Developing a Model of Nutritional Amenorrhea in Rhesus Monkeys
http://www.100md.com
《内分泌学杂志》
Departments of Physiology (M.E.L., D.A.V.V.) and Obstetrics and Gynecology (A.A.K., R.L.R., D.A.V.V.), Queen’s University, Kingston, Ontario, Canada K7L 3N6
Abstract
Nutritional amenorrhea is defined as cessation of menstrual cycles resulting from a chronic negative energy balance. Although it is agreed that nutritional amenorrhea results from reduced secretion of GnRH, the neuroendocrine mechanisms leading to GnRH inhibition are poorly defined. Because the invasiveness of many neuroendocrine experimental approaches precludes its use in the clinical setting, we set out to establish a model of nutritional amenorrhea in rhesus monkeys. Studies were conducted in four normal-weight and one obese female rhesus monkey. Dietary intake was gradually reduced with the goal of achieving a 15–20% weight reduction. Dietary restriction inhibited ovulation in all animals. The weight loss required to inhibit ovulation ranged from 2–11% in the four normal-weight animals and was achieved with a 23% reduction in dietary intake. The time of initiating reduced food intake to first missed ovulation was 62 ± 13 d. Greater weight loss (46% reduction) over a longer period (10 months) was required to inhibit ovulation in the obese monkey. The onset of anovulation was not preceded by changes in menstrual cycle length or progesterone secretion. Realimentation initiated ovulation at a weight that approximated the animal’s weight at the time of the last ovulatory cycle during dietary restriction. By contrast, caloric intake at the return of ovulation during realimentation was 28% greater. This is the first demonstration that chronic dietary restriction can inhibit ovulation in rhesus monkeys. This model will be useful for studying the neuroendocrine mechanisms involved in diet-induced anovulation in primates.
Introduction
CHRONIC NEGATIVE ENERGY balance results in the cessation of menstrual cycles in primates and of estrous cycles in nonprimate species (1). Anorexia nervosa and hypothalamic amenorrhea resulting from strenuous exercise are examples of nutritional amenorrhea caused by chronic energy deficits. Although amenorrhea is no longer considered a diagnostic criteria of anorexia nervosa, it remains a hallmark of anorexia nervosa, occurring in the vast majority of subjects when weight is less than 80–85% of ideal body weight (2). A high incidence of amenorrhea also occurs in women whose lifestyle includes strenuous exercise (3). Although the etiology and degree of negative energy balance differ in anorexia nervosa and exercise-induced amenorrhea, similarities between the two conditions exist. Both are characterized by decreased gonadotropin pulse frequency leading to suppression of ovarian cyclicity and impaired gonadal steroid production (3, 4, 5, 6, 7). Moreover, women whose life style includes a rigorous exercise component have a higher incidence of disordered eating (8).
Although many studies have described in detail the hypothalamic-pituitary-ovarian axis parameters associated with nutritional amenorrhea (3, 4, 6, 9, 10, 11) these studies cannot address the neuroendocrine events that lead to amenorrhea. Investigations of the neuroendocrine mechanisms of nutritional amenorrhea in women are limited by the inability to conduct prospective studies. In addition, a thorough investigation of neuroendocrine mechanisms involves a level of invasiveness unacceptable in the clinical setting. Although inhibition of the reproductive axis by reduced energy intake has been demonstrated in several rodent species (12), accurate translation of results to humans is uncertain given the many differences between the primate menstrual cycle and rodent estrous cycles. Nonhuman primate studies in this area are limited and primarily acute in nature. It has been extensively documented that acute fasting or glucoprivation can inhibit gonadotropin secretion in rhesus monkeys (13, 14, 15, 16). However, the few chronic studies conducted in nonhuman primates to date demonstrated different menstrual cycle responses to a chronic metabolic challenge. Chronic negative energy balance achieved by strenuous exercise inhibited ovulation in cynomolgus monkeys (17), whereas a 30% reduction in dietary intake did not inhibit menstrual cycles in rhesus monkeys (18).
Given the similarities between the reproductive axes of humans and nonhuman primates and the inhibitory effects of negative energy balance on reproductive cycles of women and nonprimate species, we hypothesized that dietary restriction would inhibit ovulation in rhesus monkeys. The purpose of the current study was to determine the degree of dietary restriction required to inhibit ovulation in rhesus monkeys. The level of dietary restriction required is important to the feasibility of establishing a model of diet-induced amenorrhea in rhesus monkeys, because the utility of such an animal model would require that ovulation was inhibited at a level of dietary restriction that did not unduly compromise animal health.
Materials and Methods
Animal husbandry
Studies were conducted in five adult female rhesus monkeys (Macaca mulatta). Monkeys ranged in age from 6–10 yr and weighed between 5.8–11.9 kg. Animals were housed in groups of two or three in a light- and temperature-controlled environment (lights on, 0700–1900 h; temperature, 22 C). Diets consisted of a twice-daily ration of Purina monkey chow (Hi Protein Monkey Diet Jumbo, Ralston Purina, St. Louis, MO) supplemented with fruits and vegetables. Water was available ad libitum. All animal husbandry practices and experimental procedures conformed to regulations of the Canadian Council on Animal Care and were approved by the Queen’s University animal care committee.
Normalizing body weight to a uniform body mass index
Group housing results in the establishment of hierarchies that can impact on food availability and consequently adiposity. The clinical impression of varying degrees of adiposity in these five monkeys was confirmed using a modified method of calculating body mass index (BMI). The weight in kilograms divided by crown-rump length in meters squared produces an index in rhesus monkeys that is highly correlated with percent body fat (19). Because BMIs were 39.3, 27.2, 25.0, 24.2, and 22.1 kg/m2, group-housed animals were separated during meal times (0900–1100 and 1400–1600 h) to standardize their weights and subsequently regulate their dietary intake. Four monkeys were initially placed on a daily ration of 15 biscuits (calculated to provide 1887 kJ/d), whereas the obese animal was restricted to 1415 kJ/d (12 biscuits). Diets were further individualized in the subsequent months to achieve a uniform BMI of 24–23 kg/m2, which translates into a lean body habitus. Adjustments to diets included 1) increasing the vegetable component, 2) modifying the number of monkey chow biscuits allotted, and 3) eliminating high carbohydrate treats to those monkeys whose BMI exceeded 24 kg/m2. Food intake was monitored daily, and monkeys were weighed monthly during the initial phase of the experiment using a transfer box that eliminated the need for sedation. Animals were trained to present a limb through an opening in the cage for blood sampling by venipuncture. Blood samples (1.5–2 ml) were drawn three times a week before the morning meal.
Effect of dietary restriction on ovulation in rhesus monkeys
Once a BMI of 24–23 kg/m2 was attained, animals were placed on a 16-wk regimen of dietary restriction with the goal of decreasing weight by 20%. This was accomplished by decreasing the amount of monkey chow allotted to each animal. The rate and extent of dietary restriction were dependent on the rate of weight loss in each animal. Every 2 wk, one cube of monkey chow was removed if weight loss had not exceeded the desired rate of decline. In each case, the aim of dietary restriction was to decrease weight by no more than 20% and dietary restriction by no more than 40% over the 16-wk period. An attempt was made to maintain the bulk of each animal’s diet by increasing the amount of low calorie vegetables given between the morning and afternoon feedings. An iron-rich vitamin supplement was given daily. Once animals became anovulatory, we attempted to maintain each animal’s weight just below (within 500 g) the animal’s weight at the time of the first missed ovulation (i.e. time of the expected luteal phase). Animals were maintained on dietary restriction for 8–10 months, during which time the effects of leptin administration on ovulation were tested (our manuscript in preparation). This was followed by a period of realimentation in which the goal was to determine the weight and dietary intake level at which ovulations returned. Throughout dietary restriction and realimentation, weights were measured weekly using a transfer box, and blood samples were drawn by venipuncture three times a week before the morning meal. Samples were assayed for progesterone and FSH. A veterinarian’s assessment was performed at the start of the study and at 6-month intervals. Hematology and blood chemistry (including liver and renal function tests) were performed at 6-month intervals.
RIAs
Blood samples were refrigerated overnight, and serum was isolated after centrifugation (2500 rpm). All serum samples were assayed for progesterone using a commercial RIA kit available from Diagnostic Products Corp. (Los Angeles, CA). The assay sensitivity was 0.02 ng/ml. The intraassay coefficient of variation (CV) was 3.6%, and the interassay CV was 3.9%. We have validated this progesterone assay for use in the nonhuman primate (20). A minimum of three consecutive serum progesterone values of 4.0 ng/ml or greater was considered indicative of the luteal phase. This cutoff was based on progesterone measurements over six consecutive cycles in five rhesus monkeys with regular ovulatory menstrual cycles (mean follicular phase progesterone, 1.8 ± 0.1 ng/ml; range, 1.0–3.7 ng/ml; mean luteal phase progesterone, 8.5 ± 0.4 ng/ml; range, 4.0–21.9 ng/ml). An integrated weekly FSH value was determined for serum pools generated by combining equal volumes of each serum sample collected during a 1-wk period. Serum pools were separated into follicular and luteal phases based on earlier progesterone measurements. Serum pools were assayed in duplicate for FSH in a single assay using reagents provided by the National Hormone and Pituitary Program. The standard curve consisted of FSH reference preparation AFP 6940A. Unknowns were incubated with recombinant FSH antibody (AFP 782594), followed by the addition of 125I-radiolabeled FSH (AFP 6940A). A sheep antirabbit -globulin (Prince Laboratories, Toronto, Canada) was used to precipitate the antigen-antibody complex. Precipitation was facilitated by adding 12.5% Carbowax (Sigma-Aldrich Corp., St. Louis, MO) before centrifugation. Assay sensitivity, defined as the amount of reference preparation required to reduce binding by 2 SD below the zero standard divided by the sample volume, was 0.1 ng/ml. The intraassay CV was 9.6%.
Statistics
A repeated measures ANOVA and Tukey’s multiple comparison test were used to determine differences in mean BMI and dietary intake at baseline, last ovulation, first missed ovulation, and first ovulation after realimentation. Menstrual cycle length, follicular phase length, luteal phase length, and mean luteal phase progesterone and FSH concentrations before, during, and after dietary restriction were compared using a repeated measures ANOVA and Tukey’s multiple comparison test. The level of significance was set at P 0.05. The effects of dietary restriction and realimentation on ovulation were assessed by visual inspection of the progesterone profiles.
Results
A moderate reduction in weight was required in three animals (monkeys 1, 2, and 3) to bring their BMI into the 23–24 kg/m2 range, whereas a moderate increase was required in one animal (monkey 4). These four monkeys were initiated on a fixed allotment of monkey chow totaling 1887 kJ/d. The effects of a regimented diet on BMI and progesterone secretion are presented in Figs. 1 and 2. BMI declined in the two monkeys depicted in Fig. 1. Monkey 2 consistently consumed less than the entire meal and consequently fell below the target BMI of 23–24 kg/m2 (Fig. 1B). This animal eventually increased her food consumption, and BMI increased to 22.1 kg/m2. Both animals exhibited a cyclical pattern of progesterone secretion indicative of ovulation. Eight ovulatory cycles were documented in each animal over the 8-month baseline period. The two monkeys depicted in Fig. 2 had discrepant responses to the fixed diet. Monkey 3 depicted in Fig. 2A had an initial BMI of 27.2 kg/m2, but underwent rapid weight loss when fed 1887 kJ/d. Two ovulations were documented in this animal before it became anovulatory. Caloric intake was increased to 2358 kJ/d, and its BMI increased from 21.3 to 24 kg/m2, at which point ovulation resumed. The second monkey depicted in Fig. 2 had an initial BMI of 22.1 kg/m2, which was below the desired BMI. This animal was anovulatory for the first 3 months of the study. BMI increased to 24.6 kg/m2 after 16 wk of individual feeding, and cyclical progesterone secretion indicative of regular ovulations ensued.
The effects of dietary restriction and subsequent realimentation on BMI and progesterone secretion in these four monkeys are presented in Figs. 3 and 4 (presented in the same sequence as in Figs. 1 and 2; data from Figs. 1 and 2 are included for added perspective). The two monkeys depicted in Fig. 3 exhibited a linear reduction in BMI during dietary restriction. Dietary restriction inhibited ovulation in both animals, as evidenced by an abrupt cessation in cyclic progesterone secretion. Despite a higher initial dietary intake and BMI, ovulation offset occurred sooner in monkey 1 than in monkey 2 (36 vs. 93 d). Dietary intake was increased subsequently in both animals to maintain their weight near the weight of ovulation offset. Therefore, dietary intake at the time of initiating realimentation in these two animals was similar to dietary intake at baseline. An additional increase in food intake during realimentation initiated ovulation. Ovulation onset occurred within a similar time frame as ovulation offset in these two animals (55 and 101 d to onset vs. 36 and 93 d to offset). Dietary intake, BMI, and progesterone secretion during food restriction and realimentation are shown in Fig. 4 for the other two monkeys (same animals as depicted in Fig. 2). Neither animal exhibited a linear reduction in BMI during dietary restriction. Monkey 3 (Fig. 4A) exhibited a delayed pattern of weight loss in response to dietary restriction. The first missed ovulation occurred after a very small decline in weight (BMI declined from 24 to 23.5 kg/m2). This monkey subsequently experienced a precipitous weight loss that required increasing the dietary intake from 13 to 18 biscuits/d over a 3-wk period to stabilize her weight. Weight declined in monkey 4 (Fig. 4B) during the initial stages of dietary restriction and then stabilized despite an additional reduction in dietary intake that resulted in anovulation. Realimentation stimulated ovulation in both animals. The times from initiating realimentation to first ovulation were 85 and 64 d (similar to the previous examples).
Weight and changes in weight associated with ovulation offset and onset are reported in kilograms and BMI units in Tables 1 and 2, respectively. Ovulation offset occurred at a mean weight of 5.8 ± 0.3 kg (BMI of 21.8 ± 0.9 kg/m2) compared with a starting weight of 6.2 ± 0.2 kg (BMI of 23.4 ± 0.4 kg/m2). This constituted an average weight loss of 0.4 ± 0.10 kg (0.1–0.7 kg). The weight at the time of the last luteal phase during caloric restriction was identical with the weight at the time of the first luteal phase during realimentation (5.9 ± 0.2 kg). Likewise, the change in BMI (BMI) was similar for ovulation offset and onset (1.6 ± 0.5 vs. 1.9 ± 0.3 kg/m2). The time from initiating dietary restriction or realimentation to ovulation offset and onset, respectively, correlated with the absolute change in weight (r = 0.93; P = 0.0009).
Dietary intake and changes in energy intake associated with ovulation offset and onset are shown in Table 3. The mean change in food consumption required to inhibit ovulation was similar to the magnitude of the increase that stimulated ovulation (443 ± 30 vs. 413 ± 102 kJ/d). However, the level of dietary intake at first ovulation during realimentation was considerably greater than the energy intake at offset [2123 ± 174 vs.1504 ± 148 kJ/d at the time of the expected luteal phase (Lutexp)]. Even when compared with the energy levels during the last follicular phase of dietary restriction, dietary intake at the time of first ovulation during realimentation was 25% greater (2123 ± 174 vs. 1710 ± 228 kJ/d). These results are shown graphically in Fig. 5. Dietary intake at first missed ovulation was significantly reduced compared with food intake before dietary restriction (Lutexp vs. basal, P < 0.01), but was not different from food intake at the time of the last ovulation during dietary restriction. Dietary intake at first ovulation during realimentation (cycle 1) was significantly greater than dietary intake at all relevant time points (most notable comparisons being the follicular phase of final ovulation during dietary restriction and the dietary intake at the time realimentation was initiated). The same comparisons were not statistically significant when performed for BMI.
The transition from ovulatory to anovulatory status (and vice versa) was abrupt. Mean menstrual cycle length, follicular phase length, luteal phase length, and luteal phase progesterone concentrations of the last two menstrual cycles observed during dietary restriction and the first two ovulatory cycles after realimentation were comparable to these same cycle parameters in cycles before and during the earlier stages of dietary restriction (Table 4).
A fifth animal included in this study was inordinately obese (BMI of 39.3 kg/m2). As shown in Fig. 6, this animal progressively lost weight and attained a lean BMI after 8–9 months on a diet of 1415 kJ/d. Cyclic progesterone secretion confirmed normal ovulatory function during this extended period of weight loss. An additional reduction in caloric intake inhibited ovulation. Anovulation occurred at a BMI of 21.5 kg/m2 (within the range of the four normal-weight monkeys). The animal remained anovulatory for several months while maintained on a diet of 1061–1533 kJ/d. She was removed from the study before realimentation due to development of a central nervous system infection.
Also shown in Fig. 6 are integrated weekly FSH concentrations for this animal. Several large excursions in FSH temporally associated with ovulations were seen. These large fluctuations were absent during the subsequent anovulatory period. Mean weekly FSH concentrations normalized to ovulation offset during dietary restriction and ovulation onset after realimentation are shown in Fig. 7. Mean FSH concentrations were more variable during the 10 wk before anovulation compared with the 15 wk of documented anovulation (see Fig. 7A). In addition, FSH levels determined at 4-wk intervals showed a significant decline in FSH secretion with anovulation (see inset). The onset of ovulation after realimentation was accompanied by a return to a variable pattern of FSH secretion (Fig. 7B).
The degree of dietary restriction necessary to inhibit ovulation was not associated with significant morbidity. Animals remained attentive to their surroundings and continued to use perches in the exercise pens. They retained their appetites throughout the study. Routine biochemical and hematological parameters remained within normal limits throughout the period of dietary restriction. No measures of bone metabolism were performed. One animal exhibited significant alopecia during dietary restriction, which was reversed by realimentation. Realimentation was accompanied by weight gain and resumption of ovulation. Animals remain in good health 2 yr after study completion.
Discussion
This study demonstrates that ovulation in rhesus monkeys was inhibited by a reduction in dietary intake. This is the first documentation that dietary restriction can inhibit ovulation in nonhuman primates. A previous study concluded that ovulation in rhesus monkeys was not inhibited by a 30% reduction in dietary intake (18). A greater degree of dietary restriction in the current study is the most likely explanation for the difference in outcomes. Anovulation occurred in our study when food intake was decreased by 21%, resulting in an average dietary intake of 1504 ± 148 kJ/d. By contrast, a 30% reduction in dietary intake in the study by Lane et al. (18) translated into a daily intake of 1709 ± 55 kJ. Although the level of dietary restriction employed was greater in terms of percent decrease from habitual levels, according to our study this absolute level of dietary intake would not be expected to inhibit ovulation.
Our results in rhesus monkeys complement those of Williams et al. (17), who observed that chronic strenuous exercise induced amenorrhea in cynomolgus monkeys. Their study concluded that anovulation resulted from a negative energy balance rather than some other factor associated with exercise, because ovulation was reinitiated in exercising monkeys by increasing food availability (21). In contrast to exercise-induced amenorrhea, our study showed that dietary restriction did not produce transitional menstrual cycles. Follicular phase length, luteal phase length, and luteal phase progesterone secretion in the last two cycles during dietary restriction were comparable to these cycle parameters before or during the early stages of dietary restriction. This observation is at variance with that by Williams et al. (17), who reported that follicular phase length was increased and plasma LH and luteal phase progesterone secretion were reduced in the menstrual cycle immediately preceding exercised-induced amenorrhea. The much shorter time span from initiation to anovulation observed with dietary restriction compared with exercise (62 ± 13 d vs. 14.3 ± 2.2 months, respectively) may account for the absence of transitional menstrual cycles in our study. A regimen of dietary restriction that produces over a 2-month period a negative energy balance that inhibits ovulation may be too short a time frame for borderline energy levels to produce transitional menstrual cycles. Because neither study quantified the degree of energy deficit, this explanation remains speculative. A second difference that merits consideration is the method by which a state of negative energy balance was achieved in the two studies. If the energy available to fuel processes such as ovulation is simply the difference between energy intake and energy expended, then a deficit, regardless of how it was achieved, should adversely affect ovulation. This premise is supported by the demonstration that LH pulse frequency in women was inhibited by a negative energy balance regardless of whether it was achieved by reducing energy intake or increasing energy expenditure (22). It is also supported by the two nonhuman primate studies combined, because ovulation was inhibited in both. However, it remains to be determined whether the subtle differences cited above are due to differences in how the energy deficits were achieved.
Our demonstration that dietary restriction reversibly inhibits ovulation in rhesus monkeys also establishes the link between energy intake and reproductive function in primates. Energy deficit has been documented previously as the root cause of reproductive axis dysfunction in women with hypothalamic amenorrhea resulting from self-imposed dietary restriction or strenuous exercise (3, 6, 23). Several studies have demonstrated that acute energy deficits resulting from short-term fasting or increased energy expenditure inhibited the hypothalamic-gonadotropic axis, as reflected by reduced LH pulse frequency in human and nonhuman primates (13, 22, 24, 25), although females may be more resistant to acute fasting-induced inhibition of LH secretion compared with males (26, 27, 28). Although the current study did not attempt to characterize changes in LH pulse frequency, inhibition of ovulation is the expected outcome if dietary restriction chronically reduced LH pulse frequency below the threshold that supports follicular maturation or that is necessary to generate an LH surge. We have demonstrated that a similar degree of dietary restriction in ovariectomized rhesus monkeys inhibits the elaboration of the estrogen/progesterone-induced gonadotropin surge (29). Anovulation was not preceded by a reduction in FSH concentrations. A reduced GnRH pulse frequency may explain the maintenance of FSH concentrations while inhibiting ovulation, because a low GnRH pulse frequency can stimulate FSH secretion while inhibiting LH secretion (30, 31, 32).
This study provides insight into the dietary intake required to support ovulation in rhesus monkeys. Three of four normal-weight monkeys ovulated when given a diet of 1887 kJ/d. Monkey 1 had a stable weight on this diet and ovulated regularly. Monkey 2 consumed less than the entire meal and lost weight. Eventually, food consumption increased to 1533 kJ/d, and weight stabilized at a BMI of 22.1 kg/m2. The decreased appetite followed by partial recovery may reflect acclimatization to being separated twice a day at mealtime. In addition, this animal was the smallest of the four, weighing approximately 10% less than the others. In contrast, monkey 4 gained weight and became ovulatory after 3 months on a diet of 1887 kJ/d. This response is probably the result of inadequate food intake in a communal setting being resolved by separating pairs at mealtime. Monkey 3 was exceptional because it underwent rapid weight loss and became anovulatory when fed 1887 kJ/d. Food intake was increased to 2358 kJ/d, and its BMI increased from 21.3 to 24 kg/m2, at which point ovulation resumed.
Although these data indicate that the level of energy intake is a critical determinant of ovulatory status, they also suggest that energy stores may play a role. Inadequate energy stores may explain why monkeys 3 and 4 did not ovulate for several months while on a diet that eventually stimulated ovulation. Both animals gained weight before ovulating. Conversely, monkey 2 continued to ovulate despite dietary intake being reduced 1 month earlier to 1179 kJ/d. Ovulation offset occurred without an additional reduction in dietary intake, but was accompanied by a modest weight loss. Much more striking was the delayed effect of dietary restriction on ovulation in the obese monkey. This animal continued to ovulate at regular intervals despite a 42% decline in weight (11.9 to 6.8 kg) over 10 months. By contrast, the normal-weight monkeys became anovulatory after an average weight loss of 0.43 ± 0.1 kg (0.1–0.7 kg). The extended period of weight loss in the obese monkey indicated that this animal was chronically exposed to a negative energy balance. Despite a chronic energy deficit, anovulation did not ensue until BMI reached a level similar to that at which the other animals became anovulatory. Our speculation that energy stores maintained ovulatory function during the first 9 months of dietary restriction is tempered by dietary intake having been further reduced just before the first missed ovulation. Therefore, anovulation may have been affected by an additional reduction in dietary intake alone or a reduction in dietary intake in combination with reduced availability of oxidizable metabolic fuels. Several reports suggest that fatness affects the response of the reproductive axis to caloric restriction. Fat hamsters were less susceptible to starvation-induced anestrus than lean hamsters (33). Lean women were more prone to disruption of the neuroendocrine axis in response to a 72-h fast compared with normal-weight women (34), whereas gonadotropin secretion in obese women was not affected by a 21-d fast (35). However, in distinction to the above findings, it has been reported that morbidly obese women who had a 25% weight loss after biliopancreatic diversion became amenorrheic despite still being obese (mean BMI of 35 kg/m2) (36).
The mean dietary intake during the follicular phase of the last menstrual cycle before ovulation offset was 1725 ± 225 kJ/d. However, there was considerable variability (1179, 1592, 1887, and 2241 kJ/d). In fact, of the three measures (weight, BMI, and dietary intake), the variability in dietary intake was considerably greater than that for weight and BMI (see SEM expressed as a percentage of the mean in Tables 1–3). Normalizing dietary intake to weight or BMI reduced variability only slightly. Normalizing dietary intake to fat-free mass or some other measure of body composition may better predict the threshold for ovulation in future studies.
The inhibitory effect of dietary restriction on ovulation was reversed by realimentation. The absolute change in food intake ( kJ) and weight ( BMI) associated with ovulation during realimentation was similar to the absolute change in these same variables associated with inhibition of ovulation during dietary restriction. However, ovulation occurred at a significantly higher energy intake level during realimentation compared with the energy consumption during the final follicular phase of dietary restriction (2123 ± 174 kJ vs. 1725 ± 225 kJ). This difference was seen in all animals. The hysteresis exhibited in our dataset is not due to a lag effect caused by a faster rate of realimentation. The rate of realimentation (kJ/d of realimentation to ovulation onset, 5.2 ± 1.1 kJ/d) was actually slower than the rate of dietary restriction (kJ/d of food restriction to ovulation offset, 7.1 ± 1.1 kJ/d). This finding may have implications for dietary requirements to reestablish menstrual cyclicity in women with anorexia nervosa and is consistent with the report that return of menses in women with anorexia nervosa occurred at a mean weight that was 2.05–3.6 kg greater than the weight at which menses were interrupted (37, 38).
In summary, moderate dietary restriction of lean rhesus monkeys reliably inhibited ovulation. The level of dietary restriction required to inhibit ovulation did not compromise the health of the animals. Ovulation was reestablished by realimentation. The level of dietary intake required to reestablish ovulation was significantly greater than the level that supported ovulation during dietary restriction. This primate model of nutritional amenorrhea affords opportunities to prospectively study the neuroendocrine mechanisms that inhibit the menstrual cycle in response to inadequate nutrition. Although it remains to be documented, it is likely that this primate model will prove useful for studying changes in bone metabolism associated with long periods of anovulation induced by dietary restriction.
Acknowledgments
We acknowledge Dr. A. F. Parlow and the National Hormone and Peptide Program for supplying FSH assay reagents.
Footnotes
This work was supported by the Canadian Institutes of Health Research Grant MOP-57934 (to D.A.V.V.).
First Published Online September 29, 2005
Abbreviations: BMI, Body mass index; CV, coefficient of variation; , change; Lutexp, time of the expected luteal phase.
Accepted for publication September 22, 2005.
References
Schneider JE 2004 Energy balance and reproduction. Physiol Behav 81:289–317
Garfinkel PE, Lin E, Goering P, Spegg C, Goldbloom D, Kennedy S, Kaplan AS, Woodside DB 1996 Should amenorrhoea be necessary for the diagnosis of anorexia nervosa Evidence from a Canadian community sample. Br J Psychiatry 168:500–506
Loucks AB, Horvath SM 1985 Athletic amenorrhea: a review. Med Sci Sports Exerc 17:56–72
Boyar RM, Katz J, Finkelstein JW, Kapen S, Weiner H, Weitzman ED, Hellman L 1974 Anorexia nervosa. Immaturity of the 24-hour luteinizing hormone secretory pattern. N Engl J Med 291:861–865
Berga SL, Mortola JF, Girton L, Suh B, Laughlin G, Pham P, Yen SS 1989 Neuroendocrine aberrations in women with functional hypothalamic amenorrhea. J Clin Endocrinol Metab 68:301–308
Laughlin GA, Yen SS 1996 Nutritional and endocrine-metabolic aberrations in amenorrheic athletes. J Clin Endocrinol Metab 81:4301–4309
Laughlin GA, Dominguez CE, Yen SS 1998 Nutritional and endocrine-metabolic aberrations in women with functional hypothalamic amenorrhea. J Clin Endocrinol Metab 83:25–32
Warren MP, Holderness CC, Lesobre V, Tzen R, Vossoughian F, Brooks-Gunn J 1994 Hypothalamic amenorrhea and hidden nutritional insults. J Soc Gynecol Invest 1:84–88
Khoury SA, Reame NE, Kelch RP, Marshall JC 1987 Diurnal patterns of pulsatile luteinizing hormone secretion in hypothalamic amenorrhea: reproducibility and responses to opiate blockade and an 2-adrenergic agonist. J Clin Endocrinol Metab 64:755–762
Williams NI, Young JC, McArthur JW, Bullen B, Skrinar GS, Turnbull B 1995 Strenuous exercise with caloric restriction: effect on luteinizing hormone secretion. Med Sci Sports Exerc 27:1390–1398
Loucks AB, Verdun M 1998 Slow restoration of LH pulsatility by refeeding in energetically disrupted women. Am J Physiol 275:R1218–R1226
Wade GN, Jones JE 2004 Neuroendocrinology of nutritional infertility. Am J Physiol 287:R1277–R1296
Cameron JL, Nosbisch C 1991 Suppression of pulsatile luteinizing hormone and testosterone secretion during short term food restriction in the adult male rhesus monkey (Macaca mulatta). Endocrinology 128:1532–1540
Schreihofer DA, Parfitt DB, Cameron JL 1993 Suppression of luteinizing hormone secretion during short-term fasting in male rhesus monkeys: the role of metabolic versus stress signals. Endocrinology 132:1881–1889
Chen MD, O’Byrne KT, Chiappini SE, Hotchkiss J, Knobil E 1992 Hypoglycemic ‘stress’ and gonadotropin-releasing hormone pulse generator activity in the rhesus monkey: role of the ovary. Neuroendocrinology 56:666–673
Heisler LE, Pallotta CM, Reid RL, Van Vugt DA 1993 Hypoglycemia-induced inhibition of luteinizing hormone secretion in the rhesus monkey is not mediated by endogenous opioid peptides. J Clin Endocrinol Metab 76:1280–1285
Williams NI, Caston-Balderrama AL, Helmreich DL, Parfitt DB, Nosbisch C, Cameron JL 2001 Longitudinal changes in reproductive hormones and menstrual cyclicity in cynomolgus monkeys during strenuous exercise training: abrupt transition to exercise-induced amenorrhea. Endocrinology 142:2381–2389
Lane MA, Black A, Handy AM, Shapses SA, Tilmont EM, Kiefer TL, Ingram DK, Roth GS 2001 Energy restriction does not alter bone mineral metabolism or reproductive cycling and hormones in female rhesus monkeys. J Nutr 131:820–827
Jen KL, Hansen BC, Metzger BL 1985 Adiposity, anthropometric measures, and plasma insulin levels of rhesus monkeys. Int J Obes 9:213–224
Lujan ME, MacTavish PJ, Krzemien AA, Bradstock MW, Van Vugt DA 2002 Estrogen-induced gonadotropin surge in rhesus monkeys is not inhibited by cortisol synthesis inhibition or hypoglycemia. Endocrine 19:169–176
Williams NI, Helmreich DL, Parfitt DB, Caston-Balderrama A, Cameron JL 2001 Evidence for a causal role of low energy availability in the induction of menstrual cycle disturbances during strenuous exercise training. J Clin Endocrinol Metab 86:5184–5193
Loucks AB 2003 Energy availability, not body fatness, regulates reproductive function in women. Exerc Sport Sci Rev 31:144–148
Warren MP 1980 The effects of exercise on pubertal progression and reproductive function in girls. J Clin Endocrinol Metab 51:1150–1157
Cameron JL, Weltzin TE, McConaha C, Helmreich DL, Kaye WH 1991 Slowing of pulsatile luteinizing hormone secretion in men after forty-eight hours of fasting. J Clin Endocrinol Metab 73:35–41
Loucks AB, Verdun M, Heath EM 1998 Low energy availability, not stress of exercise, alters LH pulsatility in exercising women. J Appl Physiol 84:37–46
Olson BR, Cartledge T, Sebring N, Defensor R, Nieman L 1995 Short-term fasting affects luteinizing hormone secretory dynamics but not reproductive function in normal-weight sedentary women. J Clin Endocrinol Metab 80:1187–1193
Soules MR, Merriggiola MC, Steiner RA, Clifton DK, Toivola B, Bremner WJ 1994 Short-term fasting in normal women: absence of effects on gonadotrophin secretion and the menstrual cycle. Clin Endocrinol (Oxf) 40:725–731
Lado-Abeal J, Robert-McComb JJ, Qian XP, Leproult R, Van Cauter E, Norman RL 2005 Sex differences in the neuroendocrine response to short-term fasting in rhesus macaques. J Neuroendocrinol 17:435–444
Lujan ME, Krzemien AA, Reid RL, Van Vugt DA 2005 Caloric restriction inhibits steroid-induced gonadotropin surges in ovariectomized rhesus monkeys. Endocrine 27:25–32
Wildt L, Hausler A, Marshall G, Hutchison JS, Plant TM, Belchetz PE, Knobil E 1981 Frequency and amplitude of gonadotropin-releasing hormone stimulation and gonadotropin secretion in the rhesus monkey. Endocrinology 109:376–385
Christman GM, Randolph JF, Kelch RP, Marshall JC 1991 Reduction of gonadotropin-releasing hormone pulse frequency is associated with subsequent selective follicle-stimulating hormone secretion in women with polycystic ovarian disease. J Clin Endocrinol Metab 72:1278–1285
Kaiser UB, Jakubowiak A, Steinberger A, Chin WW 1997 Differential effects of gonadotropin-releasing hormone (GnRH) pulse frequency on gonadotropin subunit and GnRH receptor messenger ribonucleic acid levels in vitro. Endocrinology 138:1224–1231
Schneider JE, Wade GN 1989 Availability of metabolic fuels controls estrous cyclicity of Syrian hamsters. Science 244:1326–1328
Alvero R, Kimzey L, Sebring N, Reynolds J, Loughran M, Nieman L, Olson BR 1998 Effects of fasting on neuroendocrine function and follicle development in lean women. J Clin Endocrinol Metab 83:76–80
Drezgic M, Penezic Z, Zarkovic M, Vujovic S, Ciric J, Trbojevic B, Bogdanovic AD, Nesovic M 1996 Influence of three-week fasting on gonadotropin pulsatility in obese menstruating women. Int J Obes Relat Metab Disord 20:608–612
Di Carlo C, Palomba S, De Fazio M, Gianturco M, Armellino M, Nappi C 1999 Hypogonadotropic hypogonadism in obese women after biliopancreatic diversion. Fertil Steril 72:905–909
Knuth UA, Hull MG, Jacobs HS 1977 Amenorrhoea and loss of weight. Br J Obstet Gynaecol 84:801–807
Golden NH, Jacobson MS, Schebendach J, Solanto MV, Hertz SM, Shenker IR 1997 Resumption of menses in anorexia nervosa. Arch Pediatr Adolesc Med 151:16–21(Marla E. Lujan, Alicja A. Krzemien, Robe)
Abstract
Nutritional amenorrhea is defined as cessation of menstrual cycles resulting from a chronic negative energy balance. Although it is agreed that nutritional amenorrhea results from reduced secretion of GnRH, the neuroendocrine mechanisms leading to GnRH inhibition are poorly defined. Because the invasiveness of many neuroendocrine experimental approaches precludes its use in the clinical setting, we set out to establish a model of nutritional amenorrhea in rhesus monkeys. Studies were conducted in four normal-weight and one obese female rhesus monkey. Dietary intake was gradually reduced with the goal of achieving a 15–20% weight reduction. Dietary restriction inhibited ovulation in all animals. The weight loss required to inhibit ovulation ranged from 2–11% in the four normal-weight animals and was achieved with a 23% reduction in dietary intake. The time of initiating reduced food intake to first missed ovulation was 62 ± 13 d. Greater weight loss (46% reduction) over a longer period (10 months) was required to inhibit ovulation in the obese monkey. The onset of anovulation was not preceded by changes in menstrual cycle length or progesterone secretion. Realimentation initiated ovulation at a weight that approximated the animal’s weight at the time of the last ovulatory cycle during dietary restriction. By contrast, caloric intake at the return of ovulation during realimentation was 28% greater. This is the first demonstration that chronic dietary restriction can inhibit ovulation in rhesus monkeys. This model will be useful for studying the neuroendocrine mechanisms involved in diet-induced anovulation in primates.
Introduction
CHRONIC NEGATIVE ENERGY balance results in the cessation of menstrual cycles in primates and of estrous cycles in nonprimate species (1). Anorexia nervosa and hypothalamic amenorrhea resulting from strenuous exercise are examples of nutritional amenorrhea caused by chronic energy deficits. Although amenorrhea is no longer considered a diagnostic criteria of anorexia nervosa, it remains a hallmark of anorexia nervosa, occurring in the vast majority of subjects when weight is less than 80–85% of ideal body weight (2). A high incidence of amenorrhea also occurs in women whose lifestyle includes strenuous exercise (3). Although the etiology and degree of negative energy balance differ in anorexia nervosa and exercise-induced amenorrhea, similarities between the two conditions exist. Both are characterized by decreased gonadotropin pulse frequency leading to suppression of ovarian cyclicity and impaired gonadal steroid production (3, 4, 5, 6, 7). Moreover, women whose life style includes a rigorous exercise component have a higher incidence of disordered eating (8).
Although many studies have described in detail the hypothalamic-pituitary-ovarian axis parameters associated with nutritional amenorrhea (3, 4, 6, 9, 10, 11) these studies cannot address the neuroendocrine events that lead to amenorrhea. Investigations of the neuroendocrine mechanisms of nutritional amenorrhea in women are limited by the inability to conduct prospective studies. In addition, a thorough investigation of neuroendocrine mechanisms involves a level of invasiveness unacceptable in the clinical setting. Although inhibition of the reproductive axis by reduced energy intake has been demonstrated in several rodent species (12), accurate translation of results to humans is uncertain given the many differences between the primate menstrual cycle and rodent estrous cycles. Nonhuman primate studies in this area are limited and primarily acute in nature. It has been extensively documented that acute fasting or glucoprivation can inhibit gonadotropin secretion in rhesus monkeys (13, 14, 15, 16). However, the few chronic studies conducted in nonhuman primates to date demonstrated different menstrual cycle responses to a chronic metabolic challenge. Chronic negative energy balance achieved by strenuous exercise inhibited ovulation in cynomolgus monkeys (17), whereas a 30% reduction in dietary intake did not inhibit menstrual cycles in rhesus monkeys (18).
Given the similarities between the reproductive axes of humans and nonhuman primates and the inhibitory effects of negative energy balance on reproductive cycles of women and nonprimate species, we hypothesized that dietary restriction would inhibit ovulation in rhesus monkeys. The purpose of the current study was to determine the degree of dietary restriction required to inhibit ovulation in rhesus monkeys. The level of dietary restriction required is important to the feasibility of establishing a model of diet-induced amenorrhea in rhesus monkeys, because the utility of such an animal model would require that ovulation was inhibited at a level of dietary restriction that did not unduly compromise animal health.
Materials and Methods
Animal husbandry
Studies were conducted in five adult female rhesus monkeys (Macaca mulatta). Monkeys ranged in age from 6–10 yr and weighed between 5.8–11.9 kg. Animals were housed in groups of two or three in a light- and temperature-controlled environment (lights on, 0700–1900 h; temperature, 22 C). Diets consisted of a twice-daily ration of Purina monkey chow (Hi Protein Monkey Diet Jumbo, Ralston Purina, St. Louis, MO) supplemented with fruits and vegetables. Water was available ad libitum. All animal husbandry practices and experimental procedures conformed to regulations of the Canadian Council on Animal Care and were approved by the Queen’s University animal care committee.
Normalizing body weight to a uniform body mass index
Group housing results in the establishment of hierarchies that can impact on food availability and consequently adiposity. The clinical impression of varying degrees of adiposity in these five monkeys was confirmed using a modified method of calculating body mass index (BMI). The weight in kilograms divided by crown-rump length in meters squared produces an index in rhesus monkeys that is highly correlated with percent body fat (19). Because BMIs were 39.3, 27.2, 25.0, 24.2, and 22.1 kg/m2, group-housed animals were separated during meal times (0900–1100 and 1400–1600 h) to standardize their weights and subsequently regulate their dietary intake. Four monkeys were initially placed on a daily ration of 15 biscuits (calculated to provide 1887 kJ/d), whereas the obese animal was restricted to 1415 kJ/d (12 biscuits). Diets were further individualized in the subsequent months to achieve a uniform BMI of 24–23 kg/m2, which translates into a lean body habitus. Adjustments to diets included 1) increasing the vegetable component, 2) modifying the number of monkey chow biscuits allotted, and 3) eliminating high carbohydrate treats to those monkeys whose BMI exceeded 24 kg/m2. Food intake was monitored daily, and monkeys were weighed monthly during the initial phase of the experiment using a transfer box that eliminated the need for sedation. Animals were trained to present a limb through an opening in the cage for blood sampling by venipuncture. Blood samples (1.5–2 ml) were drawn three times a week before the morning meal.
Effect of dietary restriction on ovulation in rhesus monkeys
Once a BMI of 24–23 kg/m2 was attained, animals were placed on a 16-wk regimen of dietary restriction with the goal of decreasing weight by 20%. This was accomplished by decreasing the amount of monkey chow allotted to each animal. The rate and extent of dietary restriction were dependent on the rate of weight loss in each animal. Every 2 wk, one cube of monkey chow was removed if weight loss had not exceeded the desired rate of decline. In each case, the aim of dietary restriction was to decrease weight by no more than 20% and dietary restriction by no more than 40% over the 16-wk period. An attempt was made to maintain the bulk of each animal’s diet by increasing the amount of low calorie vegetables given between the morning and afternoon feedings. An iron-rich vitamin supplement was given daily. Once animals became anovulatory, we attempted to maintain each animal’s weight just below (within 500 g) the animal’s weight at the time of the first missed ovulation (i.e. time of the expected luteal phase). Animals were maintained on dietary restriction for 8–10 months, during which time the effects of leptin administration on ovulation were tested (our manuscript in preparation). This was followed by a period of realimentation in which the goal was to determine the weight and dietary intake level at which ovulations returned. Throughout dietary restriction and realimentation, weights were measured weekly using a transfer box, and blood samples were drawn by venipuncture three times a week before the morning meal. Samples were assayed for progesterone and FSH. A veterinarian’s assessment was performed at the start of the study and at 6-month intervals. Hematology and blood chemistry (including liver and renal function tests) were performed at 6-month intervals.
RIAs
Blood samples were refrigerated overnight, and serum was isolated after centrifugation (2500 rpm). All serum samples were assayed for progesterone using a commercial RIA kit available from Diagnostic Products Corp. (Los Angeles, CA). The assay sensitivity was 0.02 ng/ml. The intraassay coefficient of variation (CV) was 3.6%, and the interassay CV was 3.9%. We have validated this progesterone assay for use in the nonhuman primate (20). A minimum of three consecutive serum progesterone values of 4.0 ng/ml or greater was considered indicative of the luteal phase. This cutoff was based on progesterone measurements over six consecutive cycles in five rhesus monkeys with regular ovulatory menstrual cycles (mean follicular phase progesterone, 1.8 ± 0.1 ng/ml; range, 1.0–3.7 ng/ml; mean luteal phase progesterone, 8.5 ± 0.4 ng/ml; range, 4.0–21.9 ng/ml). An integrated weekly FSH value was determined for serum pools generated by combining equal volumes of each serum sample collected during a 1-wk period. Serum pools were separated into follicular and luteal phases based on earlier progesterone measurements. Serum pools were assayed in duplicate for FSH in a single assay using reagents provided by the National Hormone and Pituitary Program. The standard curve consisted of FSH reference preparation AFP 6940A. Unknowns were incubated with recombinant FSH antibody (AFP 782594), followed by the addition of 125I-radiolabeled FSH (AFP 6940A). A sheep antirabbit -globulin (Prince Laboratories, Toronto, Canada) was used to precipitate the antigen-antibody complex. Precipitation was facilitated by adding 12.5% Carbowax (Sigma-Aldrich Corp., St. Louis, MO) before centrifugation. Assay sensitivity, defined as the amount of reference preparation required to reduce binding by 2 SD below the zero standard divided by the sample volume, was 0.1 ng/ml. The intraassay CV was 9.6%.
Statistics
A repeated measures ANOVA and Tukey’s multiple comparison test were used to determine differences in mean BMI and dietary intake at baseline, last ovulation, first missed ovulation, and first ovulation after realimentation. Menstrual cycle length, follicular phase length, luteal phase length, and mean luteal phase progesterone and FSH concentrations before, during, and after dietary restriction were compared using a repeated measures ANOVA and Tukey’s multiple comparison test. The level of significance was set at P 0.05. The effects of dietary restriction and realimentation on ovulation were assessed by visual inspection of the progesterone profiles.
Results
A moderate reduction in weight was required in three animals (monkeys 1, 2, and 3) to bring their BMI into the 23–24 kg/m2 range, whereas a moderate increase was required in one animal (monkey 4). These four monkeys were initiated on a fixed allotment of monkey chow totaling 1887 kJ/d. The effects of a regimented diet on BMI and progesterone secretion are presented in Figs. 1 and 2. BMI declined in the two monkeys depicted in Fig. 1. Monkey 2 consistently consumed less than the entire meal and consequently fell below the target BMI of 23–24 kg/m2 (Fig. 1B). This animal eventually increased her food consumption, and BMI increased to 22.1 kg/m2. Both animals exhibited a cyclical pattern of progesterone secretion indicative of ovulation. Eight ovulatory cycles were documented in each animal over the 8-month baseline period. The two monkeys depicted in Fig. 2 had discrepant responses to the fixed diet. Monkey 3 depicted in Fig. 2A had an initial BMI of 27.2 kg/m2, but underwent rapid weight loss when fed 1887 kJ/d. Two ovulations were documented in this animal before it became anovulatory. Caloric intake was increased to 2358 kJ/d, and its BMI increased from 21.3 to 24 kg/m2, at which point ovulation resumed. The second monkey depicted in Fig. 2 had an initial BMI of 22.1 kg/m2, which was below the desired BMI. This animal was anovulatory for the first 3 months of the study. BMI increased to 24.6 kg/m2 after 16 wk of individual feeding, and cyclical progesterone secretion indicative of regular ovulations ensued.
The effects of dietary restriction and subsequent realimentation on BMI and progesterone secretion in these four monkeys are presented in Figs. 3 and 4 (presented in the same sequence as in Figs. 1 and 2; data from Figs. 1 and 2 are included for added perspective). The two monkeys depicted in Fig. 3 exhibited a linear reduction in BMI during dietary restriction. Dietary restriction inhibited ovulation in both animals, as evidenced by an abrupt cessation in cyclic progesterone secretion. Despite a higher initial dietary intake and BMI, ovulation offset occurred sooner in monkey 1 than in monkey 2 (36 vs. 93 d). Dietary intake was increased subsequently in both animals to maintain their weight near the weight of ovulation offset. Therefore, dietary intake at the time of initiating realimentation in these two animals was similar to dietary intake at baseline. An additional increase in food intake during realimentation initiated ovulation. Ovulation onset occurred within a similar time frame as ovulation offset in these two animals (55 and 101 d to onset vs. 36 and 93 d to offset). Dietary intake, BMI, and progesterone secretion during food restriction and realimentation are shown in Fig. 4 for the other two monkeys (same animals as depicted in Fig. 2). Neither animal exhibited a linear reduction in BMI during dietary restriction. Monkey 3 (Fig. 4A) exhibited a delayed pattern of weight loss in response to dietary restriction. The first missed ovulation occurred after a very small decline in weight (BMI declined from 24 to 23.5 kg/m2). This monkey subsequently experienced a precipitous weight loss that required increasing the dietary intake from 13 to 18 biscuits/d over a 3-wk period to stabilize her weight. Weight declined in monkey 4 (Fig. 4B) during the initial stages of dietary restriction and then stabilized despite an additional reduction in dietary intake that resulted in anovulation. Realimentation stimulated ovulation in both animals. The times from initiating realimentation to first ovulation were 85 and 64 d (similar to the previous examples).
Weight and changes in weight associated with ovulation offset and onset are reported in kilograms and BMI units in Tables 1 and 2, respectively. Ovulation offset occurred at a mean weight of 5.8 ± 0.3 kg (BMI of 21.8 ± 0.9 kg/m2) compared with a starting weight of 6.2 ± 0.2 kg (BMI of 23.4 ± 0.4 kg/m2). This constituted an average weight loss of 0.4 ± 0.10 kg (0.1–0.7 kg). The weight at the time of the last luteal phase during caloric restriction was identical with the weight at the time of the first luteal phase during realimentation (5.9 ± 0.2 kg). Likewise, the change in BMI (BMI) was similar for ovulation offset and onset (1.6 ± 0.5 vs. 1.9 ± 0.3 kg/m2). The time from initiating dietary restriction or realimentation to ovulation offset and onset, respectively, correlated with the absolute change in weight (r = 0.93; P = 0.0009).
Dietary intake and changes in energy intake associated with ovulation offset and onset are shown in Table 3. The mean change in food consumption required to inhibit ovulation was similar to the magnitude of the increase that stimulated ovulation (443 ± 30 vs. 413 ± 102 kJ/d). However, the level of dietary intake at first ovulation during realimentation was considerably greater than the energy intake at offset [2123 ± 174 vs.1504 ± 148 kJ/d at the time of the expected luteal phase (Lutexp)]. Even when compared with the energy levels during the last follicular phase of dietary restriction, dietary intake at the time of first ovulation during realimentation was 25% greater (2123 ± 174 vs. 1710 ± 228 kJ/d). These results are shown graphically in Fig. 5. Dietary intake at first missed ovulation was significantly reduced compared with food intake before dietary restriction (Lutexp vs. basal, P < 0.01), but was not different from food intake at the time of the last ovulation during dietary restriction. Dietary intake at first ovulation during realimentation (cycle 1) was significantly greater than dietary intake at all relevant time points (most notable comparisons being the follicular phase of final ovulation during dietary restriction and the dietary intake at the time realimentation was initiated). The same comparisons were not statistically significant when performed for BMI.
The transition from ovulatory to anovulatory status (and vice versa) was abrupt. Mean menstrual cycle length, follicular phase length, luteal phase length, and luteal phase progesterone concentrations of the last two menstrual cycles observed during dietary restriction and the first two ovulatory cycles after realimentation were comparable to these same cycle parameters in cycles before and during the earlier stages of dietary restriction (Table 4).
A fifth animal included in this study was inordinately obese (BMI of 39.3 kg/m2). As shown in Fig. 6, this animal progressively lost weight and attained a lean BMI after 8–9 months on a diet of 1415 kJ/d. Cyclic progesterone secretion confirmed normal ovulatory function during this extended period of weight loss. An additional reduction in caloric intake inhibited ovulation. Anovulation occurred at a BMI of 21.5 kg/m2 (within the range of the four normal-weight monkeys). The animal remained anovulatory for several months while maintained on a diet of 1061–1533 kJ/d. She was removed from the study before realimentation due to development of a central nervous system infection.
Also shown in Fig. 6 are integrated weekly FSH concentrations for this animal. Several large excursions in FSH temporally associated with ovulations were seen. These large fluctuations were absent during the subsequent anovulatory period. Mean weekly FSH concentrations normalized to ovulation offset during dietary restriction and ovulation onset after realimentation are shown in Fig. 7. Mean FSH concentrations were more variable during the 10 wk before anovulation compared with the 15 wk of documented anovulation (see Fig. 7A). In addition, FSH levels determined at 4-wk intervals showed a significant decline in FSH secretion with anovulation (see inset). The onset of ovulation after realimentation was accompanied by a return to a variable pattern of FSH secretion (Fig. 7B).
The degree of dietary restriction necessary to inhibit ovulation was not associated with significant morbidity. Animals remained attentive to their surroundings and continued to use perches in the exercise pens. They retained their appetites throughout the study. Routine biochemical and hematological parameters remained within normal limits throughout the period of dietary restriction. No measures of bone metabolism were performed. One animal exhibited significant alopecia during dietary restriction, which was reversed by realimentation. Realimentation was accompanied by weight gain and resumption of ovulation. Animals remain in good health 2 yr after study completion.
Discussion
This study demonstrates that ovulation in rhesus monkeys was inhibited by a reduction in dietary intake. This is the first documentation that dietary restriction can inhibit ovulation in nonhuman primates. A previous study concluded that ovulation in rhesus monkeys was not inhibited by a 30% reduction in dietary intake (18). A greater degree of dietary restriction in the current study is the most likely explanation for the difference in outcomes. Anovulation occurred in our study when food intake was decreased by 21%, resulting in an average dietary intake of 1504 ± 148 kJ/d. By contrast, a 30% reduction in dietary intake in the study by Lane et al. (18) translated into a daily intake of 1709 ± 55 kJ. Although the level of dietary restriction employed was greater in terms of percent decrease from habitual levels, according to our study this absolute level of dietary intake would not be expected to inhibit ovulation.
Our results in rhesus monkeys complement those of Williams et al. (17), who observed that chronic strenuous exercise induced amenorrhea in cynomolgus monkeys. Their study concluded that anovulation resulted from a negative energy balance rather than some other factor associated with exercise, because ovulation was reinitiated in exercising monkeys by increasing food availability (21). In contrast to exercise-induced amenorrhea, our study showed that dietary restriction did not produce transitional menstrual cycles. Follicular phase length, luteal phase length, and luteal phase progesterone secretion in the last two cycles during dietary restriction were comparable to these cycle parameters before or during the early stages of dietary restriction. This observation is at variance with that by Williams et al. (17), who reported that follicular phase length was increased and plasma LH and luteal phase progesterone secretion were reduced in the menstrual cycle immediately preceding exercised-induced amenorrhea. The much shorter time span from initiation to anovulation observed with dietary restriction compared with exercise (62 ± 13 d vs. 14.3 ± 2.2 months, respectively) may account for the absence of transitional menstrual cycles in our study. A regimen of dietary restriction that produces over a 2-month period a negative energy balance that inhibits ovulation may be too short a time frame for borderline energy levels to produce transitional menstrual cycles. Because neither study quantified the degree of energy deficit, this explanation remains speculative. A second difference that merits consideration is the method by which a state of negative energy balance was achieved in the two studies. If the energy available to fuel processes such as ovulation is simply the difference between energy intake and energy expended, then a deficit, regardless of how it was achieved, should adversely affect ovulation. This premise is supported by the demonstration that LH pulse frequency in women was inhibited by a negative energy balance regardless of whether it was achieved by reducing energy intake or increasing energy expenditure (22). It is also supported by the two nonhuman primate studies combined, because ovulation was inhibited in both. However, it remains to be determined whether the subtle differences cited above are due to differences in how the energy deficits were achieved.
Our demonstration that dietary restriction reversibly inhibits ovulation in rhesus monkeys also establishes the link between energy intake and reproductive function in primates. Energy deficit has been documented previously as the root cause of reproductive axis dysfunction in women with hypothalamic amenorrhea resulting from self-imposed dietary restriction or strenuous exercise (3, 6, 23). Several studies have demonstrated that acute energy deficits resulting from short-term fasting or increased energy expenditure inhibited the hypothalamic-gonadotropic axis, as reflected by reduced LH pulse frequency in human and nonhuman primates (13, 22, 24, 25), although females may be more resistant to acute fasting-induced inhibition of LH secretion compared with males (26, 27, 28). Although the current study did not attempt to characterize changes in LH pulse frequency, inhibition of ovulation is the expected outcome if dietary restriction chronically reduced LH pulse frequency below the threshold that supports follicular maturation or that is necessary to generate an LH surge. We have demonstrated that a similar degree of dietary restriction in ovariectomized rhesus monkeys inhibits the elaboration of the estrogen/progesterone-induced gonadotropin surge (29). Anovulation was not preceded by a reduction in FSH concentrations. A reduced GnRH pulse frequency may explain the maintenance of FSH concentrations while inhibiting ovulation, because a low GnRH pulse frequency can stimulate FSH secretion while inhibiting LH secretion (30, 31, 32).
This study provides insight into the dietary intake required to support ovulation in rhesus monkeys. Three of four normal-weight monkeys ovulated when given a diet of 1887 kJ/d. Monkey 1 had a stable weight on this diet and ovulated regularly. Monkey 2 consumed less than the entire meal and lost weight. Eventually, food consumption increased to 1533 kJ/d, and weight stabilized at a BMI of 22.1 kg/m2. The decreased appetite followed by partial recovery may reflect acclimatization to being separated twice a day at mealtime. In addition, this animal was the smallest of the four, weighing approximately 10% less than the others. In contrast, monkey 4 gained weight and became ovulatory after 3 months on a diet of 1887 kJ/d. This response is probably the result of inadequate food intake in a communal setting being resolved by separating pairs at mealtime. Monkey 3 was exceptional because it underwent rapid weight loss and became anovulatory when fed 1887 kJ/d. Food intake was increased to 2358 kJ/d, and its BMI increased from 21.3 to 24 kg/m2, at which point ovulation resumed.
Although these data indicate that the level of energy intake is a critical determinant of ovulatory status, they also suggest that energy stores may play a role. Inadequate energy stores may explain why monkeys 3 and 4 did not ovulate for several months while on a diet that eventually stimulated ovulation. Both animals gained weight before ovulating. Conversely, monkey 2 continued to ovulate despite dietary intake being reduced 1 month earlier to 1179 kJ/d. Ovulation offset occurred without an additional reduction in dietary intake, but was accompanied by a modest weight loss. Much more striking was the delayed effect of dietary restriction on ovulation in the obese monkey. This animal continued to ovulate at regular intervals despite a 42% decline in weight (11.9 to 6.8 kg) over 10 months. By contrast, the normal-weight monkeys became anovulatory after an average weight loss of 0.43 ± 0.1 kg (0.1–0.7 kg). The extended period of weight loss in the obese monkey indicated that this animal was chronically exposed to a negative energy balance. Despite a chronic energy deficit, anovulation did not ensue until BMI reached a level similar to that at which the other animals became anovulatory. Our speculation that energy stores maintained ovulatory function during the first 9 months of dietary restriction is tempered by dietary intake having been further reduced just before the first missed ovulation. Therefore, anovulation may have been affected by an additional reduction in dietary intake alone or a reduction in dietary intake in combination with reduced availability of oxidizable metabolic fuels. Several reports suggest that fatness affects the response of the reproductive axis to caloric restriction. Fat hamsters were less susceptible to starvation-induced anestrus than lean hamsters (33). Lean women were more prone to disruption of the neuroendocrine axis in response to a 72-h fast compared with normal-weight women (34), whereas gonadotropin secretion in obese women was not affected by a 21-d fast (35). However, in distinction to the above findings, it has been reported that morbidly obese women who had a 25% weight loss after biliopancreatic diversion became amenorrheic despite still being obese (mean BMI of 35 kg/m2) (36).
The mean dietary intake during the follicular phase of the last menstrual cycle before ovulation offset was 1725 ± 225 kJ/d. However, there was considerable variability (1179, 1592, 1887, and 2241 kJ/d). In fact, of the three measures (weight, BMI, and dietary intake), the variability in dietary intake was considerably greater than that for weight and BMI (see SEM expressed as a percentage of the mean in Tables 1–3). Normalizing dietary intake to weight or BMI reduced variability only slightly. Normalizing dietary intake to fat-free mass or some other measure of body composition may better predict the threshold for ovulation in future studies.
The inhibitory effect of dietary restriction on ovulation was reversed by realimentation. The absolute change in food intake ( kJ) and weight ( BMI) associated with ovulation during realimentation was similar to the absolute change in these same variables associated with inhibition of ovulation during dietary restriction. However, ovulation occurred at a significantly higher energy intake level during realimentation compared with the energy consumption during the final follicular phase of dietary restriction (2123 ± 174 kJ vs. 1725 ± 225 kJ). This difference was seen in all animals. The hysteresis exhibited in our dataset is not due to a lag effect caused by a faster rate of realimentation. The rate of realimentation (kJ/d of realimentation to ovulation onset, 5.2 ± 1.1 kJ/d) was actually slower than the rate of dietary restriction (kJ/d of food restriction to ovulation offset, 7.1 ± 1.1 kJ/d). This finding may have implications for dietary requirements to reestablish menstrual cyclicity in women with anorexia nervosa and is consistent with the report that return of menses in women with anorexia nervosa occurred at a mean weight that was 2.05–3.6 kg greater than the weight at which menses were interrupted (37, 38).
In summary, moderate dietary restriction of lean rhesus monkeys reliably inhibited ovulation. The level of dietary restriction required to inhibit ovulation did not compromise the health of the animals. Ovulation was reestablished by realimentation. The level of dietary intake required to reestablish ovulation was significantly greater than the level that supported ovulation during dietary restriction. This primate model of nutritional amenorrhea affords opportunities to prospectively study the neuroendocrine mechanisms that inhibit the menstrual cycle in response to inadequate nutrition. Although it remains to be documented, it is likely that this primate model will prove useful for studying changes in bone metabolism associated with long periods of anovulation induced by dietary restriction.
Acknowledgments
We acknowledge Dr. A. F. Parlow and the National Hormone and Peptide Program for supplying FSH assay reagents.
Footnotes
This work was supported by the Canadian Institutes of Health Research Grant MOP-57934 (to D.A.V.V.).
First Published Online September 29, 2005
Abbreviations: BMI, Body mass index; CV, coefficient of variation; , change; Lutexp, time of the expected luteal phase.
Accepted for publication September 22, 2005.
References
Schneider JE 2004 Energy balance and reproduction. Physiol Behav 81:289–317
Garfinkel PE, Lin E, Goering P, Spegg C, Goldbloom D, Kennedy S, Kaplan AS, Woodside DB 1996 Should amenorrhoea be necessary for the diagnosis of anorexia nervosa Evidence from a Canadian community sample. Br J Psychiatry 168:500–506
Loucks AB, Horvath SM 1985 Athletic amenorrhea: a review. Med Sci Sports Exerc 17:56–72
Boyar RM, Katz J, Finkelstein JW, Kapen S, Weiner H, Weitzman ED, Hellman L 1974 Anorexia nervosa. Immaturity of the 24-hour luteinizing hormone secretory pattern. N Engl J Med 291:861–865
Berga SL, Mortola JF, Girton L, Suh B, Laughlin G, Pham P, Yen SS 1989 Neuroendocrine aberrations in women with functional hypothalamic amenorrhea. J Clin Endocrinol Metab 68:301–308
Laughlin GA, Yen SS 1996 Nutritional and endocrine-metabolic aberrations in amenorrheic athletes. J Clin Endocrinol Metab 81:4301–4309
Laughlin GA, Dominguez CE, Yen SS 1998 Nutritional and endocrine-metabolic aberrations in women with functional hypothalamic amenorrhea. J Clin Endocrinol Metab 83:25–32
Warren MP, Holderness CC, Lesobre V, Tzen R, Vossoughian F, Brooks-Gunn J 1994 Hypothalamic amenorrhea and hidden nutritional insults. J Soc Gynecol Invest 1:84–88
Khoury SA, Reame NE, Kelch RP, Marshall JC 1987 Diurnal patterns of pulsatile luteinizing hormone secretion in hypothalamic amenorrhea: reproducibility and responses to opiate blockade and an 2-adrenergic agonist. J Clin Endocrinol Metab 64:755–762
Williams NI, Young JC, McArthur JW, Bullen B, Skrinar GS, Turnbull B 1995 Strenuous exercise with caloric restriction: effect on luteinizing hormone secretion. Med Sci Sports Exerc 27:1390–1398
Loucks AB, Verdun M 1998 Slow restoration of LH pulsatility by refeeding in energetically disrupted women. Am J Physiol 275:R1218–R1226
Wade GN, Jones JE 2004 Neuroendocrinology of nutritional infertility. Am J Physiol 287:R1277–R1296
Cameron JL, Nosbisch C 1991 Suppression of pulsatile luteinizing hormone and testosterone secretion during short term food restriction in the adult male rhesus monkey (Macaca mulatta). Endocrinology 128:1532–1540
Schreihofer DA, Parfitt DB, Cameron JL 1993 Suppression of luteinizing hormone secretion during short-term fasting in male rhesus monkeys: the role of metabolic versus stress signals. Endocrinology 132:1881–1889
Chen MD, O’Byrne KT, Chiappini SE, Hotchkiss J, Knobil E 1992 Hypoglycemic ‘stress’ and gonadotropin-releasing hormone pulse generator activity in the rhesus monkey: role of the ovary. Neuroendocrinology 56:666–673
Heisler LE, Pallotta CM, Reid RL, Van Vugt DA 1993 Hypoglycemia-induced inhibition of luteinizing hormone secretion in the rhesus monkey is not mediated by endogenous opioid peptides. J Clin Endocrinol Metab 76:1280–1285
Williams NI, Caston-Balderrama AL, Helmreich DL, Parfitt DB, Nosbisch C, Cameron JL 2001 Longitudinal changes in reproductive hormones and menstrual cyclicity in cynomolgus monkeys during strenuous exercise training: abrupt transition to exercise-induced amenorrhea. Endocrinology 142:2381–2389
Lane MA, Black A, Handy AM, Shapses SA, Tilmont EM, Kiefer TL, Ingram DK, Roth GS 2001 Energy restriction does not alter bone mineral metabolism or reproductive cycling and hormones in female rhesus monkeys. J Nutr 131:820–827
Jen KL, Hansen BC, Metzger BL 1985 Adiposity, anthropometric measures, and plasma insulin levels of rhesus monkeys. Int J Obes 9:213–224
Lujan ME, MacTavish PJ, Krzemien AA, Bradstock MW, Van Vugt DA 2002 Estrogen-induced gonadotropin surge in rhesus monkeys is not inhibited by cortisol synthesis inhibition or hypoglycemia. Endocrine 19:169–176
Williams NI, Helmreich DL, Parfitt DB, Caston-Balderrama A, Cameron JL 2001 Evidence for a causal role of low energy availability in the induction of menstrual cycle disturbances during strenuous exercise training. J Clin Endocrinol Metab 86:5184–5193
Loucks AB 2003 Energy availability, not body fatness, regulates reproductive function in women. Exerc Sport Sci Rev 31:144–148
Warren MP 1980 The effects of exercise on pubertal progression and reproductive function in girls. J Clin Endocrinol Metab 51:1150–1157
Cameron JL, Weltzin TE, McConaha C, Helmreich DL, Kaye WH 1991 Slowing of pulsatile luteinizing hormone secretion in men after forty-eight hours of fasting. J Clin Endocrinol Metab 73:35–41
Loucks AB, Verdun M, Heath EM 1998 Low energy availability, not stress of exercise, alters LH pulsatility in exercising women. J Appl Physiol 84:37–46
Olson BR, Cartledge T, Sebring N, Defensor R, Nieman L 1995 Short-term fasting affects luteinizing hormone secretory dynamics but not reproductive function in normal-weight sedentary women. J Clin Endocrinol Metab 80:1187–1193
Soules MR, Merriggiola MC, Steiner RA, Clifton DK, Toivola B, Bremner WJ 1994 Short-term fasting in normal women: absence of effects on gonadotrophin secretion and the menstrual cycle. Clin Endocrinol (Oxf) 40:725–731
Lado-Abeal J, Robert-McComb JJ, Qian XP, Leproult R, Van Cauter E, Norman RL 2005 Sex differences in the neuroendocrine response to short-term fasting in rhesus macaques. J Neuroendocrinol 17:435–444
Lujan ME, Krzemien AA, Reid RL, Van Vugt DA 2005 Caloric restriction inhibits steroid-induced gonadotropin surges in ovariectomized rhesus monkeys. Endocrine 27:25–32
Wildt L, Hausler A, Marshall G, Hutchison JS, Plant TM, Belchetz PE, Knobil E 1981 Frequency and amplitude of gonadotropin-releasing hormone stimulation and gonadotropin secretion in the rhesus monkey. Endocrinology 109:376–385
Christman GM, Randolph JF, Kelch RP, Marshall JC 1991 Reduction of gonadotropin-releasing hormone pulse frequency is associated with subsequent selective follicle-stimulating hormone secretion in women with polycystic ovarian disease. J Clin Endocrinol Metab 72:1278–1285
Kaiser UB, Jakubowiak A, Steinberger A, Chin WW 1997 Differential effects of gonadotropin-releasing hormone (GnRH) pulse frequency on gonadotropin subunit and GnRH receptor messenger ribonucleic acid levels in vitro. Endocrinology 138:1224–1231
Schneider JE, Wade GN 1989 Availability of metabolic fuels controls estrous cyclicity of Syrian hamsters. Science 244:1326–1328
Alvero R, Kimzey L, Sebring N, Reynolds J, Loughran M, Nieman L, Olson BR 1998 Effects of fasting on neuroendocrine function and follicle development in lean women. J Clin Endocrinol Metab 83:76–80
Drezgic M, Penezic Z, Zarkovic M, Vujovic S, Ciric J, Trbojevic B, Bogdanovic AD, Nesovic M 1996 Influence of three-week fasting on gonadotropin pulsatility in obese menstruating women. Int J Obes Relat Metab Disord 20:608–612
Di Carlo C, Palomba S, De Fazio M, Gianturco M, Armellino M, Nappi C 1999 Hypogonadotropic hypogonadism in obese women after biliopancreatic diversion. Fertil Steril 72:905–909
Knuth UA, Hull MG, Jacobs HS 1977 Amenorrhoea and loss of weight. Br J Obstet Gynaecol 84:801–807
Golden NH, Jacobson MS, Schebendach J, Solanto MV, Hertz SM, Shenker IR 1997 Resumption of menses in anorexia nervosa. Arch Pediatr Adolesc Med 151:16–21(Marla E. Lujan, Alicja A. Krzemien, Robe)