Effects of nerve growth factor (NGF) on blood vessels area and expression of the angiogenic factors VEGF and TGFbeta1 in the rat ovary

Ovaries from 2 day old rats were dissected under aseptic conditions, placed on sterile lens paper, and cultured on plastic supports at the interface of air/culture medium [28], under an atmosphere of 5% CO2. For histochemical studies, one ovary per well was utilized, and for mRNA studies, three glands per well were cultured. Considering this, for an n = 5, 20 pups were utilized for each incubation point. Ovaries were cultured in 24-well plates (Becton Dickinson, USA); each well contained 1 ml DMEM: F-12 (50% vol/vol, Sigma Chemicals, St Louis, MO, USA) medium supplemented with sodium bicarbonate (600 mg/l, Sigma Chemicals, St Louis, MO, USA), penicillin (50 mg/l, Laboratorio Chile, Santiago, Chile), gentamycin (80 mg/l, Andrómaco, Santiago, Chile), streptomycin (50 mg/l, Laboratorio Chile, Santiago, Chile) and ketoconazol (5 mg/l, Laboratorio Chile, Santiago, Chile). The ovaries were stimulated with NGF 100 ng/ml (Sigma Chemicals, St Louis, MO, USA). The optimal dose of NGF was established by previous studies from our group [17]. The contralateral ovaries cultured with medium alone were used as controls. The times of culture were 2, 4, 8 and 24 h. In the case of mRNA studies, the ovaries were collected and stored at -80°C until RNA extraction. The ovaries collected for immunohistochemistry were fixed in Bouin's fixative.

We used 16 prepuberal rats of 23–24 day old (body weight 60 g aprox.). They were anaesthetized with isoflurane 1% v/v, 2.5 l/min (Baxter Healthcare Co, Guayama, Puerto Rico). The dissection of the superior ovarian nerve (SON) was performed in both ovaries under aseptic conditions through a single dorsal midline incision. This procedure has shown to induce an increase of NGF within the gland [29]. Because denervation induces hypertrophy of the contralateral ovary [30], controls consisted in different animals submitted to a simulated operation. The ovaries were collected three days later (from a total of 6 rats; 3 controls and 3 denervated) or six days later (from a total of 10 rats; 5 controls and 5 denervated), and stored at -80°C for mRNA studies, or fixed in Bouin's fixative for histochemical procedures. The time for ovary harvesting after denervation was established by previous studies from our group [29].

Immunohistochemical detection of VEGF and TGFβ1 in denervated ovaries and cultured tissues was performed as follows: the ovaries were fixed by immersion in Bouin's fixative, embedded in paraffin, serially sectioned at 4 μm and processed for immunohistochemistry using polyclonal antibodies from Santa Cruz Biotechnology (anti-VEGF [C-1], sc-7269 and anti-TGFβ1 [V] sc-146, both in a 1:50 dilution). Tissue sections were incubated overnight at 4°C with the antibody and the immunoreaction was developed the next day using the NBT-BCIP procedure (Sigma Chemicals, St Louis, MO, USA) for VEGF in neonatal rat ovaries, or the DAB procedure (LabVision Co, Freemont CA, USA) for the rest of the detections. Controls consistent of adjacent sections incubated without the primary antibody, or sections incubated with the antibody preabsorbed with the corresponding ligand showed no staining, proving the specificity of the immunoreaction. Whenever a well defined mark was obtained, cell counting was performed by three independent observers and the data were expressed as H-Score. The H-Score is the sum of the proportion of cells showing different degrees of reactivity, and was calculated as follows: 0 times the % of negative cells + 1 time the % of weakly positive cells + 2 times the % of moderately positive cells + 3 times the % of strongly positive cells. The H-Score ranges from 0–300, being the maximum score representative of a 100% of cells with strong positivity [31]. In the case of diffuse mark, staining intensity was evaluated with an automated digitizing system (UN-SCAN-IT gel 4.1, Silk Scientific Co, USA) and the data were expressed as number of pixels.

Immunohistochemical detection of endothelial cell marker PECAM-1 in denervated ovaries was performed using an antibody from Santa Cruz Biotechnology (anti-PECAM-1 [M-20] in a 1:100 dilution). Tissue sections were incubated with the antibody overnight at 4°C and the immunoreaction was developed the next day using the DAB procedure (LabVision Co, Freemont CA, USA). Controls consisted of adjacent sections incubated without the primary antibody. The area of positive staining was evaluated with an automated system (Image J 1.36b, NIH, USA) and the data were expressed as % area of the ovary section.

Total RNA was extracted from rat ovaries using guanidine isothiocyanate and phenol-chloroform extraction (Trizol Reagent, Invitrogen, Foster City CA, USA) following the manufacturer's protocol. Concentration and purity of RNA were measured using a spectrophotometer at 260 and 280 nm. First strand cDNA was synthesized in a 20 ml reaction mixture using 1 μg of total RNA. The reaction tubes contained 0.5 μl random hexamers (500 ng/μl, Invitrogen, Foster City CA, USA), 1 μl dNTPs (10 mM, Invitrogen, Foster City CA, USA), 4 μl 5× reaction buffer (250 mM Tris-HCl pH 8.3, 375 mM KCl, 15 mM MgCl2, Invitrogen, Foster City CA, USA), 2 μl DTT (0.1 M, Invitrogen, Foster City CA, USA), 1 μl ribonuclease inhibitor (10 U/μl, Invitrogen, Foster City CA, USA) and 1 μl M-MLV reverse transcriptase (200 U/μl, Invitrogen, Foster City CA, USA). Reactions were incubated at 37°C for 60 minutes and inactivated by freezing.

The specific primer sequences for the examined genes and the predicted PCR product sizes are shown in Table 1. The cDNA was amplified in a 25 μl reaction mixture using 1 μl of single-strand cDNA. PCR conditions were as follows: 2.5 μl 10 × reaction buffer (200 mM Tris-HCl pH 8.4, 500 mM KCl, Biotools, Madrid, Spain), 1 μl MgCl2 (50 mM), 0.5 μl dNTPs (10 mM each), 1.25 μl each primer (10 μM, Polyscience, Santiago, Chile), 0.15 μl DNA polymerase (5 U/μl, Biotools, Madrid, Spain). Reaction mixtures were incubated in a thermal cycler (Eppendorf, Foster City CA, USA) for 2 minutes at 94°C. Then 23, 27, or 28 cycles were performed for cyclophilin, VEGF and TGFβ1 respectively including: denaturation at 94°C for 45 seconds, annealing at 62°C for 1 minute and extension at 72°C for 1 minute. The optimal number of cycles was experimentally determined by analyzing the exponential phase of amplification reaction (data not shown). The PCR products were separated on Tris-borate-EDTA 2% agarose gels containing 200 ng/ml ethidium bromide (Invitrogen, Foster City CA, USA). DNA size markers were run in parallel to validate the predicted sizes of the amplified bands (100-bp DNA ladder, Biotools, Madrid, Spain). The gels were visualized under UV light, photographed using a capturing program (Ultra Violet Products Doc-It), and analyzed with an automated digitizing system (UN-SCAN-IT gel 4.1, Silk Scientific Corporation, Orem UT, USA). Differences between the experimental versus control conditions, were obtained by the ratio of the intensity between specific gene and constitutive gene of each sample, determined by densitometry.

The differences in mRNA levels, immunoreactivity and area of positive staining were analyzed using Mann Whitney test.

The photograph of 2% agarose gel electrophoresis of representative PCR products for VEGF120, VEGF164, TGFβ1 and Cyclophilin (as constitutive gene) are shown in figure 1A-C. The exponential phase of amplification for each target sequence was determined by performing PCR under different number of cycles (ranging from 20 to 35 cycles). The PCR products were separated in agarose gels, stained with ethidium bromide, and the intensity of the bands was analyzed with an automated digitalizing program (UN-SCAN-IT gel 4.1, Silk Scientific Corporation, Orem UT, USA). Results were plotted to visualize the exponential region of the curve, and thus the optimal number of cycles was selected for each target sequence (not shown). A dose/response curve was performed for each target sequence by adding increasing amounts of cDNA to the respective PCR mix. The response showed to be linear under working conditions (see figure 1D-F).

Neonatal rat ovaries were cultured with NGF (100 ng/ml) for 2, 4, 8 and 24 h. The mRNA levels for VEGF 120 and VEGF 164 – the most relevant VEGF isoforms in ovarian physiology – exhibited a significant increase in NGF-stimulated tissues compared to controls, in the two-hour culture, as shown in figure 2. No change on TGFβ1 mRNA content was found at any assayed stimulation time.

Regarding localization of VEGF protein in neonatal rat ovaries, although two controls for specificity were used, immunoreactivity was present more specifically in granulosa cell layer of primary follicles (figures 3A and 3B) and also in other types of ovarian cells (figures 3C and 3D). Densitometric analysis of the immune signal revealed that after 24 hour incubation, NGF induced an increase in VEGF staining in neonatal rat ovaries, which was significant both for the overall staining of the ovarian tissue and for staining of primary follicles (figure 3F). No significant differences in VEGF staining were found in shorter incubations (2, 4 or 8 h). As to TGFβ1, a specific mark was detected in oocytes (figure 4A and 4B). The same as for mRNA, H-Score of TGFβ1 immunoreactivity was not modified by NGF (figure 4D).

It was important to determine whether this neurotrophin would increase VEGF in an in vivo condition previously informed to provoke an increase in NGF in the rat ovary. We performed a superior ovarian nerve (SON) sectioning in prepubertal rats and confirmed that NGF protein levels, evaluated by densitometry of immunohistochemistry-processed tissues, were increased in theca cell layers of ovaries obtained from denervated animals when compared to control ones (figure 5). Nevertheless, after 6 days of denervation, no change was found in mRNA level of VEGF and TGFβ1 (figure 6). Although there were no differences in mRNA content, H-Score evaluation indicated that the immunoreactivity for VEGF was significantly increased in the inner granulosa cell layer of denervated ovaries as compared with control animals (Figure 7A,B and 7D). TGFβ1 immunoreactivity did not change after denervation (results not shown). To evaluate if there was an earlier change in VEGF and TGFβ1 mRNA levels or in TGFβ1 protein, the same studies were performed on day 3 of SON sectioning. We found no significant difference for any of the markers studied (results not shown).

We wanted to evaluate if denervated rats, which have higher ovarian NGF levels, present an increase in ovarian blood vessels. Immunohistochemical studies revealed that the amount of ovarian small vessels and capillaries, evaluated as % area of PECAM-1 positive staining was increased in rats denervated for 6 days, when compared to control animals (figure 8A-D and 8G). A similar result was found in denervated animals as early as three days after SON sectioning (figure 8F).