(Received for publication, October 18, 1996, and in revised form, February 28, 1997)
From the Institute of Human Nutrition and the
§ Department of Medicine, Columbia University, New York,
New York 10032, the ¶ Department of Veteran Affairs and the
Department of Medicine, Stanford University School of Medicine,
Palo Alto, California 94305, and the
Dipartimento di
Biochimica e Biotecnologie Mediche, Università di Napoli, I-80131
Napoli, Italia
Adipose tissue is an important storage depot for
retinol, but there are no data regarding retinol mobilization from
adipose stores. To address this, dibutyryl cAMP was provided to murine BFC-1 adipocytes and its effects on retinol efflux assessed. High
performance liquid chromatography analysis of retinol and retinyl
esters in adipocytes and media indicated that cAMP stimulated, in a
time- and dose-dependent manner, retinol accumulation in the culture media and decreased cellular retinyl ester concentrations. Study of adipocyte retinol-binding protein synthesis and secretion indicated that cAMP-stimulated retinol efflux into the media did not
result from increased retinol-retinol-binding protein secretion but was
dependent on the presence of fetal bovine serum in the culture media.
Since our data suggested that retinyl esters can be hydrolyzed by a
cAMP-dependent enzyme like hormone-sensitive lipase (HSL),
in separate studies, we purified a HSL-containing fraction from
BFC-1
adipocytes and demonstrated that it catalyzed retinyl
palmitate hydrolysis. Homogenates of Chinese hamster ovary cells
overexpressing HSL catalyzed retinyl palmitate hydrolysis in a time-,
protein-, and substrate-dependent manner, with an apparent
Km for retinyl palmitate of 161 µM,
whereas homogenates from control Chinese hamster ovary cells did
not.
Adipose tissue plays an important role in retinoid (vitamin A and its analogs) storage and metabolism (1, 2). After the liver, adipose tissue is the site of the next largest tissue store of retinoid in the rat (2). It has been estimated that adipose tissue retinoid may account for as much as 15% of the total retinoid present in the body of a rat (2). Within adipose tissue, the adipocyte is the cellular site of retinoid storage (1-3). Levels of total retinol (retinol + retinyl ester) in rat epididymal adipocytes are very substantial and rank behind only those found in hepatic stellate cells (also called fat-storing, Ito cells, or lipocytes) and retinal pigment epithelial cells (2, 4, 5). Approximately 60% of the total retinol present in adipocytes is in the form of retinyl esters (3, 6). The level of total retinol present in adipocytes is directly influenced by dietary retinoid intake and it has been suggested that the enzyme lipoprotein lipase plays a role in facilitating adipocyte uptake of dietary retinoid (6). Adipocytes synthesize and secrete retinol-binding protein (RBP),1 the sole plasma transport protein for retinol (2, 7, 8). In addition, adipocytes express cellular retinol-binding protein, an intracellular protein which is thought to play a central role in retinol uptake and metabolism (esterification to retinyl esters and oxidation to retinoic acid) (1-3, 9-11). Hence, adipocytes not only accumulate retinoid from the circulation but also express proteins whose presence suggests that they are able to mobilize it back into circulation.
There is presently no information available regarding retinoid mobilization from adipocytes or the factors which regulate this process. This is unlike fatty acid mobilization from adipocyte triglyceride stores, which is well understood. Both uptake and release of fatty acids by adipocytes is under acute hormonal and neural control and is largely regulated by factors that modulate the activity of hormone-sensitive lipase (HSL) (12-14). To gain a better understanding of retinoid mobilization from adipocytes, we have investigated retinyl ester hydrolysis in and retinol efflux from adipocytes. In addition, we have studied in vitro whether HSL can catalyze retinyl ester hydrolysis.
Dulbecco's modified Eagle's medium (DMEM),
Coon's F-12/DMEM, penicillin-streptomycin, gentamycin, biotin,
pantothenate, and pyruvate were purchased from Life Technologies, Inc.
Insulin, dexamethasone, dibutyryl cAMP, 3,3,5-triiodothyronine,
collagenase type IV, collagen type I, and Medium 199 were purchased
from Sigma. Fatty acyl chlorides were obtained from Nu-Chek-Prep, Inc.
(Elysian, NM). Fetal bovine serum was purchased from Intergen
Laboratories (Purchase, NY). Guanidinium thiocyanate and
-mercaptoethanol were purchased from Fluka Chemical Corp.
(Ronkonkoma, NY). Nick translations were performed using a kit
purchased from Boehringer Mannheim Biochemicals. All radiolabeled
nucleotides, [3H]retinol, [14C]triolein,
and cholesteryl [14C]oleate were purchased from DuPont
NEN Research Products (Boston, MA) and XAR-2 autoradiography film was
from Eastman Kodak Co. (Rochester, NY). Authentic
all-trans-retinol was a kind gift of Dr. A. Levin of Hofmann
LaRoche (Nutley, NJ).
BFC-1 preadipocytes cells were kindly
provided by Dr. Gerard Ailhaud and Dr. Claude Forest (Nice, France)
through Dr. Daryl K. Granner (Vanderbilt University, Nashville, TN)
(15, 16). The BFC-1
preadipocytes were cultured in DMEM containing
glucose (1000 mg/liter), penicillin (200 IU/ml), streptomycin (50 mg/ml), pyruvate (110 mg/liter), sodium bicarbonate (1.2 g/liter),
biotin (8.0 mg/liter), pantothenate (4 mg/liter), 15 mM
HEPES, pH 7.4, and 10% fetal calf serum (3, 17). Cells were inoculated
into 75-cm2 plastic culture flasks and grown at 37 °C in
a humidified atmosphere of 5% CO2, 95% air. Confluent
BFC-1
preadipocytes were differentiated through exposure to culture
medium supplemented with 10
8 M insulin and
2 × 10
9 M triiodothyronine
(differentiation medium) (3, 15-17). After induction to differentiate,
the BFC-1
cells were continuously maintained in the medium
supplemented with insulin and triiodothyronine, with changes in media
every 2 or 3 days. The cells used for our studies were taken 27 days
after confluence when, as assessed by the presence of intracellular
lipid droplets, approximately 80% of the preadipocytes had
differentiated into adipocytes. Under these conditions, each flask of
differentiated BFC-1
adipocytes contained approximately 3 × 106 cells.
We have previously demonstrated that BFC-1 adipocytes will take up
retinol from the culture medium and convert greater than 50% of this
retinol to retinyl ester (3). Since we wished to study retinol
mobilization from adipocytes, the BFC-1
adipocytes were preloaded
with retinol. For this purpose, differentiation medium was supplemented
with 1 µM all-trans-retinol bound to 1% (w/v)
fatty acid free bovine serum albumin (BSA). A small aliquot (<100
µl) of a concentrated solution of retinol in ethanol was added to 5 ml of a 1% fatty acid-free BSA solution in PBS (10 mM
sodium phosphate, pH 7.4, 150 mM sodium chloride) to a
final concentration of 0.1 mM. The 0.1 mM
retinol-BSA solution was sterilized by passage through a sterile filter
and added to 500 ml of the differentiation medium. This gave a
concentration of retinol in the differentiation medium of 1 µM. The retinol supplemented medium was placed over the
BFC-1
adipocytes 22 days after confluence (5 days before the start
of our experiments). Two days later this medium was exchanged for fresh
retinol-supplemented medium. The BFC-1
adipocytes remained in this
medium until they were taken for study on day 27 after confluence.
On the day of study, the medium was removed from over the adipocytes and the adipocytes were washed with 10 ml of ice-cold PBS. The adipocytes were washed 2 additional times with 10 ml of ice-cold PBS to remove residual retinol supplemented medium. Twenty ml of differentiation medium, either supplemented with dibutyryl cAMP or unsupplemented, was then placed over the cells. The dibutyryl cAMP was first dissolved in small volumes of differentiation media, and this was added to appropriate final concentrations directly to the 500 ml of differentiation medium. The adipocytes were returned to the cell culture incubator until taken for analysis.
At different time intervals, adipocytes and culture medium were taken
for determination of retinol and retinyl ester levels. For this
purpose, the culture medium was removed, cells were washed 3 times with
10 ml of ice-cold PBS and scraped from the plate into 10 ml of ice-cold
PBS. The 20 ml of medium and the 10 ml of scraped adipocytes were
stored at 70 °C for up to 4 weeks prior to analysis. For
measurement at zero time, culture medium was placed over the cells and
then immediately removed as described above.
Cellular and media levels of retinol and individual retinyl esters were measured by reverse phase HPLC using previously described procedures employed for rat liver retinol and retinyl ester measurements (18). Authentic standards of retinyl palmitate, retinyl oleate, retinyl stearate, and retinyl linoleate were synthesized from corresponding fatty acyl chlorides using standard procedures (19).
Radioimmunoassay for RBPMedia and cellular levels of RBP
for BFC-1 adipocytes were measured by using a specific
radioimmunoassay for rat RBP which has been used previously to
quantitate mouse RBP (20, 21). The rabbit anti-RBP used for these
assays does not recognize bovine RBP and thus does not detect the
bovine RBP present in adipocyte culture media (20-22). The details of
this procedure have been described (22).
Free fatty acid levels in the culture medium were analyzed using a kit and the accompanying instructions obtained from Wako Chemicals (Dallas, TX).
RNA ExtractionTotal RNA was extracted from the
preadipocyte and adipocyte cultures using a modification of the
procedure described by Chomczynski and Sacchi (23). For this purpose,
the culture medium was removed from over the cells in the tissue
culture flask and 5 ml of a solution containing 4 M
guanidinium thiocyanate, 25 mM sodium citrate, pH 7.0, 0.1 M -mercaptoethanol, and 0.5% Sarcosyl was added
directly onto the cells. The resulting cell lysate was collected from
the flask and transferred to a Corex tube. Sequentially, 0.5 ml of 2 M sodium acetate, pH 4.0, 5 ml of TNE saturated phenol, and
1.0 ml of a chloroform/isoamyl alcohol mixture (49:1) were added to the
cell lysate. The lysate was mixed vigorously upon the addition of each
of the solutions. The final solution was cooled on ice for 30 min and
centrifuged at 10,000 × g for 20 min at 4 °C. All
subsequent steps for RNA isolation were carried out exactly as
described by Chomczynski and Sacchi (23).
Relative levels of RBP and HSL
mRNAs were determined by Northern blot analysis. For this purpose,
total RNA was resolved by agarose gel electrophoresis in the presence
of formaldehyde, and transferred to nylon membranes. The loads of total
RNA applied to each gel were demonstrated to be equal by ethidium
bromide staining of the 28 S and 18 S ribosomal RNAs. For all RNA
samples, the ratio of intensities of the 28 S and 18 S ribosomal RNA
bands was approximately 2. These procedures were all carried out
according to standard descriptions (24). All blots were washed at high stringency in 0.1 × SSC, 1.0% SDS at 65 °C. Specific
mRNAs of interest were identified by hybridization with
32P-labeled probes prepared by nick translation of cDNA
clones (25). 32P-Labeled probes were made using a kit
supplied by Boehringer Mannheim according to the manufacturer's
instructions. Hybridized membranes were exposed to Kodak XAR-2 film at
80 °C. Relative mRNA levels were quantitated by scanning
densitometry using a LKB Laser Densitometer (Pharmacia Biotech Inc.,
Piscataway, NJ).
For hepatocyte isolation, mouse livers were perfused in situ through the portal vein at a rate of 10 ml/min with Kreb's buffer (NaCl, 6.75 g/liter; NaHCO3, 2 g/liter; KCl, 0.345 g/liter; MgSO4, 0.287 g/liter; gentamycin, 10 mg/liter; and glucose, 1 g/liter) for 4 min immediately followed by perfusion for 12 min with Medium 199 (Sigma) containing 0.25 mg/ml collagenase type IV (Sigma) and 10 ng/ml soybean trypsin inhibitor (Sigma). The liver then was rapidly excised from the body cavity and transferred into a Petri dish. The disassociated hepatocytes were resuspended in Medium 199 and centrifuged at 20 × g for 3 min at 25 °C. The resulting cell pellet was washed with DMEM containing glucose (4.5 g/liter), sodium pyruvate (100 mg/liter), sodium bicarbonate (3.7 g/liter), penicillin (200 IU/ml)/streptomycin (50 mg/ml), 25 mM HEPES, pH 7.4, and 10% fetal calf serum and plated in the same medium on dishes coated with collagen type I (Sigma). This medium was changed after 4 h, after viable hepatocytes had attached to the plate. Plated hepatocytes were used for our studies the morning following isolation.
Partial Purification of a HSL-containing Fraction from BFC-1A HSL-containing fraction was isolated from BFC-1
adipocytes according to procedures described by Fredrikson et
al. (26). At day 27 of differentiation, the culture medium was
removed from over the adipocytes from three 75-cm2 tissue
culture flasks. Adipocytes were washed three times in ice-cold PBS and
scraped from the plates into a total of 5 ml of 0.25 M
sucrose, 1 mM EDTA, and 5 mM Tris-HCl, pH 7.4. The scraped adipocytes were homogenized with a Polytron homogenizer and
centrifuged in a swinging bucket rotor at 100,000 × g
for 45 min. The floating fat layer was removed by slicing the
centrifuge tube, and the supernatant was filtered through glass wool to
remove the remaining fat. The pH of the filtered supernatant was
decreased to pH 5.2 by dropwise addition of 0.2 M acetic
acid to the supernatant with gentle stirring on a magnetic stir plate.
After incubating at pH 5.2 for 20 min, the precipitate was collected by
centrifugation for 30 min at 1,000 × g at room
temperature. The resulting pellet was resuspended in 2 ml of 20 mM Tris-HCl, pH 7.4, 1 mM EDTA, and 1 mM dithioerythritol and stored at
70 °C prior to use.
The protein contents of the cells and partially purified HSL fraction were analyzed using the Coomassie Protein Assay Reagent and the accompanying instructions obtained from the Pierce Chemical Co. (Rockford, IL).
CHO cells were grown in Coon's F-12/Dulbecco's modified Eagle's media containing 1 g/liter glucose (50:50) and supplemented with 10% fetal calf serum, penicillin (10,000 units/ml), and streptomycin (10,000 µg/ml). The CHO cells were transfected with pCEP4-HSL as described previously (27). pCEP4-HSL contains nucleotides 581-2933 of the rat HSL cDNA (a kind gift of Dr. M. Schotz of the University of California, Los Angeles) and encompasses the entire HSL coding region. Transfected cells were selected by hygromycin resistance, and cells overexpressing HSL were subcloned. HSL protein was detected by immunoblotting cell extracts with anti-rat HSL/fusion protein IgG and visualized by chemiluminescence as described previously (28).
Assay for HSLDepending on specific experimental needs, HSL activity was measured using three different substrates, triolein, cholesteryl oleate, and retinyl palmitate. For identification and confirmation of CHO cell clones overexpressing HSL, HSL activity was determined using a cholesterol [14C]oleate-containing emulsion as described previously (28). To assess whether HSL catalyzes retinyl ester hydrolysis, CHO cells overexpressing HSL and wild-type CHO cells were cultured to confluence in DMEM containing 10% fetal calf serum. The CHO cells were washed three times with ice-cold PBS and scraped from the plates into the ice-cold PBS. After pelleting the cells by centrifugation at 20 × g for 10 min at 4 °C, the cells were resuspended in 20 mM Tris-HCl, pH 7.4, containing 1 mM EDTA, 1 mM dithiothreitol, and 0.02% bovine serum albumin and homogenized using 10 strokes of a tight fitting Dounce homogenizer. For assay of HSL, whole CHO cell homogenate containing known amounts of protein was mixed with an emulsion (29) containing either [14C]triolein (1.25 mM final in assay concentration) or unlabeled retinyl palmitate (routinely at a final in assay concentration of 1.25 mM) or, to mimic adipocyte concentrations of triolein and retinyl palmitate, as a mixture of [14C]triolein (20 mM final in assay concentration) and unlabeled retinyl palmitate (10 µM final in assay concentration) to a final volume of 0.2 ml. Assays were carried out at 37 °C with gentle shaking for periods of up to 2 h. For determination of an apparent Km of HSL for retinyl palmitate-containing emulsions providing final in assay retinyl palmitate concentrations ranging between 0.01 and 1.25 mM were employed. To terminate reactions employing retinyl palmitate as substrate, 0.2 ml of ice-cold absolute ethanol was added to the tube followed by 2 ml of hexane to extract retinoids. Retinol arising from retinyl palmitate hydrolysis was determined by reverse HPLC analysis of the extracted retinoids employing the HPLC procedures described above. Assays employing triolein as substrate were terminated and quantitated by measure of [14C]oleate formation as described by Strålfors et al. (29).
The ability of the pH 5.2 precipitate prepared from BFC-1 adipocytes
(see above) to catalyze retinyl palmitate hydrolysis was assessed
exactly as described above for CHO cell homogenates except that
resuspended pH 5.2 precipitate was substituted for CHO cell homogenate.
As with the CHO cell homogenates, both [14C]triolein
(1.25 mM final in assay concentration) and unlabeled retinyl palmitate (1.25 mM final in assay concentration)
were used as substrates in the enzyme assay.
Statistically significant changes in response to experimental treatments were determined by Student's t test according to standard procedures (30).
Since these studies were undertaken to investigate retinol
mobilization from adipocytes, we needed first to load the murine BFC-1 adipocytes with retinol. In earlier studies, we showed that
when BFC-1
adipocytes are incubated in media containing physiologic
levels of retinol, the cultured adipocytes take up and accumulate the
retinol (3). Thus, five days before the start of the experiments,
BFC-1
adipocytes were cultured in our standard differentiation
medium (described under "Experimental Procedures") supplemented
with 1 µM retinol bound to 1% (w/v) fatty acid-free BSA.
(Plasma retinol levels in mice are approximately 1 µM
(21).) The differentiation medium was supplemented with retinol
beginning 22 days after the start of induction of adipocyte differentiation, at which point approximately 80% of the cells present
in the culture flask were adipocytes (as assessed by the characteristic
lipid droplets in the cells). After 5 days of exposure to media
containing 1 µM retinol, the BFC-1
adipocytes
contained approximately 425 ng of total retinol (retinol + retinyl
ester)/106 cells. Tsutsumi et al. (2) have
reported that primary adipocytes freshly isolated from rat epididymal
fat pads contain approximately 800 ng of total retinol/106
cells. At the start of our studies, approximately 55% of the retinol
present in the BFC-1
adipocytes was present as retinyl ester.
Retinyl linoleate, retinyl palmitate, and retinyl oleate accounted for
over 85% of the total retinyl ester present in the adipocytes. This
total retinol level and retinyl ester composition are in agreement with
those previously reported for BFC-1
adipocytes provided physiologic
levels of retinol in the medium (3).
Initially, we explored whether dibutyryl cAMP treatment of BFC-1
adipocytes will bring about retinol mobilization from the cells. Since
it is known that treatment of cultured adipocytes with hormones or
other factors which elevate intracellular cAMP levels stimulates efflux
of free fatty acids from the cells, we reasoned that increases in
intracellular cAMP concentrations might also stimulate retinol efflux.
Cultures of BFC-1
adipocytes, which had been preloaded with retinol,
were exposed to the standard differentiation media (unsupplemented with
retinol-BSA) containing 1 mM dibutyryl cAMP. At 8, 24, 48, and 72 h after addition of the dibutyryl cAMP supplemented media,
the media were taken for determination of retinol, retinyl ester and
free fatty acid levels. As seen in Fig. 1, panel
A, treatment of the adipocytes with 1 mM dibutyryl
cAMP results in a statistically significant elevation of media retinol
levels at 24 and 48 h, as compared with unsupplemented adipocyte
media. No retinyl ester was present in medium from either dibutyryl
cAMP supplemented or unsupplemented adipocytes. The dibutyryl cAMP
effect on media retinol levels was time-dependent, with the
effect reaching a maximum within 24 h after addition of the
dibutyryl cAMP to the media. By 24 h, the concentration of retinol
released by the treated adipocytes represented approximately 45% of
the level of retinol present in the adipocytes at 0 h. As
expected, the concentration of free fatty acids in the media taken from
over dibutyryl cAMP-treated adipocytes was also much greater than that
present in media from unsupplemented adipocytes (data not shown); thus,
indicating that the dibutyryl cAMP treatment, as would be expected,
stimulated free fatty acid mobilization from the adipocytes. Addition
of 1 mM butyrate to the adipocyte media did not influence
the levels of either retinol or fatty acids present in the BFC-1
adipocyte media.
We also asked whether dibutyryl cAMP treatment influences adipocyte
levels of retinol and retinyl esters. Fig. 1 also shows the
time-dependent effects of 1 mM dibutyryl cAMP
treatment on adipocyte retinol (panel B) and retinyl ester
levels (panel C). As seen in Fig. 1, panel B,
dibutyryl cAMP treatment brings about a time-dependent
increase in cellular retinol concentration. Cell-associated retinol
levels were found to be significantly elevated in dibutyryl cAMP-treated cells 24 h after treatment over those observed in unsupplemented cells. For the same cells, treatment with dibutyryl cAMP
also results in a time-dependent decrease in the cellular concentration of retinyl ester (Fig. 1, panel C). This
suggests that adipocyte retinyl esters are being hydrolyzed to retinol, thus giving rise to an enlarged cellular pool of retinol and a decreased pool of retinyl ester. This inverse trend between cellular retinol and retinyl ester levels upon treatment with dibutyryl cAMP is
also seen in Fig. 2. Fig. 2 shows the effects of
dibutyryl cAMP concentration on adipocyte retinol and retinyl ester
levels and on media retinol levels. As can be seen from this figure, there is a trend such that cellular retinol concentrations are relatively high and retinyl ester levels relatively low for dibutyryl cAMP concentrations of 1 and 5 mM. This is not the case in
the absence of dibutyryl cAMP addition, where retinol concentrations tend to be relatively low and retinyl ester levels are relatively high.
The dibutyryl cAMP effect on retinol accumulation in the adipocyte
culture media was also concentration-dependent. As seen in Fig. 2, for
retinol-loaded BFC-1 adipocytes which were exposed for 24 h to
0, 0.1, 1, or 5 mM dibutyryl cAMP, retinol levels in the
media of cells treated with 1 and 5 mM dibutyryl cAMP were approximately 40-50% greater than those of untreated cells or cells
treated with 0.1 mM dibutyryl cAMP. Similarly, media free fatty acid levels were found to be highest for adipocytes cultured in 1 and 5 mM dibutyryl cAMP (data not shown).
Since these data indicate that dibutyryl cAMP brings about an increase
in intracellular retinol levels at the expense of intracellular retinyl
ester, we asked whether specific retinyl esters were being preferentially hydrolyzed upon dibutyryl cAMP treatment of the adipocytes. Retinyl linoleate, palmitate, and oleate account for greater than 85% of the retinyl esters present in adipocytes, we
determined the levels of these retinyl esters 0, 8, 24, 48, and 72 h after dibutyryl cAMP treatment and, similarly, in untreated adipocytes. The results of these determinations are shown in Fig. 3. With increasing time after dibutyryl cAMP exposure,
less retinyl linoleate and retinyl palmitate are present in dibutyryl
cAMP-treated cells while retinyl oleate levels remain relatively
constant. The differences between retinyl linoleate and retinyl
palmitate present in treated and untreated adipocytes are highly
significant (p < 0.01) 48 h after the change of
media. This suggests that the increased cellular retinol concentration
observed upon dibutyryl cAMP treatment of the BFC-1 adipocytes
arises primarily at the expense of intracellular retinyl linoleate and
retinyl palmitate.
In other experiments, we asked whether the stimulatory effect of cAMP on retinol efflux from adipocytes was generalized to other cell types. Since the hepatocyte is the major cellular site for the mobilization of retinol from the liver, we investigated the effects of dibutyryl cAMP treatment on retinol efflux from primary cultures of mouse hepatocytes. Hepatocytes were isolated from 3-month-old female wild-type mice maintained on a chow diet. Immediately after isolation, the hepatocytes were allowed to stabilize on plastic Petri dishes overnight prior to use for our studies. (Since these primary hepatocytes contained endogenous retinol and retinyl esters it was not necessary to preload the hepatocytes with retinol.) The hepatocyte cultures were exposed to standard media (see "Experimental Procedures") either supplemented with 1 mM dibutyryl cAMP or unsupplemented for 4 time intervals of up to 24 h. We observed no differences in cellular or media retinol concentrations between the dibutyryl cAMP supplemented and unsupplemented hepatocyte cultures for any time interval examined (data not shown). These data from primary mouse hepatocytes indicate that the observed cAMP effect on retinol efflux from adipocytes is not generalizable to all retinol stores within the body.
Since the increase in cellular retinol was accompanied by a decrease in
retinyl ester levels, we investigated whether hydrolytic activities
present in BFC-1 adipocytes are able to hydrolyze retinyl esters. We
first prepared an acid precipitate, pH 5.2, of proteins present in a
BFC-1
adipocyte crude homogenate prepared from cells. This acid
precipitation is a standardly used first step in the purification of
HSL from adipocytes (26). Addition of aliquots of the resuspended
BFC-1
, pH 5.2, precipitate to emulsions containing either 1.25 mM [3H]triolein or 1.25 mM
[3H]retinyl palmitate (final concentrations) resulted in
time-dependent hydrolysis of the triolein and the retinyl
palmitate. This indicates that retinyl esters are either a substrate
for HSL or for another "HSL-like" activity present in
adipocytes.
To confirm that HSL can catalyze retinyl ester hydrolysis, we generated
a cell line which overexpresses HSL and asked whether retinyl palmitate
is hydrolyzed by homogenates prepared from these cells. The cells used
for these studies (termed 3F9 cells) were generated by stably
transfecting CHO cells with a cDNA clone for rat HSL (see
"Experimental Procedures"). As seen in Fig. 4, 3F9 cells express high levels of HSL mRNA (panel A) and
contain high levels of immunoreactive rat HSL (panel B).
Moreover, homogenates prepared from 3F9 cells actively catalyze the
hydrolysis of cholesteryl oleate, whereas homogenates from
sham-transfected CHO cells do not (panel C). As seen in Fig.
4, panel C, 3F9 cell homogenates possess a specific activity
toward cholesteryl oleate hydrolysis which is over 600-fold greater
than that of homogenates prepared from sham-transfected control CHO
cells. As can be seen from Fig. 5, 3F9 cell homogenate
is also able to catalyze the hydrolysis of retinyl palmitate; whereas
homogenate from sham-transfected control CHO cells catalyzes this
hydrolysis at a very slow rate. Hydrolysis of retinyl palmitate by 3F9
cell homogenate was found to occur in a time-, protein-, and retinyl
palmitate-dependent manner. The apparent
Km of the activity present in 3F9 cell homogenate
under our assay conditions for retinyl palmitate was determined to be
161 µM. Measures of specific activities of 3F9 cell
homogenates for triolein and retinyl palmitate indicated that HSL
catalyzes the hydrolysis of triacylglycerols at a rate which is
approximately 8.5-fold greater than the corresponding rate for retinyl
esters.
We further questioned whether HSL will catalyze retinyl ester hydrolysis under conditions which resemble those found in adipocytes. For this purpose, an emulsion containing both [14C]triolein, at a final assay concentration of 20 mM, and unlabeled retinyl palmitate, at a final assay concentration of 10 µM, was prepared and incubated with HSL expressed in 3F9 cells for 90 min at 37 °C. Under these conditions, where triolein and retinyl palmitate concentrations resemble those which might be found in adipocytes, some retinyl palmitate was hydrolyzed. Approximately 3% of the triolein fatty acyl groups present in the assay mixture and 0.4% of the added retinyl palmitate were hydrolyzed by the exogeneously added HSL.
Since all retinol in the circulation is found bound to RBP (7), we also
investigated whether RBP expression and/or RBP secretion from BFC-1
adipocytes are affected by dibutyryl cAMP treatment. Essentially, we
were interested in understanding whether the increased concentrations
of retinol observed in the media of dibutyryl cAMP-treated adipocytes
resulted from increased synthesis and/or secretion of retinol-RBP from
the adipocytes. Possible effects of dibutyryl cAMP treatment on
BFC-1
adipocyte RBP mRNA expression were examined by Northern
blot analysis. As shown in Fig. 6, dibutyryl cAMP
supplementation for time intervals of up to 72 h resulted in no
change in adipocyte RBP mRNA levels as compared with unsupplemented
cells. Media RBP levels also were measured by radioimmunoassay for
cAMP-treated and untreated adipocytes. The mean rate of RBP
accumulation in media from adipocytes supplemented with 1 mM dibutyryl cAMP, for up to 72 h, was 57.1 ± 18.7 ng of RBP/106 cells/24 h (n = 6) as
compared with 63.6 ± 23.8 ng of RBP/106 cells/24 h
(n = 6) for unsupplemented cells. These rates of RBP accumulation in the media are not statistically different. Hence, dibutyryl cAMP treatment does not influence either RBP synthesis or
secretion from cultured BFC-1
adipocytes.
The overall goal of our investigations was to explore the mechanisms responsible for retinoid mobilization from adipocyte stores. It is well established that adipocytes possess high levels of both retinol and retinyl esters and express proteins which are essential for retinol mobilization (RBP) and metabolism (cellular retinol-binding protein) (1-4, 6, 7). However, there are no data available regarding either how retinoid is mobilized from adipocytes or the factors which regulate mobilization. In the absence of such data, we reasoned that adipocytes either must mobilize retinol through processes which are very similar to those used for fatty acid mobilization or, alternatively, through processes different from those employed in fatty acid mobilization. Since much is known regarding fatty acid mobilization and its regulation in adipocytes, we chose to examine first whether retinol mobilization from adipocytes involves processes and factors which are important for fatty acid mobilization. Intracellular cAMP concentration is a key regulator of fatty acid levels and efflux from adipocytes; hence, we asked whether cAMP concentration also influences adipocyte retinol and retinyl ester levels and retinol efflux from adipocytes. Our data demonstrate that treatment of adipocytes with cAMP markedly increases retinol levels and at the same time decreases retinyl ester levels within adipocytes. This suggests to us that cAMP is able to stimulate retinyl ester hydrolysis to retinol within adipocytes. Furthermore, the increase in cellular retinol concentration is accompanied by increased retinol efflux from the cells; suggesting that the hydrolysis of retinyl esters is important to the process of retinol efflux from adipocytes. For none of our experiments were retinyl esters observed in the adipocyte culture medium; implying that retinol is the sole form of retinoid mobilized from adipocytes. We conclude from these data that retinol mobilization from adipocytes is mediated through cAMP responsive pathways and that, in this regard, retinol mobilization from adipocytes is very similar to fatty acid mobilization.
Dibutyryl cAMP treatment of adipocytes brought about increased efflux of retinol into the media but no change to either adipocyte RBP mRNA expression or media RBP levels. The amount of retinol released into the media of dibutyryl cAMP-treated adipocytes was in great molar excess (>100-fold) over the amount of RBP accumulated in the media. Hence, it is not possible to account for the elevated media retinol levels through enhanced secretion of retinol bound to RBP. This raises a question as to how retinol is released from adipocytes unaccompanied by RBP. The differentiation media does contain fetal bovine serum and consequently two bovine proteins, albumin and RBP, which are able to bind retinol with high affinity. It is likely that retinol is drawn from the adipocytes into the media by the presence of bovine albumin and RBP in the media. These proteins provide acceptor sites for retinol effluxing from the adipocytes into the media. This possibility is supported by our observation that dibutyryl cAMP-treated adipocytes must be exposed to media containing fetal bovine serum in order for retinol efflux to occur (data not shown). When we exposed retinol-loaded adipocytes to differentiation medium supplemented with dibutyryl cAMP but totally lacking fetal bovine serum for periods of up to 24 h, we observed no accumulation of retinol in the media but cellular retinol levels were much higher than those measured for cAMP-supplemented cells maintained in media containing fetal bovine serum. In the literature, it has been proposed that both retinol entry into cells and efflux from cells are dependent on the relative concentrations of high affinity binding (acceptor) sites for retinol available on the two sides of the cell's plasma membrane (7, 31, 32). Reconciling our data with this model for cellular retinol uptake and efflux, it would seem reasonable that dibutyryl cAMP directly acts to stimulate the hydrolysis of adipocyte retinyl ester to retinol. The intracellular retinol thus formed soon saturates the available high affinity binding sites for retinol within the adipocytes (i.e. cellular retinol-binding protein). The presence of available binding sites for retinol in the differentiation media draws the newly formed retinol from the adipocytes into the media. It also should be pointed out that other investigators, exploring retinol efflux from cultured chicken Müller cells, observed similar effects of fetal bovine serum on retinol efflux from these cells (33). Like our findings, these investigations with Müller cells showed that culture media supplemented with fetal bovine serum was effective, in a protein- and time-dependent manner, for drawing retinol from the Müller cells into the culture media (33).
We also have demonstrated that retinyl palmitate is a substrate for HSL in vitro. It is well established that hormone-sensitive lipase has a broad substrate specificity which includes cholesteryl esters and steroid esters, in addition to tri- and diacylglycerols (26, 34-38) but it had not been recognized previously that HSL catalyzes the hydrolysis of retinyl esters. Our studies of retinyl ester hydrolysis by homogenates prepared from CHO cells which overexpress HSL (Figs. 4 and 5) show convincingly that HSL is able to catalyze in vitro the hydrolysis of retinyl esters. From measures of the specific activities of 3F9 CHO cell homogenates toward triolein and retinyl palmitate, it is clear though that triolein is a much better substrate for HSL than is retinyl palmitate.
In summary, our studies demonstrate that intracellular cAMP concentrations influence retinol efflux from adipocytes and they document that HSL has retinyl ester hydrolase activity. This observation that HSL catalyzes retinyl ester hydrolysis places HSL together with other neutral lipid hydrolases like lipoprotein lipase and cholesteryl ester lipase which are also retinyl ester hydrolases. The actions of cAMP on adipocyte retinyl ester hydrolysis and retinol efflux appear to be specific to the adipocyte since cAMP treatment of primary mouse hepatocytes does not give rise to either increased efflux of retinol from the cells or increased cellular retinol at the expense of cellular retinyl ester levels.
We acknowledge the excellent technical support of Tres Mertz.