Role of soy protein with normal or reduced isoflavone content in reversing bone loss induced by ovarian hormone deficiency in rats

Role of soy protein with normal or reduced isoflavone content in
reversing bone loss induced by ovarian hormone deficiency in
rats1–3
Bahram H Arjmandi, Mary J Getlinger, Noopur V Goyal, Lee Alekel, Clare M Hasler, Shanil Juma, Melinda L Drum,
Bruce W Hollis, and Subhash C Kukreja
ABSTRACT
Soy protein, a rich source of isoflavones, fed
and bisphosphonates (11). Although these agents may prevent
immediately after an ovariectomy prevents bone loss in rats.
further bone loss in established osteoporosis, they cannot restore
Reports of the effectiveness of natural and synthetic isoflavones
bone mass that has already been lost (12, 13). Furthermore, the
in preventing or treating osteoporosis led us to examine the
compounds that are reported to have bone-forming properties
effect of soy protein in reversing established bone loss. Seventy-
(eg, sodium fluoride, growth hormone, parathyroid hormone,
two 95-d-old female Sprague-Dawley rats were assigned to 6
and anabolic steroids) are either associated with undesirable side
groups. The rats were either sham operated (SHAM; 2 groups) or
effects or they do not produce bone mass of desirable quality (3,
ovariectomized (OVX; 4 groups) and then fed a casein-based,
5, 6, 14). Naturally occurring compounds or synthetic substances
semipurified diet. Thirty-five days after surgery, 1 SHAM and 1
that could substitute for or reduce the use of these types of drugs
OVX group were killed to examine the occurrence of bone loss.
and that have fewer undesirable side effects are needed.
Thereafter, the other SHAM and 1 OVX groups continued to
Kalu et al (15) observed that feeding soy protein instead of
receive the casein-based diet. Whereas the remaining 2 OVX
casein to old male rats prevented age-associated bone loss. This
groups received diets in which casein was replaced by soy
positive effect of soy protein on bone was credited to its amino
protein with normal (OVX+SOY) or reduced (OVX+SOY?)
acid pattern. Omi et al (16) reported that rats fed soymilk had
isoflavone content for 65 days. The OVX control group had
greater bone mineral density and mechanical bone strength than
significantly lower femoral and fourth lumbar vertebral bone
did control rats fed casein. However, the authors speculated that
densities than the SHAM group. Femoral density of rats fed SOY
this beneficial effect might have been due to enhanced intestinal
or SOY? diets were not significantly different from SHAM or
absorption of calcium.
OVX controls. This suggests a slight reversal of cortical bone
Recently, we found that feeding soy protein to ovariectomized
loss that may be partially due to higher femoral insulin-like
rats prevented bone loss (17). Furthermore, in a follow-up study
growth factor I mRNA transcripts resulting from both the SOY
Goyal et al (18) showed that this positive effect was due to the
and SOY? diets. The ovariectomy-induced increases in indexes
isoflavones in the soy. In support of these animal data, it was
of bone turnover were not ameliorated by either of the soy diets,
shown recently that a diet of soy protein with added isoflavones
suggesting that any positive effect of soy was achieved through
(2.2 mg/g protein) significantly (P < 0.05) improved lumbar
enhanced bone formation rather than slowed bone resorption.
spine bone mineral content and density in postmenopausal
Long-term consumption of soy or its isoflavones may be needed
women compared with a casein control diet (JW Erdman et al,
to produce small but continued increments in bone mass.
Am
personal communication, 1996).
J Clin Nutr 1998; 68(suppl):1358S–63S.
Our view that isoflavones in soy produce beneficial effects on
KEY WORDS
Soy protein, isoflavones, ovariectomy,
osteoporosis, rats, casein, bone density, messenger RNA
1 From the Department of Nutritional Sciences, Oklahoma State Univer-
sity, Stillwater; the Department of Human Nutrition and Dietetics and the
School of Public Health, Division of Epidemiology/Biostatistics, University
INTRODUCTION
of Illinois at Chicago; the Department of Food Science and Human Nutrition,
Ovarian hormone deficiency is a major risk factor for osteo-
Iowa State University, Ames; Functional Foods for Health Programs; the
porosis in postmenopausal women (1–4). Preventive measures
Department of Pediatrics, Medical University of South Carolina, Charleston;
and the Department of Medicine, West Side Veterans Administration Medical
are the best and the most cost effective means to minimize osteo-
Center, Chicago.
porosis (5, 6). However, a large segment of the population is
2 Supported in part by a grant from the Illinois Soybean Operating Board.
already afflicted with this disease and requires immediate inter-
The complementary DNA probe was provided by CR Roberts.
vention; osteoporosis is a threat to ?25 million Americans, 80%
3 Address reprint requests to BH Arjmandi, Department of Nutritional Sci-
of whom are women (3). Current therapies emphasize the use of
ences, 425 Human Environmental Sciences, Oklahoma State University, Still-
antiresorptive agents, such as estrogen (7, 8), calcitonin (9, 10),
water, OK 74078–6141. E-mail: arjmand@okstate.edu.
1358S
Am J Clin Nutr 1998;68(suppl):1358S–63S. Printed in USA. © 1998 American Society for Clinical Nutrition

---

ROLE OF SOY PROTEIN IN REVERSING BONE LOSS
1359S
bone is strengthened by numerous reports that indicate that ipri-
mRNA was analyzed with a complementary DNA (cDNA) probe
flavone, a synthetic iso?avone, prevents bone loss (19, 20). Because
for rat IGF-I, generously provided by CR Roberts (27), and
ipri?avone was also reported to enhance bone formation (21, 22),
adjusted for the amount of total RNA measured with a cDNA
we proposed that soy with iso?avones would promote bone forma-
probe for 28S RNA. For Northern blot analysis, 30 ?g of total cel-
tion as well. Therefore, we examined the effects of feeding soy pro-
lular RNA was fractionated electrophoretically on 1.2% agarose
tein with normal iso?avone (SOY) or reduced iso?avone (SOY?)
gel containing 2.2 mol/L formaldehyde. All RNA samples were
content to ovariectomized rats with established bone loss.
denatured before loading by incubation for 20 min at 65 ?C in a
solution of 10X MOPS, 37% formaldehyde, and deionized for-
mamide. Electrophoresis was performed in 1X MOPS (pH 7.0) at
MATERIALS AND METHODS
80V for ?2 h. The sample was stained with ethidium bromide and
ribosomal 28S RNA was visualized under ultraviolet light to con-
Animals and diets
?rm the integrity of RNA and ensure equal loading.
Seventy-two 90-d-old, female Sprague Dawley rats were pur-
The RNA was transferred overnight from the gel to Hybond-
chased from Harlan Sprague-Dawley (Indianapolis) and used for
N nylon membrane (Amersham) by capillary action in 10X stan-
this study when they were 95 d old. On arrival at our institution,
dard saline citrate (SSC; 3 mol NaCl/L, 0.3 mol trisodium cit-
they were housed in an environmentally controlled animal labora-
rate/L), then prehybridized for 6 h at 65 ?C in hybridization
tory. Rats were acclimated for 5 d to a standard laboratory nonpu-
buffer containing 1 mol Na HPO /L (pH 7.0), 0.5 mol EDTA/L,
2
4
ri?ed diet (diet no. 8640; Teklad, Madison, WI) and randomly
0.5% bovine serum albumin, and 20% sodium dodecyl sulfate,
divided into 6 groups of 12 rats each and were either sham-oper-
and probed overnight at 65 ?C with 4 ? 106 cpm/L 32P-labeled
ated (SHAM; 2 groups) or ovariectomized (OVX; 4 groups). Ini-
cDNA. The cDNA was labeled with [32P]dCTP by using Primer-
tially after the surgery, all the animals were fed a standard, semi-
it-II Random Primer labeling system (Stratagene Cloning Sys-
puri?ed, casein-based diet (CASEIN) (Harlan Teklad diet #88190,
tems, La Jolla, CA) and purified by using a push column (Strat-
Madison, WI). Thirty-?ve days from the date of surgery, 1 SHAM
agene). The blots were washed under the conditions of increased
and 1 OVX group was killed to verify that bone loss had occurred.
stringency, up to 0.1? SSC and 0.1% SDS at 65 ?C. The mem-
Of the remaining groups, the SHAM and 1 of the OVX groups
branes were placed on Kodak X-Omat AR-5 film (Eastman
received a casein-based diet. The other 2 OVX groups received
Kodak, Rochester, NY) and exposed at ?80 ?C (28, 29). The
diets in which casein was replaced with soy protein isolate with
Northern blots were quantitated using a phosphorimager (Model
normal (SOY) or reduced (SOY?) isoflavone content. The
445-S1; Molecular Dynamics, Sunnyvale, CA).
iso?avone content of SOY diet (mg/kg soy protein isolate), as
determined by Protein Technologies International (St Louis),
Blood analysis
included genistin, 1462; genistein, 25.1; daidzin, 590; and daidzein
Serum 17?-estradiol was measured by using a double anti-
11.3. The iso?avone content of the SOY? was ?10% that of the
body kit purchased from Diagnostics Products (Los Angeles).
SOY diet. Treatment for these remaining groups continued for 65 d
Blood ionized calcium and pH were measured by using a NOVA
and the animals were killed at the end of that period.
8 Calcium Ion Analyzer (Nova Biomedicals, Waltham, MA)
immediately after blood collection. The values were quantified
Bone analysis
by using an internal standard. Serum alkaline phosphatase, an
The right femur and fourth lumbar vertebra were cleaned and
index of bone formation, and tartrate-resistant acid phosphatase,
freed of surrounding soft tissue. The length of each femur was
an index of bone resorption, were measured by using standard
measured with a vernier caliper. Bone volume and density were
kits (Sigma Diagnostics). Serum 25-hydroxycholecalciferol (cal-
measured by Archimedes’ principle (23). Briefly, each bone was
cidiol) and 1,25-dihydroxycholecalciferol (calcitriol) metabo-
placed in an unstoppered vial filled with deionized water, and the
lites were measured according to the methods of Hollis (30, 31).
vial was placed under a vacuum for 90 min to ensure that all the
trapped air diffused out of the bone. Each bone was removed
Urinary analysis
from its vial, blotted, weighed, and returned to the vial contain-
Urinary hydroxyproline was measured as an index of bone
ing deionized water. The bone was weighed again in water and
resorption. Aliquots of 12-h nighttime urine were hydrolyzed
the density was calculated (in g/cm3 bone vol). Bones were
overnight with 6 mol HCl/L and then treated with chloramine T to
brought to constant weight by placing them at 95 °C for 12 h and
convert hydroxyproline to pyrocarboxylic acid. The latter was
weighing. The dried bones were ashed at 700 °C for 12 h, cooled,
extracted with toluene and measured colorimetrically by using the
and weighed to assess the percentage mineral content. Bone cal-
modi?ed method of Kivirikko et al (32). Hydroxyproline is reported
cium was measured by using a calcium determination kit (587A;
as nanomoles of hydroxyproline per 12-h nighttime urine volume.
Sigma Diagnostics, St Louis). Phosphorus was determined by
the methods of Taussky and Shorr (24).
Statistical analysis
Statistical analysis (INSTAT version 2.00, GraphPad, 1993,
RNA extraction and Northern blot analysis
San Diego) involved estimation of means, SDs, and SEMs of all
The distal half of the left femur was quickly excised, cleaned
variables for each of the groups (33). Analysis of variance
of soft tissue, and frozen in liquid nitrogen. The frozen femurs
(ANOVA) was performed on the means to determine statistically
were homogenized in a tissue homogenizer and total cellular
significant (P < 0.05) differences among the groups. When
RNA was then extracted by the modified method of Chomczyn-
ANOVA indicated significance, the Tukey-Kramer follow-up
ski and Sacchi (25) and quantified by ultraviolet absorbance. The
multiple comparisons test was used. Unless otherwise stated,
modified method has been described elsewhere (26). The yield of
results are presented as mean ± SD.
total cellular RNA was ?1 ?g/mg bone. The abundance of IGF-I

---

1360S
ARJMANDI ET AL
TABLE 1
Effects of ovariectomy and soy protein with normal and reduced iso?avone content on blood and urinary indexes in rats1
SHAM
OVX+CASEIN
OVX+SOY
OVX+SOY?
Measure
(n = 12)
(n = 12)
(n = 12)
(n = 10)
Serum
17?-Estradiol (pmol/L)
54.7 ± 18.0a (5.14)
12.1 ± 7.0b (2.13)
15.8 ± 5.8b (1.87)
6.5 ± 1.5b (0.51)
Alkaline phosphatase (U/L activity)
85.5 ± 15.4 (4.4)
81.9 ± 14.1 (4.3)
87. 5 ± 15.0 (4.3)
92.6 ± 13.5 (4.3)
Acid phosphatase (U/L activity)
6.0 ± 1.4a (0.4)
7.4 ± 1.1ab (0.4)
7.5 ± 1.8ab (0.5)
8.0 ± 1.3b (0.4)
Calcidiol (nmol/L)
56.8 ± 12.3 (3.54)
58.2 ± 10.5 (3.02)
57.7 ± 9.5 (2.75)
67.5 ± 10.6 (3.54)
Calcitriol (pmol/L)
56.3 ± 14.2b (4.11)
71.1 ± 20.4ab (5.88)
74.5 ± 16.4a (4.74)
61.8 ± 10.3ab (3.44)
Blood
Calcium ions (mmol/L)
0.75 ± 0.11 (0.03)
0.78 ± 0.09 (0.03)
0.86 ± 0.19 (0.06)
0.73 ± 0.15 (0.05)
pH
7.48 ± 0.05 (0.02)
7.58 ± 0.10 (0.03)
7.51 ± 0.12 (0.04)
7.53 ± 0.05 (0.02)
Urine
Urinary hydroxyproline (nmol/12 h)
252 ± 111c (32.0)
412 ± 162bc (46.8)
478 ± 246ab (71.0)
631 ± 124a (35.8 )
1x– ± SD. SEMs in parentheses. Values with different superscript letters are signi?cantly (P < 0.05) different from each other. SHAM, sham-operated fed
casein-based diet; OVX + CASEIN, ovariectomized fed casein-based diet; OVX+SOY, ovariectomized fed soy protein with normal iso?avone content;
OVX+SOY?, ovariectomized fed soy protein with reduced iso?avone content.
RESULTS
scripts were significantly (P < 0.001) greater in OVX+SOY and
Thirty-?ve days after surgery, 1 SHAM group and 1 of the OVX
OVX+SOY? treatments in comparison with OVX+CASEIN
groups were killed. The femoral densities in rats killed 35 d after
(Figure 3). The SOY diet was more effective (P < 0.001) in
surgery showed that the ovariectomized animals had signi?cantly
increasing both IGF-I mRNA transcripts than SOY?.
lower bone density than did the SHAM animals (1.460 ± 0.048
Serum 25-hydroxycholecalciferol (calcidiol) was unaffected
compared with 1.499 ± 0.053 g/cm3). Although all animals started
by treatments (Table 1). Although the mean calcidiol values were
at nearly equal mean weights (210.8–212.1 g), the SHAM animals
not significantly different among the groups, the mean 1,25-
ate less food (x
– ± SD, in g/d: SHAM, 13.0 ± 0.5; OVX+CASEIN,
dihydroxycholecalciferol (calcitriol) concentrations were lower
14.8 ± 1.9; OVX+SOY, 15.3 ± 0.5 g/d; and OVX+SOY?,
in the SHAM group than in the OVX+SOY group; the OVX and
15.1 ± 0.5) and gained less weight than the OVX animals regard-
OVX+SOY? groups had intermediate mean values compared
less of the treatments (x
– ?nal body wt ± SD, in g: SHAM,
with both the SHAM and OVX+SOY groups. Mean blood ion-
282.8 ± 13.4; OVX+CASEIN, 345.9 ± 20.3; OVX+SOY
ized calcium concentrations and pH values were similar among
339.7 ± 20.2; and OVX+SOY?, 343.3 ± 19.9).
all 4 groups. Although the mean serum alkaline phosphatase
Sixty-?ve days of soy feeding (either of the soy diets) did not
activities were similar among the treatment groups, serum tar-
alter the ovariectomy-induced declines in serum 17?-estradiol con-
trate-resistant acid phosphatase activity was significantly higher
centrations (Table 1) or the uterine atrophy (g/100 g body wt, in
in the OVX+SOY? than in the SHAM groups; the OVX and
x
– ± SD: SHAM, 0.197 ± 0.046; OVX+CASEIN, 0.034 ± 0.012;
OVX+SOY groups had intermediate values, not significantly dif-
OVX+SOY, 0.011 ± 0.029; and OVX+SOY?, 0.035 ± 0.012) indi-
ferent from either the SHAM or OVX+SOY? groups. Likewise,
cating a lack of true estrogenic activity of soy or its iso?avones.
urinary hydroxyproline, the other biochemical marker for bone
The right femoral bone lengths were greater for the
resorption, was significantly higher in both the OVX+SOY and
OVX+CASEIN, OVX+SOY, and OVX+SOY? than the SHAM
OVX+SOY? treatment groups (Table 1).
groups (Table 2). Fourth lumbar vertebral densities in all the
ovariectomized animals, regardless of treatment, were lower than
in the SHAM animals. Figure 1 shows that the OVX groups had
signi?cantly lower mean femoral bone density values than the
SHAM group. The soy-treated groups tended to have higher (albeit
not signi?cant) mean femoral densities than did the OVX group.
However, the right femoral bone densities in both soy treatment
groups were not different from either the SHAM or OVX groups.
There were no significant differences among the groups in
femoral calcium or phosphorus content, femoral dry bone
weight, or femoral ash weight (Table 2). However, the SHAM
group had significantly higher mineral contents of the right
femur and fourth lumbar vertebra than the other groups. In addi-
tion, mean ash weights of the fourth lumbar vertebrae were
significantly lower in the SOY? than in the SHAM group but
FIGURE 1. Mean (+ SE) femoral (white) and fourth lumbar vertebral
not significantly different from the other OVX groups (Table 2).
(black) densities. After the occurrence of bone loss (35 d postovariec-
Northern blot analysis of total cellular RNA isolated from
tomy), 2 of the ovariectomized (OVX) groups were fed diets in which
femurs detected 7.5 and 0.8 kilobase (kb) IGF-I mRNA tran-
casein was replaced with either soy protein with normal (SOY) or
scripts (Figure 2). The differences between OVX and SHAM
reduced (SOY?) isoflavone content. Rats were killed 65 days later.
rats were significant for both 7.5 kb (P < 0.04) and 0.8 kb (P <
*Significantly different from sham-operated groups (SHAM), P < 0.05;
0.005) for IGF-I transcripts. Both 7.5 kb and 0.8 kb mRNA tran-
n = 12 rats/group except the OVX+SOY? group in which n = 10.

---

ROLE OF SOY PROTEIN IN REVERSING BONE LOSS
1361S
TABLE 2
Effects of ovariectomy and soy protein with normal and reduced iso?avone on bone indexes in rats1
SHAM
OVX+CASEIN
OVX+SOY
OVX+SOY?
Measure
(n = 12)
(n = 12)
(n = 12)
(n = 10)
Right femur
Length (cm)
3.57 ± 0.04b (0.01)
3.67 ± 0.06a (0.02)
3.67 ± 0.08a (0.02)
3.72 ± 0.08a (0.02)
Calcium (mmol/cm3 bone vol)
2.71 ± 0.90 (0.26)
2.87 ± 0.79 (0.25)
2.38 ± 1.45 (0.44)
2.06 ± 1.04 (0.33)
Phosphorus (mmol/ cm3 bone vol)
5.28 ± 0.85 (0.24)
5.96 ± 0.72 (0.21)
5.54 ± 1.03 (0.30)
6.15 ± 0.76 (0.24)
Dry bone wt (g)
0.529 ± 0.027 (0.008)
0.535 ± 0.025 (0.007)
0.544 ± 0.029 (0.009)
0.550 ± 0.026 (0.008)
Ash wt (g)
0.356 ± 0.018 (0.005)
0.351 ± 0.018 (0.005)
0.359 ± 0.02 (0.006)
0.363 ± 0.018 (0.006)
Mineral content [%(ash wt/dry wt ? 100)]
67.40 ± 0.77a (0.22)
65.59 ± 1.24b (0.36)
65.92 ± 0.59b (0.17)
65.97 ± 0.77b (0.24)
Fourth lumbar vertebra
Dry bone wt (g)
0.195 ± 0.014 (0.004)
0.200 ± 0.014 (0.004)
0.199 ± 0.014 (0.004)
0.204 ± 0.011 (0.004)
Ash wt (g)
0.119 ± 0.013a (0.004)
0.106 ± 0.009ab (0.003)
0.103 ± 0.010ab (0.003)
0.095 ± 0.023b (0.007)
Mineral content [%(ash wt/dry wt ? 100)] 58.30 ± 0.59a (0.19)
53.12 ± 1.46b (0.42)
51.35 ± 4.62b (1.46)
48.85 ± 7.48b (2.49)
1x– ± SD. SEMs in parentheses. Values with different superscript letters are signi?cantly (P < 0.05) different from each other. SHAM, sham-operated fed
casein-based diet; OVX+CASEIN, ovariectomized fed casein-based diet; OVX+SOY, ovariectomized fed soy protein with normal iso?avone content;
OVX+SOY?, ovariectomized fed soy protein with reduced iso?avone content.
DISCUSSION
gate for growth-hormone production (33), which increases bone
The present study indicated that SOY or SOY? diets were
size but not necessarily bone density. Whereas the percentage
somewhat effective in reversing the femoral but not the fourth lum-
mineral content for each of the 4 groups paralleled the bone-den-
bar bone-density loss. This may be partially due to increased
sity values of each group, the calcium and phosphorus contents
femoral IGF-I mRNA transcripts by both the SOY and SOY?
of the femur, corrected for bone volume, did not parallel either
diets, albeit the SOY more potently than the SOY? diet. IGF-I is
the percentage mineral content or the femoral density. This sug-
known to correlate with both bone mineral density and the rate of
gests that mineral content reflects bone density more so than
bone formation (26, 34). Overall, the ?ndings of this study support
either calcium or phosphorus content.
the notion that the best and most effective approach to combating
The ovariectomized animals, regardless of treatment, showed
osteoporosis is through prevention (35). Once bone loss has
uterine atrophy, indicating that the isoflavones in soy, unlike
occurred, its reversal may be dif?cult. Considering both the slightly
estrogen, are not uterotrophic. These observations were sup-
higher femoral densities and the signi?cantly higher femoral IGF-I
ported by the finding that 17?-estradiol concentrations were not
mRNA concentrations in animals that received either SOY or
influenced by the soy diets in the ovariectomized animals. Addi-
SOY? diets, we suggest that long-term consumption of soy or its
tionally, significantly lower serum 17?-estradiol concentrations
iso?avones produce small but continued increments in bone mass.
in the OVX groups were not affected by the either soy diet. It is
Although it was recently shown that soy isoflavones were
widely known that estrogen replacement maintains or increases
effective in reversing lumbar spine bone loss in postmenopausal
both uterine weight and serum 17?-estradiol concentrations (40,
women (JW Erdman et al, personal communication, 1996), this
41). Therefore, these findings suggest that the mechanism of
animal study did not show similar effects. This may be due to a
action for soy isoflavones is different from that of estrogen in
reduced potency of the soy isoflavones used in this study. The
uterine tissue and blood.
SOY and SOY? diets were stored for a prolonged period (?1 y
in 45 ?C) before their use in this experiment. In support of this
statement, Gallaher et al (36) observed that aged soy protein
isolate was less effective in reducing the number of aberrant
crypts in a colon cancer study. The findings of a recent study
also indicated that soy protein, poor or rich in isoflavone, could
not prevent ovariectomy-induced bone loss in ovariectomized
monkeys (37). This may be partly due to the fact that
isoflavones present in soy protein may exert their positive
effects on bone in a biphasic manner.
Anderson and Garner (38) and Fanti et al (39) showed that the
positive effect of genistein on bone is dose dependent. Anderson
and Garner (38) found that higher doses of genistein did not pre-
vent ovariectomy-induced bone loss. This dose-dependent effect
of isoflavones on bone can be substantiated by the recent findings
FIGURE 2. Effects of ovariectomy (OVX), soy protein diet with nor-
that adding a higher dose of isoflavone (2.2 mg/g protein) to the
mal isoflavone content (SOY), or reduced isoflavone content (SOY?) on
percentage changes in 0.8 kilobase (kb) (white) and 7.5 kb (black)
diets of postmenopausal women increased lumbar bone-mineral
femoral IGF-I messenger RNA (mRNA) transcripts obtained by North-
content and density whereas a lower dose (1.39 mg/g protein) had
ern blot analysis. IGF-I mRNA was quantitated by phosphorimaging the
no effect (JW Erdman et al, personal communication, 1996).
membranes and was normalized to 28S ribosomal RNA to correct for
However, in their study, the optimal dose was not established.
differences in total RNA. Bars are mean + SE; n = 4 animals per treat-
The increase in femoral length is expected because ovariec-
ment group. Bars with different letters are significantly (P < 0.05) dif-
tomy in this animal model enhances synthesis of IGF-I, a surro-
ferent from each other.

---

1362S
ARJMANDI ET AL
REFERENCES
1. Avioli LV. Significance of osteoporosis: a growing international
health care problem. Calcif Tissue Int 1991;49(suppl):S11–4.
2. Bonnick SL. AMWA position statement on osteoporosis. J A Med
Womens Assoc 1990;45:75–9.
3. Mundy GR. Visions for the future in osteoporosis research. Osteo-
poros Int 1993;2(suppl):S29–S34.
4. Tolstoi LG, Levin RM. Osteoporosis: the treatment controversy.
Nutr Today 1992;27:6–12.
5. Lindsay R. Prevention and treatment of osteoporosis. Lancet
1993;341:801–5.
6. Raisz LG. Bone cell biology: new approaches and unanswered
questions. J Bone Miner Res 1993;8(suppl 2):S457–65.
FIGURE 3. Effects of ovariectomy (OVX) with a casein diet
7. Gallagher JC. The pathogenesis of osteoporosis. Bone Miner
(CASEIN), a soy protein diet with normal isoflavone content (SOY), and
1990;9:215–27.
a soy protein diet with depleted isoflavone content (SOY?) on Northern
8. Eriksen EF, Axellrod DW, Melsen F. Bone histomorphometry. New
blots of femur total cellular RNA probed with 32P-labeled complemen-
York: Raven Press, 1994.
tary DNA for rat IGF-I (left panel) and 28S (right panel) RNA. Each lane
9. Hunt SM, Groff JL. Advanced nutrition and human metabolism.
represents RNA extracted from a femur of one animal.
New York: West Publishing Co, 1990.
10. Teitelbaum SL. Skeletal growth and development. In: Primer of
metabolic bone diseases and disorders of mineral metabolism. New
Serum alkaline phosphatase activity, an index of bone for-
York: Raven Press, 1991:7–11.
mation, tended to be higher (albeit not significantly) in the soy-
11. Watts NB, Harris ST, Genant HK, et al. Intermittent cyclical
fed animals compared with the OVX+CASEIN group, suggest-
etidronate treatment of postmenopausal osteoporosis. N Engl J Med
1990;323:73–9.
ing a modest positive effect on bone formation, which is
12. Turner CH. Toward a cure for osteoporosis: reversal of excessive
supported by the femoral bone density data. The greater urinary
bone fragility. Osteoporos Int 1991;2:12–19.
hydroxyproline concentrations in the OVX rats, including the
13. Grisanti JM. Some new views on preventing osteoporosis. Contemp
soy-treated groups, showed that bone resorption was also
Ob/Gyn 1989;33:60–78.
greater in these animals. Hence, our data suggest that treatment
14. Wardlaw GM. Putting osteoporosis in perspective. J Am Diet Assoc
with soy, regardless of isoflavone content, does not slow down
1993;93:1000–6.
the ovariectomy-induced higher rates of bone turnover. Interest-
15. Kalu DN, Masoro EJ, Byung PY, Hardin RR, Hollis BW. Modula-
ingly, it was the OVX+SOY? group that had the highest excre-
tion of age-related hyperparathyroidism and senile bone loss in Fis-
tion of hydroxyproline, indicating that there was an isoflavone
cher rats by soy protein and food restriction. Endocrinology
1988;122:1847–54.
effect beyond that of soy protein. We have no explanation for
16. Omi N, Aoi S, Murata K, Ezawa I. Evaluation of the effect of soy-
the seemingly separate effect of soy protein and isoflavones on
bean milk and soybean milk peptide on bone metabolism in the rat
bone resorption. Nonetheless, soy protein may indeed moder-
model with ovariectomized osteoporosis. J Nutr Sci 1994;
ately increase bone turnover, with bone formation exceeding
40:201–21.
bone resorption as reflected in the slightly greater femoral bone
17. Arjmandi BH, Alekel L, Hollis BW, et al. Dietary soybean protein
densities of the soy groups. In this sense, isoflavones appear to
prevents bone loss in an ovariectomized rat model of osteoporosis.
act differently from estrogen, which is known to attenuate the
J Nutr 1996;126:161–7.
dramatic rise in bone turnover that normally accompanies
18. Goyal N, Getlinger MJ, Sun P, et al. Effect of soy protein, with and
menopause. Soy protein, rich in isoflavones, may stimulate
without isoflavonoids, on bone in ovariectomized rats. J Bone Miner
Res 1995;10(suppl 1):S453 (abstr).
growth hormone, which in turn may stimulate (42, 43) bone for-
19. Petilli M, Fiorelli G, Benvenuti S, Frediani U. Interactions between
mation. The bone-forming properties of these compounds have
ipriflavone and the estrogen receptor. Calcif Tissue Int
not been adequately evaluated.
1995;56:160–5.
The results of this relatively short-term study suggest that for
20. Benvenuti S, Petilli M, Frediani U, Tanini A. Binding and bioeffects
soy protein to reverse bone loss, long-term consumption may be
of irpriflavone on a human preosteoclastic cell line. Biochem Bio-
necessary. Longer-term and dose-response studies using freshly
phys Res Commun 1994;201:1084–9.
produced soy products are needed to evaluate whether soy protein
21. Civitelli R, Abbasi-Jarhomi SH, Halstead LR, Dimarogonas A. Ipri-
or its iso?avones are effective in restoring bone after loss has
flavone improves bone density and biomechanical properties of
occurred as a result of ovarian hormone de?ciency. Furthermore, to
adult male rat bones. Calcif Tissue Int 1995;56:215–9.
establish the role of soy protein in bone formation, consideration of
22. Valente M, Bufalino L, Castiglione GN, D’Angelo R. Effects of 1-
year treatment with ipriflavone on bone in postmenopausal women
the practical, safe, and effective amounts of soy and iso?avones for
with low bone mass. Calcif Tissue Int 1994;54:377–80.
inclusion in a daily diet regimen (or supplement) is warranted.
23. Kalu DN, Liu CC, Salerno E, Hollis B, Echon R, Ray M. Skeletal
Bone-density assessments do not represent the whole picture.
response of ovariectomized rats to low and high doses of 17?-estra-
Therefore, it is also important to assess the effects of soy on the
diol. Bone Miner 1991;14:175–87.
biomechanical properties of bones as well to judge the quality of
24. Taussky HH, Shorr E. A microcolorimetric method for the determi-
the bone that is produced. The implication for human use would
nation of inorganic phosphorus. J Biol Chem 1953;202:675–85.
be to examine whether estrogen could be used in conjunction
25. Chomczynski P, Sacchi N. Single-step method of RNA isolation by
with a diet containing soy products.
guanidinium thiocyanate-phenol-chloroform extraction. Anal
Biochem 1986;162:156–9.
We gratefully acknowledge the assistance of J Artwohl, D Amin, and T
26. Barengolts EI, Kouznetsova T, Segalene A, Lanthon P, Odvina C,
Kouznetsova.
Kukreja SC, Unterman TG. Effects of progesterone on serum levels

---

ROLE OF SOY PROTEIN IN REVERSING BONE LOSS
1363S
of IGF-I and on femur IGF-I mRNA in ovariectomized rats. J Bone
growth factor-I and insulin-like growth factor binding protein-3.
Miner Res 1996;11:1406–12.
Bone Miner 1994;24:135–48.
27. Roberts CT Jr, Lasky SR, Lowe WL Jr, Seaman WT, Leroith D.
35. Kanis JA. Estrogens, the menopause, and osteoporosis. Bone
Molecular cloning of rat insulin-like growth factor 1 complimentary
1996;19:185S–90S.
deoxyribonucleic acid: Differential messenger ribonucleic acid pro-
36. Gallaher DD, Gallaher CM, An Z, Hoffman RM. Soy protein iso-
cessing and regulation by growth hormone in extrahepatic tissues.
lates and genistein: Effects on initiation, promotion, and progres-
Mol Endocrinol 1987;1:243–8.
sion of colon cancer. Am J Clin Nutr 1998;68(suppl):1524S (abstr).
28. Unterman TG, Jentel JJ, Oehler D, Lacson RG, Hofert JF. Effects of
37. Jayo MJ, Anthony MS, Register TC, Rankin SE, Vest T, Clarkson
glucocorticoids on circulating levels and hepatic expression of
TB. Dietary soy isoflavones and bone loss: a study in ovariec-
insulin-like growth factor-I binding proteins and IGF-I in the
tomized monkeys. J Bone Miner Res 1996;11:S228 (abstr).
38. Anderson JJ, Garner SC. The effects of phytoestrogens on bone.
adrenalectomized streptozotocin-diabetic rat. Endocrinology
Nutr Res 1997;17:1617–32.
1993;133:2531–9.
39. Fanti O, Faugere MC, Gang Z, Schmidt J, Cohen D, Malluche HH.
29. Ney DM, Yang H, Smith S, Unterman TG. High-calorie total par-
Systematic administration of genistein partially prevents bone loss
enteral nutrition reduces hepatic insulin-like growth factor-I mRNA
in ovariectomized rats in a nonestrogen-like mechanism. Am J Clin
and alters serum levels of insulin-like growth factor binding protein-
Nutr 1998;68(suppl):1517S (abstr).
1, -3, -5, and -6 in the rat. Metabolism 1995;44:152–60.
40. Kalu DN, Salerno E, Liu CC, et al. A comparative study of the
30. Hollis BW. Assay of circulating 1,25 dihydroxy vitamin D involv-
action of tamoxifen, estrogen and progesterone in ovariectomized
ing a novel single-cartridge extraction and purification procedure.
rat. Bone Miner 1991;15:109–24.
Clin Chem 1986;32:2060–3.
41. Ettlinger B, Golditch IM, Friedman G. Gynecologic consequences
31. Hollis BW, Kamerud JQ, Selvaag SR, Lorenz JD, Napoli JL. Deter-
of long-term unopposed estrogen replacement therapy. Maturitas
mination of vitamin D status by radioimmunoassay with an 125I-
1988;10:271–82.
labeled tracer. Clin Chem 1993;39:529–33.
42. Andreassen TT, Jorgensen PH, Flyvbjerg A, Orskov H, Oxlund H.
32. Kivirikko KI, Laitinen O, Prockop DJ. Modification of a specific
Growth hormone stimulates bone formation and strength of cortical
assay for hydroxyproline in urine. Anal Biochem 1967;19:249–55.
bone in aged rats. J Bone Miner Res 1995;10:1057–67.
33. Daniel WE. Biostatistics: a foundation for analysis in the health sci-
43. Wright NM, Renault J, Hollis B, Bell NH, Key LL. Effect of growth
ences. New York: John Wiley & Sons, 1991.
hormone on bone: bone mineral density, trabecular bone volume,
34. Kalu DN, Arjmandi BH, Liu CC, Salih MA, Birnbaum RS. Effects
and alkaline phosphatase improve or are restored in the dwarf rat
of ovariectomy and estrogen on the serum levels of insulin-like
treated with growth hormone. J Bone Miner Res 1995;10:127–31.

---