Excessive ovarian production of nerve growth factor facilitates development of cystic ovarian morphology in mice and is a feature of polycystic ovarian syndrome in humans

Abstract

Although ovarian nerve growth factor (NGF) facilitates follicular development and ovulation, an excess of the neurotrophin in the rodent ovary reduces ovulatory capacity and causes development of precystic follicles. Here we show that ovarian NGF production is enhanced in patients with polycystic ovarian syndrome (PCOS) and that transgenically driven overproduction of NGF targeted to the ovary results in cystic morphology, when accompanied by elevated LH levels. NGF levels are increased in the follicular fluid from PCOS ovaries and in the culture medium of granulosa cells from PCOS patients, as compared with non-PCOS patients. Ovaries from transgenic mice carrying the NGF gene targeted to thecal-interstitial cells by the 17alpha-hydroxylase gene promoter produce more NGF than wild-type (WT) ovaries and are hyperinnervated by sympathetic nerves. Antral follicle growth is arrested resulting in accumulation of intermediate size follicles, many of which are apoptotic. Peripubertal transgenic mice respond to a gonadotropin challenge with a greater increase in plasma 17-hydroxyprogesterone, estradiol, and testosterone levels than WT controls. Transgenic mice also exhibit a reduced ovulatory response, delayed puberty, and reduced fertility, as assessed by a prolonged interval between litters, and a reduced number of pups per litter. Sustained, but mild, elevation of plasma LH levels results in a heightened incidence of ovarian follicular cysts in transgenic mice as compared with WT controls. These results suggest that overproduction of ovarian NGF is a component of polycystic ovarian morphology in both humans and rodents and that a persistent elevation in plasma LH levels is required for the morphological abnormalities to appear.

Figure 1

NGF content is increased in ovarian follicular fluid from PCOS patients, and granulosa cells from PCOS ovaries produce more NGF than granulosa cells from non-PCOS ovaries A, NGF levels are greater in the ovarian follicular fluid of PCOS patients than the fluid of non-PCOS subjects. *, P < 0.05 (Mann-Whitney rank sum test) vs. non-PCOS group. B, Granulosa cells from PCOS ovaries produce more NGF than granulosa cells from non-PCOS ovaries. The cells, collected during follicle aspiration, were placed in culture for 3–4 d in serum-containing medium. At this time, the medium was replaced with serum-free medium and the cells were incubated for an additional 24 h period before NGF measurement. NGF values are expressed as picograms per arbitrary unit (AU) of ATP (an indirect assessment of viability). *, P < 0.05 (t test) vs. normal group. Numbers on top of bars are numbers of subjects per group.

Figure 2

Transgenic mice overexpressing NGF in the ovary exhibit delayed puberty and reduced fertility; estrous cyclicity is disrupted and the ovulatory response to gonadotropins is reduced in 17NF mice. A, Vaginal opening (mean ± sem shown on graph) was delayed in mice with enhanced ovarian NGF expression. ***, P < 0.001 (t test) vs. WT mice. B, After vaginal opening, daily vaginal lavages were used to determine the age at first estrus, which was also delayed in transgenic mice. ***, P < 0.001 (Mann-Whitney rank sum test) vs. WT mice. The discrepancy in the number of WT animals in A and B was because some the animals were mistakenly identified as being in estrus and thus were not included in B. C, Transgenic and WT (B6D2) females were set up in breeding cages with WT (B6D2) males at 30 d of age. The interval from cage setup to first litter was delayed (***, P < 0.001; t test) as was the intervals between subsequent litters. **, P < 0.01 (t test) vs. WT. D, The number of litters born per female during 3-month intervals in transgenic mice was reduced compared with WT mice (*, P < 0.05; one-way repeated measures ANOVA) (inset, summary of data for 12 months; ***, P < 0.001; t test). E, The number of pups/litter per female was reduced in transgenic mice compared with WT animals. *, P < 0.05 (Friedman repeated measures ANOVA). F, The total number of pups born per female was reduced in the transgenic mice both quarterly (***, P < 0.001; one way repeated measures ANOVA) and for the entire year (inset; ***, P < 0.001; t test). G, Young adult (60 d of age) 17NF mice spend more days in estrus (**, P < 0.01; t test) and fewer days in diestrus. *, P < 0.05 (Mann-Whitney rank sum test) vs. age-matched WT mice. H, The ovulatory response of 17NF mice to a priming PMSG injection followed by a low dose of hCG (1 IU) was lower than that of WT animals. *, P < 0.05 (Mann-Whitney rank sum test). This reduced response was not apparent at higher doses of hCG (2.5 and 5 IU hCG). Numbers on top of bars are number of subjects per group.

Figure 3

The ovaries of 17NF mice contain more antral follicles of an intermediate size than the ovaries of WT mice. A, Microphotograph of a WT ovary. B, Microphotograph of a 17NF ovary. The ovaries were collected from 32-d-old mice, fixed in Kahle’s solution, embedded in paraffin, sectioned at 6 μm, stained with Weigert’s iron hematoxylin, and counterstained with picric acid-methyl blue. Bars, 300 μm. C, The total number of follicles per ovary was greater in 17NF mice than WT controls. *, P < 0.05 (t test). D, The number of antral atretic follicles but not that of healthy follicles was also increased in 17NF mice. *, P < 0.05 (ANOVA, Student-Newman-Keuls multiple comparison of means). E, The ovaries of 17NF mice tended to have more healthy follicles measuring 101–200 μm in diameter than WT ovaries. F, The transgenic ovaries had significantly more atretic follicles in the range of 101–200 and 201–300 μm than WT controls. *, P < 0.05 (ANOVA, Student-Newman-Keuls multiple comparison of means).

Figure 4

The ovaries of 17NF transgenic mice exhibit more apoptotic follicles than WT controls, as assessed by TUNEL assay. A, WT ovary. B, 17NF ovary. Bars, 300 μm. C, Example of follicle from a WT ovary showing some TUNEL-positive granulosa cells. D, Typical atretic follicle from a 17NF ovary exhibiting numerous TUNEL-positive granulosa cells. E, Another example of granulosa cell apoptosis in a 17NF ovary. Bars, 100 μm. F, Ovaries from 17NF mice have a similar number of healthy antral follicles but more antral apoptotic follicles than WT ovaries. **, P < 0.01 (t test).

Figure 5

Basal serum levels of sex steroids and gonadotropins are similar in juvenile (J; 28–30 d old) 17NF and WT mice but differ substantially after PMSG stimulation. A, Basal serum P4 levels are modestly increased in 17NF mice (*, P < 0.05; t test), but the P4 response to PMSG is decreased (**, P < 0.01; t test) in these animals in comparison with WT controls. B, The 17-OHP4 response to PMSG is greater in 17NF mice than in WT controls. *, P < 0.05 (t test). C, Serum Δ4 levels are similar in both groups stimulated with PMSG, but the T4 response is increased in 17NF mice. ***, P < 0.001 (Mann-Whitney rank sum test). D, The E2 response to PMSG is also greater in 17NF mice than WT controls. ***, P < 0.001 (Mann-Whitney rank sum test). E, Serum LH levels are similar in the two groups. F, FSH levels are also similar. Numbers on top of bars are number of subjects per group.

Figure 6

The ovaries from 17NF transgenic mice respond more pronouncedly than WT ovaries to a persistent, but mild elevation, in circulating levels of LH-like activity. A sustained elevation in circulating LH-like activity was achieved by treating juvenile (24–27 d old) 17NF and WT mice with a low dose of hCG (50 mIU/h) delivered for 7 d via osmotic minipumps. To simulate a peripubertal condition preceding the exposure to hCG, all animals received a single dose of PMSG (5 IU/mouse) before initiating the hCG treatment. A–C, Total number of antral follicles (examples denoted by arrows in A and B) is increased in 17NF ovaries. ***, P < 0.001 (t test). D–F, The number of precystic type III follicles (examples denoted by arrows in D and E) is also increased. **, P < 0.01 (t test). G–I, Likewise, ovaries from 17NF mice have more follicular cysts (arrows in G and H) than WT controls. *, P < 0.05 (t test). Bars in microphotographs, 250 μm. Numbers on top of bars are number of subjects per group.