Twenty age-matched (approximately 7 months old) chinchillabreed female rabbits (Oryctolagus cuniculus) were housed in individual stainless-steel cages and kept at 20˚C±2˚C under artificial lighting conditions (16 light and 8 hours dark, lights on at 06:00 hours). Rabbits are reflex ovulators that under these housing conditions are considered to be in an early proestrous phase [21]. They were provided daily with pellet food (Conejina, Purina, México) and had continuous access to water. Rabbits were allocated into 2 groups: a control group composed of nulliparous rabbits (N; n=10) and an experimental group of primiparous rabbits (P3; n=10). Rabbits in the P3 group were mated at 6 months of age. The rabbits in both groups were euthanized with an overdose of sodium pentobarbital (100 mg/kg of weight, intraperitoneally) when they reached the established age (N) or at 3 days postpartum (P3). The pups in the latter group were decapitated at 1-day postpartum. The Ethics Committee of the Universidad Autónoma de Tlaxcala approved all experimental procedures described below.

The vaginal tract of 20 rabbits (10 nulliparas and 10 primiparas) was excised as described elsewhere [15,16]. For 8 rabbits (4 nulliparas and 4 multiparas), the vagina was cut lengthwise in half to separate out the dorsolateral region in which the paravaginal ganglia are embedded [15]. Considering that the number of ganglion neurons is similar on the left and right sides [15], the latter group of neurons was stored at −80˚C to perform Western blot assays. Otherwise, the vaginas of 6 nulliparous and 6 primiparous rabbits were harvested, and the pelvic vagina was excised, washed with saline, immersed in Bouin-Duboscq fixative for 18 hours, and embedded in Paraplast X-tra (Sigma-Aldrich, St. Louis, MO, USA).

Paraplast-embedded samples from the pelvic vagina were cut on a microtome (Leica, Wetzlar, Germany) to obtain 7-μm-thick transverse sections. For each animal, the slides were separated in 4 series, and 1 of them was stained with Masson trichrome, covered with mounting medium and a coverslip, and observed under light microscopy using an Axio Imager A1 microscope (Carl Zeiss AG, Oberkochen, Germany). Images were acquired with a digital camera (ProgRes CT5, Jenoptik AG, Jena, Germany) with a resolution of 5.1 megapixels.

Total protein extracts were prepared from frozen vaginal tissue as described elsewhere [11]. Equal amounts of protein (100 μg) were denatured in Laemmli sample buffer and resolved through sodium dodecyl sulfate-polyacrylamide (15% for GDNF; 10% for GFRα-1) gels and electro-blotted to nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA, USA). Blots were stained with Indian ink (Pelikan Holding AG, Schindellegi, Switzerland) diluted in phosphate-buffered saline (PBS) to confirm that the protein content was equal in all lines [23]. The membranes were soaked in PBS and incubated in 5% dried skim milk diluted in PBS containing 0.2% tween-20 (PBST) for 60 minutes to block nonspecific protein binding sites. Membranes were incubated overnight at 4˚C with mouse monoclonal anti-GDNF (sc-13147, Santa Cruz Biotechnology; 1:100) or mouse monoclonal anti-GFRα-1 (sc-271546, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA; 1:100), followed by secondary antibodies (1:1,000, goat anti-mouse IgG-biotinylated, sc-2039, Santa Cruz Biotechnology Inc.). The sequence of epitopes used to raise anti-GDNF (human; NCBI accession version CAG46721.1) and anti-GFRα-1 (rat; NCBI accession NP_037091.1) share a homology of 100% (NCBI accession XP_008260349.1) and 96.8% (NCBI accession XP_002718804.1) with rabbit proteins. Immunoreactive polypeptides were detected using a chemiluminescence kit (SuperSignal West Pico Chemiluminescent Substrate Kits, Thermo Fisher Scientific, Waltham, MA, USA) and exposed to Kodak X-OMAT film. The expression of GDNF and GFRα-1 was measured by densitometry and normalized against the signal obtained from the India inkstained lane, which was used as a loaded control [24]. The ImageJ software (National Institutes of Health, Bethesda, MD, USA) was used for the densitometric analysis. The relative expression levels of GDNF and GFRα-1 were calculated by dividing the density of each band by the density of the matching India ink-stained lanes. Data were expressed as the percentage (%) of change.

The location of GDNF and GFRα-1 expression was assessed using immunohistochemistry, as reported elsewhere [11]. Briefly, homologous slides were incubated in a humidified chamber for 72 hours at 4˚C with anti-GDNF (1:100) or anti-GFRα-1 (1:100) antibodies, and afterwards incubated with a biotinylated secondary antibody (sc-2039, Santa Cruz Biotechnology Inc.; 1:250) for 2 hours at room temperature. Immunostaining was performed following the Vectastain ABC kit directions (Vector Laboratories Inc., Burlingame, CA, USA). Some sections were incubated with the secondary antibody alone and were considered as negative controls. Immunostained sections were counterstained with Mayer hematoxylin. The proportion of rabbits showing positive results for each immunostaining assay is indicated inside parentheses. Both GDNF and GFRα-1 immunostaining in the vaginal layers were categorized in accordance to the intensity of each marker as negative (-), weak (+), moderate (++), or strong (+++) by 2 independent observers (VGV and LGHA). These categories were then assigned numerical values instead of signs to perform a semiquantitative analysis [25]. Sections of the cranial portion of the pelvic wall of the vagina for nulliparas and multiparas were simultaneously processed to obtain a qualitative estimate of GDNF and GFRα-1 immunoreactivity based on the intensity of each marker in neurons and SGCs, as reported elsewhere [12]. To obtain representative images of vaginal GDNF and GFRα-1 immunostaining, sections were observed under light microscopy using a Ni-NU microscope (Nikon, Tokyo, Japan) coupled to a digital camera with a resolution of 16.25 megapixels (DS-Ri2, Nikon). Selected photographs were postprocessed using the automated color normalization function of the program GIMP (v. 2.8.6. for MacOS).

The data for morphometric variables are presented as means±standard error of the mean. Significant differences (P≤0.05) were identified using the unpaired 2-tailed Student t-test. Data from western blot assays and immunohistochemistry for GDNF and GFRα1 are presented as medians±maximum and minimum values. Significant differences (P≤0.05) were identified using the Wilcoxon matched-pairs signed-rank test. Statistical analyses were done with the program Prism 4c for PCs (GraphPad Software Inc., La Jolla, CA, USA).

The morphology of the paravaginal ganglia was inspected in Masson trichrome sections (Fig. 1A, B). The area covered by the ganglia in primiparous rabbits and nulliparous rabbits was similar (Fig. 1C). The same was true for the average number of ganglionic neurons (Fig. 1D) and the cross-sectional area (CSA) of the neuronal somata (Fig. 1E). When the CSA was separated into 3 bins, a significant reduction in the percentage of neurons with larger (650–1,400 μm²) soma CSA was measured in primiparous rabbits (Fig. 1F). The percentages of small (50–350 μm²) and medium (351–650 μm²) neuronal soma CSA were similar between nulliparas and primiparas. The average number of SGCs per neuron soma was also similar for nulliparous and primiparous rabbits (Fig. 1E). The same was true for the percentage of paravaginal neurons associated with SGCs (Fig. 1F).

Western blot assays showed the presence of a single GDNF immunoreactive band with an approximate molecular weight of 17 kDa in nulliparas and primiparas (Fig. 2A). The GDNF expression measured in the pelvic vagina in the N and P3 groups was highly similar (Fig. 2B).

In the vaginal sections of nulliparous rabbits, GDNF immunostaining was observed in the epithelium (4 of 6; Fig. 2C), submucosa (2 of 6), smooth muscle cells of the muscle layer (4 of 6; Fig. 2F); and in the skeletal myofibers of the external layer (5 of 6; Fig. 2I). A highly similar pattern of GDNF staining was seen in the epithelium (5 of 6; Fig. 2D), submucosa (1 of 6; Fig. 2D), smooth muscle (5 of 6; Fig. 2G), and skeletal muscle (5 of 6; Fig. 2J) of primiparous rabbits. Such staining was highly specific in comparison to the negative control results (Fig. 2E, H, K). In a semiquantitative analysis of the intensity of GDNF immunostaining, there were no significant differences between the nulliparous and primiparous rabbits for any of the inspected vaginal regions (Fig. 2L–O).

Two major GFRα-1 immunoreactive bands were detected between 53 and 79 kDa in the total protein extracts from the pelvic vagina of nulliparas and primiparas (Fig. 3A). Both bands were considered to determine the relative expression of GFRα-1. The relative GFRα-1 expression was similar between the N and P3 groups (Fig. 3B).

GFRα-1 immunostaining in the nulliparous rabbits (Fig. 3) was seen in the epithelium (5 of 6; Fig. 3C), submucosa (2 of 6), smooth muscle of the muscle layer (4 of 6; Fig. 3E), and skeletal myofibers of the external layer (6 of 6; Fig. 3G). The distribution was highly similar in the primiparous rabbits, with GFRα-1 immunostaining observed in the epithelium (6 of 6; Fig. 3D), submucosa (2 of 6; Fig. 3D), smooth muscle (5 of 6; Fig. 3F), and skeletal muscle (5 of 6; Fig. 3H). In a semiquantitative analysis, no significant differences between the nulliparous and primiparous rabbits were found (Fig. 3I–L).

Weak GDNF immunostaining was observed in the paravaginal ganglia, particularly in paravaginal neurons; notably, such immunostaining was seen as discrete clusters that were restricted to the cytoplasm near the nuclear perimeter (Fig. 4A). GDNF immunostaining was also observed in SGCs (Fig. 4A). Stronger GDNF immunostaining was noted in the cytoplasm of both the neurons and SGCs of primiparous rabbits (Fig. 4B).

GFRα-1 immunostaining was also noted in the cytoplasm of paravaginal neurons and SGCs (Fig. 4C). Such labeling was stronger in the neurons and SGCs of primiparous rabbits (Fig. 4D). No immunostaining was seen in sections incubated with the secondary antibody alone (Fig. 4E).