| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Department of Pharmacology (J.M.F., C.A.P.) and Department of Radiology (R.M.K.), New York Medical College, Valhalla, New York 10595
Address all correspondence and requests for reprints to: C. Andrew Powers, Ph.D., Department of Pharmacology, New York Medical College, Valhalla, New York 10595. E-mail: andrew_powers{at}nymc.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
T3 regulates growth as well as energy and lipid and bone metabolism; E interplay with T3 may provide a versatile mechanism for metabolic adaptations related to reproductive biology. The physiological relevance of such interplay has been explored in ovariectomized, hypothyroid female rats (14, 15, 16, 17). The results identified a subset of E responses that are T3 dependent; this includes effects on growth, bone mass, and triglycerides. T3-independent E responses include uterine growth and induction of pituitary PRL and kallikrein. Moreover, TM exhibited high agonist efficacy in T3-dependent responses but acted more like an antagonist in most T3-independent responses. These findings suggest that estrogens may mediate two distinct types of biological regulation. In T3-independent responses, E can be viewed as a primary driver that fully activates the ER mechanisms mediating the response, and SERMs appear to poorly activate such mechanisms. In T3-dependent responses, however, E can be viewed as a modulator whose responses rely on interplay with T3 signaling mechanisms; such modulator actions are also activated by SERMs.
The concept of distinct primary driver and modulator actions may clarify E actions in complex physiological systems regulated by multiple hormones. Moreover, targeted activation of modulator responses by SERMs may benefit novel groups of patients. For example, androgen deprivation therapy used in men with disseminated prostate cancer can increase the risk of osteoporosis and bone fractures (18, 19, 20, 21). Estrogens can induce a state of androgen deprivation that can avoid such risks (22), but doses needed for satisfactory tumor responses are often toxic (23), and even low doses evoke breast growth. On the other hand, use of SERMs during androgen deprivation therapy may well prevent osteoporosis with little risk of toxicity or breast growth. However, there is little information about SERM effects in men.
Animal studies of SERMs or their interplay with T3 have also focused on females. Thus, it is unclear if males exhibit equivalent responses, and there are reasons why such equivalence should not be presumed. In particular, the GH-IGF-I axis is targeted by both T3 and estrogens, and some estrogen effects on growth and metabolism have been proposed to reflect changes in the GH-IGF-I axis (24, 25, 26, 27). However, an irreversible sex difference in GH secretory patterns develops soon after birth (24, 28), and this may influence the nature of E effects in males and females. In addition, the association between E effects on the GH-IGF-I axis and E effects on growth and metabolism has not been rigorously analyzed in either male or female rats. Specifically, correlation/regression analyses of data sets of individual rats have not been reported. Thus, it is unclear if measures of growth or metabolism in a given rat during E treatment are significantly related to the GH or IGF-I levels in that rat.
To address the above considerations, male rats were used 1) to study the role of T3 in E and SERM effects and their relation to GH and IGF-I, and 2) to evaluate the therapeutic potential of SERMs in androgen-deprived males. The results show that interplay with T3 contributes to multiple E actions in male rats and imply that SERMs may benefit men on androgen deprivation therapy. In addition, the correlation/regression analyses suggest novel roles for GH and IGF-I in E effects on specific parameters. Finally, the data revealed three pharmacologically distinct classes of E responses that presumably reflect the diversity of ER effector mechanisms in vivo.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
It should be noted that the male rats used in this study had ages similar to those of female rats used in previous studies of T3 interactions. Both male and female rats still have active epiphyseal growth plates in the age range studied, unlike the bones of adult men and women or aged adult rats. Nonetheless, young adult rats respond to gonadectomy and E or SERM treatment with robust changes in cancellous bone turnover and BMD that are qualitatively similar to those of aged rats with closed epiphyses (29, 30).
Experimental design and drug treatments
A gonadectomized, hypothyroid rat model was used which allows
TM, E2B, and T3 effects to be separately
identified and hormonal interactions studied. The model was designed to
selectively detect E2B and TM effects on the pharmacodynamic actions of
T3 and avoid other types of hormonal
interactions. Thus, use of methimazole eliminates changes in the
synthesis or release of T3 or
T4 as a mechanism of E2B or TM effects. Use of
T3 eliminates interactions related to
5'-deiodination of T4 to
T3. Use of T3 also avoids
interactions due to E2B or TM effects on plasma
T4-binding proteins (transthyretin and
T4-binding globulin). T3
has only 1/10 the affinity of T4 for such
proteins, and unlike T4, the distribution of
T3 in the rat is not restricted by plasma protein
binding (volume of distribution = 1.65 liter/kg for
T3 vs. 0.16 liter/kg for
T4) (31).
Six groups of orchidectomized rats received the following drug and hormone treatments starting 3 d after surgery: 1) vehicle solutions alone (n = 8); 2) TM (trans isomer, free base) at 0.5 mg/kg via sc injection, once daily (n = 7); 3) 17ß-E2–3-benzoate (E2B) at 50 µg/kg via sc injection three times weekly (n = 6); 4) T3 (sodium salt) at 10 µg/kg via ip injection once daily (n = 7); 5) T3 plus TM (n = 7); and 6) T3 plus E2B (n = 8). The group of four euthyroid, testis-intact rats received daily injections of vehicle solutions. During the treatment phase of the study, body weight was measured three times per week and food intake was measured once a week. Rats were gang-caged (3–4 per cage) to decrease risk of hypothermia due to hypothyroidism; thus, total food intake per group was measured during a 24-h period, and average intake per rat interpolated. Body temperature was measured weekly with a rectal thermocouple probe and digital thermometer to provide an index of T3 effects on thermogenesis.
The doses of E2B, TM, and T3 used in this study have been found to yield maximal target organ and growth responses in dose-response studies in ovariectomized or ovariectomized, hypothyroid female rats (15, 32). E2B and TM were dissolved in sesame oil with 2% benzyl alcohol; T3 was dissolved in 0.9% NaCl with 5 mM NaOH.
Tissue processing
After 6 wk of treatment, rats were euthanized in random order
with 100 mg/kg sodium pentobarbital (ip) between 1030 and 1530 h.
Blood, anterior pituitaries, and right tibias were collected within 5
min of pentobarbital injection as previously described (14, 15). Blood samples were allowed to clot for 5 min at room
temperature, cooled on ice, and then refrigerated at 5 C; serum was
collected the next day, aliquoted, and stored at -80 C until assay.
The right tibia was stripped of most muscle and connective tissue and
stored in 70% ethanol for subsequent measurements of tibia length (an
index of longitudinal growth) and bone mineral density (BMD). Anterior
pituitaries were sonicated in 400 µl of 10 mM sodium
phosphate buffer (pH 7.5) containing 150 mM NaCl and 0.1%
Triton X-100; samples were then aliquoted and stored at -20 C until
assay. Seminal vesicles were dissected and weighed. The ventral
prostate lobes were visually inspected for any gross differences in
size. The left kidney and heart were dissected and weighed to examine
the effect of TM or E2B on T3 actions that yield
increases in heart and kidney mass that exceed changes in overall
somatic growth (33, 34).
Analysis of BMD
Tibia BMD was measured by dual-energy x-ray absorptiometry with
a QDR-1000 (Hologic, Inc., Waltham, MA) as previously
described (14, 15). Image analysis software
(Hologic, Inc.) automatically calculated bone mineral
content (g), cross-sectional area (cm2), and BMD
(g/cm2) in two regions: the proximal tibia and
the diaphysis. The proximal tibia (upper 1/3 of tibia length) is
relatively enriched in cancellous bone and has a higher ratio of
cancellous to cortical bone than the diaphysis (middle 1/3 of tibia
length) (35).
Metabolic analyses
Lipid metabolism was assessed by measuring total cholesterol,
triglycerides, and ß-hydroxybutyrate (an index of fatty acid
ß-oxidation) in serum using colorimetric kits from Sigma
(St. Louis, MO). Serum glucose was measured as an index of carbohydrate
metabolism using a colorimetric kit from Stanbio Laboratories (San
Antonio, TX). In the rat, 70–80% of serum cholesterol is bound to
high-density lipoprotein, and estrogens and SERMs have equivalent
effects on high density lipoprotein- or low density
lipoprotein-cholesterol (36).
Assay of serum GH and IGF-I
Serum levels of GH and IGF-I were used to assess changes in the
GH-IGF-I axis. Total serum levels of IGF-I were determined using a RIA
kit from Nichols Institute Diagnostics (San Juan
Capistrano, CA) after extraction of IGF-I from serum. IGF-I in rat
serum was separated from IGF-binding proteins using the acid-ethanol
extraction method described by Crawford et al.
(37). Serum samples (100 µl) were mixed with 900 µl
acid-ethanol solution (12.5% 2 M HCl: 87.5%
ethanol, vol/vol) and incubated 30 min at room temperature. Following
centrifugation at 1,500 x g for 30 min at 5 C, 200
µl of the supernate was mixed with 100 µl 0.855
M Tris base (pH 11) and incubated 30 min at room
temperature. After centrifugation at 1,500 x g for 30
min (5 C), 100 µl of the final supernate was mixed with 1.4 ml
phosphate buffer (pH 7.5), and used for RIA. This acid-ethanol
extraction method has previously been validated for RIA of total IGF-I
in male and female rat serum by comparison with results using HPLC
methodology (37). Serum GH was determined using a rat GH
RIA kit from Amersham Pharmacia Biotech (Piscataway,
NJ).
Anterior pituitary analyses
Induction of anterior pituitary PRL and GH by E and
T3, respectively, was qualitatively analyzed
using denaturing PAGE as previously described (15, 16).
Such analyses have previously been validated with GH and PRL RIAs
(15, 16). Anterior pituitary levels of glandular
kallikrein (an E-induced protease in lactotrophs) was assayed using
D-Val-Leu-Arg-p-nitroanilide as previously
described (38). Total protein provided an index of
E-induced anterior pituitary hyperplasia (38).
Statistical analysis of grouped data
Data were subjected to ANOVA followed by a posthoc analysis with
Duncan’s new multiple range test; P < 0.05 was the
criterion of significance. Serum GH and triglyceride data values were
log-transformed to normalize variances.
The change in a parameter due to T3 (‘
due to
T3') was calculated by subtracting the mean of
the vehicle, TM, or E2B group without T3 from
each data point in the matching group with T3;
transformed data were then analyzed as noted above. This facilitated
comparison of E2B and TM interplay with T3 in
diverse parameters and was used to calculate the % efficacy of TM
relative to E2B (100%) in responses involving interplay with
T3.
Correlation/regression analysis
Pearson’s product-moment correlation (r) and multiple linear
regression analysis was used to identify significant associations of
serum GH or IGF-I to other parameters. Correlations/regressions were
calculated using sets of individual rat data (single units of
expression) that captured the effects of the four distinct hormonal
manipulations studied (T3 manipulation with and
without an estrogenic background, and estrogenic manipulation with and
without a T3 background). Correlations capturing
the effect of T3 manipulation without an
estrogenic background were calculated using individual rat data from
groups 1 (vehicle alone) and 4 (T3 alone) (n
= 15 data pairs per parameter). Correlations during
T3 manipulation with an estrogenic background
used data from groups 2 (TM alone), 3 (E2B alone),
5(T3 plus TM) and 6 (T3
plus E2B) (n = 27). Correlations during estrogenic manipulation
without T3 used data from groups 1 (vehicle
alone), 2 (TM alone), and 3 (E2B alone) (n = 21). Correlations
during estrogenic manipulation in the presence of
T3 used data from groups 4
(T3 alone), 5 (TM plus T3)
and 6 (E2B plus T3) (n = 21).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). This
probably reflects the lack of testicular androgens
(39, 40, 41, 42). T3 alone had no effect on
BMD but potentiated TM and E2B effects: TM evoked 4.2% increases in
rats lacking T3, and 9.8% increases in rats
given T3. E2B evoked 12.8% and 17.6% increases
in BMD in the absence or presence of T3,
respectively. In T3-treated rats, TM prevented
bone loss associated with orchidectomy, while E2B yielded BMD values
exceeding euthyroid, testis-intact rats (Fig. 1).
|
). This suggests that E effects on BMD in male
rats involved a T3-independent component in which
TM had low efficacy, and a T3-dependent component
in which TM and E2B had equivalent efficacy.
It is of note that rats lacking T3 displayed no
net growth (see Fig. 3), and E2B increased proximal tibia BMD in such
rats without altering tibia length (see Fig. 4) or cross-sectional area
(not shown). E2B effects on BMD were thus largely independent of
changes in bone growth.
|
|
).
The diaphysis has a lower ratio of cancellous bone to cortical bone,
and the data are consistent with differences in the turnover rates and
E sensitivity of cancellous bone and cortical bone (29, 30, 35).
TM, E2B, and T3 effects on serum levels of cholesterol
and triglycerides
TM lowered cholesterol by 38 mg/dl in orchidectomized, hypothyroid
rats lacking T3 (Fig. 2). Cholesterol was also reduced by
T3 (-24 mg/dl), and T3 and
TM effects were additive. Thus, TM and T3 affect
cholesterol independently of one another. Remarkably, E2B had no effect
on cholesterol, even at doses as high as 250 µg/kg (unpublished
data). Euthyroid, testis-intact male rats had slightly lower
cholesterol levels (-18 mg/dl) than orchidectomized, hypothyroid rats
treated with T3, but the difference was not
significant (Fig. 2).
|
).
T3, TM, or E2B alone did not alter triglycerides
in hypothyroid rats, although TM tended to lower levels. However, TM
and E2B increased triglycerides in T3-treated
rats (Fig. 2); the TM increase was 65% that evoked by E2B
(inset, Fig. 2).
TM, E2B, and T3 effects on male sex organs
TM had no effect on seminal vesicle weight in orchidectomized,
hypothyroid rats with or without T3 (Fig. 2);
this was also the case with the ventral prostate (based on visual
inspection). E2B in combination with T3 evoked a
small rise in seminal vesicle weight (Fig. 2), but seminal vesicles in
rats given E2B plus T3 were still minuscule
compared with testis-intact rats.
TM, E2B, and T3 effects on somatic growth
In the first 10 d, orchidectomized, hypothyroid controls lost
3% of body weight while E2B- or TM-treated rats lost 9% (Fig. 3). Weight then stabilized in control and
TM-treated rats, but a slow weight loss continued in E2B-treated rats
(-4% in 32 d) (Fig. 3). T3 increased
weight, and this gain was inhibited 78% by TM and 65% by E2B
(inset, Fig. 3).
In rats lacking T3, TM or E2B did not alter
tibia length. However, increases in tibia length evoked by
T3 were similarly inhibited by TM (-69%) and
E2B (-81%) (inset, Fig. 4).
During the 6-wk treatment phase of the study, euthyroid, testis-intact
rats gained 70% more weight than orchidectomized, hypothyroid rats
treated with T3 alone (not shown). This
presumably reflects a stimulation of somatic growth by testicular
androgens (43), which may also partly account for the
greater tibia lengths of testis-intact rats (Fig. 4).
TM, E2B, and T3 effects on serum and pituitary GH
Serum GH in orchidectomized, hypothyroid rats was only 8% that of
euthyroid, testis-intact rats (Fig. 4), and this was associated with a
loss of pituitary GH content (Fig. 5).
Although T3 alone increased serum GH, it yielded
only 25% the level of euthyroid, testis-intact males even though
pituitary GH content was almost fully restored. In rats lacking
T3, TM and E2B evoked 4.0-fold and 6.9-fold
increases in serum GH, respectively, whereas T3
alone yielded a 3.2-fold increase (Fig. 4). However, TM and E2B evoked
only weak T3-like effects on pituitary GH content
(Fig. 5). Thus, TM and E2B effects on serum GH do not simply reflect
changes in pituitary GH content; alterations in neuroendocrine
regulation are evident.
|
). In either the presence or absence of
T3, TM displayed half the efficacy of E2B on
serum GH. On the other hand, T3 induction of
pituitary GH content was clearly inhibited by TM, whereas E2B had
little effect (Fig. 5). Thus, although TM mimicked E2B effects on serum
GH in T3-treated rats, it displayed a distinct
effect on pituitary GH. The pituitary GH data of male rats match the
results of similar studies in female rats (14, 15, 16, 17). TM or
E2B effects to increase serum GH, however, were much smaller in female
rats, and were not enhanced by T3
(15). Only orchidectomized, hypothyroid rats given T3 in combination with TM or E2B had serum GH levels approaching those of euthyroid, testis-intact males. However, serum GH increases evoked by E2B and TM did not yield corresponding increases in somatic growth.
TM, E2B, and T3 effects on serum IGF-I
Orchidectomized, hypothyroid rats had serum IGF-I levels
that were half that of testis-intact, euthyroid rats (Fig. 4).
T3 alone fully restored IGF-1 to the levels of
euthyroid, testis-intact rats despite its relatively modest effect on
serum GH. In contrast, E2B or TM did not yield IGF-I increases similar
to T3 even though they caused greater increases
in serum GH. Interestingly, E2B and TM had differing effects on IGF-I
in the absence of T3; TM modestly increased IGF-I
(+20%), whereas E2B markedly lowered IGF-I (-40%) (Fig. 4).
Rats treated with E2B plus T3 also had much lower
IGF-I levels than those given T3 alone, due in
large part to a smaller IGF-I response to T3. TM
also displayed a tendency to lower IGF-I in
T3-treated rats (-12%) that did not reach
significance. However, unlike E2B, TM alone modestly increased IGF-I
(+20%), and this partly masked inhibition of T3
effects. When IGF-I increases due to T3 were
calculated, it was evident that this T3 effect
was significantly inhibited by TM (-42%) as well as by E2B (-73%)
(inset, Fig. 4). TM effects on IGF-I in male rats resemble
results in female rats (14): this includes both TM’s
tendency to increase IGF-I in the absence of T3,
and the inhibition of T3 effects on IGF-I.
Effects on food intake, body temperature, serum glucose, and serum
ß-hydroxybutyrate
Orchidectomized, hypothyroid male rats lacking
T3 consumed 35% less food per kg body weight
than euthyroid, testis-intact males (Fig. 7). T3
increased food intake to that of euthyroid, testis-intact rats. E2B and
TM did not alter food intake in rats with or without
T3 (inset, Fig. 7). TM and E2B effects
on growth and metabolism are unlikely to reflect changes in food
intake. The data further document that not all T3
effects are sensitive to E2B or TM.
|
). T3 produced a
1.5 C increase in temperatures of orchidectomized, hypothyroid rats.
T3 effects on body temperature were slightly
inhibited by TM and E2B (-11% and -33%, respectively), but only E2B
had a significant effect. Only a small fraction of the thermogenic
effects of T3 are sensitive to E2B, and TM has
only 1/3 the efficacy of E2B.
Euthyroid, testis-intact rats had significantly lower body temperatures
(-1.0 C) than orchidectomized, hypothyroid rats given
T3 (Fig. 7). This temperature difference is due
to testicular androgens; it does not reflect use of an excessive
T3 dose (our manuscript in
preparation).
TM and E2B tended to lower serum glucose in orchidectomized,
hypothyroid rats lacking T3, but the changes were
insignificant (Fig. 7). T3 alone had no effect on
glucose but enhanced TM and E2B effects over 2-fold. Decreases in
glucose evoked by E2B (-28%) were significant in
T3-treated rats, but not those for TM (-13%).
Serum ß-hydroxybutyrate levels were less than 2.5 mg/dl in all groups
(data not shown); thus, no major treatment effects on fatty acid
oxidation were evident. The glucose and ß-hydroxybutyrate data did
not reveal alterations in energy metabolism likely to explain E2B and
TM effects on growth.
TM, E2B, and T3 effects on relative heart and
kidney weights
Calculation of organ weight relative to body weight
(relative weight, g/kg) can detect changes in organ weight that are
disproportionate from changes in body weight. T3
significantly increased relative heart weight in hypothyroid male rats
(Fig. 8); this may reflect direct actions on the heart as well as
effects secondary to changes in vascular tone and autonomic nervous
system function (33). TM and E2B did not alter relative
heart weight in rats lacking T3, and increases
evoked by T3 were insensitive to E2B or TM
(inset, Fig. 8). Relative kidney weight was slightly
increased by TM or E2B in rats lacking T3 (+0.5
and +0.4 g/kg, respectively). T3 caused greater
increases (+1.2 g/kg). T3 and TM had additive
effects on relative kidney weight, but increases with E2B plus
T3 were larger than expected (inset,
Fig. 8).
|
).
Unlike E2B, TM slightly decreased pituitary protein in the presence or
absence of T3.
|
). T3 alone
evoked only small PRL increases, but markedly enhanced E2B actions.
However, PRL increases with E2B plus T3 were
coupled to nonspecific increases pituitary proteins characteristic of
tissue hyperplasia (Figs. 5 and 6). TM did not alter PRL either with or
without T3.
T3 alone slightly decreased glandular kallikrein
(Fig. 6), an E-induced lactotroph protease (32). E2B
induced total kallikrein activity (nmol/min·pituitary) by 6-fold in
the absence of T3, and 18-fold in the presence of
T3. However, E2B induction of kallikrein was
unaffected by T3 when expressed as specific
activity (nmol/min·mg protein) to adjust for changes in total protein
due to tissue hyperplasia (Fig. 6). TM lacked effect on glandular
kallikrein.
Overall, the data indicate that E2B induction of PRL and kallikrein was T3 independent, whereas E2B actions to evoke pituitary hyperplasia were T3 dependent. The hyperplastic effect of E is due to a selective induction of lactotroph proliferation (45, 46), and this amplifies E2B effects on PRL and kallikrein by increasing lactotroph numbers.
Correlation of serum GH and serum IGF-I with other physiological
parameters
Group means suggested that E2B and TM effects on GH and IGF-I may
be related to changes in four other parameters: body weight, tibia
length, proximal tibia BMD and serum triglycerides.
Correlation/regression analysis was used to quantitatively assess these
associations.
E2B and TM effects on a parameter might reflect changes in T3 regulation of GH or IGF-I, changes in GH or IGF-I independent of T3, effects unrelated to GH or IGF-I, or combinations thereof. To separately evaluate these alternatives, correlation analysis was performed on four unstructured sets of individual rat data capturing the relationships associated with T3 manipulations with or without an estrogenic background (E2B or TM treatment), and estrogenic manipulations with or without a T3 background (see Materials and Methods for details).
Relative heart weight illustrates results for a parameter that was
unaffected by E2B or TM and was unlikely to be related to GH or IGF-I.
Relative heart weights during T3 manipulation in
the absence of E2B/TM were strongly correlated with GH and IGF-I, but
these correlations were absent from T3
manipulations with E2B/TM (Table 1).
Because E2B or TM altered GH and IGF-I without affecting relative heart
weight (Figs. 4 and 7), the correlations
during T3 manipulation without E2B/TM must be
coincidental. Indeed, heart weight was
not significantly correlated with GH or IGF-I in estrogenic
manipulations (Table 1).
|
). Thus, an estrogenic background
impaired GH induction of IGF-I during T3
manipulation, as expected from the grouped results (Fig. 4, inset). GH and IGF-I were poorly related during estrogenic
manipulations without T3, but a surprising
negative correlation was seen during estrogenic manipulation of
T3-treated rats (Table 1). This inverse relation
suggests the operation of an IGF-I negative feedback loop that inhibits
GH release and is T3 dependent. These results
suggest that decreases in IGF-I (or reduced sensitivity of the feedback
loop to IGF-I) may contribute to E2B and TM effects on GH in
T3-treated rats.
Body weight change during T3 manipulation without
E2B or TM was positively correlated with GH and IGF-I (Table 1). E2B
and TM degraded such correlations during T3
manipulation; this suggests that a portion of E2B/TM effects on body
weight involve T3 actions unrelated to changes in
GH and IGF-I levels per se (see below). On the other hand,
weight change in estrogenic manipulations was positively correlated
with IGF-I and negatively correlated with GH regardless of
T3 background. Given the essential role of IGF-I
in growth (25), the results suggest that IGF-I decreases
also partly contribute to E effects on weight.
Tibia length was not correlated with GH during T3
manipulation without E2B/TM but displayed a strong positive correlation
with IGF-I (Table 1). This IGF-I correlation was obliterated in the
presence of E2B/TM. Thus, changes in serum IGF-I poorly explain E2B and
TM effects on T3-evoked bone growth; effects on
other T3 actions that promote bone growth seem
likely. This might include T3 actions to enhance
tissue sensitivity to IGF-I because hypothyroidism eliminated growth
but lowered IGF-I only 50% (see Figs. 3 and 4). On the other hand,
tibia length was positively correlated with IGF-I and negatively
correlated with GH during estrogenic manipulations in the presence of
T3, but not in its absence. Given the role of
IGF-I in longitudinal bone growth (25), this suggests that
decreases in IGF-I may also contribute to E2B and TM effects on tibia
length. It is also noteworthy that, unlike body weight change, tibia
length displayed an association with IGF-I and GH only in the presence
of T3.
Proximal tibia BMD was positively correlated with serum GH during T3 manipulations in the presence of E2B or TM, but not in their absence. The BMD and GH correlation was even stronger in estrogenic manipulations, and T3 had little influence on the association. Unlike body weight or tibia length, BMD was not significantly correlated with IGF-I during T3 manipulations, and BMD was negatively correlated with IGF-I during estrogenic manipulations regardless of T3. In view of evidence that GH can increase BMD (47), the correlation analyses suggest that E2B and TM effects on BMD may partly reflect their effects on serum GH.
Triglycerides displayed a significant positive correlation with GH only during T3 manipulations in the presence of E2B/TM. Triglycerides and IGF-I were poorly associated during T3 manipulations. Triglycerides were also significantly correlated with GH (positive) and IGF-I (negative) during estrogenic manipulations with T3, but not in its absence. In view of evidence that GH can increase triglycerides (48, 49), the results suggest that GH may contribute to E2B and TM effects on triglycerides. Unlike BMD, the triglyceride response to E2B and TM displayed a T3 requirement that appears unrelated to effects on GH and IGF-I.
Multiple linear regression analysis
During estrogenic manipulations with T3,
serum GH and IGF-I were inversely related, and it wasn’t surprising
that GH and IGF-I correlations with a parameter were often the inverse
of one another. However, GH and IGF-I were not correlated during
estrogenic manipulations without T3; thus, it was
notable that in such manipulations BMD was positively correlated with
GH and negatively correlated with IGF-I, whereas the opposite occurred
with body weight change (Table 1). This suggested that GH and IGF-I may
independently contribute to E effects on BMD and body weight. Multiple
linear regression analysis was used to evaluate this possibility.
Multiple linear regression analysis can be used to test hypotheses that two or more independent variables (GH and IGF-I in this case) make significant, independent contributions to regressions predicting the observed levels of a dependent variable (BMD, etc.). This involves fitting a regression equation to the data using the method of least-squares. The residual sum squares (total, regression, and error) are then used to determine the significance of the regression coefficients contributed by each independent variable (null hypothesis: coefficients = 0), and the regression correlation coefficient (r). For the present analysis, the significance of the regression coefficient of each independent variable was of primary interest.
Proximal tibia BMD was the only parameter where GH and IGF-I both made significant, independent contributions to regressions predicting observed values. In estrogenic manipulations without T3, BMD was well correlated (r = 0.78; P < 0.01) with regression predictions [BMD = 0.258 + 0.0242(logGH) - 0.000064(IGF-I)], and the regression coefficients for GH and IGF-I were each significant (P < 0.01). In estrogenic manipulations with T3, BMD displayed an r = 0.71 (P < 0.01) with regression predictions [BMD = 0.270 + 0.0172(logGH) - 0.0000373(IGF-I)]; the IGF-I coefficient was significant (P = 0.04) and the GH coefficient neared significance (P = 0.067). Regression predictions from analysis of all orchidectomized rat data [BMD = 0.239 + 0.0271(logGH) - (0.0000234 x IGF-I)] yielded an r = 0.72 (P < 0.01), with significant regression coefficients for both GH (P < 0.01) and IGF-I (P = 0.02).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Influence of T3 on E2B and tamoxifen effects, and
relation to the GH-IGF-I axis
The hypothyroid, gonadectomized rat model enables the
detection of target-selective T3 and E interplay
arising from pharmacodynamic (receptor-mediated) mechanisms
(14, 15, 16, 17). The present data further indicate that some
actions of E2B and TM involve interplay with T3
and also imply some sex differences. Moreover, correlation analysis
provided a rigorous test of the association of serum GH and IGF-I with
specific E responses in individual rats. The significant correlations
found in data sets capturing the effects of estrogenic manipulation do
not prove a cause and effect linkage. Nonetheless, the significant
correlations satisfy an essential requirement of hypotheses that
changes in GH and IGF-I contribute to E effects on growth and
metabolism, and also imply distinct roles for GH and IGF-I in different
E responses.
Interactions related to the GH-IGF-I axis and somatic
growth
The GH-IGF-I axis has a major role in growth and metabolism
(25, 27, 47, 55), and is a likely target of E and
T3 interplay. T3 promotes
GH synthesis and secretion, enhances GH induction of hepatic IGF-I
synthesis, and induces IGF-I receptors and binding proteins
(56, 57, 58). Estrogens and TM also increase GH secretion but
lower serum IGF-I as well as hepatic expression of IGF-I mRNA
(26, 59, 60, 61, 62, 63, 64, 65).
In the present study, increases in serum GH evoked by T3 were coupled to significant increases in IGF-I. In contrast, TM or E2B failed to increase IGF-I despite striking effects to increase GH. Interestingly, estrogen effects on serum IGF-I appear to involve two components. A T3-independent component was evident in E2B effects to lower IGF-I by 40% in rats lacking T3; TM lacked agonist efficacy in this component and slightly raised IGF-I. A T3-dependent component was evident in E2B and TM inhibition of T3 induction of IGF-I; TM had 58% the efficacy of E2B in this component. These data are consistent with findings in hypothyroid female rats (14) in which TM slightly elevated IGF-I in rats lacking T3, but inhibited T3 effects on IGF-I. It should be noted that TM did not inhibit ovine GH effects on somatic growth or IGF-I in hypothyroid rats (14), nor do estrogens consistently inhibit GH effects on serum IGF-I or hepatic IGF-I mRNA in hypophysectomized rats (61, 62). Overall, E effects on serum IGF-I are complex and seem likely to target T3 actions that enhance GH effects, as well as mechanisms independent of T3 or GH.
Multiple mechanisms also seem to mediate E effects on T3-evoked growth. Weight gain and tibia length were correlated with IGF-I during estrogenic manipulations, suggesting that decreases in IGF-I contribute to E2B and TM effects to inhibit growth. In T3-treated rats, the net decrease in IGF-I caused by E2B (-55%) seemed large enough to impair growth, but the 12% decrease with TM was unlikely to do so. This might indicate that IGF-I production in targets such as bone is more relevant to E2B and TM effects on growth (66, 67). Alternatively, decreases in tissue sensitivity to IGF-I due to E effects on IGF-I receptors or binding proteins may contribute to the inhibition of somatic growth (63, 64, 65). Indeed, IGF-I feedback inhibits GH release (68, 69), and TM effects on serum GH in T3-treated rats seem consistent with reduced sensitivity of the feedback mechanism to IGF-I. Decreases in IGF-I sensitivity may also be relevant to E2B and TM effects to degrade the correlation of body weight change and tibia length with serum GH and IGF-I during T3 manipulations.
Interactions related to BMD
In ovariectomized-hypothyroid female rats,
T3 decreased proximal tibia BMD, and E2B and TM
effects seemed to reflect inhibition of T3
actions (14, 15). This appeared consistent with
T3 actions directly on bone to stimulate
resorption (70, 71), the increased risk of osteoporosis
due to thyrotoxicosis (72, 73), and E prevention of
T4-evoked bone loss in postmenopausal women
(74, 75). Nonetheless, T3 alone
lacked effect on BMD in the present male rats, and E2B increased BMD in
the absence of T3. This suggests a sex difference
in T3 effects on BMD; direct comparisons of the
sexes are needed to clarify this issue.
TM had less effect on BMD than E2B in the present male rats, but TM has been shown to fully mimic E effects on BMD in female rats (76, 77), again hinting at a sex difference. In this regard, it is of note that E effects on BMD seemed to involve two components; TM had efficacy equal to E2B in a T3-dependent component but only 33% efficacy in a T3-independent component. Given the T3-dependence of E effects on BMD in female rats (14), the T3-independent component may be most relevant to any sex differences.
GH stimulates bone remodeling associated with longitudinal growth and also participates in the remodeling of mature bone (47). Although GH stimulates both bone resorption and formation, an anabolic effect eventually emerges due to greater bone formation. In the present study, BMD was positively correlated with serum GH during estrogenic manipulations in either the presence or absence of T3; such results suggest that GH increases contribute to E2B and TM effects on BMD. This association is unlikely to reflect changes in growth-related bone remodeling since somatic growth was absent in rats lacking T3. Given the sex difference in GH secretion patterns (see Introduction), GH might also be relevant to possible sex differences in T3 and E effects on BMD. Nonetheless, it is important to note that GH release evoked by T3 alone failed to increase BMD in this and other studies of hypothyroid, gonadectomized rats (14, 15). This may be relevant to evidence that E can act directly on bone to alter remodeling (1, 78); such actions may enhance the net anabolic effect of GH. Indeed, TM enhanced ovine GH effects on BMD in hypothyroid rats (14), and transgenic mice overexpressing human GH require ovarian estrogens to develop elevated BMD (79).
BMD was also negatively correlated with IGF-I during estrogenic manipulations. Bone growth and remodeling involves a complex interplay between T3, GH, IGF-I as well as other hormones and cytokines, and is not fully understood (27, 47, 78). Many GH effects result from IGF-I released from the liver or target tissue (the somatomedin theory of GH action). However, effector mechanisms unrelated to IGF-I also contribute in some tissues, including bone (the dual-effector theory of GH action) (47, 80, 81). An implication of the dual-effector theory is that GH and IGF-I may make distinct contributions to GH responses involving dual effectors. Indeed, multiple linear regression analysis showed that GH and IGF-I each made significant, independent contributions to regression equations predicting BMD during estrogenic manipulations; this was true of no other physiological parameters. This result appears to satisfy a major requirement of the hypothesis that GH actions via multiple effector mechanisms (IGF-I dependent and independent) may be relevant to E effects on BMD.
Interactions related to serum triglycerides and cholesterol
Estrogens and SERMs can increase serum triglycerides in rats and
man (50, 54, 82, 83). In the male rats used in this
study, such increases were T3 dependent, and TM
had 65% the efficacy of E2B; female rats gave similar results
(14, 15, 17). Moreover, triglycerides were positively
correlated with GH during estrogenic manipulations in
T3-treated rats. This may be relevant to reports
that GH can stimulate the hepatic synthesis or release of triglycerides
(48, 49). Although T3 alone had
little effect on triglycerides, it is well known to stimulate multiple
pathways involved in fatty acid and triglyceride synthesis, storage,
mobilization, and oxidation. E interplay with a subset of such
T3 actions might alter the balance between
opposing pathways to increase triglycerides.
The role of T3 in E and SERM effects on serum cholesterol had not been previously studied. TM evoked cholesterol decreases that were T3 independent, but E2B had no effect despite powerful actions on other parameters in the same rats. Others have reported similar findings. Thus, ethinyl E2 and SERMs lowered cholesterol at doses affecting the uterus, somatic growth and bone metabolism, but 17ß-E2 had no effect on cholesterol at doses evoking maximal effects on other targets (36, 84). Moreover, this phenomenon is not dependent on the route of drug dosing (36). Classic estrogens such as 17ß-E2, E2B, and ethinyl E2 have been thought to have equivalent pharmacodynamic actions despite differing pharmacokinetics, and act as potent agonists in most T3-independent E responses. However, the cholesterol response reveals a more complex pharmacology and defines a novel T3-independent E response. LH suppression in ovariectomized rats is another unusual T3-independent response because E2B and TM exhibit similar agonist efficacy (15).
Interactions related to lactotroph hyperplasia
Pituitary hyperplasia evoked by E2B represents another distinct E
response: it was T3 dependent, but TM lacked
agonist efficacy (unlike other T3-dependent
responses). E2B and T3 interplay to evoke cell
proliferation is tissue-specific since induction of uterine growth does
not require T3 (15). The interplay
is also gene specific because PRL and kallikrein induction in
lactotrophs was largely T3-independent.
Pharmacological classification of in vivo E responses
E2 binding to ER
or ERß is now recognized to trigger a
cascade of interactions among diverse effector molecules that
ultimately evoke an E response. It is also evident that SERMs can
activate ER effector mechanisms mediating some responses, while
inhibiting those mediating other responses (see Refs. 4, 5, 6
for review). This differential sensitivity of ER effector mechanisms is
widely recognized to provide the molecular foundation for the disparate
pharmacological characteristics of different E responses. However, the
pharmacological features of different in vivo E responses
have not been systematically analyzed, and it’s unclear how many
pharmacologically distinct classes are operative in animals and man.
Such information may identify groups of in vivo E responses
mediated by similar effector mechanisms, and indicate how many distinct
effector mechanisms regulate physiological systems in vivo.
Also, as ER ligands become available that target selective ER subtypes
or effectors, it may be possible to link the pharmacological classes to
specific ER effector mechanisms. Conversely, classification efforts may
identify sets of in vivo E responses amenable to selective
pharmacological targeting, and thus facilitate efforts to design new ER
ligands with novel therapeutic actions.
Based on the present results and prior reports (16, 17, 36, 76, 77, 84), three pharmacologically distinct classes of in
vivo E responses can currently be identified in the rat (Classes
A, B, and C) (Table 2). In Class A
responses, classic estrogens (E2B, 17ß-E2, ethinyl E2) act as potent
agonists, and TM acts more like an antagonist (<33% agonist
efficacy). In Class B responses, classic estrogens are potent agonists,
and TM acts more like an agonist (
50% agonist efficacy). In Class C
responses, TM and ethinyl E2 are potent agonists, but E2B and 17ß-E2
are inactive at doses evoking maximal Class A effects.
|
and ERß exhibit Class A
pharmacology on model genes regulated via E response elements
(85, 86, 87, 88, 89). However, in models regulated by transcription
factor AP-1, ER exhibits Class B pharmacology while ERß exhibits
Class C pharmacology (88); ERß also exhibits Class C
pharmacology in certain other models (89). Systematic
characterization of in vivo E responses using a larger
number of distinctive ER ligands may further refine the pharmacological
classification.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received March 27, 2001.
Accepted for publication June 27, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ß. Biochem Biophys Res Commun 236:140–145[CrossRef][Medline]
and estrogen
receptor ß to partial estrogen agonists/antagonists. Mol Pharmacol 54:105–112 and ERß at
AP1 sites. Science 277:1508–1510-1 promoter
in response to tamoxifen and other estrogen receptor antagonists, but
not in response to estrogen. Mol Endocrinol 13:418–430This article has been cited by other articles:
 |
|
 |
|
  J. M. Fitts, R. M. Klein, and C. A. Powers Comparison of Tamoxifen and Testosterone Propionate in Male Rats: Differential Prevention of Orchidectomy Effects on Sex Organs, Bone Mass, Growth, and the Growth Hormone-IGF-I Axis J Androl, July 1, 2004; 25(4): 523 - 534. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Vasudevan, S. Ogawa, and D. Pfaff Estrogen and Thyroid Hormone Receptor Interactions: Physiological Flexibility by Molecular Specificity Physiol Rev, October 1, 2002; 82(4): 923 - 944. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
