https://orcid.org/0000-0002-8249-263X
The common hyperandrogenic, subfertile, and obesogenic polycystic ovary syndrome (PCOS) takes approximately 2 years to diagnose, often after several changes of physicians. Such delayed care delivers symptom relief but provides no cure. Progress toward a cure and timely clinical intervention, however, has been hindered by a scarcity of animal models that reliably replicate PCOS (1). Several PCOS animal models, nevertheless, implicate hyperandrogenism, likely acting through the androgen receptor (AR), as a common denominator for PCOS etiopathogenesis, involving genetic, epigenetic, and developmental interactions (1). Consistent with this notion, flutamide treatment of women with PCOS can normalize episodic release of progesterone-regulated pituitary luteinizing hormone and likely hypothalamic gonadotropin-releasing hormone (2), inducing ovulation (3).
Using transgenic mouse models, Walters et al examined AR’s functional involvement in the etiopathogenesis of PCOS by employing genetically manipulated mice in which AR gene expression was diminished or knocked out (KO) globally or in specific cell types of discrete organ systems (4). When such AR gene manipulations are combined with transient prenatal or persisting peripubertal-onset dihydrotestosterone (DHT) administration, these ARKO models provide unique insight into how ARs in specific cells contribute to the origins of PCOS-like traits (4). Specifically, absence of subsets of DHT-induced PCOS-like traits within AR gene-silenced mouse cells expressing a specific molecule, such as calcium/calmodulin-dependent protein kinase II alpha (CamKIIalpha) in neuron-specific ARKO (NeuroARKO) (4), identify contributions made by forebrain neurons to androgen-mediated etiopathogenesis of PCOS-like traits.
In this issue, Walters et al (5) utilized their mouse genetics toolbox to cross ARflox mice with transgenic fatty acid-binding protein 4–Cre (FABP4-Cre) mice, thus limiting excision of AR exon 3 (and AR gene KO) to neurons, neuroglia, adipocytes (white and brown), and macrophages that all express the intracellular fatty acid chaperone FABP4 (white and brown adipose and brain-specific ARKO mice, AdBARKO), thereby preventing peripubertal DHT-induced PCOS-like reproductive and metabolic traits (Fig. 1). Detection of intact AR gene expression in the pituitary, ovary, uterus, liver, skeletal muscle, and kidney contrasts with its absence in the brain and white subcutaneous (inguinal) and brown adipocytes, indicating successful selective cell type ARKO; its partial absence in visceral white adipocytes (parametrial and mesenteric) may reflect inclusion of stromal-vascular cells in parametrial and mesenteric fat cell preparations (5). Macrophages and neuroglia were not examined for AR expression.
Diagrammatic representation of the potential sites in female mice enabling induction of PCOS-like traits from peripubertal onset of androgen action. In contrast, selective genetic knockout of AR in neurons and adipocytes (subcutaneous white adipocytes, SC-WAT; visceral white adipocytes, V-WAT; beige and brown adipocytes, BAT) and likely macrophages and neuroglia in AdBARKO female mice contribute (solid line arrows) and likely contribute (dashed line arrows) protection against most PCOS-like reproductive and metabolic traits induced by peripubertal onset of androgen action. AR, androgen receptor; DHT, dihydrotestosterone; PCOS, polycystic ovary syndrome.
When examining the role of AdBARKO in mediating DHT-induced PCOS-like reproductive traits, Walters et al (5) confirmed the brain (not pituitary or ovary) as the site for AR-mediated, PCOS-like phenotype. Moreover, they unexpectedly pinpointed AR action in adipocytes and/or macrophages as contributing to DHT-mediated disruption of ovarian folliculogenesis that NeuroARKO alone failed to prevent (4), potentially linking lipotoxicity with ovarian dysfunction, apart from AR-mediated disruption of hypothalamic-pituitary estradiol (E2) and progesterone-mediated negative feedback. The FABP4-Cre recombinant driver used to develop AdBARKO (5), however, may be more comprehensive in excising AR exon 3 in brain cells than CamKII-Cre in NeuroARKO (4).
Care was taken to describe strain-typical ovarian morphology, ovulatory estrus cycles, fertility, and fecundity in AdBARKO adult female mice and, without circulating ovarian and pituitary hormone values, one assumes strain-typical levels, as found in forebrain neuronal ARKO (NeuroARKO) females (4). In contrast to previous NeuroARKO results (4), when exposed to peripubertal DHT onset, AdBARKO females failed to exhibit any PCOS-like reproductive traits, perhaps adding adipocytes, as well as macrophages and neuroglia, to the central nervous system as yet additional sites of AR-induced PCOS-like phenotypic expression (Fig. 1).
If, however, Walters et al had utilized peripubertal-onset aromatizable testosterone (T), instead of nonaromatizable DHT, with AdBARKO female mice, as they recently reported for whole-body ARKO mice (6), they may have found PCOS-like ovarian disruption identical to that in DHT-exposed WT females above, implicating PCOS-like conversion beyond AR, likely involving multiple target organ aromatization, with PCOS-like traits mediated mostly through estrogen receptor alpha (ERα) (6). E2-mediated contributions to the PCOS-like phenotype exist in sheep and rodent models (1) and are readily understandable given (a) the central nervous system as the site of PCOS-like conversion in female mice (4) and (b) prior findings implicating rodent brain aromatization of T to E2 and subsequent inflammation-mediated hypothalamic E2 neuronal reprogramming mediated mostly by ERα (7). Such estrogenic contributions, however, lie downstream of AR action in nonhuman primates NHPs and humans since nonaromatizable DHT contributes to PCOS-like conversion and brain masculinization in NHP, and gene variant-mediated absence of AR gene expression in XY humans prevents brain masculinization, unlike in rodents (1, 7).
When examining the role of AdBARKO in mediating DHT-induced PCOS-like metabolic traits, Walters et al (5) added adipocytes, macrophages, and neuroglia as potential sites for AR-mediated, PCOS-like phenotypic expression beyond the central nervous system (4). The additional selective AdBARKO completely prevented DHT-mediated visceral adipocyte hypertrophy and hepatic steatosis that were only partially protected by NeuroARKO. Selective AdBARKO also prevented DHT-mediated increase in gene expression of a chemokine macrophage attractant, MCP1/CCL2, potentially diminishing localized sites of inflammation. Moreover, AdBARKO, as found with NeuroARKO (4), prevented both hyperlipidemia and increased total body as well as visceral white adipose tissue (WAT) depot weights (5). While not previously reported for NeuroARKO, AdBARKO also prevented an increase in subcutaneous WAT and brown adipose tissue (BAT) depot weights (5). Whether the site for developing PCOS-like traits resides in adipocytes of visceral versus subcutaneous WAT or BAT, or perhaps in conjunction with stromal-vascular cell interactions, will be determined by future selective ARKO mouse models. In women who are normal-weight with PCOS, however, adipose insulin resistance positively and negatively correlates with serum triglyceride levels and insulin sensitivity by intravenous glucose tolerance testing, respectively, and remains so, adjusting for hyperandrogenemia. This raises possible epigenetic, developmentally programmed contributions to PCOS-related metabolic dysfunction (8).
Discussion of PCOS-like metabolic traits is incomplete without considering glucoregulatory dysfunction. Perhaps surprisingly, both NeuroARKO (4) and AdBARKO (5) fail to replicate whole-body ARKO protection against peripubertal onset, DHT-induced hyperglycemia (4). It may require a hepatocyte-specific ARKO to prevent AR-mediated dysregulation of hepatic glycogen, as proposed by Walters et al (4) and further suggested in another DHT-treated mouse model showing reduced basal hepatic glucose uptake and lower glucose uptake in WAT compared with control mice (9).
In sum, the multiple AR-mediated, PCOS-like conversion sites identified by Walters et al (4, 5), which may include downstream ERα action (6), agreed with accumulating evidence for a combined genetic, epigenetic, and developmental etiopathogenesis of PCOS in women, as well as of PCOS-like traits in rodent, sheep, and NHP models (1). With the etiopathogenesis of PCOS likely occurring at a prepubertal developmental stage, thus preceding the peripubertal time used by Walters et al 4-6), systematic developmental molecular analysis of discrete reproductive and metabolic dysfunctions that resemble PCOS traits offer new directions toward improved understanding of PCOS, leading to its prevention and cure.
Abbreviations
- AR
androgen receptor
- BAT
brown adipose tissue
- CamKII
calcium/calmodulin-dependent protein kinase II
- DHT
dihydrotestosterone
- E2
estradiol
- ERα
estrogen receptor alpha
- FABP4
fatty acid-binding protein 4
- KO
knockout
- Neuro
neuron-specific
- NHP
nonhuman primate
- PCOS
polycystic ovary syndrome
- T
testosterone
- WAT
white adipose tissue
Acknowledgments
Financial Support: D.H.A. and J.E.L.: NIH DK121559, HD102172, OD011106; D.A.D.: NIH HD071836.
Additional Information
Disclosure Summary: The authors have nothing to disclose.
Data Availability: Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
References
