Genes encoding the cellular retinoid-binding

CRBP-I
Several whole animal studies have examined the effects
of oral administration of retinyl acetate or retinoic
acid on CRBP-I mRNA levels in various tissues.
Rajan et al. (1990) reported that the levels of CRBP-I
mRNA in certain tissues (lung, testis, spleen and small
intestine) of vitamin A-defi cient rats were lower than
those in normally-fed rats. Oral repletion with retinyl
acetate restored the levels to normal. Retinoid defi -
ciency did not affect the levels of CRBP-I mRNA in
the three tissues with the highest content of CRBP-I
protein and CRBP-I mRNA (proximal epididymis,
liver and kidney). Haq & Chytil (1988) showed that
administration of retinoic acid to vitamin A-defi cient
rats elicited a more rapid response than observed with
retinyl acetate, with a two- to three-fold increase in
CRBP-I mRNA in lung tissue within 1 hour. Increased
expression of CRBP-I was also demonstrated in the
lung tissue of normally-fed, vitamin A-replete rats
after oral administration of retinoic acid (Rush et al.,
1991). Consistent with these studies was the demonstration
that topical application of retinol or retinoic
acid to adult human skin led to increased levels of
CRBP-I mRNA and protein (Fisher et al., 1995). These
experiments indicate that, in certain tissues at least,
dietary retinol (or more directly retinoic acid) directly
induces the expression of the CRBP-I gene. Thus organs
in which CRBP-I expression is depressed during
vitamin A defi ciency are able to respond rapidly to
retinoids as soon as they become available.
The expression of CRBP-I can be modulated in
the vitamin A-replete animal by glucocorticoid hormones.
The administration of dexamethasone to
vitamin A-replete rats decreased the levels of CRBP-I
mRNA levels in lung and liver tissues. In addition,
the increase in CRBP-I mRNA induced by oral administration
of retinoic acid was blocked when dexamethasone
was administered at the same time (Rush
et al., 1991). This effect may be maintained by putative
glucocorticoid-response elements identifi ed in the
promoter of the CRBP-I gene.
The hormonal changes that occur during pregnancy
and lactation also affect the expression of CRBP-I.
Towards the end of pregnancy in the rat, the level of
CRBP-I mRNA in the maternal liver rises markedly
then drops abruptly at term (Levin et al., 1987). This
event coincides with a signifi cant mobilization of hepatic
vitamin A stores so that the milk of the lactating
rat can meet the vitamin A needs of the suckling pups.
Wei et al. (1989) demonstrated that treatment of
tissue cultures with retinoic acid induced an increase
in CRBP-I mRNA followed by the protein itself. The
3-hour lag period and the nondependence on concomitant
protein synthesis identifi ed the CRBP-I
gene as a primary response gene. CRBP-I induction
occurred both at 10–7 and 10–9 M retinoic acid in EC
cells that differentiate into neurons and extra-embryonic
endoderm and also in a non-EC cell that differentiates
into glial cells.
The rapid synthesis of the CRBP-I protein in response
to retinoic acid, observed for the whole animal
as well as in cultured cells, suggested that the CRBP-I
gene would contain a RARE. Such a response element
has been demonstrated in the mouse (Smith et al.,
1991) and the rat (Husmann et al., 1992). In both
species, the RARE is located 1 kilobase upstream
of the transcription start site. The RARE in the rat
CRBP-I gene is a DR2 type element and is activated
in vitro by the RAR–RXR heterodimer. The presence
of a RARE in the CRBP-I gene indicates that CRBP-I
protein levels are regulated primarily at the level of
transcription.
162 Vitamins: their role in the human body
CRBP-II
In the adult, expression of the CRBP-II gene is essentially
confi ned to the small intestinal enterocytes.
In contrast to CRBP-I, whose expression is decreased
in various organs during vitamin A defi ciency, expression
of intestinal CRBP-II is actually increased.
Rajan et al. (1990) reported an increase in the CRBPII
mRNA level of 42% in the small intestine of the
vitamin A-defi cient rat. The contrasting expression
profi les of these two similar proteins refl ect the
physiological needs of the animal. During vitamin A
defi ciency, the animal conserves its precious supply
of retinoids to maintain only essential functions, allowing
the animal to hunt for food. In addition, the
intestine is kept at the ready to receive dietary vitamin
A as soon as the animal fi nds it.
During the last 5 days of pregnancy in the rat,
mRNA for CRBP-II in the small intestine shows a
four-fold increase, accompanying the increase observed
for CRBP-I in the maternal liver. Intestinal
CRBP-II levels remain elevated during the suckling
period and return to normal by 1 week post-weaning
(Quick & Ong, 1989). This increase in production of
intestinal CRBP-II, like that of hepatic CRBP-I, is an
adaptive response, ensuring that suffi cient vitamin A
reaches the mammary gland during lactation.
Nakshrati & Chambon (1994) showed that the
mouse CRBP-II gene contains three response elements,
which they designated RE1, RE2 and RE3. The
elements RE2 and RE3, which have DR2 and DR1
sequence motifs, respectively, are putative RAREs
and are conserved in the rat and mouse CRBP-II gene
5´-fl anking regions. The more distal RE1 element,
which also has a DR1 motif, is a truncated form of an
element previously found in the rat and designated
an RXRE by Mangelsdorf et al. (1991). Nakshatri &
Chambon (1994) showed that RE2 and RE3 were
required for maximal RAR–RXR-mediated retinoic
acid inducibility to the mouse CRBP-II promoter
in transfection experiments. RE2 had no effect on
its own, but co-operated with RE3. RE1 was not
involved in the induction; if anything, it exerted a
slight inhibitory effect. Thus RE3 appeared to be the
only element within the mouse CRBP-II promoter
which responded on its own to RAR and RXR. The
RE3 and RE1 response elements have a much higher
affi nity for HNF-4 and ARP-1 than for RAR–RXR heterodimers
and RXR homodimers. HNF-4 and ARP-1
are members of the nuclear receptor superfamily and
are expressed in liver and intestinal cells. Nakshatri &
Chambon (1994) showed that HNF-4 constitutively
activates the mouse CRBP-II promoter, whereas ARP-
1 competitively represses its activation by RAR, RXR
and HNF-4. Although the data reported by Nakshatri
& Chambon (1994) and also by Mangelsdorf et al.
(1991) show that the mouse and rat CRBP-II promoters
can respond to over-expressed RXR and RAR in
co-transfection experiments, for these results to be
physiologically relevant, it must be demonstrated that
expression of the CRBP-II gene can be stimulated by
retinoic acid treatment in vivo. There are no available
data concerning a possible retinoic acid induction of
CRBP-II mRNA or protein in adult intestine and in
prenatal liver or intestine, which are also known to
express the CRBP-II gene (Nakshatri & Chambon,
1994). It appears that HNF-4 is the major transcriptional
activator of the CRBP-II gene in vivo. Inducement
of the CRBP-II gene by retinoic acid may only
take place in tissues where HNF-4 and ARP-1 are
lacking or present in low amounts.
CRABP-I
CRABP-I is ubiquitously expressed in adult animal
tissues at a low basal level, except in several retinoic
acid-sensitive tissues such as the eye and the testis
where the expression is highly elevated. In adult rats,
neither CRABP-I protein nor mRNA levels were affected
by dietary vitamin A defi ciency (Rajan et al.,
1991), suggesting that the binding protein is not
regulated by overall vitamin A status. CRABP-I is also
constitutively expressed at a required level in certain
regions of the developing embryo and at certain times
of development. In the developing mouse, CRABP-I
mRNA has been detected in the neural crest, the eye
region and the craniofacial region (Perez-Castro et al.,
1989) and in rudimentary limbs (Dollé et al., 1989).
Wei et al. (1989), using tissue cultures, showed
that expression of the CRABP-I gene was induced by
10–7 M (but not 10–9 M) retinoic acid in a manner that
required the synthesis of new protein. A detectable
increase in CRABP-I mRNA was not observed until
12 hours after retinoic acid treatment. These observations,
and the absence of a RARE with any known
consensus sequence, place the CRABP-I gene in the
category of secondary response genes. Induction occurred
in cells, which differentiate into neurons, but
not in cells which differentiate into extra-embryonic
endoderm or glial cells.
Vitamin A: retinoids and carotenoids 163
CRABP-II
Transcription of the human CRABP-II gene was
markedly induced by retinoic acid in skin in vivo and
in skin fi broblasts in vitro. Retinoic acid had no such
effect on cultured lung fi broblasts, demonstrating
cell-specifi c regulation of the CRABP-II gene (Åström
et al., 1991). In the cultured skin fi broblasts, the increase
in CRABP-II mRNA was detected 1 hour after
addition of retinoic acid. The level peaked at 2 hours
and returned to basal levels within 6 hours. Ongoing
protein synthesis was required for this transient
increase of transcription (Åström et al., 1992). The
human CRABP-II contains one far upstream RARE of
the DR5 type that binds the RAR–RXR heterodimer
(Åström et al., 1994). The binding of receptor to the
response element and the early response are consistent
with CRABP-II being a primary response gene, but
the ongoing protein synthesis is inconsistent with this
defi nition. Interestingly, the mouse CRABP-II gene
contains two co-operating response elements, a DR2
RARE and a DR1 RXRE, mediating differential induction
of transcription in response to both all-trans and
9-cis retinoic acid (Durand et al., 1992).