Cellular uptake of circulating retinol

Because of the disruptive effects of free (uncomplexed)
retinol on membrane structure and function
(Roels et al., 1969), it is desirable to discourage
non-specifi c uptake of retinol by cells that have no
particular requirement for it. This might be achieved
by means of specifi c RBP receptors on the plasma
membrane of vitamin A-requiring cells, which recognize
holoRBP, bind it, and facilitate the transfer of
holoRBP or free retinol across the membrane and into
the cytoplasm.
RBP binding to cell membranes has been reported
in the following specifi c locations: the choroidal
surface of retinal pigment epithelial cells, but not in
Vitamin A: retinoids and carotenoids 147
rod outer segments (Heller, 1975); interstitial cells
of testes, but not on epithelial cells of seminiferous
tubules (McGuire et al., 1981); human placental cells
(Sivaprasadarao & Findlay, 1988a) and epithelial cells
of the choroid plexus, signifying movement of retinol
across the blood–brain barrier (MacDonald et al.,
1990). Evidence of RBP receptor-mediated uptake of
retinol has also been reported in enterocytes of the
small intestine (Rask & Peterson, 1976) and testicular
Sertoli cells (Shingleton et al., 1989). Among these locations,
cellular uptake of retinol without a concomitant
uptake of the RBP has been reported to take place
in retinal pigment epithelium (Chen & Heller, 1977),
enterocytes (Rask & Peterson, 1976) and Sertoli cells
(Shingleton et al., 1989). The latter authors found that
the amount of retinol accumulated by cultured Sertoli
cells from holoRBP was approximately equal to the
cellular content of CRBP-I. This implied that the ligand
saturation of CRBP-I may be the factor regulating
the cellular uptake of retinol.
There appears to be two entirely different mechanisms
for receptor-mediated uptake of circulating
retinol by the cells of target tissues. The fi rst mechanism
applies to the tissues mentioned above, at least
some of which have been shown to take up retinol unaccompanied
by RBP. Upon recognition and binding
by a cell-surface receptor, holoRBP releases its retinol
molecule and, as a result, undergoes a conformational
change which reduces its affi nity for both the receptor
and transthyretin. The transthyretin portion of
the holoRBP–transthyretin complex is not involved
in the recognition process. Indeed, transthyretin has
been found to inhibit the binding of RBP to plasma
membranes in vitro (Sivaprasadarao & Findlay,
1988a). Retinol is internalized into the cytoplasm
where it interacts with its specifi c cytoplasmic binding
protein, CRBP-I. The low-affi nity form of apoRBP
is circulated to the kidneys where it is fi ltered at the
glomerulus, endocytosed in the cells of the proximal
convoluted tubules, and degraded in lysosomes
within the tubular cells.
Sivaprasadarao & Findlay (1988b) suggested that
the ratio of apo- to holoRBP levels in the plasma can
regulate retinol distribution among various tissues.
Vitamin A defi ciency can effect a change in this ratio:
holoRBP levels decrease to almost zero while apoRBP
levels remain unchanged. Since apoRBP has poor
affi nity for transthyretin, most of it will exist in the
free state. As both holo- and apoRBP bind to the cellsurface
receptor, the high ratio of apo- to holoRBP in
the plasma would not only result in the decreased uptake
of retinol by most tissues, but actually stimulate
the secretion of retinol by extrahepatic tissues. The
retinol so liberated might then be used for critically
dependent tissues such as the eye and gonads.
The second mechanism of retinol uptake, exemplifi
ed in the liver and kidney, also involves the interaction
of holoRBP with a specifi c cell-surface receptor,
but the entire retinol–RBP complex is internalized by
receptor-mediated endocytosis. Evidence for this uptake
mechanism was provided by Senoo et al. (1990),
who studied the in vivo uptake of RBP in rat liver cells
by immunocytochemistry at the electron microscopic
level using ultra-thin cryosections. Rats were injected
intravenously with human RBP and simultaneously,
for comparison, with asialo-orosomucoid, a protein
known to be taken up by hepatocytes by receptor-mediated
endocytosis. The native, unmodifi ed RBP was
subsequently identifi ed in the liver sections by a sheep
anti-human RBP antibody, which does not recognize
endogenous rat RBP. Ten minutes after injection, RBP
was found in close contact with the plasma membrane
of parenchymal and stellate cells. Ten minutes later,
RBP was also found attached to the membranes of
small vesicles near the cell surface. At 2 hours after
injection, RBP was detected in larger vesicles deeper
in the cytoplasm, remote from the vesicle membranes
as though dissociated from its putative receptor.
Asialo-orosomucoid was also localized in these larger
vesicles.
Blomhoff’s research group (Gjøen et al., 1987)
studied the uptake of RBP in various organs at different
times. RBP was labelled with 125I-tyramine cellobiose
(125I-TC-RBP) and injected intravenously into
rats. (The advantage of using 125I-TC-RBP is that the
radioactive degradation products do not escape from
the cells in which the protein is degraded, as is the
case for proteins radioiodinated directly.) Of all the
organs tested (liver, kidneys, intestine, spleen, heart,
lungs, etc.) the liver contained the most radioactivity,
followed by the kidneys. After 1 hour, approximately
20% and 10% of the injected dose was recovered in
the liver and kidneys, respectively; other organs contained
less than 3%. Of the liver cell types, parenchymal
and stellate cells took up about equal amounts
of 125I-TC-RBP when calculated per gram of liver;
Kupffer cells and endothelial cells accumulated insignifi
cant amounts. The liver was shown to be the main
148 Vitamins: their role in the human body
organ for tissue catabolism of plasma RBP. The same
laboratory (Senoo et al., 1990) labelled RBP and several
other proteins with 125I-TC and compared their
in vivo uptake in different rat liver cell types. RBP was
the only protein that was taken up selectively by parenchymal
and stellate cells. The data from these two
studies suggest that hepatic parenchymal and stellate
cells and cells in kidneys contain receptors for RBP,
and that RBP is internalized in these cells, probably by
receptor-mediated endocytosis.
Proof of an RBP receptor-mediated uptake of circulating
retinol requires isolation and characterization
of an RBP receptor. An abundant 63-kDa terminally
glycosylated membrane protein which specifi cally
binds RBP has been identifi ed as the RBP receptor in
microsomal membranes of retinal pigment epithelial
cells (Båvik et al., 1991) and further characterized
(Båvik et al., 1992, 1993). The receptor cannot discriminate
between the apo- and the holo-form of
RBP. Particularly large numbers of receptor-binding
sites were also found in microsome fractions of liver
and kidney, whereas lung and muscle contained few,
if any. The abundance of RBP receptors in retinal pigment
epithelium and in liver and kidney cells is in
accordance with the responsibility of these cells for
the uptake and transport of large amounts of retinol
for functional purposes. In contrast, it is expected that
only relatively small amounts of retinol are needed in
cells where accumulated retinol is used mainly for the
synthesis of retinoic acid.