(Received for publication, January 2, 1997, and in revised form, April 21, 1997)
From the Hubrecht Laboratory, Netherlands Institute for Developmental Biology, 3584 CT Utrecht, The Netherlands
Previous studies have shown that
all-trans-retinoic acid (RA) inhibits in vitro
proliferation of hormone-dependent human breast cancer
cells but not the growth of hormone-independent cells. Here we report
on RA metabolism in breast cancer cells as examined by high performance
liquid chromatography analysis and found a correlation with sensitivity
to growth inhibition by RA. RA-sensitive T-47D and MCF-7 cells
exhibited high rate metabolism to polar metabolites, whereas
RA-resistant MDA-MB-231 and MDA-MB-468 cells metabolized RA to a much
lesser extent, and almost no polar metabolites could be detected. The
high metabolic rate in RA-sensitive cells appears to be the result of
autoinduction of RA metabolism, whereas RA-resistant cells showed no
such induction of metabolism. We observed furthermore that transfection
with retinoic acid receptor- expression vectors in RA-resistant
MDA-MB-231 cells resulted in increased RA metabolism and inhibition of
cell proliferation. Metabolism of RA, however, seems not to be required
to confer growth inhibition of human breast cancer cells. The
biological activity of the polar metabolites formed in RA-sensitive
cells was found to be equal or lower than that of RA, indicating that RA itself is the most active retinoid in these cells. Together our data
suggest that RA-sensitive cells contain mechanisms to activate strongly
the catabolism of RA probably to protect them from the continuous
exposure to this active retinoid.
Retinoids are a group of naturally occurring (e.g. all-trans-retinoic acid; RA1) and synthetic analogs of vitamin A which play an important role in cellular growth and differentiation (1, 2). The actions of retinoids are mediated by two types of receptors, the retinoic acid receptors (RARs) and retinoid X receptors (RXRs) (3, 4). Both receptor types belong to the steroid-thyroid hormone receptor superfamily and regulate transcription in the presence of their ligands. RARs can be activated both by RA and 9-cis-RA, whereas only 9-cis-RA binds to RXRs (5, 6).
Retinoids are highly effective in preventing chemically induced carcinogenesis in experimental animals (7) and can inhibit proliferation of a large variety of normal and neoplastic cell types in vitro (8). More recently the effectiveness of retinoids in the treatment and prevention of a number of human cancers has been established (9-15).
Unfortunately, lack of response to retinoid treatment and relapse of tumors are commonly observed. It is becoming increasingly clear that variations in metabolic rates of retinoids may be involved in the differences in retinoid response. Interindividual variation in the pharmacokinetics of retinoids has been reported for several malignancies, and a recent study suggested that high rate metabolism of RA is linked to an increased risk of squamous or large cell lung cancer (16-18). In acute promyelocytic leukemia patients, relapse and resistance to RA have been associated with a rapid and marked decrease of retinoid levels in the plasma (19, 20).
Retinoid resistance has also been documented in vitro in
cell lines derived from various tumor types, including breast cancer (21, 22). We and others have demonstrated that RA strongly inhibits
proliferation of estrogen receptor-positive human breast cancer cell
lines but not the growth of estrogen receptor-negative cell lines
(23-25). To investigate whether the RA resistance of human breast
cancer cells is associated with enhanced retinoid breakdown, we
examined metabolism in two estrogen receptor-positive and RA-sensitive
(T-47D and MCF-7) and two estrogen receptor-negative and RA-resistant
(MDA-MB-468 and MDA-MB-231) cell lines. We found that the RA-sensitive
cells exhibited high rate metabolism of RA to polar metabolites,
whereas the RA-resistant cells metabolized RA to a much lesser extent,
and almost no polar metabolites were detected. We furthermore observed
that transfection with RAR expression vectors in RA-resistant
MDA-MB-231 cells resulted in increased RA metabolism and inhibition of
cell proliferation. These observations suggested that metabolism of RA
may be required to confer growth inhibition of human breast cancer
cells. To investigate this possibility we examined the biological
activity of the polar metabolites formed in RA-sensitive cells. By
using retinoid-sensitive reporter cells, we found that only some RA
derivatives (e.g. 4-oxo-RA) are active, whereas most others
are inactive catabolic products, indicating that RA itself is the most
active retinoid in these cells. Together our data suggest that
RA-sensitive cells contain mechanisms to activate catabolism of RA to
protect them from the continuous exposure to this active retinoid.
DF medium (a 1:1 mixture of Dulbecco's modified
Eagle's medium and Ham's F-12, buffered with 44 mM
NaHCO3) and fetal calf serum (FCS) were obtained from Life
Technologies, Inc. DCC-FCS was prepared by treatment of FCS with
dextran-coated charcoal (DCC) to remove retinoids and steroids, as
described previously (26). Liarozole fumarate (27) was a kind gift from
Dr. W. Wouters (Janssen Research Foundation, Beerse, Belgium),
dissolved in ethanol at a concentration of 10 mM and stored
at 80 °C. [11,12-3H]RA (50 Ci/mmol) was obtained
from NEN Life Science Products. RA was purchased from Sigma, and
4-hydroxy-RA and 4-oxo-RA were kindly provided by Drs. M. C. Hsu and
L. H. Foley (Hoffman-LaRoche Laboratories, Nutley, NJ). Retinoids were
dissolved in dimethyl sulfoxide at a concentration of 10 mM
and stored in the dark at
80 °C.
MCF-7 cells were supplied by Dr. C. Quirin-Stricker (Strasbourg, France) and T-47D cells by Dr. R. L.
Sutherland (Sydney, Australia). MDA-MB-231 and MDA-MB-468 cells were
purchased from the American Type Culture Collection (Rockville, MD).
The RAR-transfected MDA-MB-231 cells have been described previously
(28). The F9-1.8 cell line is a derivative of F9 embryonal carcinoma
cells stably transfected with a 1.8-kilobase mouse RAR
2
promoter-lacZ reporter construct as described (29). All cell lines were
maintained in DF medium supplemented with 7.5% FCS. In case of F9-1.8
cells, geneticin (200 µg/ml) was added to the medium.
Cells were seeded into a 24-well plate in DF medium supplemented with 7.5% FCS. After 24 h of attachment, cells were treated with test compounds for the indicated times. Control cells were treated with solvent alone. The total DNA content/well was determined by fluorescent staining with Hoechst as described previously (26). Experiments were performed in triplicate.
Retinoid Extraction and FractionationCell lines were
cultured in 100-mm dishes in DF medium supplemented with 7.5% FCS at
different densities (ranging from 4 to 7 × 106 cells)
to obtain equal amounts of protein/dish at the start of RA treatment.
Cells were then treated with 10 nM [3H]RA or
1 µM RA for the indicated times. After incubation, medium was removed, and plates were rinsed with ice-cold phosphate-buffered saline. Cells were scraped in 1 ml of ice-cold phosphate-buffered saline and collected by centrifugation. Cell pellets were stored at
80 °C until extraction. Retinoids were extracted essentially as
described previously (30). In short, cells were lysed in 0.8 ml of
distilled water, and the lysate was added to 3 ml of methanol:dichloromethane (2:1). After addition of the internal standard
Ro10-1670 (31) (Hoffmann-LaRoche Laboratories), the mixture was
vortexed for 1 min and filtered through a glass sinter. The residue was
washed once with 3 ml of methanol:dichloromethane (2:1) and once with 5 ml of dichloromethane. The filtrate and washes were combined, and 1.75 ml of 0.9% NaCl was added. The mixture was vortexed for 1 min, and the
aqueous phase was removed. The organic phase was evaporated using a
stream of nitrogen, and the lipids were dissolved in 100 µl of
methanol, 60 mM ammonium acetate, pH 5.75 (9:1). In some
experiments retinoids were extracted from the medium by using the same
method.
To examine the biological activity of RA metabolites present in T-47D
extracts, near confluent cultures of 10 150-mm dishes were treated with
1 µM RA for 4 h. One dish was incubated also with 10 nM [3H]RA for 4 h to check whether peaks
detected by UV measurement corresponded to derivatives of
[3H]RA. Cell pellets of the dishes were pooled and stored
at 80 °C. Retinoid extracts were prepared and analyzed by HPLC.
One-min fractions were collected and extracted by partition between
distilled water (0.9 volume), methanol (1 volume), and dichloromethane
(1 volume) to remove mobile phase salts. The organic phase was
evaporated, and the residue was dissolved in dimethyl sulfoxide.
Aliquots of the fractions were tested in the reporter cell system
described below.
Retinoid extracts were injected into a reverse phase HPLC system containing a Spherisorb S50DS2 column (25 × 0.46 cm; Phase Separations Inc., Franklin, MA) and separated by gradient elution with solvent A (60 mM ammonium acetate, pH 5.75) and solvent B (methanol). The gradient program with a flow rate of 1 ml/min was 5 min isocratic at 65% B followed by a convex gradient (no. 4; Waters Associates, Brussel, Belgium) to 85% B in 15 min, a linear gradient to 99% B in 10 min, and 10 min isocratic at 99% B. Retinoids and the internal standard were detected by measuring the absorbance at 350 nm in a model 481 UV flow spectrometer (Waters Associates). After UV measurement radiolabeled retinoids were detected on-line with a LB506 radiochromatography monitor (Berthold, Bad Wildbad, Germany) equipped with a Z-1000 flow cell and a scintillant flow rate of 2 ml/min. Radioinert retinoids were run as standards.
Reporter Cell SystemF9-1.8 cells were seeded into a
96-wells plate in DF medium supplemented with 10% DCC-FCS. The
following day, retinoid fractions were added to the wells. After
24 h of incubation, cells were rinsed twice with
phosphate-buffered saline and lysed by overnight shaking in 50 µl of
buffer containing 100 mM potassium phosphate, pH 7.8, 0.2%
Triton X-100, and 1 mM dithiothreitol. The
-galactosidase activity in the cell lysates was determined by using
the Galacto-Light chemiluminescent reporter assay system (Tropic Inc.,
Bedford, MA). Chemiluminescence was measured in a Topcount
scintillation counter (Packard Instrument Company, Meriden, CT).
Whole cell extracts were prepared as
described previously (28). Fifteen µg of protein was run on a 10%
sodium dodecyl sulfate-polyacrylamide gel and transferred
electrophoretically to a nitrocellulose sheet. Membranes were treated
with blocking buffer containing 4% non-fat powdered milk, 10 mM Tris-HCl, pH 8.0, and 150 mM NaCl and then incubated for 2 h with anti-RAR rabbit polyclonal antiserum
RP
(F) (32) diluted at 1/500 in blocking buffer. The RP
(F)
antiserum was kindly provided by Drs. M. P. Gaub and C. Rochette-Egly
(Institut de Génétique et de Biologie Moleculaire et
Cellulaire, Illkirch, France). After washing in 10 mM
Tris-HCl, pH 8.0, 150 mM NaCl, and 0.05% Tween 20, the
membranes were immunostained using the ECL Western blotting system
(Amersham International plc, Little Chalfont, Bucks, U. K.).
Our previous
in vitro experiments have demonstrated that T-47D and MCF-7
human breast cancer cells are sensitive and that MDA-MB-468 and
MDA-MB-231 cells are resistant to growth inhibition by 1 µM RA (25). To investigate the role of RA metabolism in RA sensitivity of these cell lines, cells were incubated with 10 nM [3H]RA for 1-24 h, and cell extracts were
analyzed by HPLC. Fig. 1 shows a representative example
of the HPLC profiles of radioactive retinoids in the cell lines after
an incubation period of 4 h. The RA-sensitive T-47D and MCF-7
cells demonstrated extensive conversion to polar metabolites (retention
times 1-20), including 4-oxo-RA (peak 2) and 4-hydroxy-RA (peak 3).
MCF-7 cells furthermore showed conversion to apolar RA derivatives
(retention times 30-40). By contrast, the HPLC profiles of
RA-resistant MDA-MB-468 and MDA-MB-231 cells showed no significant
conversion of RA to polar or apolar derivatives.
Fig. 2 shows that the RA-sensitive cell lines
metabolized RA at a high rate. Most of the cell-associated RA was
converted within 4 h to polar metabolites detected in the organic
phase and very polar derivatives present in the aqueous phase of cell extracts. The majority of the very polar metabolites were discharged by
the cells into the medium. After 24 h of treatment more than 80%
of the initial amount of RA was recovered in the medium as derivatives
with very high polarity.
In Fig. 3 the rate of metabolism in RA-sensitive (T-47D
and MCF-7) and RA-resistant (MDA-MB-468 and MDA-MB-231) cells is
compared. In contrast to the RA-sensitive cell lines, the RA-resistant
cell lines metabolized trace amounts of RA (open circles) at
a much slower rate. The amount of cell-associated RA in MDA-MB-231
cells remained almost at a constant level during 24 h of
treatment, and MDA-MB-468 cells exhibited a low rate metabolism.
Although most of the intracellular RA in MDA-MB-468 cells was
metabolized after 16 h, the amount of polar metabolites was very
low because of conversion to very polar derivatives that were
discharged into the medium (data not shown).
To investigate RA metabolism at the same conditions under which RA-mediated growth inhibition has been studied, the human breast cancer cell lines were also incubated with 1 µM RA for 1 to 24 h (Fig. 3, closed circles). Both RA-sensitive cell lines show a strong reduction of cell-associated RA with kinetics similar to that observed during incubation with 10 nM [3H]RA. After 24 h, the amounts of RA in T-47D and MCF-7 cells were decreased to undetectable levels. The RA-resistant cell lines exhibited no rapid depletion of intracellular RA levels. MDA-MB-231 cells retained more than 50% of the initial amount of RA even after 24 h of treatment as has been found during incubation with 10 nM [3H]RA. MDA-MB-468 cells showed a decrease of intracellular RA to about 45% of the initial amount after 24 h of incubation. This is in contrast with the observations in experiments with 10 nM treatment, where most of the intracellular RA was cleared from the MDA-MB-468 cells after 24 h.
In summary, RA-sensitive breast cancer cells show a high rate of retinoid metabolism, whereas RA-resistant cells metabolize RA to a much lesser extent.
Effect of 4-Oxo- and 4-Hydroxy-RA on the Proliferation of Breast Cancer CellsFormation of the metabolites 4-oxo-and 4-hydroxy-RA
have long been considered as an inactivating pathway of RA (33).
However, recent studies have demonstrated that these metabolites
exhibit strong biological activity (29, 30, 34, 35), suggesting that
their formation could also be an activation step. Since we have found
that RA-sensitive and not RA-resistant cells convert RA to 4-oxo-and
4-hydroxy-RA (Fig. 1), it can be argued that conversion of RA to these
metabolites may be required to confer growth inhibition. To investigate
the effect of 4-oxo and 4-hydroxy-RA on breast cancer cell
proliferation, we treated the cell lines with various concentrations of
retinoid for 3 days. Fig. 4 shows that both retinoids
could inhibit the proliferation of RA-sensitive T-47D and MCF-7 cells
in a concentration-dependent fashion and were equipotent in
growth inhibition compared with RA. By contrast, the RA-resistant
MDA-MB-231 and MDA-MB-468 cells were not inhibited in their growth by
4-oxo or 4-hydroxy-RA, demonstrating that the resistant phenotype is
not due to the inability of the cells to convert RA to these polar
metabolites.
Liarozole Increases the Sensitivity of Breast Cancer Cells to Growth Inhibition by Retinoids
Enzymes of the cytochrome P450
system play an active role in the oxidative metabolism of RA (33, 36).
It has been demonstrated that P450 inhibitors, such as ketoconazole and
liarozole, inhibit RA metabolism in various in vivo and
in vitro systems, including human breast cancer cells (27,
37, 38). To investigate whether inhibition of RA metabolism by
liarozole alters the sensitivity of breast cancer cells to growth
inhibition by RA, MCF-7 cells were treated with various concentrations
RA for 7 days, in the presence or absence of 10 µM
liarozole. Fig. 5 shows that the combination of
liarozole and RA significantly enhanced the growth-inhibitory effect of
RA, particularly at lower concentrations. When 10 nM RA was
used in combination with liarozole, an effect was found similar to that
when cells were treated with 100 nM RA alone. Treatment
with 10 µM liarozole alone had no effect on the
proliferation of MCF-7 cells (data not shown). Since 10 µM liarozole strongly inhibited the proliferation of
T-47D cells, we could not examine the effect of liarozole on
retinoid-mediated growth inhibition in these breast cancer cells.
We also examined the effect of 4-oxo-RA on the proliferation of MCF-7 cells in the presence or absence of 10 µM liarozole. Fig. 5 shows that liarozole significantly increased the growth-inhibitory effect of 4-oxo-RA. This increase, however, was less pronounced than the enhancement of growth inhibition by RA. We furthermore observed that 4-oxo-RA was somewhat less potent than RA to inhibit the growth of MCF-7 cells in this 7-day proliferation assay. This is in contrast with the results in the 3-day proliferation assay, where 4-oxo-RA was almost equipotent to RA (Fig. 4).
To check whether RA metabolism indeed was inhibited by liarozole in the proliferation experiments, MCF-7 cells were pretreated with 10 µM liarozole for 24 h and subsequently incubated with 10 nM [3H]RA and 10 µM liarozole for 4 h. HPLC analysis of the MCF-7 extracts revealed that conversion of RA to polar metabolites was reduced by more than 75% (data not shown).
Together the results indicate that enhancement by liarozole of the antiproliferative effects of RA and 4-oxo-RA is due to inhibition of their metabolism.
Biological Activity of RA Metabolites Extracted from Breast Cancer CellsThe above results suggest that RA does not require prior
metabolism for its antiproliferative effects on breast cancer cells, even though some of the metabolites formed are active as well and can
contribute to growth inhibition (e.g. 4-oxo-RA, Fig. 4). To
investigate further to what extent RA metabolites formed in breast
cancer cells are bioactive retinoids, we incubated T-47D cells with 1 µM RA for 4 h and tested the extracted retinoids for
their ability to transactivate the retinoid-inducible 1.8-kilobase RAR2 promoter stably transfected into F9-1.8 reporter cells. This
reporter system has been demonstrated to be very useful in determining
biological activity of retinoids, and the response of the RAR
2
promoter correlated well with other effects of retinoids in the F9-1.8
cells (29). Fig. 6A shows the HPLC profile of retinoids extracted from T-47D cells after 4 h of incubation. At
this time point, T-47D cells contain high levels of polar metabolites, as has been found previously (Fig. 2). Fig. 6B shows that
most of the fractions containing polar metabolites were unable to
activate the RAR
2 promoter. Only fractions 13, 14, 15, and 17 showed
activation of the RAR
2 promoter. Activity in fraction 13 and
fraction 14 is very probably due to the presence of 4-oxo- and
4-hydroxy-RA, respectively, as these are the positions where the
respective retinoids coelute under these conditions. The identities of
the active retinoids present in fraction 15 and 17, however, are
presently unknown. The RA-containing fractions (fractions 26 and 27)
exhibited most of the activity in the T-47D cell extract. The formation of a limited amount of biological active metabolites, as measured in
this assay system, again suggests that in T-47D cells metabolism of RA
to polar derivatives is not required for antiproliferative activity.
RA Treatment Induces RA Metabolism in RA-sensitive Breast Cancer Cells
The strong increase of RA metabolism in T-47D and MCF-7
cells observed after 4 h of incubation with either 10 nM or 1 µM RA (Fig. 3) suggests that RA
metabolism in these cell lines is induced by RA. To investigate this
possibility, cells were pretreated with either 10 nM or 1 µM RA for various times and subsequently incubated with
10 nM [3H]RA for 2 h. Fig.
7 shows that RA metabolism in T-47D cells was induced
already after 2 h of pretreatment, with no further increase by
prolonged pretreatment. A stronger induction of RA metabolism was
observed by pretreatment with 1 µM than with 10 nM RA. The difference in induction of RA metabolism is most
pronounced after 24 h of pretreatment, where strongly increased
metabolism was seen in the presence 1 µM RA but not in
the presence of 10 nM RA. Similar effects of pretreatment
with RA on metabolism of RA were obtained in MCF-7 cells (data not
shown). By contrast, RA could not induce RA metabolism in RA-resistant
MDA-MB-231 (Fig. 7) and MDA-MB-468 cells (data not shown). These
results suggest that the low rate of RA metabolism in RA-resistant
cells is due to their inability to induce metabolism in the presence of
RA.
Increased RAR
Retinoid action is mediated by the nuclear
retinoid receptors (RARs and RXRs), which are known to activate
transcription of target genes in response to nanomolar concentrations
of retinoids (4). The RA-mediated induction of RA metabolism by low
doses of RA in breast cancer cells suggests a receptor-mediated
process. Since T-47D and MCF-7 cells express higher levels of RAR
than MBA-MD-468 and MDA-MB-231 cells (25), we investigated whether increased expression of RAR
in MDA-MB-231 cells could influence the
rate of RA metabolism. RAR
-transfected derivatives of MDA-MB-231 cells, clones
8,
18, and
26, expressing different levels of RAR
, were incubated with 1 µM RA for 2-24 h, and
retinoid extracts were analyzed by HPLC. Fig. 8 shows
that clones
8 and
18 containing high levels of RAR
protein
demonstrated an increased rate of RA metabolism, whereas clone
26
with low level expression of RAR
showed a metabolic rate similar to
that of parental MDA-MB-231 cells.
We investigated RA metabolism in human breast cancer cells and found a correlation with sensitivity to growth inhibition by RA. The time course experiments showed that most of the RA added to the cultures of RA-sensitive T-47D and MCF-7 cells was converted to more polar derivatives within 4 h, whereas RA-resistant MDA-MB-231 and MDA-MB-468 cells retained more than 45% of the initial amount of RA after 24 h of incubation. We could demonstrate that RA induces its own metabolism in RA-sensitive cells. Two hours of pretreatment with RA dramatically increased the rate of metabolism, whereas no further enhancement was observed after prolonged pretreatment. These observations were made during incubation with both 10 nM and 1 µM RA. RA-induced metabolism of RA was not observed in the RA-resistant cells, suggesting that the low rate of RA metabolism in these cells may at least in part be explained by the lack of this autoinduction mechanism.
RA-induced metabolism of RA appears to be regulated at the transcriptional level. It was shown recently that actinomycin D prevented the induction of RA metabolism in T-47D cells (39). We have repeated these experiments with similar results (data not shown). Since RA induced metabolism in RA-sensitive cells within 2 h, transcription and translation of the appropriate target genes must occur rapidly. In an efficient cloning strategy used to identify RA-responsive genes in embryonal carcinoma cells, it was found that some of the cloned genes were rapidly induced by RA (40). An increase in RNA levels of these genes was detectable 30 min after the RA addition, and the RNA levels further increased up to 6 h and subsequently decreased between 12 and 24 h. Genes involved in the induction of RA metabolism in breast cancer cells presumably belong to this class of early RA-responsive genes. So far individual candidate genes have not been identified.
A category of proteins relevant to RA metabolism is the cytochrome P450 enzyme system. The P450 enzymes play an active role in RA metabolism, and ketoconazole and liarozole were shown to inhibit the P450-dependent oxidation of RA (27, 37, 38). Various purified isozymes of microsomal cytochrome P450, such as P450 1A2, 2C7, and 2C8, have been shown to be capable of metabolizing RA (36, 41, 42) and other retinoids (43) in a reconstituted system. More recently, it has been reported that expression of P450 enzymes can be regulated by retinoids (44-46). Furthermore P450 enzymes have been shown to be expressed in normal and tumor breast tissue (47). Although the identities of the P450s involved in metabolism of RA in the breast are unknown, it is clear from our experiments with liarozole that RA-metabolizing P450 enzymes are present in the RA-sensitive cells.
Takatsuka et al. (48) recently reported a similar relationship between RA metabolism and inhibition cell proliferation in various cell lines, including breast cancer cells. However, the investigators obtained these results using serum-free assay conditions and could not obtain similar results in 10% serum-containing medium. Under the latter conditions, RA was not rapidly taken up by the cells, and no difference in metabolism between RA-sensitive and -resistant cells nor an effect on cell growth was detected (48). We performed our experiments in 7.5% serum-containing medium but did not encounter these problems. The authors furthermore suggested that a metabolite of RA, rather than RA itself, may be responsible for the observed growth inhibition. However, our data do not support this idea as will be discussed below.
We demonstrated that in contrast to RA-sensitive breast cancer cells,
the RA-resistant cell lines exhibit no significant metabolism of RA,
and polar metabolites were present in very low amounts. The most simple
explanation for the RA resistance of MDA-MB-231 and MDA-MB-468 cells
would be that metabolism of RA to polar metabolites is required for the
antiproliferative effect of RA on breast cancer cells. Although it has
been found that metabolites of RA exhibit strong biological activity
in vivo and in vitro (29, 30, 34, 35, this
study), our data do not support this hypothesis. In the first place,
examination of the biological activity of retinoid fractions from
metabolizing T-47D cells revealed that RA-containing fractions exhibited more activity than fractions containing polar metabolites. Some of the polar metabolites may have a higher transactivation capacity than RA, but their concentration in the fractions is too low
to give strong activation of the RAR2 promoter. Second, we
demonstrated that the RA-resistant breast cancer cells were not
growth-inhibited by the metabolites 4-oxo- and 4-hydroxy-RA, whereas
these retinoids were potent growth inhibitors of RA-sensitive cells.
These results show that the resistant phenotype of breast cancer cells
is not due to their inability to convert RA to the 4-oxygenated
metabolites. Third, we demonstrated that the enhancement by liarozole
of the antiproliferative effects of RA on MCF-7 cells appears to be due
to decreased metabolism. We also found that growth inhibition by
4-oxo-RA was enhanced by liarozole, suggesting that the breakdown of
4-oxo-RA is also inhibited. It has been reported recently that
liarozole indeed is capable of inhibiting both the metabolism of RA and
4-oxo-RA (49). Finally, we demonstrated that 4-oxo-RA is somewhat less
potent than RA in growth inhibition of MCF-7 cells during a 7-day
proliferation experiment, whereas treatment for 3 days appeared to be
too short to observe the difference between the growth-inhibitory
activity of these retinoids. This difference may be explained by the
fact that RA is converted first to metabolites, such as 4-oxo-RA, prior
to its catabolism.
A second possible explanation for RA resistance of breast cancer cells is that impairment in the retinoid signaling has occurred at the level of the cellular RA-binding proteins (CRABPs). CRABPs are thought to play a negative role in retinoid signaling by sequestering RA in the cytoplasm and facilitating its catabolism such that lower RA concentrations reach the nucleus (50-52). It has been demonstrated that cells from patients with acute promyelocytic leukemia at relapse contained high levels of CRABP-I, whereas this protein was not detected prior to RA administration (53). Furthermore, overexpression of CRABP-I in F9 embryonal carcinoma cells resulted in decreased sensitivity to RA (51). Our preliminary data show that both RA-sensitive and -resistant breast cancer cells display low or no expression of CRABP-I mRNA (data not shown). Similar observations have recently been reported for other breast cancer cell lines (54). Together these results indicate that the RA-resistant phenotype of human breast cancer cells seems not to be due to high levels of CRABP-I.
The third possibility to explain RA resistance of breast cancer cells
is that impairment in the retinoid signaling occurs at the level of
nuclear receptors. We and others have demonstrated that RAR expression
is of crucial importance for the cellular response of breast cancer
cells and that RA-resistant cells expressed low levels of RAR
compared with RA-sensitive cells (25, 55). It was furthermore shown
that retinoic acid-responsive element-dependent transcription induced by RA was higher in RA-sensitive than in -resistant cells, indicating that resistance could be due to
underexpression of functional RARs (25). More recently, our laboratory
and others demonstrated that elevation of RAR expression levels in
RA-resistant MDA-MB-231 and MDA-MB-468 cells by stable transfection
with RAR
or RAR
expression vectors restored growth inhibition by
RA (27, 56-59). Here we demonstrated that RAR
-transfected
MDA-MB-231 cells showed increased RA metabolism, which indicates that
at least RAR
is involved in RA-mediated growth inhibition and in
retinoid metabolism. Retinoid receptors are known to activate
transcription in response to nanomolar concentrations of RA. Since RA
metabolism can be induced by 10 nM RA and appears to be
directly at the transcriptional level, it is tempting to speculate that
RA may directly regulate transcription of the metabolizing enzyme(s)
through RAR
. The exquisite sensitivity of this (negative) regulatory
loop probably emphasizes the necessity for the cell to control strictly
the intracellular levels of RA as a potential growth and
differentiation regulatory molecule. Another autoregulatory loop has
been described recently in human keratinocytes, whereby RA regulates
its own biosynthesis from retinol through regulation of retinol
esterification, indicating that the availability of RA in these cells
is controlled strictly (60).