Spontaneous Metritis Related to the Presence of Vaginal Septum in Pregnant Sprague Dawley Crl:CD(SD) Rats

Abstract

Recently, 6% of 1,176 Sprague Dawley rats examined in our reproductive toxicity studies presented with dark-red uterine contents with or without fetuses demonstrating delayed development. Sometimes, a high proportion of the litter was found dead, and dystocia with death or preterminal euthanasia of the dam occurred. Microscopic findings in the uterus consisted of necrohemorrhagic and suppurative periplacentitis associated with the presence of bacterial colonies identified as Escherichia coli. In the vagina, similar findings were observed that were associated with mucus accumulation and the presence of a transverse occlusive or partially occlusive thin membrane identified as a vaginal septum. Microscopically, this septum consisted of a thin band of connective tissue covered on both sides by a mucous epithelium that was continuous with vaginal epithelium. In some cases, there was only mucus accumulation retained by a septum in the vagina without evidence of bacterial infection. Serological and histological examinations did not reveal any specific pathogenic agent. The presence of these septa in the vagina most likely favored mucus accumulation, nonspecific ascending bacterial infection, and dystocia.

This colony of rats presented with an unusually high incidence of vaginal septa as it was described in different strains of mice and rats in the past. We hypothesized that the use of an impedance meter by the breeder—to determine the phase of the estrous cycle by introducing a probe in the vagina—likely facilitated gestation by perforating the vaginal septum in some cases.

Vaginal malformations in humans (eg, atresia, persistent hymen, hypoplasia, dysgenesis) are generally rare and can result from genetic factors or prenatal exposure to a toxicant, or they may have an unknown etiology. ⁵ Transverse (occlusive or partially occlusive) and longitudinal vaginal septa have been described in various species, including different strains of mice, ¹⁰ Wistar rats, ³ one Sprague Dawley rat, ¹ and in humans. ⁹

The presence of transverse septa usually results in the accumulation of mucus and cellular debris in the anterior part of the vagina and in the uterus (mucometra), which can lead to mucopurulent vaginitis and endometritis in rats. ³ In humans, the absence of perforation could result in retention of menstrual blood and cervical mucus, and hematocolpos may develop (vagina filled with menstrual blood). Furthermore, in women with a septum, dyspareunia (painful sexual intercourse) and obstructed and painful labor can occur. ⁹

Since 2006, in our facilities, an unusually high incidence of transverse vaginal septum associated with metritis was noted in mated Sprague Dawley Crl:CD(SD) female rats and thereby interfered with the interpretation of our reproductive toxicity studies (embryofetal and pre- and postnatal developmental studies). The presence of the vaginal septa did not result in a lower fertility rate, possibly because of the procedure of identification of the estrous cycle in female rats (use of an impedance meter) that resulted in perforation, thereby enabling fertilization.

Material and Methods

During the past few years (2006-2009), 1,176 Sprague Dawley Crl:CD (SD) rats (10-12 weeks old) originating from the same supplier were used in our laboratory for embryofetal and pre- and postnatal developmental studies. In embryofetal toxicity studies, dams were treated with test compound between gestation day (GD) 6 to GD17 (main organogenesis period). On GD21, the dams were euthanized, and the uterus was examined macroscopically for identification of implantation sites, resorptions, and live or dead fetuses; live fetuses were weighed, sexed, and examined for external, visceral, and skeletal defects. In pre- and postnatal developmental studies, dams were treated from GD6 throughout lactation period to weaning (21 days after parturition).

In all of these studies, mated female rats were supplied by Charles River Laboratories France. The supplier selected females for pairing using an impedance meter to determine the estrous cycle phase. A probe that was 1.5 cm long and 3 mm in diameter was introduced into the vagina to measure the local electrical resistance. ⁸ Females in the estrous phase were then placed with a male, and mating was confirmed by the presence of the vaginal plug after a 4-hour pairing.

In one study, an in-house natural mating (without the use of an impedance meter) was performed using approximately 200 mature females and 25 breeder males to obtain 96 Crl:CD (SD) mated female rats (pairing ratio: 1 male for 3 females maximum overnight) to evaluate the potential impact of the use of an impedance meter. The mating was evidenced by the presence of sperm in the vagina using cytological examination.

At necropsy, macroscopic examination of the entire genital tract was performed. In some studies, if abnormal uterine and/or vaginal contents were observed, the ovaries, uterus, placentae, fetuses, and vagina were sampled and fixed in 10% buffered formalin. The liver, heart, brain, spleen, kidney, and lung of some affected pregnant females were also sampled and fixed in formalin. These tissues were processed to paraffin block, sectioned, stained with hematoxylin-eosin (H&E) or hematoxylin-eosin-saffron (HES) and Periodic Acid-Schiff (to identify a possible fungal infection) and examined microscopically. Cytological samples of the uterine and vaginal contents were stained with May-Grünwald-Giemsa. Samples from selected uteri presenting with abnormal content were taken for microbiological examination, and serum samples from selected affected dams and from sentinel animals were taken for serological evaluations for rat coronavirus (RCV/Sendai), parvovirus (RPV 1&2, PARVO NS1, TOOLAN H1, Kilham RV), pneumonia virus of mice (PVM), reovirus (REO3), Hantaan, Lymphocytic choriomeningitis virus (LCMV), Theiler, Clostridium piliforme, Mycoplasma spp., Encephalitozoon cuniculi, and CAR Bacillus; these samples were shipped to Charles River and Vébiotel Laboratories, France. These analyses were performed in 3 different studies during the period 2006-2009. Each time, one to three rats were analyzed.

All studies and the procedures of our laboratories complied with the French regulation (Decree 87-848, revised 2001) implementing European Directive 86/609 and European Convention ETS123. Furthermore, the studies were performed in accordance with the standards of the ILAR Guide^4bis (1996) and the sanofi-aventis Charter on the Humane Care and Use of Laboratory Animals, including ethical review.

Results

Anatomic Pathology Observations

Macroscopic examination at GD21 of pregnant SD rats from these studies revealed the presence of an abnormal number of rats with red hemorrhagic or dark fluid in the uterus (Fig. 1) with an abnormal development of fetuses (Fig. 2). When compared to a normal fetus (Fig. 3), fetuses from affected pregnant females were pale, with delayed development (smaller size), and with a hemorrhagic fibrin clot often seen at the surface of the corresponding placentae (Fig. 4, arrow). In one of the females found dead before the end point of the study, the uterus was filled with a dark liquid containing tissue debris, and only one macerated fetus was identifiable (Fig. 5).

Fig. 1. Uterus, rat, gestation day 21; abnormal hemorrhagic uterine contents.

Fig. 2. Litter from an affected dam, rat; gestation day 21; fetuses with different sizes (delayed development).

Fig. 3. Normal fetus and placenta, rat, gestation day 21.

Fig. 4. Pale rat fetus with delayed development from an affected dam, gestation day 21; hemorrhagic placenta (arrow).

Fig. 5. Macerated rat fetus from an affected dam found dead at gestation day 19.

Fig. 6. Normal vagina, rat.

Fig. 7. Vagina, rat, vaginal septum with necro-hemorrhagic fluid accumulation.

Fig. 8. Vagina, rat, thin vaginal septum.

Fig. 9. Vagina, rat, thick vaginal septum.

Fig. 10. Vagina, rat, mucus plug accumulation in the proximal part of the vagina.

Fig. 11. Normal placenta, rat, hematoxylin-eosin (15×).

Fig. 12. Placenta, rat, hemorrhagic and suppurative periplacentitis, hematoxylin-eosin (15×).

Fig. 13. Placenta, rat, higher magnification of Fig. 12. Presence of bacterial colonies (arrow), hematoxylin-eosin (200×).

Fig. 14. Cytological smear of the vaginal contents, rat. Presence of mucus, bacteria, and neutrophils, May-Grünwald-Giemsa (400×).

Fig. 15. Vagina, rat, vaginal septum, hematoxylin-eosin-saffron (15×).

Fig. 16. Vagina, rat, vaginal septum (higher magnification of Fig. 15), presence of myofiber (arrow), hematoxylin-eosin-saffron (100×).

When compared to a normal vagina (Fig. 6), the presence of a transverse occlusive or partially occlusive membrane was identified as a vaginal septum of variable thickness (Figs. 7–9). Rarely, the vaginal septum was associated only with the presence of a dense mucous plug in the anterior part of the vagina, without evidence of uterine and fetal anomalies (Fig. 10). (Cervix on the top for Figs. 6–10.)

Microscopic findings of all organs sampled from affected dams and fetuses were limited to the female genital tract and placentae. In all uteri examined, the lumen and glands were filled with variable amount of cellular debris, mucus and neutrophils. When compared to a normal placenta (Fig. 11), placentae from affected dams were covered by a dense accumulation of fibrin, red blood cells mixed with degenerated neutrophils, and bacterial colonies (Figs. 12 and 13). Large areas of ischemic necrosis were observed in a few of the affected placentae. Cytological examination of vaginal contents confirmed the presence of bacteria, mucus, and neutrophils (Fig. 14).

In the vagina of affected uteri, the septum consisted of a thin band of vascular connective tissue sometimes associated with some myofibers (Fig. 16, arrow), and covered by a mucous and/or squamous epithelium that was continuous with vaginal epithelium (Figs. 15 and 16).

Serological examinations did not demonstrate the presence of any specific rat pathogen. The negative PAS stain ruled out fungal infections. All bacterial cultures obtained from the abnormal uterine contents identified the presence of a large amount of Escherichia coli bacteria only.

Incidence

In all embryofetal toxicity studies evaluated, the incidence of vaginal septum associated with metritis ranged from 0 to 10% per study (Table 1: nos. 1 to 10). Approximately 6% of mated females from all embryofetal toxicity studies (785 mated rats) had vaginal septum/mucus accumulation and/or abnormal uterine content in both control and treated groups.

Table 1.

Detailed Incidence of Vaginal/Uterine Findings and/or Dystocia by Studies From Both Control and Treated Groups

Study number	Number of Mated Female Rats	Number of Mated Female Rats With Affected Genital Tract a
Study 1	100	10
Study 2	96	4
Study 3	96	0
Study 4	96	7
Study 5	60	5
Study 6	106	6
Study 7	16	2
Study 8	24	0
Study 9	96	8
Study 10	95	5
Study 11	100	10
Study 12	159	5
Study 13	132	8
Total	1176	70 (6%)

^a Presence of a septum associated with mucus, metritis, and/or dystocia

In pre- and postnatal developmental toxicity studies (Table 1: nos. 11 to 13), 23 of 391 mated females presented with a vaginal septum that was associated most often with dystocia, representing an incidence of 5.9% of mated females received from the supplier. For example, in one pre/postnatal study, vaginal septa were observed in 10/100 rats; dystocia, death of dams, and loss of the entire litter near the time of parturition were observed in 8 of the 10 affected rats.

In the embryofetal toxicity study in which an impedance meter was not used (in-house natural matings), an unusually high incidence of nonpregnant females (19%) was observed. No macroscopic examination of the vagina was performed at this time to check for the presence of a septum; however, in the past, the background incidence of nonpregnant female rats for this study type was approximately 10%.

Investigations into the origin of the mated female rats (breeding room) in the supplier facilities revealed that most of the females with genital tract anomalies were from a specific breeding area.

Discussion

It was concluded that the high incidence of metritis that was observed in pregnant females in our facilities was a consequence of an abnormal number of female rats with a vaginal septum (partial or complete) that possibly reflected a genetic drift in the supplier facilities. In the literature, the incidence of vaginal septa in Wistar rats is reported to be approximately 2%. ³ In our studies with Sprague Dawley Crl:CD(SD) rats, the overall incidence was 6%, with a maximum of 10% in some reproductive toxicity studies.

A high incidence of metritis associated with the presence of vaginal septum in reproductive toxicity studies may seriously interfere with the evaluation of data. Indeed, vaginal septum can lead to mucus accumulation in the vagina, vaginitis, metritis, placentitis, reduced fetal weight, pale fetuses, growth retardation, and, in some cases, fetal death, as observed in our studies. In pre- and postnatal developmental toxicity studies, dystocia may occur that can be associated with the death or euthanasia of the dam and entire litter losses. As a consequence, the evaluation of the developmental toxicity of a compound could be influenced. Vaginal septa and their sequelae may also have consequences on fertility studies by having an impact on fertility, fecundity, and gestation indices. A systematic examination of the vagina must be done to identify vaginal septa to cull affected dams from reproductive toxicity studies or to take this finding into account during interpretation of results.

It is hypothesized that the use of an impedance meter by the breeder to determine the phase of the estrous cycle could cause perforation of the vaginal septum, which may favor pregnancy of females that would not become pregnant by natural mating. Furthermore, in the study in which the female rats were mated naturally, the incidence of nonpregnant females was unusually high, which suggested that the use of an impedance meter likely allowed pregnancy to occur in rats with a vaginal septum. The hypothesis that vaginal trauma caused by the impedance meter can induce the presence of vaginal septa is highly unlikely; indeed, these septa are very well organized structures with connective tissue, myofibers and a well-defined epithelium on both sides that do not fit with scar or healing tissue.

In humans, transverse and longitudinal vaginal septa have been described. Transverse septa are approximately 1 cm thick and may be complete or have a small central or off center perforation. They are usually isolated defects, but they are also a component of the McKusick-Kaufman syndrome, which includes congenital heart malformation and polydactyly. ⁶ In our studies, no other morphological anomalies were reported in affected females, but careful examination of other organs was not systematically performed, and we are not able to rule out the presence of possible other congenital malformations. In different mouse strains (BALB/c, B6, DBA, etc.), the presence of vaginal septa was suspected to be inherited as a complex recessive genetic defect; ⁷ in a Balb/cJ mouse colony, 38% of females presented with longitudinal septa that were likely induced by more than one recessive gene, the expression of which was influenced by the genetic background. ²

Morphologically, in humans, the transverse septum is distinguished from an imperforate hymen by the identification of a hymen separated from the septum. In the rat, a membrane that closes the distal part of the vagina and that is observable by the naked eye is present during the prepubertal period. ⁴ In Sprague Dawley rats, this membrane degenerates between postnatal days 30 and 38, at the time of puberty. In the present studies, the presence of a septum in the middle of the vagina allowed us to rule out a possible persistence of this prepubertal membrane; the presence of a septum was likely related to an incomplete canalization of the vaginal sinus during organogenesis.

The absence of perforation in humans results in retention of menstrual blood and cervical mucus (hematocolpos) and, in rare cases, obstructed labor may occur. ⁹ In different strains of mice (BALB/c, B6, DBA, etc.), imperforate vagina leads to a marked accumulation of mucus or fluid in the uterus (mucometra, hydrometra). In one study, the authors isolated 3 types of bacteria in 10% of the affected mice that were considered as contaminants, as none of the uteri had microscopic findings consistent with an infection. ¹⁰ In contrast, in Wistar rats ³ as in our studies in Sprague Dawley Crl:CD(SD) rats, a transverse septum (occlusive and/or partially occlusive) can cause frequent mucopurulent metritis and vaginitis. The presence of these septa most likely caused mucus accumulation in the vagina followed by secondary nonspecific bacterial infection resulting from proliferation of E. coli bacteria. With regard to our observations, it appeared that most bacterial proliferation associated with metritis and placentitis likely occurred during the third part of gestation.

In an attempt to reduce the incidence of longitudinal septa in a colony of Balb/cJ mice, the breeding of only nonseptate female Balb/cJ mice failed. This result suggests that responsible genes were fixed within this strain of mice in this colony. ² It is therefore possible that the unusually high number of rats with vaginal septa in our studies may have been related to genetic drift. Eventually, after discussion with the supplier, new Sprague Dawley breeder males from another colony were recently introduced into the specific breeding area to try to decrease the incidence of vaginal septum and to balance the suspected genetic drift.