Tong 1998

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

European Journal of Cancer, Vol. 34, No. 13, pp.

2119±2125, 1998
# 1998 Elsevier Science Ltd. All rights reserved
Pergamon Printed in Great Britain
0959-8049/98/$Ðsee front matter

PII: S0959-8049(98)00267-6

Original Paper

Growth Regulation of Human Colon Cancer Cells by


Epidermal Growth Factor and 1,25-Dihydroxyvitamin D3 is
Mediated by Mutual Modulation of Receptor Expression

W.-M. Tong, E. KaÂllay, H. Hofer, W. Hulla, T. Manhardt, M. Peterlik and H.S. Cross
Department of General and Experimental Pathology, University of Vienna Medical School, Waehringer Guertel
18-20, A-1090 Vienna, Austria

The human colon adenocarcinoma-derived cell line Caco-2 was used as a model system to study the
interaction of epidermal growth factor (EGF) and 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) in control
of colorectal cancer cell growth. The mitogenic stimulus of EGF was rapidly transduced via apical and
basal membrane receptors alike into elevation of c-myc expression, causing a shift of Caco-2 cells
from the G0/G1 into the S phase of the cell cycle. The stimulatory eVect of EGF on cell division was
eVectively counteracted by 1,25(OH)2D3: the presence of the steroid hormone prevents the negative
eVect of EGF on vitamin D receptor abundance and concurrently minimises ligand-occupied
EGF receptor numbers on both sides of Caco-2 cell monolayers. Our data suggest that EGF and
1,25-(OH)2D3 actions on mutual receptor levels represent a speci®c feature of the potent anti-
mitogenic eVect of the steroid hormone on colon cancer cells. # 1998 Elsevier Science Ltd. All rights
reserved.

Key words: colon adenocarcinoma, Caco-2 cells, vitamin D, epidermal growth factor receptor, vita-
min D receptor, c-myc expression, cell cycle, apical/basal receptor distribution
Eur J Cancer, Vol. 34, No. 13, pp. 2119±2125, 1998

INTRODUCTION rediVerentiation of the human colon cancer cell line, Caco-2


Epidermal growth factor (EGF) is a 6000 kDa polypeptide [7±9]. The eVects of 1,25(OH)2D3 on cell growth and dif-
which is produced predominantly in the gastrointestinal tract ferentiation are mediated by binding to the nuclear vitamin D
by salivary glands, Brunner's glands and Paneth cells. The receptor (VDR) which then acts as a transactivation factor for
peptide hormone is released into the digestive ¯uids as well as various genes.
into the circulation and, hence, can locally as well as sys- Regulation of receptor abundance may be an important
temically regulate cell growth and function in a variety of mechanism for modulating the extent of target cell respon-
tissues [1]. In the gastrointestinal tract, EGF inhibits acid siveness to hormones and growth factors. In this respect,
secretion by gastric parietal cells [2] and, importantly, stimu- Krishnan and Feldman [10] showed that in mouse ®broblasts
lates ornithine decarboxylase [3], DNA synthesis and, thus, and human breast cancer cells, VDR gene expression can be
proliferation of small and large intestinal epithelial cells [4]. upregulated by serum and growth factors. EGF has also been
There is also evidence that EGF plays a signi®cant role in found to increase VDR in the small intestine of the neonatal
autocrine growth stimulation of colon tumour cells [5]. rat [11]. Conversely, 1,25(OH)2D3 has been reported to
1,25-Dihydroxyvitamin D3 (1,25(OH)2D3) inhibits pro- modulate EGF receptor (EGFR) expression, although in
liferation and induces diVerentiated cell functions in normal opposite directions depending on the cell type involved. In
and, notably, also in malignant cells [6]. In this respect, human T-47D breast cancer cells, growth inhibition induced
vitamin D compounds have been shown, for example, to by 1,25(OH)2D3 is associated with a decrease in EGFR
inhibit eVectively cell growth and to support phenotypic numbers [12], whereas Desprez and colleagues [13] reported
an increase in EGFR expression as the result of 1,25(OH)2D3
Correspondence to H.S. Cross. treatment in BT-20 breast cancer cells. Because no respective
Received 6 Mar. 1998; revised 26 May 1998; accepted 1 Jun. 1998. data are available for human colon cancer cells, we investigated

2119
2120 Wei-Min Tong et al.

the action and interaction of EGF and 1,25(OH)2D3 in 15 min at room temperature, 2.0 ml of STC buVer with
growth control and mutual receptor regulation using the 0.5 mg/ml trypsin inhibitor (Sigma) and 0.6 mg/ml RNase
human colon adenocarcinoma-derived cell line Caco-2 as a were added and the cells were incubated for 15 min at room
model system. temperature. The nuclei were collected by centrifugation at
The cell line Caco-2, despite its malignant origin, has 1,000 g for 20 min after addition of 20 ml of STC buVer
retained the potential for spontaneous rediVerentiation in containing additional STC (1.6 mg/ml ®nal concentration).
culture [14]. Caco-2 cells also display two distinct classes of Nuclear c-myc protein was assessed using a monoclonal
EGFR, at the apical and basolateral plasma membrane, anti-c-myc antibody (®nal concentration 25 mg/ml), which
respectively [15, 16] and, importantly, express the VDR at was detected with a ¯uorescein isothiocyanate-labelled sec-
the mRNA and protein level [17, 18]. ond antibody (25 mg/ml ®nal concentration).
For cell cycle analysis, the nuclei were incubated with 50 ml
MATERIALS AND METHODS of antibody solution on ice, washed twice with STC buVer
Materials and resuspended in 300 ml STC buVer containing 0.42 mg/ml
1,25(OH)2D3 was a gift from HoVmann-LaRoche (Basle, propidium iodide (Sigma). The stained nuclei were analysed
Switzerland). EGF (tissue culture grade) was obtained from in a ¯ow cytometer (FACStar; Becton Dickinson, Mountain
Sigma (Deisenhofen, Germany). An anti-c-myc monoclonal View, California, U.S.A.) using linear ampli®cation for pro-
antibody was purchased from Oncogene Science (Cambridge, pidium iodide ¯uorescence and logarithmic ampli®cation of
Massachusetts, U.S.A.). The anti-VDR monoclonal antibody ¯uorescein isothiocyanate ¯uorescence. Cell cycle distribu-
was from Chemicon (Temecula, California, U.S.A.). tion was analysed using CELLFIT software (Becton Dick-
inson), ®tting G0/G1 and G2/M in normal distributions and S
Cell culture phase cells outside 3 standard deviations (S.D.) of the mean
Caco-2 cells (provided by A. Quaroni, Cornell University, for G0/G1 and G2/M.
Ithaca, New York, U.S.A.) were routinely cultured in Costar
vented tissue culture ¯asks at 37 C in a humidi®ed atmo- EGFR radioligand assay
sphere of 95% air and 5% CO2 in Dulbecco's modi®ed Binding of iodinated EGF to apical or basolateral mem-
Eagles medium (DMEM) containing 4.0 mM glutamine, branes of Caco-2 cells grown on ®lter units was determined
10% fetal calf serum (heat-inactivated at 56 C for 30 min), essentially as described by Cross and Quaroni [15]. Brie¯y,
20 mM HEPES, 50 U/ml penicillin and 50 mg/ml streptomy- 1.0 ng/ml mouse 125I-EGF (speci®c activity 100 mCi/mg,
cin. The cultures were fed every 48 h and subcultured serially Amersham International, Buckinghamshire, U.K.) in 2 ml
when approximately 80% con¯uent. They were used between binding medium (DMEM with 0.1% bovine serum albumin
passages 10 and 30. (BSA)) was applied selectively to the apical or basal ®lter
For experimentation, 15 000 cells/ml were routinely seeded compartment for 3 h at 4 C. Non-speci®c binding was deter-
per well in 24 well Falcon plastic tissue culture dishes. For mined in the presence of a 100-fold excess of unlabelled
the investigation of selective stimulation of apical and baso- EGF. After incubation, the monolayers were washed, trypsi-
lateral EGFR, 60 000 cells/ml were seeded on to 25 mm Fal- nised and counted for radioactivity in an automated gamma
con polycarbonate ®lter units (0.4 mm pore size). Electron counter (1277 GammaMaster, LKB).
microscopy had been used previously to verify monolayer
formation of con¯uent ®lter grown Caco-2 cells [15, 19]. Western blotting
The medium was changed and new treatments were added The cells were homogenised on ice with a Polytron
to Caco-2 cell cultures every 24 h, if not indicated otherwise. (Brinkmann Instruments, Westbury, New York, U.S.A.) in
If appropriate, 1,25(OH)2D3 was added to the cultures in 10 mM Tris±HCl (pH 7.2) buVer containing a protease inhi-
ethanolic solution. The vehicle concentration in all cultures bitor mix: 1 mM phenylmethylsulphonyl ¯uoride (PMSF),
containing the steroid hormone was 0.01%. 50 mg/ml leupeptin, 50 mg/ml antipain, 0.1 mg/ml aprotinin
(Boehringer Mannheim, Vienna, Austria). The homogenate was
Cell proliferation assay and cell counting centrifuged at 4 C for 10 min at 1280 rpm. The supernatant
For the determination of [3H]thymidine incorporation into was centrifuged for 60 min at 106 000 g and stored at ÿ 80 C.
cellular DNA, the cells were incubated for 4 h at 37 C in For immunoblotting, the samples were run on 12%
DMEM containing 4 mCi/ml of [3H]thymidine (speci®c sodium dodecylsulphate±polyacrylamide gel electrophoresis
activity 70 Ci/mmol, American Radiolabeled Chemicals, St (SDS±PAGE) gels (total protein per lane 120 mg) and were
Louis, Missouri, U.S.A.). The cells were washed with phos- blotted to a Hybond ECL nitrocellulose membrane (Schleicher
phate buVered saline (PBS), ®xed and extracted twice with & Schuell, Dassel, Germany) with a Hoefer Transblot2 Unit
5% trichloroacetic acid. The cells were solubilised in 1 ml of (Pharmacia Biotech., San Francisco, California, U.S.A.).
0.1 N NaOH and assayed for protein (BCA protein assay kit, The gels were checked for equal protein loading by staining
Pierce, Rockford, Illinois, U.S.A.) and counted for radio- with Coomassie Blue. Non-speci®c binding was blocked in
activity. Trypsinised cells were counted in a Coulter Counter 3% BSA in PBS/0.1% Tween for 2 h at room temperature.
(Model D, Coulter Electronics, Luton, U.K.). The membranes were incubated with rat anti-VDR antibody
(IgG2b, 1:1000, Chemicon International, Temecula, Cali-
Cell cycle analysis and c-myc protein determination fornia, U.S.A.) in PBS/Tween with 1% BSA at 4 C over-
Cell nuclei were prepared as described elsewhere [20]. In night. Incubation of the membrane with a horseradish
brief, trypsinised cells were suspended in 2.0 ml citrate buVer, peroxidase-conjugated antirat IgG (Amersham, 1:10 000 in
pH 7.6 (containing 0.1% Nonidet-P 40, 1.5 mM spermidine PBS/0.1% Tween) was performed for 2 h at room tempera-
tetrahydrochloride (STC), 0.5 mM Tris/STC buVer and ture, with subsequent detection by the SuperSignal TM CL-
0.03 mg/ml trypsin (all from Sigma). After incubation for HRP substrate system (Pierce, Rockford, Illinois, U.S.A.).
EGF and Vitamin D Receptors in Caco-2 Cells 2121

RNA isolation and Northern blot hybridisation analysis sperm DNA at 42 C, and then hybridised for 12±24 h in the
The cells were lysed in 3.0 ml guanidinium thiocyanate same buVer containing the radiolabelled probe. The mem-
buVer (4 M guanidinium thiocyanate, 25 mM sodium citrate, branes were washed twice with 2SSC/0.1% SDS for 5 min
0.5 N-lauroylsarcosine, 0.1 M 2-mercaptoethanol, pH 7) and at room temperature, then twice with 0.2SSC/0.1% SDS
total RNA was isolated. RNA was quanti®ed by absorbance under the same conditions, and twice at 42 C with 0.2SSC/
at 260 nm. The intactness of the RNAs was visualised on a 0.1% SDS for 15 min and then exposed to a Kodak
non-denaturing, ethidium bromide stained 1% agarose gel. X-OMAT AR ®lm at ÿ 70 C for 1±5 days. To normalise the
Fifteen micrograms of total RNA were then run on a signals for gel loading variations, the membranes were
denaturing (3% formaldehyde), 0.8% agarose gel and blotted stripped by boiling in 0.1% SDS and rehybridised using a
with 20SSC (3.0 M sodium chloride, 0.3 M sodium citrate) radiolabelled probe for b-actin. Autoradiographs were eval-
on to Nylon (Immobilon-N, Millipore, Bedford, Massachu- uated using a laser densitometer (LKB UltroScan XL and
setts, U.S.A.) or nitrocellulose (Schleicher & Schuell) mem- GelScan XL software).
branes. A 358 bp Cla1-Msp1 fragment from c-myc exon 3, a The c-myc probe was obtained from an original American
cDNA probe for the human VDR, as well as b-actin were Type Culture Collection clone. VDR cDNA was kindly pro-
32
P-labelled by random priming (Promega, Madison, Wis- vided by J. W. Pike (Ligand Pharmaceuticals, San Diego,
consin, U.S.A.) and used for hybridisation. After baking for California. U.S.A.). The b-actin probe was reverse transcrip-
2 h at 80 C, the membranes were prehybridised in a solution tion±polymerase chain reaction (RT±PCR)-generated from
of 25 mM phosphate buVer (pH 7.4), 5SSC, 5 Caco-2 RNA using commercially available primers from
Denhardt's solution, 50% formamide and 100 mg/ml salmon Clontech (Palo Alto, California, U.S.A.).

Figure 1. EVect of epidermal growth factor (EGF; 25 ng/ml) on c-myc mRNA (a), nuclear c-myc (b) cell cycle distribution (c)
and [3H]thymidine incorporation (d) of con¯uent Caco-2 cells. One representative experiment (out of three) is shown. (b±d)
data are means ‹ standard error of the mean (SEM) from four measurements.
2122 Wei-Min Tong et al.

Data presentation and statistical analyses EGF, we investigated the mutual eVects of the growth factor
Data are presented as means ‹ standard error of the mean and of 1,25(OH)2D3 on expression levels of their respective
(SEM), if appropriate. One-way analysis of variance was used receptors.
for statistical evaluations. Groups versus control group were
compared by Bonferroni's method. DiVerential expression of apical and basolateral EGFR: eVects of
EGF and 1,25(OH)2D3
RESULTS Because Caco-2 cells display an unequal distribution of
EGF aVects proliferation and c-myc expression EGFR on their apical and basolateral plasma membrane
The time-course of the mitogenic eVect of EGF on Caco-2 [15], which might also re¯ect functional diVerences [16], the
cells is illustrated in Figure 1, showing the relationship site-speci®c expression and function of EGFR was studied on
between [3H]thymidine incorporation into DNA, shifts in ®lter grown Caco-2 cells. Analysis of selective 125I-EGF
cell cycle distribution and expression of c-myc mRNA and of binding to either the apical or basolateral surface of Caco-2
nuclear c-myc protein. Within 6 h after the addition of EGF, monolayers revealed a particularly high density of basolateral
c-myc mRNA was increased 2-fold and then sharply declined membrane EGFR in untreated Caco-2 cells, whereas only
to a level still above that of untreated controls (Figure 1a). approximately one-®fth of the total number was found at the
Nuclear c-myc protein also showed a transient change inas- apical side (Table 1). Treatment with 25 ng/ml EGF, either
much as its initial value doubled, reaching a maximum after at the apical or basolateral side of the ®lter chamber, initiated
24 h and then steadily declined to control values at 96 h a decrease of EGFR numbers within 30 mins. Homologous
(Figure 1b). This initial increase in nuclear association of the downregulation was very eYcient on the basolateral side,
c-myc protein is exactly mirrored by a shift of Caco-2 cells where receptor density was reduced to approximately 20% of
out of the resting, i.e. G0/G1, phase of the cell cycle and a its initial level, whereas reduction of EGFR numbers was less
corresponding maximum of DNA synthetic activity conspicuous on the apical side (Table 1). Steady state EGFR
(Figure 1c, d). levels were achieved after little more than 2 h of EGF treat-
In order to investigate from which cell side the EGFR was ment and were practically in the same order of magnitude on
stimulated, we added EGF to either the apical or basal cell the apical and the basolateral side (Table 1).
aspect of ®lter grown Caco-2 cell monolayers. This resulted Treatment of Caco-2 cells with 10ÿ8 M 1,25(OH)2D3
in a signi®cant rise in c-myc mRNA which became apparent reduced EGFR numbers in the steady state on the apical side
after 2 h, reached a peak level 4±5 times above control at 4 h by approximately 50% and on the basolateral side by around
and the rise was similar regardless of the direction of the sti- 25%. When combined with EGF, the steroid hormone
mulation (Figure 2). Thereafter, c-myc mRNA expression induced a signi®cant further decrease in EGFR numbers,
declined to a steady state level which was still approximately leading to extremely low receptor density on both sides of
twice as high as in zero time controls (cf. Figure 1a). Caco-2 cells (Table 1).
Combined stimulation of apical and basolateral EGFR
populations had no additional eVect on c-myc mRNA (not VDR mRNA and protein expression: eVect of EGF and
shown). 1,25(OH)2D3
Selective stimulation of apical or basolateral EGFR by
EGF-induced proliferation: attenuation by 1,25(OH)2D3 addition of 25 ng/ml EGF to one of the respective cell
Figure 3 shows that growth of Caco-2 cells exposed to compartments on the ®lter units suppressed VDR mRNA
EGF was signi®cantly reduced by 1,25(OH)2D3 at a con-
centration as low as 10ÿ10 M. In contrast, in the absence of
EGF, the steroid hormone inhibited cell growth only at a 10-
fold higher concentration. In order to explain the increased
sensitivity of Caco-2 cells to 1,25(OH)2D3 in the presence of

Figure 3. Antimitotic eVect of 1,25-dihydroxyvitamin D3


(1,25(OH)2D3) on con¯uent Caco-2 cells in the absence (*) or
presence of epidermal growth factor (EGF; 25 ng/ml) (*).
Figure 2. Selective stimulation of apical or basolateral epi- 1,25(OH)2D3-free vehicle controls contained 0.01% ethanol.
dermal growth factor receptor (EGFR) and c-myc mRNA Cell counts in triplicate after 4 days treatment. *Statistically
expression. EGF (25 ng/ml) was added on day 2 after con¯uence. signi®cant diVerence from respective EGF-free group (at least
Data are from one representative experiment (out of three). P < 0.05).
EGF and Vitamin D Receptors in Caco-2 Cells 2123

Table 1. Homologous and heterologous regulation of epidermal growth factor receptor (EGFR) density on luminal and basolateral
(Caco-2 cell membranes)

EGFR density (fmol/mg protein)


Apical (luminal) Basolateral
Additions to luminal side Additions to basolateral side

Time (h) EGF 1,25(OH)2D3 EGF plus 1,25(OH)2D3 EGF 1,25(OH)2D3 EGF plus 1,25(OH)2D3

0 847 ‹ 12 847 ‹ 12 847 ‹ 12 2,433 ‹ 57 2,433 ‹ 57 2,433 ‹ 57


0.5 392 ‹ 31* N.D. N.D. 897 ‹ 57* N.D. N.D.
2.0 256 ‹ 28* N.D. N.D. 567 ‹ 20* N.D. N.D.
48 214 ‹ 48* N.D. N.D. 386 ‹ 49* N.D. N.D.
168 314 ‹ 20* 413 ‹ 86* 81 ‹ 7*,y 435 ‹ 17* 1,776 ‹ 203* 51 ‹ 13*y

EGF (25 ng/ml) was added to either the luminal or basolateral compartment of ®lter grown Caco-2 cells. 1,25(OH)2D3 (10ÿ8 M) was added
on both sides. Control cultures contained 0.01% ethanol as the vehicle. EGFR was quanti®ed by duplicate radioligand assays. Data are
means ‹ standard error of the mean (SEM) from two separate experiments. *Statistically signi®cant diVerence (at least P  0.05) from control.
yStatistically signi®cant diVerence (at least P  0.05) from EGF treated group at the same time point. N.D., not determined; 1,25(OH)2D3,
1,25-dihydroxyvitamin D3.

abundance within 4 h to 40% of initial values. Further expo- modulation of a regulatory site downstream or independent
sure to EGF for up to 96 h maintained VDR mRNA expres- of c-myc action. This could probably involve an appropriate
sion at a similarly low level (Figure 4). Combined stimulation eVect of vitamin D on the expression of other cycle reg-
of apical and basolateral EGFR had no additional eVect on ulators, such as cyclins, cyclin-dependent kinases or their
VDR expression (data not shown). inhibitors, as observed by Kawa and colleagues [21] in pan-
Heterologous downregulation of the VDR by EGF was creatic cancer cells or by Wang and associates [22] in human
also observed at the protein level (Figure 5). 1,25(OH)2D3 leukaemia HL-60 cells.
has a positive eVect on receptor abundance, even in the pre- The observation in the present study that EGF-stimulated
sence of EGF (Figure 5). Caco-2 cells responded particularly well to growth inhibition
by 1,25(OH)2D3 (cf. Figure 3), could be explained, at least in
DISCUSSION part, if one assumes that, apart from cell cycle regulatory
The present study strongly suggests that, regardless of events, changes in receptor abundance may also be an
whether the stimulatory eVect of EGF on the growth of important mechanism for modulating the extent of target cell
human colon adenocarcinoma-derived Caco-2 cells is elicited
via luminal or basolateral EGFR activation, it eventually
involves upregulation of the cell cycle regulatory protein c-myc.
This would facilitate Caco-2 cells to enter into and progress
through the G1 phase of the cell cycle in response to EGF.
Until now the mechanism underlying the antimitogenic
action of 1,25(OH)2D3 in Caco-2 cells has remained an
enigma, since, unlike in other tumour cells, c-myc expression
in Caco-2 cells is resistant to modulation by 1,25(OH)2D3
[18]. It had to be assumed, therefore, that the antimitotic
eVect of the steroid hormone results from VDR-mediated

Figure 5. Western blot analysis of vitamin D receptor (VDR)


density in Caco-2 cells. Epidermal growth factor (EGF; 25 ng/
ml) or/and 1,25-dihydroxyvitamin D3 (1,25(OH)2D3; 10ÿ8 M)
Figure 4. Selective stimulation of apical or basal epidermal were added 2 days after con¯uence for 96 h. Untreated controls
growth factor receptor (EGFR): eVect on vitamin D receptor contained 0.01% ethanol as vehicle. Data from one repre-
(VDR) mRNA expression. Treatment with 25 ng/ml EGF api- sentative experiment (out of three) are shown. Means
cally or basally for the time period indicated. Data shown are ‹ standard error of the mean (SEM). Statistically signi®cant
from one representative experiment (out of three). diVerence from the controls. *P < 0.05; yP < 0.01.
2124 Wei-Min Tong et al.

responsiveness to growth factors and hormones. Bishop and 3. Feldman EJ, Aures D, Grossman MI. Epidermal growth
Wen [16] had suggested that Caco-2 cell proliferation is dri- factor stimulates ornithine decarboxylase activity in the
ven exclusively by ligand-activated basolateral membrane digestive tract of the mouse. Proc Soc Exp Biol Med 1978, 159,
400±402.
EGFR because they could observe an appropriate increase in 4. Scheving LA, Yeh YC, Tsai TH, Scheving LE. Circadian phase-
intrinsic EGFR tyrosine kinase activity and [3H]thymidine dependent stimulatory eVects of epidermal growth factor on
uptake only when Caco-2 cells were exposed to 1±5 ng/ml deoxyribonucleic acid synthesis in the duodenum, jejunum,
EGF at their basolateral but not at their apical aspect. How- ileum, caecum, colon and rectum of the adult male mouse.
Endocrinology 1980, 106, 1498±1503.
ever, the present study shows that by raising the EGF con- 5. Gross ME, Zorbas MA, Daniels YJ, et al. Cellular growth
centration to 25 ng/ml signalling via apical EGFR can be response to epidermal growth factor in colon carcinoma cells
eVectively transduced (cf. Figures 2 and 4). Since Cross and with an ampli®ed epidermal growth factor receptor derived from
Quaroni [15] found no diVerence in the KD or ligand binding a familial adenomatous polyposis patient. Cancer Res 1991, 51,
1452±1459.
between apical and basolateral EGFR, it remains to be seen
6. Frampton RJ, Omond SA, Eisman JA. Inhibition of human
whether the lack of sensitivity of apical membrane EGFR to cancer cell growth by 1,25-dihydroxyvitamin D3 metabolites.
low ligand concentrations can be attributed to any structural Cancer Res 1983, 43, 4443±4447.
diVerence outside the ligand binding domain. 7. Cross HS, Huber C, Peterlik M. Antiproliferative eVect of 1,25-
There exists, however, a signi®cant diVerence between the dihydroxyvitamin D3 and its analogs on human colon adeno-
carcinoma cells (Caco-2): in¯uence of extracellular calcium.
two populations of EGFR with respect to modulation of their Biochem Biophys Res Comm 1991, 179, 57±62.
density. Whereas expression of apical membrane EGFR 8. Cross HS, Pavelka M, Slavik J, Peterlik M. Growth control of
remains relatively unchanged in the presence of the ligand, human cancer cells by vitamin D and calcium in vitro. J Natl
their basolateral counterparts undergo rapid and eYcient Cancer Inst 1992, 84, 1355±1357.
reduction in numbers upon ligand binding (Table 1). In view 9. Cross HS, Farsoudi KH, Peterlik M. Growth inhibition of
human colon adenocarcinoma-derived Caco-2 cells by 1,25-
of the preponderance of EGFR on the basolateral membrane, dihydroxyvitamin D3 and two synthetic analogs: relation to in
eVective homologous downregulation can be seen as a vitro hypercalcemic potential. Arch Pharmacol 1993, 347, 105±
mechanism of self-attenuation of the response to EGF, which 110.
could protect colonocytes from mitogenic overstimulation. 10. Krishnan AV, Feldman D. Stimulation of 1,25-dihydroxyvitamin
D3 receptor gene expression in cultured cells by serum and
Our data indicate a pattern of EGF and 1,25(OH)2D3 growth factors. J Bone Mineral Res 1991, 6, 1099±1107.
interaction at the level of receptor expression that seems to be 11. Bruns DE, Krishnan AV, Feldman D, et al. Epidermal growth
characteristic for human colon carcinoma cells. Unlike, for factor increases intestinal calbindin-D9k and 1,25-dihydroxy-
instance, in normal enterocytes [11], in Caco-2 cells expres- vitamin D receptors in neonatal rats. Endocrinology 1989, 125,
sion of the VDR can be substantially suppressed by EGF 478±485.
12. Koga M, Eisman JA, Sutherland RL. Regulation of epidermal
(Figures 4 and 5). A similar eVect of EGF on VDR levels has growth factor receptor levels by 1,25-dihydroxyvitamin D3 in
only been observed in osteoblast-like cells [23]. Importantly, human breast cancer cells. Cancer Res 1988, 48, 2734±2739.
in the presence of 1,25(OH)2D3, Caco-2 cells are partially 13. Desprez PY, Poujol D, Falette N, Lefebvre MF, Saez S. 1,25-
resistant to the inhibitory eVect of EGF on VDR expression dihydroxyvitamin D3 increase epidermal growth factor receptor
gene expression in BT-20 breast carcinoma cells. Biochem Bio-
(cf. [24] and Figure 5). In addition, in Caco-2 cells, phys Res Commun 1991, 176, 1±6.
1,25(OH)2D3 decreases the abundance of EGFR, especially 14. Pinto M, Robine-Leon S, Appay MD, et al. Enterocyte-like dif-
in the presence of its ligand (Table 1), whereas in breast ferentiation and polarization of the human colon carcinoma cell
cancer cells the steroid hormone induces a 3-fold rise in line Caco-2 in culture. Biol Cell 1983, 47, 323±330.
15. Cross HS, Quaroni A. Inhibition of sucrase-isomaltase expres-
EGFR numbers [25].
sion by EGF in the human colon adenocarcinoma cells Caco-2.
The present study points at a pivotal role of the steroid Am J Physiol 1991, 261, C1173±C1183.
hormone 1,25(OH)2D3 in counteracting mitogenic stimula- 16. Bishop WP, Wen JT. Regulation of Caco-2 cell proliferation by
tion by EGF in Caco-2 cells and, thus, lends experimental basolateral membrane epidermal growth factor receptors. Am J
support for the notion derived from epidemiological data that Physiol 1994, 267, G892±G900.
17. Giuliano AR, Franceschi RT, Wood RJ. Characterization of the
vitamin D insuYciency is a contributing factor to the devel- vitamin D receptor from the Caco-2 human colon carcinoma cell
opment of colorectal carcinomas in humans [26]. We suggest line: eVect of cellular diVerentiation. Arch Biochem Biophys 1991,
that this is a condition which would facilitate downregulation 285, 261±269.
of the VDR by EGF (cf. Figures 4 and 5) which, in turn, 18. Hulla W, KaÂllay E, Krugluger W, Peterlik M, Cross HS. Growth
control of human colon adenocarcinoma-derived Caco-2 cells by
would further diminish the antimitotic potential of any residual
vitamin D compounds and extracellular calcium: relation to
1,25(OH)2D3 and at the same time alleviate VDR-mediated c-myc proto-oncogene and vitamin D receptor expression. Int J
control of EGFR levels (Table 1) favouring the transduction Cancer 1995, 62, 711±716.
of mitogenic signals via the EGF/EGFR pathway. In contrast, 19. Hidalgo IJ, Raub TJ, Borchardt RT. Characterization of the
if vitamin D is also available, for example, from endogenous human colon carcinoma cell line (Caco-2) as a model system for
intestinal epithelial permeability. Gastroenterology 1989, 96, 736±
production by colonocytes [27] in suYcient amounts to protect 749.
VDR expression from downregulation by EGF (cf. [24] and 20. Pedrazolli P, Bains MA, Watson R, Risher J, Hoy TG, Jacobs A.
Figure 5), the steroid hormone could then eVectively attenu- c-myc and c-myb oncoproteins during induced maturation of
ate the proliferative impulse from EGF signalling and thereby myeloid and erythroid human leukemic cell lines. Cancer Res
reduce the risk of neoplastic transformation of colonocytes. 1989, 49, 6911±6916.
21. Kawa S, Nikaido T, Aoki Y, et al. Vitamin D analogues up-reg-
ulate p21 and p27 during growth inhibition of pancreatic cancer
1. Carpenter G. Epidermal growth factor. Handb Exp Pharmacol cell lines. Br J Cancer 1997, 76, 884±889.
1981, 57, 90±126. 22. Wang QM, Jones JB, Studzinski GP. Cyclin-dependent kinase
2. Bower JM, Camble R, Gregory H, Gerring EL, Willshire IR. inhibitor p27 as a mediator of the G1-S phase block induced by
The inhibition of gastric acid secretion by epidermal growth fac- 1,25-dihydroxyvitamin D3 in HL60 cells. Cancer Res 1996, 56,
tor. Experimentia (Basel) 1975, 31, 825±826. 264±267.
EGF and Vitamin D Receptors in Caco-2 Cells 2125

23. Leeuwen JPTM van, Pols HAP, Schilte JP, Visser TJ, Birken- expression of 25-hydroxyvitamin D3-1a-hydroxylase activity and
haÈger JC. Modulation by epidermal growth factor of the basal regulation of side-chain metabolism. J Steroid Biochem Mol Biol
1,25(OH)2D3 receptor level and the heterologous up-regulation 1997, 62, 21±28.
of the 1,25(OH)2D3 receptor in clonal osteoblast-like cells. Calcif
Tissue Int 1991, 49, 35±42.
24. Cross HS, Hulla W, Tong W-M, Peterlik M. Growth regulation of
human colon adenocarcinoma-derived cells by calcium, vitamin D AcknowledgementsÐThese investigations were supported by
and epidermal growth factor. J Nutr 1995, 125, 2004S±2008S. grants from the Austrian Science Foundation (Project nos. P 09917-
25. Falette N, Frappart L, Lefebvre MF, Saez S. Increased epidermal MED and P 10133-MED). from the Herzfelder Foundation, from
growth factor receptor level in breast cancer cells treated by 1,25- the Oncology Research Fund and from the Anton Dreher Memorial
dihydroxyvitamin D3. Mol Cell Endocrinol 1989, 63, 189±198. Fund (Project no. 225/93) of the University of Vienna Medical
26. Garland C, Shekelle RB, Barrett-Connor E, Criqui MH, Rossof School. Drs Wolfgang Hulla and Wei-Min Tong are presently aYli-
AH, Paul O. Dietary vitamin D and calcium and risk of color- ated with the International Agency for Research on Cancer, 150
ectal cancer: a 19-year prospective study in men. Lancet 1985, 1, cours Albert Thomas, 69008 Lyon, France. Dr W. Hulla acknowl-
307±309. edges gratefully personal support from the Hans Moser Stiftung,
27. Cross HS, Peterlik M, Reddy GS, Schuster I. Vitamin D metab- Austria. The authors thank Mrs Erika Bajna for skilful technical
olism in human colon adenocarcinoma-derived Caco-2 cells: assistance.

You might also like