Background
Polycystic ovary syndrome (PCOS) is one of the most common endocrine conditions, affecting 9–18% of women of reproductive age, depending on diagnostic criteria [
1]. PCOS is characterized by at least two of the three following criteria: clinical or biochemical hyperandrogenism, oligo/anovulation, and polycystic ovaries [
2]. Clinically, these manifestations are associated with reduced fertility, due to dysfunctional follicular maturation and consequent anovulation, and hyperandrogenism, causing acne and hirsutism [
3]. Both chronic anovulation and androgen excess are linked to disturbed folliculogenesis, that is expressed by multiple cystic follicles between 2 and 9 mm and increased ovarian volume in women with PCOS [
4].
Clinical PCOS may have its onset before or during puberty [
5]. Clinical studies have suggested that girls with PCOS exhibit increased gonadotropin releasing hormone (GnRH) pulse frequency, leading to excess luteinizing hormone (LH) secretion. This causes premature acquisition of LH receptor expression by growing ovarian follicles at excessively early stages, leading to increased ovarian androgen production [
6,
7] and, probably, to the arrested antral follicle development of PCOS [
8]. Indeed, this follicular arrest is consistent with the polycystic ovary morphology found on ultrasound examination in human PCOS [
2]. Moreover, PCOS-affected ovaries exhibit an increase in the number of growing preantral and antral follicles, which leads to antrum expansion, increased granulosa cell degeneration, and development of cystic follicles, with thin granulosa cell walls and a thicker surrounding layer of theca cells [
8].
Since Stein and Leventhal first described PCOS in the mid-1930s [
9], the search began to identify the etiologic mechanisms associated with its development. Yet, despite its prevalence and health impact, the etiology of PCOS remains poorly understood. Even whether reproductive hormone abnormalities are primary or secondary remains enigmatic. Hypotheses for the origins of this pathology include hormonal imbalances, genetic abnormalities, and lifestyle and environmental factors [
3]. Due to logistic and ethical limitations on human experimentation, several animal models to simulate PCOS have been developed in recent decades. These experimental models can improve understanding of the pathophysiology of PCOS and have the potential to support development of innovative and curative treatments.
Within this context, the aim of this paper was to systematically review the available rodent and non-human primate animal models of PCOS-like conditions and summarize ovarian function and hormonal disturbances linked to reproductive damage, as well as discuss the advantages and possible limitations of applying these models to human PCOS.
Methods
Search strategy and study selection
We performed a systematic review in accordance with the PRISMA guidelines. Briefly, we searched the MEDLINE database, via PubMed, for literature published in English and Spanish until October 2016. The search strategy consisted of the following Medical Subject Headings (MeSH): “Animal Model” OR “Animal Models” OR “Model, Animal” OR “Laboratory Animal Models” OR “Animal Model, Laboratory” OR “Animal Models, Laboratory” OR “Laboratory Animal Model” OR “Model, Laboratory Animal” OR “Models, Laboratory Animal” OR “Experimental Animal Models” OR “Animal Model, Experimental” OR “Animal Models, Experimental” OR “Experimental Animal Model” OR “Model, Experimental Animal” OR “Models, Experimental Animal” OR Rodents OR Rodentias OR Rodent OR Rat OR Rattus OR “Rattus norvegicus” OR “Rats, Norway” OR “Rats, Laboratory” OR “Laboratory Rat” OR “Laboratory Rats” OR “Rat, Laboratory” OR Mus OR Mouse OR “Mus musculus” OR “Mice, House” OR “House Mice” OR “Mouse, House” OR “House Mouse” OR “Mus domesticus” OR “domesticus, Mus” OR “Mus musculus domesticus” OR “domesticus, Mus musculus” OR “musculus domesticus, Mus” OR “Mice, Laboratory” OR “Laboratory Mice” OR “Mouse, Laboratory” OR “Laboratory Mouse” OR “Mouse, Swiss” OR “Swiss Mouse” OR “Swiss Mice” OR “Mice, Swiss” OR AND “Ovary Syndrome, Polycystic” OR “Syndrome, Polycystic Ovary” OR “Stein-Leventhal Syndrome” OR “Stein Leventhal Syndrome” OR “Syndrome, Stein-Leventhal” OR “Sclerocystic Ovarian Degeneration” OR “Ovarian Degeneration, Sclerocystic” OR “Sclerocystic Ovary Syndrome” OR “Polycystic Ovarian Syndrome” OR “Ovarian Syndrome, Polycystic” OR “Polycystic Ovary Syndrome 1” OR “Sclerocystic Ovaries” OR “Ovary, Sclerocystic” OR “Sclerocystic Ovary”.
The selection criteria were as follows: experimental studies of interventions in female rats, mice, guinea pigs, or non-human primates aiming to induce PCOS-like reproductive characteristics, namely, follicles with cystic appearance and altered serum testosterone levels. We also performed a hand search of the reference lists of full-text articles. Studies were excluded from the analysis if the outcome was not induction of PCOS symptoms in experimental animals and if reproductive outcomes were nor described.
Two investigators (LP and RBR) independently reviewed the titles and abstracts and selected articles for full-text review. Disagreements were resolved by a third reviewer (PMS) or by consensus discussion. The selected articles were independently read in full to confirm eligibility and extract data. The information extracted from each individual study was as follows: name of first author, publication year, animal species, number of animals, age at start and at intervention time, intervention type, experimental period, and three variables of interest: serum levels of sex hormones, serum levels of gonadotropin, and ovarian morphology.
Discussion
Animal models are regarded as valuable tools to investigate pathophysiological processes of human diseases. Indeed, in most cases, because of obvious ethical concerns, some relevant queries cannot be answered by directly studying affected patients. In PCOS, additional complicating issues are its heterogeneous clinical presentation and the fact that the etiology is still not well defined [
49]. Therefore, in the present review, we specifically selected studies focusing on two main endocrine traits of PCOS: ovarian morphology changes and circulating levels of sex hormones and gonadotropins.
Reviews about animal models of PCOS have been published previously, and have addressed various features in different animal models of PCOS-like phenotypes [
14,
50‐
54]. However, this is the first systematic review to provide a full list of rodents and non-human primate models generated by distinct interventions, specifically focusing on two main reproductive features present in women with PCOS: hyperandrogenism and polycystic ovaries.
We included 39 experimental studies which used distinct procedures to induce PCOS-like models of ovarian abnormalities and androgen excess, stratified into those using androgens [
10‐
28], estrogens and endocrine disruptors [
27,
29‐
36], or other interventions [
37‐
48]. Overall, there were broad differences among the studies concerning hormonal interventions, animal species, and developmental stage at the time of the experiments. Most resulted in ovarian morphological changes, mainly increases in the number of antral and cystic follicles and decreases in the corpus luteum. However, while a hyperandrogenic status could be induced by using androgens [
10‐
27] and other drugs [
37‐
47] as stimulatory agents, studies using drugs with estrogenic effect did not measure androgen levels or did not observe changes in circulating androgens [
27,
29‐
36]. Therefore, hormonal interventions using androgens seem to promote the most consistent features of a PCOS-like phenotype in animals, as previously suggested by Abbott et al. [
49].
Among studies generating a hyperandrogenic state, prenatal exposure of non-human primates to androgens resulted in the most suitable animal model, displaying both metabolic and reproductive characteristics of PCOS [
55,
56]. However, these models are expensive and are not readily adaptable to genetic manipulation. In turn, rodent models provide a versatile tool for investigating biological mechanisms associated with the development of PCOS. Among the advantages of using rodent species, their stable genetic backgrounds, ease of handling and maintenance, shorter reproductive lifespan, and short estrous cycles are the most important. However, some aspects that limit the use of rodents for investigation of reproductive features should be taken into consideration. First, rodents are polyovulatory, while women are mono-ovulatory, suggesting that, despite similarities in the hypothalamic-pituitary-ovarian axis, the FSH-dependent follicle selection process in rodents differs from that in women [
57,
58]. Second, although the initial stages of follicular growth (from primordial to preantral stage) seem to be comparable between humans and rodents, differences in regulation by intra-ovarian growth cannot be ruled out [
59]. Finally, there are marked differences in the timing of onset of folliculogenesis between rodents and women. While the primordial follicle pool and initiation of follicle growth may arise during the later stages of fetal development in humans, these processes occur only during the early postnatal period in rodents [
60]. Thus, results obtained from mice and rats may not translate directly to women.
Interestingly, differences in the generation of reproductive phenotypes are observed according to the developmental period in which androgen treatment is started [
11]. In this sense, either starting testosterone treatment postnatally [
14‐
18] or administering DHT treatment during the prepubertal period [
19,
20] leads to the development of cyst-like follicles. However, postnatal exposure to DHT results in reprogramming of the hypothalamic-pituitary-ovarian-axis [
61]. Thus, comparisons between different intervention models may be useful to define the timing of reproductive PCOS phenotypes in experimental animal models. One example is the study of Ota et al. [
17], in which, after treatment of 5-day-old female rats with a single dose of testosterone propionate, various reproductive characteristics of PCOS – such as cystic follicles, anovulation, and imbalances in gonadotropins and sex hormones – were later found over a 200-day observation period. These results suggest androgen induction may have indirectly promoted a pathologic elevation of FSH [
17] that blocked ovulation and induced cystic formation.
Estrogens and drugs with estrogenic effects have been used to induce a PCOS-like syndrome in animals [
52] because of their ability to induce continuous estrus and cystic follicles, with morphologic characteristics resembling those observed in women. However, few studies using these treatments have demonstrated high androgen levels in blood. This was the case in a study in which neonatal female rats were treated with the endocrine disruptor bisphenol A [
32], and exhibited high testosterone levels and numerous cystic and atretic follicles later in life. Possibly, acute exposure to estrogen could lead to changes in follicular enzyme activity and subsequent suppression of androgen production by theca cells. BPA is a potential agonist of endocrine estrogen, acting differentially depending on tissue estrogen receptor expression [
62]. This suggests BPA could have less of an effect on regulation of hypothalamic-pituitary-ovarian axis negative feedback and, consequently, on ovarian androgen secretion. Therefore, estrogen-induced intervention may not be the optimal experimental models for study of PCOS.
Other experimental interventions using external physical stressors have been reported to induce reproductive features similar to the PCOS phenotype. Chronic cold stress may produce such changes in ovarian morphology by marked central activation of the sympathetic nerves to the ovary [
42]. Such activation by cold stress is probably mediated through a regulatory mechanism on the hypothalamic-pituitary-adrenal axis by the locus coeruleus [
42]. In addition, continuous light exposure could lead to changes in the estrous cycle, such as continuous estrous and cystic follicles, by altering the circadian system [
41].
Letrozole, an aromatase inhibitor, was another hormonal intervention that induced high androgen levels and ovarian cysts [
40] by inhibiting androgen conversion to estrogen and promoting alteration of the hypothalamic-pituitary-gonadal axis and high LH levels. In addition, by similar mechanisms, a transgenic mouse model successfully generated reproductive abnormalities by promoting recombination of the LHβ gene and hCG β-subunit [
37], thus inducing chronic elevation of LH levels as well as increased testosterone and estrogen levels and cystic follicles.
Finally, although genetic rodent models cited in the present review also fail to fully replicate the reproductive phenotype of PCOS, the use of different transgenic animals may be useful to identify potential pathways involved in alterations in reproductive and endocrine aspects in these animals, which, in turn, may lead to important clinical insights into the etiology of human PCOS.
One limitation of the present systematic review is that we did not search for animal models related to the metabolic abnormalities often associated with PCOS in women. However, although insulin resistance is frequently found in PCOS, it is not considered its primary etiology. Another limitation is that we did not search for studies using other animal species than rodents and non-human primates, as these animals require no unusual laboratory facilities. A third limitation is that we did not perform meta-analysis, as the great heterogeneity in animal species and experimental procedures precluded clustering of different studies to test the efficacy of each model in producing the expected characteristics.