Although the negative effects of obesity on reproductive function were first documented over 2000 years ago by Hippocrates (Bray 1990), the etiology of this unfavorable relationship has not been studied thoroughly in women. Nevertheless, it is becoming increasingly important to understand this association because ~one fourth of all women in the U.S. are overweight and, in some subgroups of women, up to 45% of women are overweight or obese (Kuczmarski 1992).

Clinical studies have revealed that obese parturients are more likely to suffer complications of pregnancy, including a higher incidence of hypertensive disease and preeclampsia as well as impaired glucose metabolism and gestational diabetes (Calandra et al. 1981, Ekblad and Grenman 1992). Obese women are more likely to experience prolonged labor (Calandra et al. 1981) and are more susceptible to complications during delivery, resulting in a higher incidence of unplanned cesarean sections (Ekblad and Grenman 1992). Additional evidence suggests that obese women have less success initiating (Richardson 1952) and continuing (Rutishauser and Carlin 1992) breastfeeding.

The studies necessary for examining these reproductive complications of obesity are difficult to perform in women. However, it is easy to induce excess body fatness in rats by offering them a selection of palatable foods (such as potato chips, salami, and peanut butter), an approach known as cafeteria feeding (Sclafani and Springer 1976). Obesity also has been induced in rats by mixing a pelleted, closed-formula diet with hydrogenated vegetable oil, casein, sugar and vitamin mix to create a more palatable food (Wehmer et al. 1979).

In addition, investigators have employed open-formula diets high in fat (Bue et al. 1989, Reynolds et al. 1984). BHE rats fed 22 g fat/100 g diet for 3 wk before mating did not weigh significantly more than their counterparts fed a control diet of 5 g fat/100 g diet (Bue et al. 1989) even though their glucose tolerance during pregnancy was affected. When this high fat diet was fed from weaning, only 20% of the rats became obese (weighed substantially more than the other rats fed this diet). Nonetheless, only half of the group fed the high fat diet conceived and bore young, and only 70% of these had litters that survived the neonatal period. Prolonged gestation was observed among the rats fed the high fat diet, and the surviving pups of the dams fed the high fat diet grew less well in the first 10 d of life. Sprague-Dawley rats (with an initial weight of 125-150 g) fed 30 g fat/100 g diet were declared to be obese after 5 wk when they weighed 278 g (Reynolds et al. 1984); their weight and reproductive performance relative to controls were not described.

Rats fed cafeteria diets have many complications of reproduction, such as decreased rates of conception and delivery (Rolls et al. 1980, Wehmer et al. 1979). Fetal weight is not significantly affected by cafeteria feeding (Lederman and Rosso 1981). Additionally, cafeteria diets negatively affect milk yield (Rolls et al. 1983), so that pups of dams fed such diets grow less well and have much higher mortality rates (Rolls and Rowe 1982, Wehmer et al. 1979) than their control counterparts.

Using the cafeteria diet as a model for obesity is problematic because animals may not obtain adequate protein from such a diet to support reproduction. Thus, their poor reproductive performance cannot be attributed solely to the diet's high fat content. Therefore, the aim of the present study was to investigate the etiology and time course of the reproductive difficulties experienced by rats fed a high fat (35 g fat/100 g), open-formula diet that was nutritionally adequate (i.e., rats would consume adequate protein, vitamins and minerals). We examined the products of conception at three stages of pregnancy, and once during early lactation. In addition, plasma insulin and glucose were measured to gain information on the effect of dietary treatment on metabolism, and plasma prolactin was measured in early lactation to obtain information on suckling stimulus provided by the pups.

MATERIALS AND METHODS

Rats were acclimated to our facilities and weaned to a purified diet of AIN-76A^TM (AIN 1977 and 1980) pellets (Dyets, Bethlehem, PA) over a 5-d period. At 27 d of age, they were randomly assigned to one of the two dietary treatment groups: control (C, n = 46, 72.0 ± 16.3 g, mean ± SD), fed AIN-76A^TM pelleted diet, or high fat (HF, n = 78, 76.7 ± 12.8 g), fed a powdered modification of AIN-76A^TM diet (Table 1). Both groups had free access to food throughout the entire study. The amount of food consumed daily during the prepregnancy period was calculated for a weekly rotating sample of 10 animals from each dietary treatment group, and for all animals during pregnancy, by weighing back the food left over from the previous day. Rats were weighed on the day of arrival and the day of dietary assignment, twice weekly thereafter during the prepregnant period, and on predetermined days during pregnancy.

Table 1. Composition of the purified diets fed during the experiment [View Table]

As they became pregnant, rats were systematically assigned, in a serial-slaughter design, to one of four subgroups within each dietary treatment group. These subgroups were designed to assess ovulation, implantation, fetal viability, and pup mortality, respectively (Fig. 1). At the end of the breeding period, all nonpregnant animals (n = 37) were removed from the study.

To determine the number of eggs ovulated on d 0 of pregnancy, the morning after mating, rats assigned to the ovulation subgroups were killed with an overdose of CO₂, their ovaries were removed and trimmed of fat, and the number of corpora lutea (a proxy for the number of eggs ovulated) was counted using a stereoscopic dissecting microscope (Reichert #569, Cambridge Instruments, Buffalo, NY). A corpus luteum was identified as a highly vascularized, opaque swollen structure on the ovary.

To determine the number of fertilized eggs that successfully implanted in the uterine wall on d 5 of pregnancy, animals assigned to the implantation subgroups were anesthetized with diethyl ether and immediately given an injection of Pontamine Blue dye (Chicago Sky Blue 6B, Sigma Chemical, St. Louis, MO) directly into the heart (0.5 mL/100 g body weight) according to the procedure outlined in Psychoyos (1971). After a 15-min equilibration period, rats were killed with an overdose of CO₂ and a necropsy was performed. The intact uterus and ovaries were removed and examined for blue bands; each band represented one implantation site.

To determine the number of fetuses (both alive and dead), as well as the number of necrosed placentas (if any) on d 18 of pregnancy, animals assigned to the fetal viability subgroups were killed with an overdose of CO₂, and the intact uterus and ovaries were removed. The intact uterus was weighed, and then reweighed after being opened and emptied to determine, by difference, the weight of the amniotic fluid. The weights of the fetuses and placentas were summed for each litter.

On the night of d 21 of pregnancy, the activities of animals assigned to the pup survival subgroups were videotaped. To document the number of pups initially born to each dam, delivery was recorded with a time-lapse video recorder (Panasonic, AG-6730, Secaucus, NJ), under a 60-W red light bulb, because rats are not sensitive to red light. For the next 3 d, pup morbidity was measured qualitatively by noting skin pallor and flakiness. The presence of milk in the pups' stomachs was noted, and pup mortality was recorded. On d 3 of lactation, dams were anesthetized with diethyl ether, and a blood sample was removed by cardiac puncture; dams and remaining pups were then killed with an overdose of CO₂.

For animals in the fetal viability and pup survival subgroups, plasma insulin values were measured using an ¹²⁵I-insulin RIA kit (Amersham Life Science, Arlington Heights, IL). Plasma glucose concentrations were measured using enzymatic (glucose oxidase) determination at 425-475 nm (Sigma Diagnostics), and plasma prolactin values were measured using an ¹²⁵I-prolactin RIA kit (Amersham Life Science).

RESULTS

Although the HF dams continued to weigh more than the C dams during pregnancy, there was no difference between the two groups in total pregnancy weight gain (171.2 ± 33.7 g, n = 9 vs. 164.0 ± 28.2 g, n = 9; data from rats in pup survival subgroup only). The HF rats tended to gain weight more rapidly and then, during the last week of gestation, to gain weight significantly (d 15-18, P < 0.05) less rapidly than C rats (Fig. 2). The distribution of weight gained until d 18 of pregnancy between maternal and fetal compartments was not significantly affected by dietary treatment (Table 2).

Table 2. Effect of consuming a high fat diet on body weight of rat dams and their products of conception at d 18 of pregnancy [View Table]

Table 3. Effect of consuming a high fat diet on reproductive outcome in rats at d 0, 5 and 18 of pregnancy (P), and d 0 of lactation (L) [View Table]

Litter number and weight at birth (Table 3) were not affected by dietary treatment. However, HF pups experienced significantly (P < 0.04) higher mortality rates (16.5 vs. 7.7%, HF and C pups, respectively) than did C pups during these first days of life (Table 3). Seventy per cent of the HF dams lost one or more pups from their litters but only 33% of the C dams experienced comparable losses (P = 0.10).

As expected in the C rats, plasma insulin concentrations were significantly (P < 0.05) higher at pregnancy d 18 (303.2 ± 175.2 pmol/L) than at lactation d 3 (139.2 ± 121.9 pmol/L (Fig. 3). Importantly, this expected drop in insulin values was not present among the HF rats (Fig. 4).

Glucose concentrations did not differ between the two dietary treatment groups at either d 18 of pregnancy or d 3 of lactation. Among the HF rats, plasma glucose concentrations were significantly (P < 0.002) higher at d 18 of pregnancy (7.8 ± 1.2 mmol/L) than at d 3 of lactation (6.2 ± 0.8 mmol/L).

After controlling for the smaller litter size of HF dams by d 3 of lactation, there was no difference in plasma prolactin values between the two dietary treatment groups (22.0 ± 20.2 vs. 20.7 ± 20.5 µg/L, control vs. HF, respectively). As expected, plasma prolactin concentrations increased significantly (P < 0.008) among the C animals between d 18 of pregnancy (9.72 ± 10.5 µg/L) and d 3 of lactation (35.8 ± 20.0 µg/L). Although this trend was also present among the HF animals, plasma prolactin values at d 3 of lactation (29.9 ± 19.6 µg/L) were not significantly higher than those in late pregnancy (13.8 ± 19.1 µg/L).

DISCUSSION

The HF rats did not continue their accelerated rate of weight gain during pregnancy. Although HF rats continued to weigh more than C rats throughout gestation, their net pregnancy weight gain did not differ from that of the C rats. This finding is similar to that of Steingrimsdottir et al. (1980), who found that weight gains attributable solely to reproduction at d 20 of pregnancy were equivalent in Osborne-Mendel rats fed a control and those fed a high fat (55% fat) diet for the 5 d before conception and during pregnancy. BHE rats fed a high fat diet for 5 wk before and during pregnancy also did not gain more weight during pregnancy than their counterparts fed a low fat diet (Bue et al. 1989).

Despite similar gestational weight gains, the pattern of gain among HF rats differed from that of the C rats. Knopp et al. (1973) found that this first phase of adipose tissue metabolism is characterized by an increase in food intake, increased plasma insulin levels, and increased hepatic conversion of glucose to fatty acids. The overall result of these metabolic changes is to direct ingested fuels to maternal stores, because fetal needs are minimal at this time. Of the weight gained during the first 2 wk of gestation, Sohlström et al. (1994) showed that a significant proportion is fat, such that most of the fat retained by rats after pregnancy had been gained during these first 2 wk.

However, during the final third of gestation, fetuses are growing and developing rapidly. The normal maternal adaptation to this increased fetal need begins the second phase of adipose tissue metabolism, which is characterized by continued hyperinsulinemia and decreased tissue responsiveness to insulin, and results in a decline in hepatic conversion of glucose to fatty acids (Knopp et al. 1973, Zammit 1985).

For peripheral insulin resistance to increase late in gestation, assisting the shunting of fuel to the feto-placental unit, there must be some insulin-antagonizing mechanism at work (Flint 1985). However, whether the plasma insulin concentrations seen during late gestation in the HF rats are important for the apparently differing pattern of gestational weight gain of these animals compared with the C rats remains to be investigated. The inability of HF rats to gain as much weight as C rats at the end of gestation, during a period critical for fetal growth and development, may have contributed to the decreased viability of their pups. This has been observed by Lederman and Rosso (1981), who induced obesity in rats before conception by cafeteria feeding. In their experiment, the obese dams gained about 20% less weight between d 12 and 21 of pregnancy and had smaller fetuses.

Despite similar litter weights at birth, the pups of HF dams were more likely to die during the first 3 d of life. Others have documented increased pup mortality among those born to dams fed a cafeteria diet (Rolls and Rowe, 1982, Rolls et al. 1980) or other high fat diets (Bue et al. 1989, Wehmer et al. 1979). The cause of this excess mortality is not known at present but could include both biological and behavioral components.

Rolls et al. (1986) postulated that the young of dams fed cafeteria diets are unable to consume sufficient amounts of their more energy-dense milk or to metabolize the longer-chain fatty acids that are more abundant in this milk. The effect of these differences in milk composition on the growth of the young is not known at present, and was not investigated here.

Another possible explanation for the poor performance of the pups of HF dams is that the dams do not allow adequate nursing and/or that the pups are not able to stimulate appropriate maternal responses. However, our measurements of plasma prolactin levels at d 3 of lactation indicate that the suckling stimulus of the litters was the same; there was no effect of dietary treatment on plasma prolactin concentration after controlling for litter size. However, this does not rule out the possibility that the suckling stimulus of pups of the dams fed the high fat diet was weak before d 3 of lactation and contributed to the early demise of some of the pups.

It has also been suggested that maternal behavior in rats fed a high fat diet is abnormal, leading to high rates of cannibalism of the young (Rolls and Rowe 1982, Wehmer et al. 1979). This cannibalistic maternal response may stem from either a failure of the dam to respond to the changes that occur during the transition from pregnancy to lactation or a failure of the litter to evoke the appropriate response. However, because we did not measure hormone concentrations during pregnancy, we cannot exclude the possibility of decreased priming of maternal behavior among our HF rats.

This model for studying dietary obesity during pregnancy shows promise. Future research should focus on documenting body composition of dams and pups fed this high fat diet and examining further the inability of these rats to maintain healthy litters in the immediate postpartum period. This should involve further examination of fuel regulation during pregnancy and lactation.

In conclusion, our results suggest that the feeding of HF diets is detrimental to reproductive performance in rats, and that it is the presence of the excess fat in the diet (and hence the body) that is problematic. We pinpointed conception and the perinatal period as being the most troublesome for the HF rats, but the exact mechanisms involved remain unknown. Nevertheless, it is likely that altered metabolic processes affect the normal maternal adaptations to the increased energy demands of pregnancy and lactation and result in the observed differences in rate and distribution of gestational weight gain. Although there also is evidence suggesting that abnormal maternal-pup interactions are responsible for the increased pup mortality among the HF rats, this was not conclusively shown in this study.