Because of the wide variation of female hormone levels observed during the estrous cycle, estrous cycle stage is emerging as an important consideration when working with female animals in clinical scenarios and in several research areas. For example, estrous synchronization and staging can improve breeding husbandry in genetically engineered mouse colonies by improving the availability of foster dams. Synchronization of estrus allows for increased reliability in arranging timed pregnancies for studies requiring fetal or neonatal tissues as well as for primary tissue culture studies. For some experimental infectious disease work, the susceptibility, intensity, and duration of infection will be dependent on estrous cycle stage and corresponding levels of endogenous female sex steroids (13, 42). Estrous cycle stage also can affect research areas besides reproduction and infectious diseases, such as brain injury outcome in rodent models of focal cerebral ischemia (10, 11, 28, 37).

The average rat lifespan is approximately 2 to 3 years, depending on the stock or strain, sex, health and disease status, and genetic background of the animal. By 2 months of age, young female rats are reproductively mature and exhibit estrous cycles and ovulation every 4 to 5 days (24). The rat estrous cycle can be further divided into four stages: proestrus, estrus, metestrus, and diestrus. In rats, the duration of proestrus is approximately 12 h, whereas estrus ranges from 9 to 15 h, metestrus from 14 to 18 h, and diestrus from 60 to 70 h (25, 47). The estrous cycle is characterized by cyclical changes in uterus, ovaries, vaginal mucosa, behavior, and hormone levels (for review, see 31, 40). Female rats exhibit regular cyclicity until middle age, when animals then undergo a transition to irregular cyclicity and finally to acyclicity by 12 to 18 months of age. Generally, by 10 to 12 months of age, animals lose the ability to conceive and are considered to be ‘reproductively senescent’ (26, 27, 30, 58).

Vaginal cytology is considered to be the ‘gold standard’ for staging female animals, as the various rat estrous cycle stages can be reliably identified from vaginal smears (36). However, the preparation and assessment of vaginal smears can be time-consuming and laborintensive, and it requires a certain level of technical expertise. The inherent electrical impedance resistance of the vaginal wall changes predictably during the different estrous cycle phases in several species, including the rat (5, 44, 51). Such resistance measurements take only seconds to obtain, are less amenable to subjective interpretation, and are technically easy. Instrumentation is currently available and simple to use. Vaginal impedance measurements have been successfully done in a number of species (7, 20, 38), including rat (4, 5, 44, 51), guinea pig (6, 29), sheep (1, 3), cattle (12, 15, 53), and macaques (16, 52), and have been used for specific clinical and research applications. For example, artificial insemination performed during periods of low vaginal resistance led to improved conception rates in buffaloes (20).

In the present study, we sought to correlate vaginal impedance measurements in intact cycling female Wistar rats with plasma estradiol and progesterone levels. Correlation of vaginal impedance measurements with female sex steroid levels has been examined in buffaloes (20), cynomolgus monkeys (52), and sheep (3) but has not been reported for the rat. The association between hormone levels and estrous stage as determined by vaginal cytology is well known (9, 31) and served as the standard for comparison and verification. In addition, correlation between vaginal impedance and estrous cycle stage as determined by vaginal cytology was examined.

Animals. This study was conducted in accordance with the National Institutes of Health guidelines for the care and use of animals in research. The Institutional Animal Care and Use Committee (IACUC) of the Johns Hopkins Medical Institutions (JHMI) approved all protocols. Sexually mature female Wistar rats (Rattus norvegicus, Hsd:WI) were purchased from Harlan Sprague Dawley (Indianapolis, Ind.) at 200 to 250 g body weight (approximate age, 64 to 72 days). Animal facilities at JHMI are accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International. However, for the duration of the study and as approved by the JHMI IACUC prior to study initiation, animals were housed separately in a study area (2) or satellite facility (41) maintained by the investigators. Animals were allowed to acclimate for 10 days before any experimental manipulations were initiated. Animals were group housed (n = 2, 4 per cage) in autoclaved filter-top covered polycarbonate cages (area, 840 cm2; height, 20 cm) containing corncob bedding (Quality Lab Products, Elkridge, Md.) in a temperature (20.0 to 22.2°C)- and humidity (30 to 70%)-controlled room on a 14:10-h light:dark cycle (lights on, 05:30 to 19:30). All animals were transferred to clean, autoclaved cages with fresh bedding twice a week in a class II biosafety cabinet (Biogard Hood, model B40-112, Baker Co., Sanford, Maine). Autoclaved rodent chow (RMH1000, PMI Feeds Inc., St. Louis, Mo.) and water were provided ad libitum from arrival until euthanasia.

According to quarterly vendor microbiologic monitoring reports, rats were serologically negative for the following pathogens: Mycoplasma pulmonis, Sendai virus, sialodacryoadenitis virus/rat coronavirus, Kilham rat virus, H-1 virus, rat parvovirus, pneumonia virus of mice, reovirus 3, Hantaan virus, Theiler's murine encephalomyelitis virus, mouse adenovirus, lymphocytic choriomeningitis virus, Clostridium piliforme, and Encephalitozoon cuniculi. As determined by the vendor, rats were negative by polymerase chain reaction for Helicobacter sp., cilia-associated respiratory bacillus, cytomegalovirus, and Mycoplasma pulmonis. Nasopharygeal and cecal cultures performed by the vendor were negative for dermatophytes and respiratory and enteric bacterial pathogens. Animals were free of internal and external parasites.

General study design and sampling. Based on chance and random sampling of the test population, 8 to 12 animals were expected to be in each estrous cycle stage. A 15-day, twice daily sampling schedule (30 sampling periods total) was used to allow evaluation of at least two complete estrous cycles per rat, based on a 4- to 5-day estrous cycle. All rats (n = 41) were sampled twice daily (morning and afternoon) for vaginal impedance and cytology measurements and were weighed daily. The morning sample collections started at 08:00, and afternoon samples were collected beginning at 16:00. All but terminal sampling measurements were done in awake, minimally restrained animals. Final sampling groups of animals (n = 3 or 4 per group) were terminally bled by cardiac puncture under halothane anesthesia (4 to 5% induction via chamber, 1 to 2% maintenance via facemask) during the last five morning and five afternoon sampling time points (10 time points total) to allow plasma hormone determination concurrently with vaginal cytology and impedance measurements. After cardiac puncture, animals were euthanized via decapitation while still deeply anesthetized (4 to 5% halothane anesthesia).

Vaginal impedance measurements. All vaginal impedance measurements were made with the Estrus Cycle Monitor EC40 (Fine Science Tools, Foster City, Calif.) per the vendor's recommendations and as previously described (44). One person restrained the animals while another inserted the vaginal probe for approximately 30 sec to take impedance measurements. All impedance measurements were done immediately before collection of vaginal swabs. The vaginal impedance probe was cleaned with 70% alcohol before each measurement.

Vaginal cytology. The appearance of vaginal papillae and any vaginal discharge was noted prior to collection of vaginal cytology. A sterile cotton-tipped swab moistened with sterile saline then was inserted in the vaginal opening. The swab was rotated gently against the dorsal vaginal wall and withdrawn. The appearance of the swab tip was recorded. The swab immediately was rolled onto a glass slide and allowed to air dry. All slides then were stained with Dif-Quick (International Medical Equipment, San Marcos, Calif.). Slides were examined for the following features: cornified epithelial cells, nucleated epithelial cells, leukocytes, and mucus. Estrous cycle stage then was determined using the following criteria (23, 31, 47, 54): 1) proestrus—predominantly nucleated epithelial cells, few cornified epithelial cells, and few leukocytes; 2) estrus—predominantly cornified epithelial cells; 3) metestrus—predominantly leukocytes and cornified epithelial cells; and 4) diestrus—predominantly leukocytes, some nucleated epithelial cells, and mucus. All slides were reviewed and classified by two separate reviewers. In cases of disagreement, a third reviewer determined final slide classification.

Blood collection for plasma hormone assays. Plasma progesterone and estradiol were measured using commercial radioimmunoassay kits (Diagnostic Products Corp., Los Angeles, Calif.). All samples were assayed in duplicate. Our inter- and intra-assay variations are 4.4% and 6.8%, respectively (22). In light of assay limits, minimal plasma volumes of 100 to 200 ml and 50 to 100 ml were needed to run a single estradiol and progesterone determinations, respectively. Therefore a minimum blood sample volume of 0.8 to 1.0 ml was required at each time point. In light of these sample limits, we elected to do a single terminal blood collection per animal and staggered animal sampling groups across one estrous cycle (5 days) for final blood sampling. Vaginal impedance and cytology measurements were collected before terminal blood collection. Terminal blood collection from animals under 4 to 5% halothane anesthesia was performed via cardiac puncture prior to decapitation.

Statistical analysis. All values are reported as mean ± the standard error of the mean unless otherwise indicated. Daily body weight (% baseline), estrous cycle stage vaginal impedance, and hormonal levels for each estrous cycle stage were subjected to one-way analysis of variance with the Tukey post hoc test. The relationship between vaginal impedance and hormone levels (estradiol, progesterone, and estradiol:progesterone ratio [E/P]) was examined by regression analysis for animals with or without respect to estrous cycle stage. A Pearson product moment correlation was done between vaginal impedance and hormone levels (estradiol, progesterone, and E/P) with or without respect to estrous cycle stage. To establish a method of correlation between vaginal cytology and impedance, each estrous cycle stage was assigned a number: proestrus = 1, estrus = 2, metestrus = 3, diestrus = 4. A Spearman's rank order correlation coefficient then was done between estrous cycle stage number and the vaginal impedance values. All statistical analyses were done using SigmaStat Version 3.0 (SPSS Inc., Chicago, Ill.). The criterion for statistical significance was P < 0.05.

All animals gained weight throughout the course of the study, with a significant (P < 0.05) increase in body weight compared with baseline levels (sampling day 1) beginning on sampling day 10 and continuing through to the end of the study on sampling day 15 (Table 1). Rats classified in estrus by vaginal cytology had significantly higher vaginal impedance values (3.6 ± 0.3 Ω, n = 37 samples; P < 0.05) than did rats in the other estrous cycle stages (proestrus, 2.2 ± 0.1 Ω, n = 40; metestrus, 2.6 ± 0.1 Ω, n = 41; diestrus, 2.4 ± 0.1 Ω, n = 36). No correlation was seen between estrous cycle stages and vaginal impedance measurements (r = 0.05439, P = 0.5029).

Table 1 Body weights in female Wistar rats during the 15-day sampling period

Estradiol and progesterone levels were evaluated in staggered sampling groups during the last 5 days of the 15-day sampling period. There were no significant differences in either estradiol or progesterone levels between estrous cycle stages (Table 2). Based on vaginal cytology, 54% (22 out of 41) of animals sampled for hormone levels were in proestrus. For proestrus animals (n = 22), no correlation was seen between vaginal impedance and estradiol (r = 0.139, P = 0.536), progesterone (r = 0.259, P = 0.245), or E/P (r = 0.109, P = 0.629). We did not attempt to correlate vaginal impedance and hormone levels for the other estrous cycle stages as the analysis would be underpowered due to small sample size (estrus, n = 7; metestrus, n = 3; diestrus, n = 6). Figure 1 depicts the correlation between vaginal impedance and hormone levels (estradiol, progesterone, and E/P) for all study animals (n = 41) regardless of estrous cycle stage. No correlations were seen between vaginal impedance and estradiol levels (r = -0.0647, P = 0.688; Fig. 1A), progesterone levels (r = 0.0929, P = 0.563; Fig. 1B), or E/P (r = -0.140, P = 0.382; Fig. 1C).

Table 2 Plasma estradiol and progesterone levels for female Wistar rats at various stages of the estrous cycle and sampled terminally on study days 11 through 15

Figure 1 Correlation between vaginal impedance and (A) estradiol levels, (B) progesterone levels, and (C) estradiol:progesterone ratio (E/P) in unstaged cycling female Wistar rats (n = 41). No correlations were seen between vaginal impedance and estradiol levels (more ...)

Several animals (5 of 41 study animals, 12%) exhibited abnormal vaginal smears or discharges at least once during the study and were submitted for necropsy (Table 3). Histological changes in the uterus included multifocal to coalescing areas with marked expansion of the luminal and glandular epithelium. The endometrium had papillary projections into the lumen and formed various sizes of cysts. The cysts were lined by simple low columnar epithelium. Multifocally the glands and the stroma contained edema, neutrophils, and scant macrophages. In three of five cases, the uterine histology was characterized as endometrial hyperplasia (cystic and papillary) with endometritis. The squamous epithelial cells lining the vagina were vacuolated and contained neutrophils (diestrus). The corresponding ovaries were predominantly luteal with no or single follicles. Two of these cases (animals 14 and 38) had a series of discrete nodules that were composed of decidual cells in the uterine wall (deciduomata), similar to the anatomical appearance of normal embryo implantation sites. These decidual cells were large, polygonal, less cohesive, multinucleated, and pleomorphic. Overall the reproductive organ histology in rats 14 and 38 was consistent with pseudopregnancy due to persistent corpora lutea as a result of cervical stimulation.

Table 3 Reproductive tract histology for female Wistar rats with abnormal vaginal smears or discharges

Beginning on sampling day 11 and continuing until the end of the study (sampling day 15), the number of animals decreased daily as animals were removed from the study for terminal blood sampling. However, the number of peak vaginal impedance measures and estrus events became less frequent even before sampling day 11. Of the 41 animals in the study, 4 (~10% of the population) did not demonstrate estrus on vaginal cytology throughout the 15-day sampling period. All animals had at least one time point at which peak vaginal impedance clearly was demonstrated. The number of peak vaginal impedance readings was highest during sampling days 1 through 5 of the study, with 26 animals (~63%) demonstrating peak vaginal impedance. Eight animals (~20%) showed peak impedance between sampling days 5 and 10. Only three animals demonstrated peak vaginal impedance during the final 5 days of the study (sampling days 11 through 15). Twenty of the rats with peak impedance values were classified as being in the estrus stage via vaginal cytology. In the remaining 17 animals, peak vaginal impedance did not occur during estrus as determined by vaginal cytology. However, in nine of these nonestrus animals, the vaginal cytology slides collected either immediately before or after the peak impedance measurement showed that the animal was in estrus. The vaginal impedance values prior to and after peak impedance did not demonstrate any appreciable pattern suggestive of an estrous cycle with four stages.

Slides demonstrating estrus were collected throughout the study. During sampling days 1 through 5, 20 animals (~49%) were established as being in estrus by vaginal cytology. According to their vaginal cytology, 9 rats were classified as being in estrus during sampling days 5 through 10, and 8 were in estrus during sampling days 11 through 15. Only 16 animals (~39%) were identified as being in estrus more than once during the course of the study. In these cases, the second estrus was associated with a smaller increase in vaginal impedance.

We designed this study to determine whether vaginal impedance measurements in intact cycling female Wistar rats could be correlated with female sex steroid levels (estradiol, progesterone). This study demonstrates several findings. First, only estrus could be readily distinguished from the other estrous cycle stages via vaginal impedance measurements in rats. However, no correlation was seen between estrous cycle stage and vaginal impedance. Second, no correlation was seen between vaginal impedance and hormone levels in proestrus animals or when all rats were considered irrespective of estrous cycle stage. Due to small sample size for three of the four estrous cycle stages, a correlation between vaginal impedance and hormone levels relative to estrous cycle stage could not be determined for estrus, metestrus, and diestrus animals. Lastly, 5% of study animals (2 of 41 rats) were determined to be pseudopregnant in light of reproductive tract pathology. Our results also suggest that vaginal impedance could be useful in distinguishing estrus from nonestrus animals but may be limited for chronic estrous cycle monitoring because of the risk of inducing pseudopregnancy.

Female rats were group-housed in this study. The induction of anestrus in females housed together is known as the Lee—Boot effect. Although this phenomenon is well known in mice, it does not occur as strongly in rats (47). However, this effect was not likely to have influenced the results of the study, as the majority of animals had higher estradiol and progesterone levels than would be expected in anestrus or noncycling females. In addition, ovarian cycle synchrony, which is known to occur in humans (8, 19, 32, 43, 45), has been reported in laboratory rats (33, 34). Estrous synchrony can develop in group-housed female rats after three or four ovarian cycles (15 to 20 days), although complete synchrony rarely is achieved (33). Animals in our study were group-housed for a total of 21 to 26 days, 10 days before beginning the study and, depending on the terminal sampling groups to which they were assigned, 11 to 15 days during the study. Therefore, it is likely that the some of the animals within each cage had synchronized estrous cycles during the terminal sampling period (sampling days 10 through 15). This phenomenon also may explain why more than half of the study animals (54%; 22 of 41 rats) were in one stage, proestrus, when blood samples were collected for hormone analysis. This type of ovarian synchrony in the rat is quite different from the synchrony described in female mice as the Whitten effect, where exposure of female mice to a male mouse or male urine induces estrus in group-housed females (57). This phenomenon in mice has been used to synchronize estrous cycles (13). The Whitten effect is not thought to occur in rats (47).

We used a twice-daily sampling schedule, similar to that of other studies that were evaluating the estrous cycle in rodent species other than Rattus (18, 38). Previous studies using Sprague-Dawley rats have reported daily vaginal cytology sampling for as long as 60 days (33, 35). Similarly, in a group of 100 rats, consecutive twice-daily vaginal impedance measurements were taken during a 15-day sampling period (51). In both studies, no adverse events or pseudopregnancies were reported and all animals demonstrated normal 4- to 5-day estrous cycles. In addition, we monitored each animal's weight on a daily basis. Despite frequent handling, the rats increased in body weight throughout the study, indicating that they likely were stressed only minimally by the chronic sampling procedures and repeated handling.

Approximately 5% of study animals (40% of those with abnormal vaginal discharges or smears) had histological findings suggestive of pseudopregnancy. Pseudopregnancy in the rat occurs when the female receives cervical stimulation, either from a vasectomized male or mechanically from insertion of a probe or swab. Once such stimulation occurs, the corpora lutea are released from a quiescent state, become functional, and secrete progesterone (21, 26). Luteal function is maintained by increased prolactin secretion that also occurs in response to cervical stimulation. Pseudopregnancy in the rat lasts over 14 days with a long diestrus (21, 50). Because the lifespan of the corpora luteum in the pseudopregnant rat is very similar to that in other mammalian species, the pseudopregnant rat has been used as a model to study luteolytic regulatory mechanisms (31). The most likely cause of pseudopregnancy in our study would have been the sampling method, which required each subject to undergo a total of six vaginal insertions daily (two for impedance measurements and one vaginal swab twice daily for 15 days).

There were no significant differences in either estradiol or progesterone levels between the estrous cycle stages sampled, likely due both to the small sampling size for three of the four estrous cycle stages and to the overlapping range of values reported in the literature for each stage. Female sex steroid levels vary in intact female rats depending on estrous cycle stage and length, reproductive status, and age. The highest levels of estradiol occur during early proestrus (approximately 40 to 188 pg/ml), with values decreasing to their lowest levels during estrus and metestrus (approximately 15 to 47 pg/ml; 9, 39, 48). With respect to progesterone, the highest levels have been observed during late proestrus and early estrus (approximately 29 to 76 ng/ml) and with lower levels described during metestrus to diestrus (approximately 1 to 27 ng/ml; 9, 39). In pregnant rats, estradiol and progesterone levels can range from < 30 to 840 pg/ml and approximately 7 to 95 ng/ml, respectively, depending on gestation day (14, 17, 46). During pseudopregnancy, estradiol levels range from approximately 50 to 285 pg/ml, whereas progesterone values have been reported as > 25 to 147 ng/ml (14, 17, 48, 55, 56). Progesterone in lactating females has been cited to be around 67 ± 6 ng/ml (49).

In our study, estradiol levels measured for all stages were within reported ranges for cycling female rats. However, mean progesterone levels for diestrus and metestrus were higher than would be expected for normal, cycling females but within the ranges reported for pseudopregnant rats. The majority of animals sampled for hormone levels were in proestrus according to their vaginal cytology, and progesterone levels for estrus and proestrus rats fell within expected ranges for cycling animals and at the low end of progesterone values for pseudopregnant rats.

The issue of linearity between hormone levels and vaginal impedance in a staged female rat remains an open question at this time. We were unable to determine such a relationship in our experiments relative to estrous cycle stage and found no association between vaginal impedance and hormone levels when estrous cycle stage was not considered. In contrast to our findings, Bartos (4) observed a transient increase in vaginal impedance that correlated strongly with changes in estradiol levels. We also found no correlation between estrous cycle stage and vaginal impedance in rats. However, Lilley and colleagues (29) were able to correlate impedance changes with cytological changes observed in vaginal smears from guinea pigs. These findings suggest that the association of vaginal impedance with respect to estrous cycle stage and hormone levels may vary from species to species or even between rodent stocks and strains. In order to further examine whether correlation between vaginal impedance and hormonal levels exists at least in estrus rats, blood samples could be collected at the estrus time point associated with greatest increase in impedance. Because of the overlap of hormone values reported for the various estrous stages, the inclusion of levels of luteinizing hormone or follicle stimulating hormone might help to further define and confirm the occurrence of estrus.

Our results do suggest that vaginal impedance could be a viable alternative for distinguishing estrus from nonestrus rats. A study that used a commercially available digital multimeter to measure vaginal wall resistance demonstrated that peak resistance occurred when the majority of vaginal epithelial cells were cornified, followed next by nucleated and then leukocytic cells, a cellular profile typical of early estrus (44). We also observed peak impedance values during estrus. In contrast with these observations, others have reported that peak vaginal resistance in rats occur during proestrus (5). In our study, in 19 cytology sampling events classified as proestrus, the vaginal impedance was higher than the measurement taken at the next day and time point when the cytology demonstrated estrus. Conflicting results in the estrous cycle stage associated with peak electrical resistance also have been reported to occur in the guinea pig (6, 29). In an earlier study, peak resistance was found to be associated with proestrus, but no vaginal cytology data were presented (6). One study, in which detailed vaginal cytology data were provided, demonstrated peak vaginal impedance during metestrus (29).

In summary, our study indicates that estrus is characterized by a marked increase in vaginal impedance, that there may be a poor correlation between vaginal impedance and hormone levels in unstaged female rats as well as between vaginal impedance and estrous cycle stage, and that vaginal impedance measurements may not be a viable option for staging cycle female rats in chronic studies because of the possible induction of pseudopregnancy. Because problems with multiple vaginal samplings of experimental rats have been underreported in the literature, it is important for investigators to be aware of sampling issues and to report such complications in vaginal impedance and cytological sampling of rats in future publications.

This study was supported by National Institutes of Health grant RR00163 (to S.J.M.) and National Aeronautics and Space Administration and National Space Biomedical Research Institute grant RE00205 (to D.L.H.). We would like to thank Erin H. Holt, B.S., M.P.H., for her assistance with the statistical analysis. We also thank the following staff members from the Anesthesia Core Research Laboratories at the Johns Hopkins Medical Institutions (Baltimore, Md.) for the their technical assistance during the course of this study: Kathleen Blizzard, Indira Debchoudhury, Jessica Gaines, Diane Solana, Judy Klaus, and Mike Sabol.