the samples were incubated for 18 hours at 40C with the appropriate dilution of monoclonal mouse anti-human CEA (DAKO-CEA, 11-7, Dako, Denmark). The samples were then rinsed four times with PBS and incubated for 60 minutes at room temperature with the appropriate dilution of peroxidase-conjugated rat anti-mouse IgG (H&L) (Boehringer Mannheim, Indianapolis, IN). The slides were rinsed with PBS and incubated for 5 minutes with
CEA Assay Different numbers of HCC cells were plated into 38-mm2 wells of 96-well plates. After 4 or 5 days, the cultures were washed, frozen in Hanks' balanced salt solution, and thawed. After cell lysis, expression of cell-associated CEA protein was analyzed by enzyme immunoassay (EIA) (Abbot CEA-EIA Monoclonal One-Step, Abbott Laboratories, North Chicago, IL). The concentration of CEA in the unknown samples was determined by comparing the optical density of the samples to the standard curve. The CEA protein level was normalized to viable cell number.
Results Steady-State CEM Expression Level of HCC Cells In the first set of experiments, we examined the steady-state expression levels of CEA mRNA and protein in cultured HCC cells and compared them with the preoperative serum CEA levels in the original patients. The KM12C (low metastatic) and KM20 (highly metastatic) cell lines were isolated from Dukes' stage B2 (poorly differentiated adenocarcinoma, serum CEA 6.7 ng/ml) and Dukes' stage D patients (moderately differentiated adenocarcinoma, serum CEA 281.0 ng/ml), respectively.22'23 Sparse monolayer cultures (50% confluence) were used for Northern blot analysis and EIA. Sparse cultures of KM12 and KM20 cells expressed similar levels of CEA mRNA (3.6 kb and 4.0 kb) (Figure 1A) and protein (Figure 18). KM12SM, a highly metastatic variant cell isolated from the low metastatic parental KM12C cell, constitutively expressed three- to fivefold the levels of CEA mRNA and protein as the KM12C and KM20 cells (Figure 1). Thus, the level of CEA expression in the sparsely growing monolayer cultures did not correlate with preoperative serum CEA levels of the original patients.
Kitadai et al
1160
AJP October 1996, Vol. 149, No. 4
B.
A.
_ 3
4"t
Cj Ao,_k r"q
hev-,
0
kb -4 .0 - 3 .6
CEA
o
2
_
c* .
.3
GAPDH
- 1.
a.
< 0
1.0
3.0 0.9
0
KMI12C KM12SM KM20
Figure 1. Expression of CEA mRNA and protein in HCC cell lines. (A) Northern blot analysis. Polyadenylated mRNA (2 gg/lane) was lused to detect CEA mRNA transcripts in KM12C, KM12SM, and KM20 cells grotving as sparse monolayers. A rat GAPDHprobe was used as an internal control. The numbers represent densitometric quantitation of the ratio of the area between the specific CEA transcripts and the GAPDH transcript with the value for KM12C cells defined as 1.0. (B) CEA cellular protein level in HCC cells assayed by an EIA technique. Tumor cells (66 cells/mm2) were plated in 96-u'ellplates and harvested after 5 days ofgrowth. The cells were washed in cold Hanks' balanced salt solution, and celllular CEA protein was assayed in cell lysates after ote freeze-thaw cycle and normalized to viable cell number. This is one representative experiment of three.
Expression Level of CEA in HCC Lines Growing in Nude Mice In the next set of experiments, we examined the CEA expression level in HCC cells growing in the subcutis or cecal wall of nude mice. Tumor cells growing in culture were used as a control for baseline CEA expression. KM12C and KM12SM cells growing in the cecal wall of nude mice produced higher levels of CEA-specific mRNA transcripts than the same cells growing in the subcutis or in culture (Figure 2). KM12SM
KM12C
KM20
0
4-* S
-0
-4P
4_C0
0v
AP
Cq
kb
CEA
GAPDH
*
-v I
-4.0
*,
!
******
*O
Cell Density-Dependent Expression qf CEA in Cultures of KM12SM Cells
Cells/mm2 at plating
Cells/mm2 at harvest
-1.3
33 66 131 262 524 1048
2210 2395 2500 2868 3895 3342
internal control. The numbers shown are the densitonietric quantitation of the ratios of the area betueen the specific CEA transcripts and the GAPDH transcript compared in each case u'ith the respective cultured cells defined as 1.0. This is one representative experiment of three.
Table 1.
-3.6
1.0 2.1 7.0 1.0 1.6 6.2 1.0 22.3 28.7 Figure 2. Steady-state expression of CEA mRNA by HCC' cells grolning in cuilture or in nude mice. Polyadenylated mRNA (2 jg/lane) was used to detect C'EA mRNA transcripts. A rat GAPDHprobe was uised as an
The CEA mRNA expression by KM20 cells was significantly increased in both orthotopic tumors (cecum) and ectopic (s.c.) tumors (Figure 2). When HCC cells growing in nude mice were adapted to growth in culture, their expression of CEA mRNA decreased to basal level (data not shown). Immunohistochemical studies confirmed these results. The production of CEA by HCC tumors growing in the wall of the cecum directly correlated with metastatic potential. The staining intensity with anti-CEA antibody was weak in the low metastatic KM12C cells (Figure 3B), stronger in the KM12SM cells (Figure 3D), and strongest in the KM20 cells (Figure 3F). Regardless of the cell line used and the tumor cell density in the lesions, s.c. tumors (Figure 3A, 3C,
Cellular CEA ng/105 cells 1.80 2.46 3.07 4.40 5.15 6.14
+
± ± + + +
0.04 0.15 0.20 0.08 0.19 0.20
KM12SM cells were plated at the indicated densities into 38mm wells of 96-well plates and harvested after 5 days. Separate plates were used for cell counts (hemocytometer) and CEA assay. Values are the mean + S.D. of triplicate cultures. For CEA assay, the harvested cells were lysed by one freeze-thaw cycle and analyzed using EIA kit. The values were normalized to viable cell numbers.
Regulation of CEA Expression in HCC Cells
1161
AJP October 1996, Vol. 149, No. 4
,41-4
\
; ; -A 1-
*
*
k
-.r¼rt WfA &
;\
'a~~~~~~~
i;
O
w ,
H
- --- `v
;'
?c
Figure 3. Imnutnohistochemical stainitng with anti-CEA antibody of HCC tumors growing in the subcutis and cecal wall of nude mice. HCC tumors from cecal nall of nude mice produced higher levels of CEA protein than HCC s.c. tumors. (A) KM12C s.c. tumor. (B) KM12C cecal tuimor. (C) KM12SM s.c. ttmor.(D) KM12SM cecal tumor. (E) KM20 s.c. tunor. (F) KM20 cecal tumor. (G) Alcian blue staining of KM12SM tumors in the subcutis. (H) Alcian blue staining of K/12SM tumors in the cecal wall.
1162 Kitadai et al A/P October 1996, Vol. 149, No. 4
and 3E) produced less CEA than cecal tumors (Figure 3B, 3D, and 3F). The s.c. tumors contained few mucin-producing alcian blue positive cells (Figure 3G), whereas the cecal tumors contained gland-like structures with many mucin producing cells (Figure 3H). CEA staining was particularly prominent in gland-forming areas of the tumors. The level of CEA mRNA and protein in the KM12C and KM20 tumors growing in the wall of the colon of nude mice directly correlated with preoperative CEA serum level in the patients from whom the surgical specimen was derived (6.7 ng/ml and 281.0 ng/ml, respectively). These results suggest that the production of CEA in the KM20 cells (Duke's stage D) may be regulated differently from that of the KM12 cells (Duke's stage B). In the KM20 cells, three-dimensional growth may influence CEA expression, whereas in the KM12 cells, CEA expression is influenced by three-dimensional growth and signals from the organ environment.
Cell Density-Dependent Expression of CEA Because cell density has been shown to influence expression of the basic fibroblast growth factor gene,28 the multidrug resistance-1 gene (D. Fan, P. J. Beltran, Y.-F. Wang, C. D. Bucana, S.-S. Yoon, and I. J. Fidler, submitted for publication), the type IV collagenase gene,29 and the integrin gene,30 and colon carcinoma cells have been shown to spontaneously differentiate and polarize when cultured under confluent conditions,31'32 we next investigated the effect of cell density on the pattern of CEA expression in KM12C, KM12SM, and KM20 cells. Steady-state CEA mRNA expression levels increased two- to fivefold in all cell lines growing as confluent monolayers or spheroids over those in their counterparts growing as sparse (50% confluence) monolayers (Figure 4). The production of cell-associated CEA protein also directly correlated with cell density (Table 1). Immunohistochemical analyses revealed that cells growing as confluent monolayers or spheroids produced mucin and high levels of CEA protein, whereas cells in sparse monolayers did not (Figure 5, A-D). In some sparse cultures we often found single CEA positive cells (Figure 5A). Whether these cells represent a nondividing, terminally differentiated population is unclear. When HCC cells from confluent monolayers or spheroids were harvested and replated at sparse densities, the level of CEA mRNA and protein decreased to basal levels within 3 days (data not shown). We also cultured KM12SM cells on monolayers of mouse skin or colon fibroblasts.15 The production of CEA in the KM12SM cells
KM12C
k
8
42 Ct,JC a
KM20
KM12SM
ct 4.
.iQ
70-
C
A
40 'o~~~~4
J.
?
.s ,
kb - 4.0 -3.6
CEA
I
GAPDH
I1
1.0 2.1 2.3
1.0 3.6 4.7
-1.3
1.0 1.5 1.6
Figure 4. Steady -state expression of CEA miRNA in spanse, conihient, and spheroid HCC cdltures. The n umbers shoun are the densitometric quantitationt of the ratio of the areas between the specific C'EA traniscripts anid the GAPDH transcript compared in each case to sparse monolavers defined as 1.0. 7This is one representative expriment of three.
was unaffected by the density of the co-culture (data not shown). Previous data from our laboratory have shown that
organ-specific fibroblasts can influence the biological behavior of human colon carcinoma cells.15 To determine whether orthotopic (colon) or ectopic (skin) murine fibroblasts can also influence production of CEA, we cultured mouse skin or colon fibroblasts15 to confluence. Dispersed KM12SM cells were seeded on top of the fibroblasts, and after 1, 2, and 4 days, we analyzed the cultures for CEA protein by immunohistochemistry. The production of CEA in KM12SM cells growing on fibroblast monolayers did not increase, indicating that heterotypic cell-cell contact did not increase CEA production (data not shown).
Effect of Culture Medium on CEM Expression in Sparse Monolayer Cultures To elucidate the mechanism for upregulation of CEA in dense cultures of HCC cells, we treated sparse monolayers with medium from HCC growing under different culture conditions. Conditioned medium taken from high density (confluent and spheroid) cultures increased the level of CEA protein (Table 2) and mRNA (Figure 6) in HCC cells growing under sparse conditions, whereas conditioned medium from sparse cultures did not (Table 2). Moreover, sparse cultures of HCC cells growing in conditioned medium of dense HCC cultures grew slowly and stained positive for CEA. Incubation of sparse cultures in serum-free medium inhibited cell replication and stimulated the production of CEA to levels that were intermediate between sparse and confluent cultures (Table 2), suggesting that inhibition of cell replication can enhance production of CEA. Sparse
Regulation of CEA Expression in HCC Cells 1163 4/P October 1996, Vol. 149, No. 4
Table 2.
Cell Division and Expression of CEA
The Effict of Conditioned Cu/ltolre Medium on CEA Expression in KM12SM Cells
Cells/mm2 Medium
at harvest
Cellular CEA ng/l05 cells
Control (10% FBS) CM from sparse cultures CM from confluent cultures CM from spheroid cultures Serum depletion (1% FBS) Serum depletion (0% FBS)
2500 1342
2.45 ± 0.15 3.65 + 0.20
895
7.55 + 0.53
710
10.0 + 0.29
1316
2.60 + 0.02
526
4.49 + 0.17
To determine whether CEA expression was inversely correlated with cell division, we inhibited cell division in cultures of KM12SM cells by incubation for 1 hour with 20 ,tg/ml of mitomycin-C.33 The cultures were then washed and harvested by a brief trypsinization. Viable KM12SM cells were then replated under sparse conditions (197 cells/mm2). Monitoring of cell viability and numbers confirmed complete inhibition of cell division. After two additional days in culture, CEA protein production was determined by EIA and immunohistochemistry. Sparse cultures of mitomycin-C-treated KM12SM cells produced a higher level of CEA protein (11.8 ng/105 cells) than sparse control (nontreated) cultures (4.2 ng/105 cells), suggesting that CEA production may be related to a decline in cell proliferation.
KM12SM cells were plated at the density of 66 cells/mm2 into
38-mm2 wells of a 96-well plate. After 18 hours, the cells were refed with medium conditioned by KM12SM sparse monolayers, confluent monolayers, or multicell spheroids (CM). The cells were harvested after 4 days. CEA protein was measured by EIA and normalized to viable cell number. Values are the mean + S.D. of triplicate cultures.
Discussion
cultures of HCC growing in conditioned medium harvested from confluent monolayers of mouse skin fibroblasts or mouse colon fibroblasts did not stain positive for CEA (data not shown).
The present results demonstrate that the expression of CEA is regulated by the organ microenvironment. The orthotopic (cecal wall) implantation of HCC cells
Q.1 _
A
Ae
'2,
0~~3p
'3Oi}m
.
~,...D
Figure 5. Imrnlozohistochernical staining oJ KIM12SM cells grown in cu/ltoire ntith anti-CEYA anitibodies. Confluent monolayers (B) spheroids (C, D) oJ/KMJ2,SM cells exhibit inttenise CEA immnunoreactivity as coiipar-ecl uith .sparse mnonolayer cultures (A).
nd ulticellutlar aticl
1164
Kitadai et al
AJP October 1996, Vol. 149, No. 4
KM12SM Oh 3h 24h 48h
CEA
kb -
GAPDH
4.0 3.6
-1.3 1.0
1.0
1.4 1.5
Figure 6. Steady-state expression of CEA mRNA by KM12SM cells growing in miediuim conditioned by conifluenit cultures. KM12SM cells were plated overnight under sparse coniditions. The sparse cultures were harvested at the beginning of the experiment(O hours) or3, 24, and 48 hours after the addition7 of medium fiom confluent cuiltuires of KM12SM cells. The numbers shoun arce the densitometric quantitation oj' the ratio of the areas betuween the specific CEA transcripts and the GAPDH traniscript compared in each case to sparse cultu.rcs at time 0 hours definetd 10. Tis is one representative experiment of tuwo. as
into nude mice yielded more differentiated tumors, which expressed high levels of steady-state CEA mRNA and protein. In contrast, ectopic (s.c.) tumors were less differentiated and produced low levels of steady-state CEA mRNA and protein. In the cecal wall of nude mice, KM20 cells (isolated from a moderately differentiated adenocarcinoma of the colon from a patient with high preoperative serum CEA levels) produced differentiated lesions with higher levels of CEA mRNA and protein than lesions produced by KM12C cells (isolated from poorly differentiated colon adenocarcinoma from a patient with a low preoperative serum CEA level). In culture, however, KM20 and KM12C cells produced similar levels of CEA mRNA and protein. The influence of the organ microenvironment on the biology of tumor cells has been recognized since Paget's "seed and soil" hypothesis, which suggested that the interaction between tumor cells and target organs determines whether metastasis will occur.3 Recent evidence supports the role of the microenvironment in regulating tumorigenesis,13 the production of degradative enzymes,14'15 angiogenic molecules,16'17 the level of P-glycoprotein associated with multidrug resistance,18 and the induction of terminal differentiation.19 In this study, HCC tumors growing in the cecal wall of nude mice contained gland-like structures and mucin-producing cells, indicating a more differentiated phenotype than their counterpart s.c. tumors, suggesting that specific organ microenvironments may influence HCC in differentiation and expression of CEA.
In patients with colorectal carcinomas, an elevated preoperative serum CEA level is associated with poor prognosis.2'3 The level of CEA in the serum of patients with colorectal carcinoma is influenced by the balance between production of CEA by tumor cells and the ability of the liver to clear CEA from the blood.2'3 Thus, elevated serum levels of CEA may be due to an increased tumor burden, especially metastatic lesions in the liver.23 Several reports have suggested that CEA can function as a homotypic intracellular adhesion molecule and thus play a role in cancer metastasis.4`8 Our data do not address this possibility but show that CEA expression is indeed regulated by cell density. When cells enter quiescence, they can exhibit major alterations in cell surface receptors, expression of transcription factors, enzymes, and cellular ultrastructure.3536 The expression of basic fibroblast growth factor gene, multidrug resistance-1 gene, and type IV collagenases have been shown to be down-regulated in confluent cultures28'29 (D. Fan, P. J. Beltran, Y-F. Wang, C. D. Bucana, S-S. Yoon, and 1. J. Fidler, submitted for publication), and HCC cells have been shown to undergo polarization and differentiation when cultured to confluence.31 32 The spontaneously differentiating Caco-2 cells and T84 cells were reported to express an increasing level of CEA when cultured to confluence,37 and similar results were reported for expression of integrins.30 In agreement with these data, we found that dense monolayer cultures or spheroids of the HCC KM12 and KM20 lines produced mucin after reaching confluence and showed increased expression of CEA. Incubation of KM12SM cells in serum-free medium decreased cell proliferation and enhanced CEA production. We also treated KM12SM cells (high production of CEA) with mitomycin-C and found that growth arrested cells (under sparse conditions) produced higher levels of CEA. This finding suggested that CEA expression is influenced by cell proliferation. Moreover, medium taken from dense cultures increased the expression of CEA in HCC cells growing as sparse monolayers, indicating that CEA expression may be regulated by an autocrine mechanism. Our data indicate that CEA expression on the protein level does not always correlate with that on the mRNA level, agreeing with a recent report showing that treatment of colon carcinoma cell lines by relatively high concentrations of interferon (IFN)-y (1000 or 2000 U/ml) can disproportionately increase the CEA mRNA level and protein production.37 Whether these results suggest that CEA expression by HCC may be regulated by both transcriptional and post-transcriptional mechanisms needs further
Regulation of CEA Expression in HCC Cells
1165
AJP October 1996, Vol. 149, No. 4
study. A recent report has identified the cis-acting elements involved in the transcriptional control of the CEA gene,12 and treatment of HCC cells with transforming growth factor-338,39 or IFN-,40'41 has stimulated expression of CEA. We cultured KM12SM cells with various concentrations of transforming growth factor-a, transforming growth factor-(3, hepatocyte growth factor, basic fibroblast growth factor, vascular endothelial growth factor, interleukin-,B, and IFN-a, -X3, or -y. Because none of the cytokines (at the concentrations used here) affected cell division or influenced CEA expression (data not shown), it remains unclear whether these or other cytokines can act as paracrine or autocrine factors that regulate CEA expression and production. In conclusion, we have shown that the production of CEA by HCC cells can be modulated by specific organ microenvironments, cell density, and autocrine and paracrine factors. Because CEA expression levels and differentiation of tumor cells growing in the cecal wall of nude mice directly correlated with the preoperative serum CEA level and pathological diagnosis of the original patient surgical specimens, we conclude that the orthotopic implantation of HCC xenografts provides a relevant model to study the regulation and role of CEA in cancer metastasis of HCC.
6.
7.
8.
9.
10.
11.
12.
Acknowledgments We thank Ricardo Sanchez, Michael R. Wilson, and Kenneth Dunner, Jr., for technical assistance, Walter Pagel for his editorial review, and Lola Lopez for her expert assistance in the preparation of this manuscript.
13.
14.
References 1. Gold P, Freedman SO: Demonstration of tumor-specific antigen in human colonic carcinoma by immunological tolerance and absorption techniques. J Exp Med 1965,
121:439-462 2. Steele G Jr, Zamcheck N: The use of carcinoembryonic antigen in the clinical management of patients with colorectal cancer. Cancer Detection Prevention 1985, 8:421-427 3. Jessup JM, Gallick GE: The biology of colorectal carcinoma. Curr Probl Cancer 1992, 16:261-328 4. Beauchemin N, Benchimol S, Cournoyer D, Fuks A, Stanners CP: Isolation and characterization of fulllength functional cDNA clones for human CEA. Mol Cell Biol 1987, 7:3221-3230 5. Barnett TR, Kretschmer A, Austen DA, Goebel SJ, Hart JT, Elting JJ, Kamarck ME: Carcinoembryonic
15.
16.
17.
18.
antigens: alternative splicing accounts for the multiple mRNAs that code for novel members of the carcinoembryonic antigen family. J Cell Biol 1989, 108:267-276 Paxton R, Mooser G, Pande H, Lee TD, Shively JE: Sequence analysis of carcinoembryonic antigen: identification of glycosylation sites and homology with the immunoglobulin super gene family. Proc Natl Acad Sci USA 1987, 84:920-924 Zimmerman W, Ortlieb B, Friedrich R, von Kleist S: Isolation and characterization of cDNA clones encoding the human carcinoembryonic antigen reveal a highly conserved repeating structure. Proc Natl Acad Sci USA 1987, 84:2960-2964 Benchimol S, Fuks A, Jothy S, Beauchemin N, Shirota K, Stanners CP: Carcinoembryonic antigen, a human tumor marker, functions as an intercellular adhesion molecule. Cell 1989, 57:327-334 Liotta LA, Kleinerman J, Saidel GM: The significance of hematogenous tumor cell clumps in the metastatic process. Cancer Res 1976, 36:889-894 Wagner HE, Toth GA, Steele GD Jr, Thomas P: Metastatic potential of human colon cancer cell lines: relationship to cellular differentiation and carcinoembryonic antigen production. Exp Metastasis 1992, 10: 25-31 Tibbetts LM, Doremus CM, Tzanakakis GN, Vezeridis MP: Liver metastasis with 10 human colon carcinoma cell lines in nude mice and association with carcinoembryonic antigen production. Cancer 1993, 71:315321 Hauck W, Stanners CP: Transcriptional regulation of the carcinoembryonic antigen gene. J Biol Chem 1995, 270:3602-3610 Radinsky R, Risin S, Fan D, Dong Z, Bielenberg D, Bucana CD, Fidler IJ: Level and function of epidermal growth factor receptor predict the metastatic potential of human colon carcinoma cells. Clin Cancer Res 1995, 1:19-31 Nakajima M, Morikawa K, Fabra A, Bucana CD, Fidler IJ: Influence of the organ environment on extracellular matrix degradative activity and metastasis of human colon carcinoma cells. J Natl Cancer Inst 1990, 82: 1890-1898 Fabra A, Nakajima M, Bucana CD, Fidler IJ: Modulation of the invasive phenotype of human colon carcinoma cells by organ-specific fibroblasts of nude mice. Differentiation 1992, 52:101-110 Singh RK, Bucana CD, Gutman M, Fan D, Wilson MR, Fidler IJ: The influence of the organ microenvironment on the expression of basic fibroblast growth factor in human renal cell carcinoma cells. Am J Pathol 1994, 145:365-374 Gutman M, Singh RK, Xie K, Bucana CD, Fidler IJ: Regulation of interleukin-8 expression in human melanoma cells by the organ environment. Cancer Res 1995, 55:2470-2475 Dong Z, Radinsky R, Fan D, Tsan R, Bucana CD, Wilmanns C, Fidler IJ: Organ-specific modulation of
1166
Kitadai et al
A/P October 1996, Vol. 149, No. 4
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
steady-state mdr-1 gene expression and drug resistance in murine colon cancer cells. J Natl Cancer Inst 1994, 86:913-920 Radinsky R, Fidler IJ, Price JE, Esumi N, Tsan R, Petty CM, Bucana CD, Bar-Eli M: Terminal differentiation and apoptosis in experimental lung metastases of human osteogenic sarcoma cells by wild-type p53. Oncogene 1994, 9:1877-1883 Toribara NW, Sack TL, Gum JR, Ho SB, Shively JE, Willson JKV, Kim YS: Heterogeneity in the induction and expression of carcinoembryonic antigen-related antigens in human colon cancer cell lines. Cancer Res 1989, 49:3321-3327 Chakrabarty S, Chen JW-J, Chen X, Trujillo JM, Lin P: Modulation of differentiation-related response in human colon carcinoma cells by protein kinase inhibitor H-7. Anticancer Res 1992, 12:97-104 Morikawa K, Walker SM, Jessup JM, Fidler IJ: In vivo selection of highly metastatic cells from surgical specimens of different human colon carcinomas implanted into nude mice: Cancer Res 1988, 1948, 48:1943 Morikawa K, Walker SM, Nakajima M, Pathak S, Jessup JM, Fidler IJ: The influence of the organ environment on the growth, selection, and metastasis of human colon cancer cells in nude mice. Cancer Res 1988, 48:68636871 Kobayashi H, Man S, Graham CH, Kapitain SJ, Teicher BA, Kerbel RS: Acquired multicellular-mediated resistance to alkylating agents in cancer. Proc NatI Acad Sci USA 1993, 90:3294-3298 Radinsky R, Kraemer PM, Raines MA, Kung HJ, Culp LA: Amplification and rearrangement of Kirsten ras oncogene in virus transformed BALB/c 3T3 cells during malignant tumor progression. Proc Natl Acad Sci USA 1987, 84:5143-5147 Fort P, Marty L, Liechaczyk M, Sabrouty SE, Dani C, Blanchard JM: Various rat adult tissues express only one major mRNA species from the glyceraldehyde-3phosphatidehydrogenase multigene family. Nucleic Acid Res 1985, 13:1431-1442 Feinberg AP, Vogelstein B: A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal Biochem 1983, 132:6-13 Murphy PR, Sato R, Sato Y, Friesen HG: Fibroblast growth factor messenger ribonucleic acid expression in a human astrocytoma cell line: regulation of serum and cell density. Mol Endonucl 1988, 2:591-598 Xie B, Bucana CD, Fidler IJ: Density-dependent induction of 92-kd type IV collagenase activity in cultures of A431 human epidermoid carcinoma cells. Am J Pathol 1994, 144:1058-1067 Hauptmann S, Denkert C, Lohrke H, Tietze L, Ott S,
31.
32.
33.
34. 35.
36.
37.
38.
39.
40.
41.
Klosterhalfen B, Mittermayer C: Integrin expression on colorectal tumor cells growing as monolayers, as multicellular tumor spheroids, or in nude mice. Int J Cancer 1995, 61:819-825 Fogh J, Fogh JM, Orfeo T: One hundred and twentyseven cultured human tumor cell lines producing tumors in nude mice. J Natl Cancer Inst 1977, 59:221225 Pinto M, Robine-Leon S, Appay M-D, Kedinger M, Triadou N, Dussaulx E, Lacroix B, Simon-Assmann P, Haffen K, Fogh J, Zweibaum A: Enterocyte-like differentiation and polarization of the human colon carcinoma cell line Caco-2 in culture. Biol Cell 1983, 47: 323-330 Tomasz M, Lipman R, Chowdary D, Pawlak J, Verdine GL, Nakamishi K: Isolation and structure of covalent crosslink adduct between mitomycin C and DNA. Science 1987, 235:1204-1208 Paget S: The distribution of secondary growth in cancer of the breast. Lancet 1889, 1:571-573 Mulder KM: Differential regulation of c-myc and transforming growth factor-a messenger RNA expression in poorly differentiated and well differentiated colon carcinoma cells during the establishment of a quiescent state. Cancer Res 1991, 51:2256-2262 Williams GM, Stromberg K, Kroes R: Cytochemical and ultrastructural alterations associated with confluent growth in cell cultures of epithelial-like cells from rat liver. Lab Invest 1973, 29:293-303 Hauck W, Stanners CP: Control of carcinoembryonic antigen gene family expression in a differentiating colon carcinoma cell line, Caco-2. Cancer Res 1991, 51:3526-3533 Chakrabarty S, Jan Y, Brattain MG, Tobon A, Varani J: Diverse cellular response elicited from human colon carcinoma cells by transforming growth factor-j3. Cancer Res 1989, 49:2112-2117 Chakrabarty S, Fan D, Varani J: Modulation of differentiation and proliferation in human colon carcinoma cells by transforming growth factor 131 and 132. Int J Cancer 1990, 46:493-499 Guadagni F, Witt PL, Robbins PF, Schlom J, Greiner JW: Regulation of carcinoembryonic antigen expression in different human colorectal tumor cells by interferon-y. Cancer Res 1990, 50:6248-6255 Kanai T, Hibi T, Hayashi A, Takashima J, Shiozawa M, Aiso S, Toda K, lwao Y, Watanabe M, Tsuchiya M: Carcinoembryonic antigen mediates in vitro cell aggregation induced by interferon-y in a human colon cancer cell line: requirement for active metabolism and intact cytoskeleton. Cancer Lett 1993, 71:109-117
