Gene regulation

The ligand-activated VDR may function to recruit
coactivators that remodel chromatin structure and
permit greater accessibility of the transcriptional machinery
to DNA. Coactivators shown to interact with
liganded VDR include mouse SUG1 (vom Baur et al.,
1996), RAC3 (Li et al., 1997), NCoA-62 (Baudino et
al., 1998; MacDonald et al., 2001) and a multiprotein
complex known as DRIP (Rachez et al., 1998, 1999).
There have been no reports as yet of co-repressor proteins
which bind to the VDR, so it is presently unclear
whether the VDR interacts with co-repressors to bring
about down-regulation of target genes. At least one
co-repressor, N-CoR, fails to interact with the VDR
(Hörlein et al., 1995). Determination of whether vitamin
D hormone action is stimulatory or inhibitory
may be tissue-specifi c or dependent on the state of
cellular differentiation. Most vitamin D-responsive
genes are up-regulated by 1α,25(OH)2D3, the human
PTH gene being one of the few that is down-regulated
(see Table 8.2).
Repression of vitamin D-induced transactivation
of the osteocalcin gene by YY1
Overlapping the recognition sequences for VDR–RXR
within the VDRE of the osteocalcin gene are two binding
motifs for the multifunctional transcription factor
YY1. This protein can either activate, repress or initiate
gene transcription (Shrivastava & Calame, 1994). In
the context of the osteocalcin promoter, YY1 represses
1α,25(OH)2D3-induced transcription by competing
with VDR–RXR for binding at the VDRE. YYI also interacts
directly with TFIIB thereby interfering with the
transactivation function of the VDR (Guo et al., 1997).
This repressive function of YY1 would prevent the precocious
induction of osteocalcin gene transcription by
VDR-mediated mechanisms under physiological conditions
in which the gene should not be expressed.
Suppression of osteocalcin gene expression
during osteoblast proliferation
The target cell for the hormonal action of vitamin
D on bone is the osteoblast. Within this cell,
1α,25(OH)2D3 enhances osteocalcin gene expression
at the three principal levels – transcription, mRNA
accumulation and protein synthesis. The enhanced
transcription is dependent upon basal levels of gene
expression (Owen et al., 1991), suggesting that there
is a coordinate transactivation involving the contribu-
Vitamin D 203
tion of activities at the VDRE and basal elements. Two
essential basal elements in the proximal promoter are
the TATA box and the osteocalcin box (OC box), a 24-
nucleotide sequence that contains a central CCAAT
motif (Lian & Stein, 1992). The VDRE is located further
upstream between nucleotides –512 and –485 in
the human osteocalcin gene promoter (Ozono et al.,
1990).
The sequential expression of vitamin D-dependent
genes associated with bone tissue development
has been studied in cultures of normal diploid rat
osteoblasts. The fi rst 10 to 12 days constitute the cell
proliferation period, characterized by the expression
of genes encoding AP-1 proteins (Jun and Fos) and
extracellular matrix proteins. The jun and fos genes
are transcribed in response to growth factors and
other cell-surface stimuli, and the protein products
are responsible for converting the transient stimuli
into a long-term transcriptional response. During the
following stages, the proliferative activity declines and
the genes encoding the AP-1 proteins and extracellular
matrix proteins are gradually down-regulated.
During the next stage of matrix maturation (days
12 through 18), alkaline phosphatase mRNA and
enzyme activity increases more than ten-fold to peak
levels, and matrix Gla protein is expressed maximally.
Days 16 through 20 are characterized by the progressive
mineralization of the extracellular matrix. As the
cellular levels of alkaline phosphatase mRNA decline,
the accumulation of calcium in the matrix increases
coordinately with the up-regulated expression of
genes encoding the calcium-binding proteins, osteocalcin
and osteopontin (Stein & Lian, 1993).
The modular organization of the human osteocalcin
gene promoter explains how gene expression
might be suppressed during cell proliferation. Overlapping
the VDRE and also within the OC box are
AP-1 binding sites which bind Jun–Fos heterodimers
and Jun homodimers. During the proliferation
of osteoblasts, the Jun and Fos proteins are present
at high levels. Occupancy of the AP-1 binding site
overlapping the VDRE by Jun and Fos dimers blocks
the binding of liganded VDR to its specifi c site and
prevents transcription of the osteocalcin gene. At the
end of the proliferation period, the levels of Fos and
Jun decline, allowing the VDR to bind to the VDRE
and the osteocalcin gene to be expressed. Thus the
competition of the VDR and the AP-1 proteins for
binding to the composite DNA element determines
the activation or suppression of the osteocalcin gene
(Lian & Stein, 1992). High levels of Jun and Fos also
suppress basal transcription of the osteocalcin gene
(Schüle et al., 1990).
Inhibitory effect of the VDR on transactivation of
the growth hormone gene by thyroid hormone
and retinoic acid
The rat growth hormone gene, located exclusively in
the somatotropic pituitary cells, contains a hormone
response element that functions both as a retinoic
acid response element (RARE) and a thyroid hormone
response element (TRE) (García-Villalba et al.,
1993). This allows thyroid hormone and retinoic acid
to interact co-operatively to stimulate transcription
of the growth hormone gene in pituitary cells (Bedo
et al., 1989). García-Villalba et al. (1996) reported that
incubation of rat pituitary cells with nanomolar concentrations
of vitamin D3 inhibits thyroid hormone
and retinoic acid transactivation of the growth hormone
gene by interference on the common response
element. The results suggested that the liganded VDR
can directly affect the pituitary response to other nuclear
receptors, thereby contributing to the growth
arrest that occurs in hypervitaminosis D. This negative
effect on the growth hormone gene is apparently
paradoxical, since a defi ciency of vitamin D produces
a rachitic state associated with a defect in growth.
Repression of VDR-mediated transcription of the
osteocalcin and osteopontin genes by the thyroid
hormone receptor
Thompson et al. (1999) demonstrated two distinct
repressive actions of the TR on VDR-mediated
transcription of the rat osteocalcin and mouse osteopontin
genes: (1) a thyroid hormone-independent
action, perhaps due to TR–RXR out-competing
VDR–RXR for binding to the VDREs and (2) a
thyroid hormone-dependent repression, likely by diversion
of limiting RXR from VDR–RXR toward the
formation of TR–RXR heterodimers. The reverse of
this phenomenon was indicated by a relatively weak
binding of VDR–RXR to the TRE of the myosin heavy
chain gene and modest repression by liganded VDR
of thyroid hormone-mediated transactivation. This
reciprocal inhibition of transactivation by VDR and
TR is permitted by the half-site homology between
the rat osteocalcin VDRE and the rat myosin heavy
chain TRE. The formation of a TR–VDR heterodimer
204 Vitamins: their role in the human body
reported by Schräder et al. (1994) was not observed
by Thompson et al. (1999) who concluded that this
heterodimer is not biochemically or physiologically
relevant. As far as is known, the only relevant heterodimerization
among nuclear receptors (excluding
RXR) is between the type I mineralocorticoid and
glucocorticoid receptors.
8.5.3 Nongenomic actions of 1α,25-
dihydroxyvitamin D3 and 24R ,25-
dihydroxyvitamin D3
Background information can be found in Sections
3.7.6 and 3.7.9.
A large number of responses occur within seconds
to minutes following addition of vitamin D metabolites
to the in vitro system – too rapid to involve
changes in gene expression controlled by the VDR.
Moreover, such responses are not blocked by inhibitors
of transcription (actinomycin D) or protein synthesis
(cycloheximide). For these reasons, the rapid
responses are described as nongenomic.
1α,25-Dihydroxyvitamin D3
Rapid nongenomic responses to 1α,25(OH)2D3 have
been observed in many cell types, including enterocytes,
colonocytes, chondrocytes, osteoblasts, hepatocytes,
skeletal and cardiac muscle cells, keratinocytes,
mammary gland epithelial cells, parathyroid cells
and pituitary cells (Norman, 1998). The responses
have been most clearly delineated in osteoblast-like
osteosarcoma cells. At a molecular level, these include
effects on membrane phospholipid metabolism
(Grosse et al., 1993), activation of voltage-sensitive
calcium channels (Caffrey & Farach-Carson, 1989)
and elevation of cytosolic (Baran et al., 1991) and nuclear
(Sorensen et al., 1993) Ca2+ concentrations.
A number of nongenomic effects of 1α,25(OH)2D3
are concerned with the regulation of intracellular
calcium, which is an important second messenger
involved in the activation of many target enzymes.
Additional effects include stimulation of prostaglandin
production (Boyan et al., 1994), increased
intra cellular pH (Jenis et al., 1993) and increased
membrane fl uidity (Brasitus et al., 1986).
Regulation of intracellular calcium
Transient increases in intracellular calcium can be
initiated in two major ways. (1) Extracellular calcium
can enter the cell down its electrochemical gradient
by the opening of voltage-gated calcium channels. (2)
Calcium can be released from intracellular storage
sites associated with mitochondria and the endoplasmic
reticulum.
The presence of L-type (dihydropyridine-sensitive)
voltage-activated calcium channels has been
demonstrated in basolateral membranes of rabbit
ileal enterocytes (Homaidan et al., 1989) and in osteosarcoma
cells (Guggino et al., 1989). The opening
of such channels is the fastest known response of
osteoblasts to treatment with nanomolar concentrations
of 1α,25(OH)2D3. The hormone enhances the
activity of voltage-gated channels in rat pituitary cells
(Tornquist & Tashjian, 1989) and rat osteosarcoma
cells (Caffrey & Farach-Carson, 1989).
The binding of 1α,25(OH)2D3 to a membrane
receptor in a wide range of cell types stimulates the
activity of phospholipase C, whose hydrolytic action
on membrane phosphoinositides results in the generation
of the two second messengers, diacylglycerol
and inositol triphosphate (see Fig. 3.31). Cell types
examined include osteosarcoma cells (Civitelli et al.,
1990), myoblasts (Morelli et al., 1993), enterocytes
(Lieberherr et al., 1989), colonocytes (Wali et al.,
1990), hepatocytes (Baran et al., 1988), parathyroid
cells (Bourdeau et al., 1990) and keratinocytes (Mac-
Laughlin et al., 1990). Diacylglycerol is an activator
of PKC, which is involved in a myriad of cellular
processes. Inositol triphosphate releases Ca2+ from
the intracellular storage sites, thereby increasing the
concentration of Ca2+ in the cytosol.
A rapid increase in calcium translocation, termed
transcaltachia, has been described using the perfused
chick duodenal loop (Nemere et al., 1984). The
involvement of 1α,25(OH)2D3 in the activation of
voltage-gated calcium channels appears to be an early
effect in transcaltachia. Introduction of a calcium
channel antagonist completely abolished the movement
of calcium, while a calcium channel agonist
mimicked the stimulatory response to 1α,25(OH)2D3
(de Boland et al., 1990). de Boland & Norman (1990)
provided evidence for the involvement of PKA and
PKC in transcaltachia. Forskolin and phorbol ester,
activators of PKA and PKC, respectively, stimulated
transcaltachia analogously to 1α,25(OH)2D3. In addition,
the transcaltachial response to 1α,25(OH)2D3
was respectively suppressed or abolished by inhibitors
of PKA and PKC. Collectively, these observations sug-
Vitamin D 205
gest that 1α,25(OH)2D3 binds to a membrane receptor
located in the basolateral membrane of the enterocyte
and signal transduction via G proteins, effector
proteins and second messengers leads to activation
of PKA and PKC. These kinases might stimulate calcium
infl ux via phosphylation-dependent activation
of voltage-gated calcium channels at the basolateral
membrane.
Evidence for a distinct membrane receptor for
1α,25(OH)2D3
There is strong evidence that the nongenomic responses
to 1α,25(OH)2D3 are mediated by a membrane
receptor that is biochemically different from the
nuclear VDR. Ligand specifi city for the rapid actions
is different from that for genomic responses (Farach-
Carson et al., 1991). Using osteoblasts from mice in
which the VDR gene had been genetically ablated,
Wali et al. (2003) showed that the 1α,25(OH)2D3-
induced rapid increases in intracellular calcium and
PKC activity are neither mediated, nor dependent
upon, a functional VDR.
A 66-kDa protein that binds vitamin D analogues
has been isolated from basolateral membranes of
chick intestinal epithelium (Nemere et al., 1994) and
from both plasma membranes and matrix vesicles
of rat chondrocytes (Nemere et al., 1998). Antibody
(Ab99) generated to a [3H]1α,25(OH)2D3 binding
protein isolated from the basolateral membrane of
chick intestinal epithelium blocked the rapid activation
of PKC by 1α,25(OH)2D3 in chondrocytes
(Nemere et al., 1998) and enterocytes (Nemere et al.,
2000). Slater et al. (1995) discovered that physiological
concentrations of 1α,25(OH)2D3 can directly activate
PKC in artifi cial membranes. This suggests that the
PKC protein itself can act as a membrane-associated
receptor for 1α,25(OH)2D3, providing an additional
signal transduction pathway to the well- established
route in which PKC is activated by diacylglycerol.
Baran et al. (2000) reported that annexin II, a 36-kDa
protein, can serve as a cell-surface receptor mediating
1α,25(OH)2D3-induced increase in intracellular
calcium in osteoblast-like ROS 24/1 cells that lack
the functional nuclear VDR. The rapid effects of
1α,25(OH)2D3 to increase intracellular calcium were
not observed in cells pre-treated with anti-annexin
II antibodies. It is likely that several membrane receptors
mediate the rapid actions of 1α,25(OH)2D3
(Brown et al., 1999).
Modulation of the genomic actions of
1α,25(OH)2D3 by nongenomic mechanisms
1β,25-Dihydroxyvitamin D3 does not interact with
the nuclear VDR and thus has no effect on basal gene
transcription. It does, however, inhibit nongenomic
effects by competing with 1α,25(OH)2D3 for the
membrane receptor. Baran et al. (1992) demonstrated
that, in osteosarcoma cells, the 1β epimer inhibits the
1α,25(OH)2D3-induced rapid rise in intracellular
calcium and accompanying increase in osteocalcin
gene transcription. The inhibition of transcription
occurred without interfering with the binding of the
1α,25(OH)2D3–VDR complex to the VDRE. Since
1β,25(OH)2D3 has no genomic effects on osteocalcin,
yet can block the transcription that accompanies the
nongenomic effect of 1α,25(OH)2D3, this study suggests
that the nongenomic effects of 1α,25(OH)2D3
can modulate the genomic actions of this hormone
in some way.
Khoury et al. (1994) showed that the 1α,25(OH)2D3-
induced transcription of osteocalcin and osteopontin
genes in osteosarcoma cells is independent of Ca2+
infl ux, suggesting that the nongenomic stimulation
of calcium channels by 1α,25(OH)2D3 is not required
for target gene activation. Jenis et al. (1993) discovered
that the nongenomic 1α,25(OH)2D3-induced
increase in intracellular pH is necessary for osteocalcin
and osteopontin expression in the osteoblast. The
effect upon pH appears to be regulated by the Na+/H+
antiporter, since the incubation of cells in a Na+-free
medium eliminated the pH effect and blocked the
hormone-induced increase in osteocalcin and osteopontin
mRNA steady-state levels.
The genomic response may be modulated by second
messengers generated at both the plasma and
nuclear membranes in response to the binding of
1α,25(OH)2D3 to membrane receptors (Baran &
Sorensen, 1994). 1α,25-Dihydroxyvitamin D3 increases
PKC activity through stimulation of plasma
membrane phosphoinositide and formation of diacylglycerol
(Wali et al., 1990). In renal epithelial cells,
1α,25(OH)2D3 specifi cally induces translocation of
PKCβ (but not PKCα) from the plasma membrane
to the nuclear membrane, an event which enhances
phosphorylation of nuclear proteins (Simboli-
Campbell et al., 1994). Phosphorylation of the nuclear
VDR by PKC has been shown to down-regulate
transcription (Hsieh et al., 1993). Sorensen & Baran
(1993) reported that 1α,25(OH)2D3 rapidly enhances
206 Vitamins: their role in the human body
phospholipase C activity in the nuclear membrane of
osteosarcoma cells, resulting in an increased level of
inositol triphosphate, which in turn releases calcium
from intranuclear storage sites.
24R,25-Dihydroxyvitamin D3
The importance of 24R,25(OH)2D3 in endochondral
bone formation is indicated by its accumulation in
growth plate cartilage when normal rats are injected
with [3H]25(OH)D3 (Seo et al., 1996).
Effect on calcium current in osteosarcoma cells
Li et al. (1996) demonstrated a dual nongenomic effect
of 24R,25(OH)2D3 on the L-type calcium channel
current in osteosarcoma cells. At a low physiological
concentration (1 × 10–8 M), 24R,25(OH)2D3 activated
the PKA signal pathway leading to an increase
in current amplitude, whereas a higher concentration
(1 × 10–5 M) reduced the current amplitude via the
PKC signal pathway. In comparison, a high concentration
of 1α,25(OH)2D3 (1 × 10–6 M) increased the
current amplitude.