This study aimed to investigate the effect of estrogen withdrawal on bone tissue in adult female marmoset monkeys. In a 1-year follow-up study we used quantitative computer tomography to measure total bone mineral density (BMD) of the proximal tibia and the second-last lumbar vertebral body (L5/L6) before and 1, 3, 6, and 12 months after ovariectomy. Body mass did not significantly change during the 1-year observation period. However, a significant decline of total BMD after ovariectomy was observed in the proximal tibia but not in L5/L6. In addition, regression analysis showed a significant positive relationship between BMD and body mass in both tibia and L5/L6. The results of our study support the idea that ovariectomized marmoset monkeys may serve as a model to investigate bone loss related to decline of estrogen production.
More than 70 years ago the link between bone loss and estrogen depletion in
women was first described (Albright et al., 1940; Reifenstein and Albright,
1947). Today we know that in about 20 % of women, osteoporosis develops
after menopause when estrogen production ceases because of high bone
turnover and excess resorption (IOF, 2018). In
preclinical osteoporosis research, the ovariectomized rat is an accepted and
valuable animal model (Levolas et al., 2008; Giardino et al., 1993; Kalu,
1991). However, in the
Among nonhuman primates, the common marmoset monkey (
The question of whether induction of estrogen deficiency would induce loss of bone mineral density in adult female marmoset monkeys was addressed in a recent study by Saltzman et al. (2018). The reproductive physiology in adult female marmosets is significantly triggered by social status. Physiologically, subordinate individuals are characterized by prolonged periods of anovulation which are, among others, associated with very low concentrations of estrogen and their reproductive activity may remain suppressed for 2 years or more (Abbott and George, 1991). In their present study Saltzman et al. (2018) show that socially or surgically induced hypoestrogenism is not associated with adverse skeletal consequences such as lower bone mineral density in the lumbar vertebrae (L5/L6) of female marmosets. This is a very interesting result as it demonstrates for the first time in female primates the conservation of bone mass despite estrogen deficiency. In line with these findings we could not observe in our present 1-year follow-up study a significant decline of total bone mineral density after ovariectomy (ovx) in L5/L6. However, a significant decline of total bone mineral density after ovariectomy was observed in the proximal tibia raising the question of which parts of the skeleton may be suitable or even better to study the effects of estrogen deficiency in female marmosets.
All experiments were performed in accordance with the European Communities Council Directive 86/609/EEC and the German legislation on animal rights and welfare and were approved by the Lower Saxony Federal State Office for Consumer Protection and Food Safety, Oldenburg, Germany. Whenever applicable, ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines were followed.
Six adult female common marmoset monkeys (
The experimental rooms and the cages were cleaned at weekly intervals and
disinfected using water and Biguacid (Antiseptica, Brauweiler (Pulheim),
Germany). The room temperature was maintained at
A pelleted marmoset diet (ssniff Spezialdiäten GmbH, Soest, Germany;
Water was offered ad libitum via a drinking bottle.
To keep the number of experimental animals at a minimum we decided not to use a control group. Instead, the animals served as their own controls and returned to the animal colony after the observation time of 12 months.
To evaluate the influence of age and nonbreeding on body mass data, 23
intact females (age 2–8 years) from the same colony housed in opposite-sex
pairs and treated with PGF2
Bilateral ovariectomy was performed under general anesthesia using a ventral
midline approach and approved standard methods. The success of the
ovariectomy was verified by measurement of serum 17
Quantitative computerized tomography (qCT) was performed 1–2 weeks before
and 1, 3, 6, and 12 months after ovariectomy. Tomography was performed in
the morning under general anesthesia after an overnight food, but not water,
withdrawal. For anesthesia, animals received intramuscular injections of
10
qCT was performed using the XCT 2000 (Type 803100, Stratec Inc., Pforzheim,
Germany). For the first scan of the proximal metaphysis of the left tibia,
the scanner was positioned at the knee bend and a coronal computed
radiograph (scout view) in the distal direction was generated. The scout
view was used to position the scanner at the site of measurement, as
illustrated in Fig. 1. Three tomographic slices, one in the reference line,
one 1.0
Tomographic scout view of the distal femur and the proximal tibia
of a marmoset monkey recorded by quantitative computer tomography. The arrow
points to the left proximal tibia. Radiographs were taken in the reference
line (solid line) and 1.0
Image acquisition, processing, and calculation of the results were performed
using the software package XCT 5.40 (Stratec Inc.). The software separates
at the outer borderline of the bone all voxels located in the soft tissue
below a defined density threshold (280
Tomographic scout view of the lower part of the body of a marmoset
monkey showing bone tissue of the pelvic, lumbar, and sacral spine. The
arrow points to the position of the reference line (solid line) which equals
the center of the second-last lumbar vertebral body. Rostral and caudal of
this line, additional cross sections were taken at a distance of 1.0
Because the number of lumbar vertebrae in marmoset monkeys is 6 or 7 (Wagner
and Kirberger, 2005; Casteleyn et al., 2012), and the individual number of
lumbar vertebrae was not known, we could not specify whether L5 or L6 was
scanned. The reference line of the scanner was positioned in the center of
the second-last vertebral body with cross sections taken 1.0
Because appositional bone growth is absent in adult marmoset monkeys (Bagi
et al., 2007), the perimeter of a given cross section is expected to remain
relatively constant over time. Thus, bone perimeter is not expected to
differ between measurements when the scanner positioning is the same. It was
postulated that reliable data could be expected when the mean bone
circumference showed no tendency to decrease or increase significantly
(
Linear regression analysis and paired
Body mass was not significantly affected during the 1-year observation period (Fig. 3). However, 1 month after ovariectomy a small, but not significant, decrease in body mass was observed. Obviously the observed variation of body mass increased with time and was highest 12 months after ovx.
Body mass of six female marmoset monkeys before and during 12 months after ovariectomy.
Analysis of body mass development in 23 intact females from the same colony
housed in opposite-sex pairs and treated with PGF2
As mentioned above, the bone perimeter was used as a measure of reproducibility of the scanning position. The perimeters of the proximal tibia at the scanning positions did not differ between months 0, 1, 3, 6, and 12 after ovariectomy (data not shown).
Total bone mineral density (BMD) of the tibia declined linearly with time
and reached its lowest value at 12 months after ovx (Fig. 4b). Total tibial
BMD was significantly (
Total mineral density and regression line in proximal tibia vs. body mass
Total mineral density in proximal tibia and in the second-last
lumbar vertebrae (L5/L6) in six female marmoset monkeys before (0 months) and 12 months after ovariectomy (
There was no significant effect on total BMD in L5/L6 before (0 months) and 12 months after ovariectomy (Figs. 4d, 5). In addition, perimeters of the
cross sections did not differ before (0 months) and 12 months after ovx (
Total BMD in the proximal tibia and in the lumbar vertebrae increased with
increasing body mass (Fig. 4a and c). Both relations, i.e., between total
BMD in proximal tibia and body mass as well as total BMD in the lumbar
vertebrae and body mass, could be described by linear regression equations.
For these linear regressions all data of total BMD in tibia or L5/L6 were related to the respective body mass data. In these one
factorial approaches the time after ovx was not taken into account. By
minimizing the squares of deviations, the best fitting regression line was
calculated. For total BMD vs. body mass this equation, for example, had the
following form: total BMD
Regression statistics of total mineral density in proximal tibia and in the second-last lumbar vertebrae (L5/L6) vs. body mass or vs. months after ovariectomy. Data plots and regression lines are shown in Fig. 4a–d.
The two main results of this investigation are that (1) female marmoset monkeys lose bone mass in the proximal tibia within 1 year after ovariectomy (Figs. 4b and 5) but not in the lumbar vertebra (Figs. 4d and 5) and that (2) body mass is a strong predictor of bone mass in the lumbar vertebrae and in the tibia.
Body mass was unaffected by ovx (Fig. 3). This is in line with a recent report showing that body mass did not change 6 to 7 months following ovariectomy (Saltzman et al., 2018). Thus, the changes in total bone mineral density observed in this study after ovx were not secondary to changes in body mass.
Our result that the lumbar vertebrae did not respond to ovariectomy within 12 months after ovx is consistent with the findings of Colman and coworkers, who reported that, even after 2 years, ovariectomy did not affect vertebral bone mineral density in marmoset monkeys (Colman et al., 1997). A similar finding was reported recently by Saltzman et al. (2018).
In women, high body mass protects to some extent from postmenopausal bone loss (Rico et al., 2002; Cifuentes et al., 2003). Because women produce increasing amounts of estrogen with increasing size of the fat depots (Baglietto et al., 2009) their body mass had a greater impact on BMD than the interval since menopause (Cifuentes et al., 2003). One-way linear regression analysis (Fig. 4a, c, Table 1) showed that a high body mass had a positive significant effect on BMD in marmoset monkeys also.
From this we conclude that lean marmoset monkeys may be more suitable for studying bone loss related to estrogen deficiency than overweight or obese animals and that weight gain during such studies should be avoided in this setting. Our current data show that estrogen withdrawal in female marmoset monkeys induces loss of bone mass in the axial skeleton as it is generally observed in postmenopausal women (Ji and Yu, 2015). The lack of loss of bone mass in the vertebral bodies differs from what is seen in postmenopausal women.
In contrast to reports in rats (Bonnet et al., 2006), we found that in adult female marmoset monkeys BMD does not correlate with age (data not shown) and that bone perimeter remained constant during the experimental period. Rat models show an increase in circumference (Kalu, 1984) and length (Waarsing et al., 2006) of the tibia with age due to appositional bone growth. This is not the case in the marmoset tibia. This view is supported by the finding that epiphyseal closure in the long bones occurs not later than the age of 1.8 years in marmoset monkeys (Kohn et al., 1997). The absence of a significant relationship between bone perimeter and time is one more similarity to human bone biology (Grohmann et al., 2012). Several studies have been carried out in ovariectomized or postmenopausal nonhuman Old World monkeys such as baboons and macaques (Brommage, 2001; Turner, 2001). These species lose bone mass after castration (Jerome et al., 1995; Hordon et al., 2006). However, osteoporosis studies in Old World monkeys are long lasting and may be confounded by age-related effects (Lundon et al., 1994; Cahoon et al., 1996; Champ et al., 1996). For example, in rhesus monkeys, closure of the epiphyseal plate takes place at about 6 years of age but further bone mass is gained up to 9–15 years of age and is lost thereafter (Jerome et al., 1995; Cahoon et al., 1996; Champ et al., 1996). Compared with macaques, marmoset monkeys have an earlier puberty and sexual maturity (Abbott et al., 2003) and presumably achieve earlier peak bone mass. In addition, the animals are easy to breed and to handle under controlled laboratory housing conditions. Therefore, they may have advantages over macaques in preclinical osteoporosis research.
In a recent study, Saltzman and coworkers reported that estrogen depletion is not associated with lower bone mass in female marmoset monkeys (Saltzman et al., 2018). However, their conclusion was drawn from studying lumbar vertebrae L5/L6. Our results support the view that the effect of estrogen withdrawal is not necessarily the same in all localizations of the skeleton. In line with their findings we also could not observe an effect of ovariectomy on total bone mineral density of L5/L6 (Fig. 5). Our study, however, shows that bone tissue of the tibia of adult female marmoset monkeys is sensitive to estrogen withdrawal and is setting up the question of whether other parts of the skeleton may also be suitable or even better to study the effects of ovariectomy.
The crucial differences between our study and that by Saltzman and coworkers are the use of two different imaging techniques: in this case qCT versus DXA (dual X-ray absorptiometry), a 12-month versus a 6–7-month duration of surgically induced estrogen depletion, and housing the animals in opposite-sex pairs versus group housing with sexually dominant and subordinate females. Depending on the imaging technique and measuring site, different information on bone quality is obtained. Amstrup et al. (2016) investigated in postmenopausal women to what extend the results from these imaging techniques correlate. They found a good correlation between the methods when assessing the same skeletal site. However, when assessing correlations between different sites, central and distal sites, the associations were only weak to moderate. This may explain why the results of our study and that by Saltzman and coworkers reveal similar results for the effect of long-term estrogen depletion on bone mineral density in lumbar vertebrae L5/L6. It would be interesting to see if a similar agreement – in this case reduction of bone mineral density – can be found when measuring the proximal tibia with DXA.
The results presented in this 1-year follow-up study support the idea that the ovariectomized marmoset monkey is a promising nonrodent model for preclinical testing of anti-osteoporosis drugs. However, for the development of a standardized preclinical model we will need more information about the marmoset skeleton, which should be studied in more detail with respect to localizations of post-castrational bone loss; extent of pre- and post-castrational remodeling; and distributions of cytokine, steroid, and peptide hormone receptors. Undoubtedly, bone of marmoset monkeys shows close similarities to human bone (Bagi et al., 2007; Grohmann et al., 2012), responds to standard therapeutics (Bagi et al., 2007), and is a valuable addition for preclinical research. Moreover, other menopausal symptoms such as hot flashes and sleep disturbances could also show to be inducible by estrogen withdrawal in the marmoset monkey (Gervais et al., 2016). These findings open an additional window for holistic approaches in preclinical drug testing which in the best case can answer more than one question about the efficacy of new therapies.
Data used in this paper are available in the Supplement.
The supplement related to this article is available online at:
Regression analysis of body mass versus age of 23 intact
PGF2
Regression analysis of body mass versus age of 23 intact
PGF2
Body mass in 23 intact PGF2
CS, DSW, and EF designed the experiments and CS and DSW carried them out. EF and CS prepared the article.
Eberhard Fuchs is coordinating editor of the journal
We thank Sabine Lüdemann and Julia Schell for their excellent technical assistance.
This research has been supported by the European Union (grant no. QLRI-CT-2002-02758 (EUPEAH)).
This paper was edited by Gerhard Weinbauer and reviewed by two anonymous referees.