Sixty female Sprague-Dawley rats (180–200 g), with regular 4-day estrus cycles were purchased from Medical Experimental Animals Center of Fudan University (Shanghai, China). The animals were housed under laminar flow in an isolated room with controlled temperature and at a 12 /12 (light /dark) schedule. Thirty-six of them underwent ovariectomy with ether anaesthesia, which were then divided randomly into three groups: ovariectomized (OVX), ovariectomized with EA in specific points (OVX+EA) and ovariectomized with EA in non-specific points (OVX+C). The rest were treated as controls, which were divided into two groups: intact (INT) and intact with EA in specific points (INT+EA). Four weeks later, OVX+EA (n = 12), OVX+C (n = 12) and INT+EA (n = 12) received EA treatment. Thirty minutes before the EA treatment, all the animals were bound as comfortably as possible, and during the EA procedure, the rats were conscious without anaesthesia. Electrical stimulation was administered via three stainless steel needles of 0.3 mm diameter inserted 5 mm in four acupoints in the belly: "Guanyuan" acupoint (RN4), in the midline of abdomen (15 mm bellow the umbilicus); "Zhongji" acupoint (RN3), 5 mm bellow "Guanyuan" acupoints (one needle was flatly punctured in the RN4 which penetrated into the RN3); bilateral "Zigongxue" acupoints (EXTRA22), 7.5 mm lateral to the "Zhongji" acupoint, and one needle inserted 3 mm at one acupoint in the hind leg, "Sanyinjiao" (SP6), near ankle joint (at the concentration of the superior border of the medial melleolus, between the posterior border of the tibia and anterior border of the Achilles tendon). These acupoints are widely applied in Oriental medicine for the treatment of gynaecological disease in women 8 and for the acupuncture mechanism research in the ovariectomized rats 4579. The control acupoints were "Waiguan" (SJ5), between the radius and ulna (5 mm above the dorsal transverse crease of the wrist) and "Huatuojiaji" (EXTRA15), in the back (5 mm lateral to the lower border of each spinous process from the first thoracic vertebra to the fifth lumbar vertebra (Fig. 1). The stimulation was generated by an EA apparatus (Model G6805-2, SMIF, Shanghai, China) and lasted for 30 min (8:00–10:00 AM), Q.D, for 3 days altogether. The stimulation parameters were 2 mA of density and a low-burst frequency of 3 Hz. Individual pulses within the burst frequency were square wave pulses with alternating polarities and pulse duration of 0.2 ms, 80 pulse per second. The intensity was adjusted to produce a slight twitch of the limbs. All experimental procedures involving the use of animals were conducted in accordance with NIH Guidelines and were reviewed and approved by the Animal Use and Care Committee for the Fudan University.

At the time of sacrifice (6 hours after the last EA), the vaginal cytology of each OVX+EA rat was first examined. The animals whose epithelial cells reappeared were adopted for the following experiments, which could be a validity indicator of the EA. The tissues of the OVX, OVX+EA and INT+EA rats were collected respectively, and those of the INT animals, during the period of proestrus. All the operations were carried out at 4°C. The rats were sacrificed by decapitation; the subcutaneous abdominal (SA) adipose and liver tissues were excised; and then snap-frozen in liquid nitrogen, stored at -80°C. The preparation of the microsomal pellet was accordance with the report by Hiroshi 10. Total tissue RNA was extracted using 'TRIzol Regent' (Biobasic Inc, Canada), and the purity and integrity of the RNA were checked spectroscopically and by gel electrophoresis before analytical procedures.

A modified version of the radiometric assay of Charles and William 11 was applied to determine the aromatase activity. The microsomal pellets were sonicated in 1:30 (w/v) phosphate buffers (10 mM KPO4, 100 mM KCl and 1 mM EDTA, pH 7.4). The incubation mixture, made of NADPH-generating system and 0.13 μM (1β-³H) androstenedione (Amersham Pharmacia Biotech, UK, 24 Ci/mmol) in 100 μl phosphate buffer (10 mM KPO4, 100 mM KCl, 10 mM dithiothreitol, 1 mM EDTA and 2.5 mM glucose-6-phosphate dehydrogenase/ml, pH 7.4), was pre-incubated at 37°C for 30 min to generate sufficient NADPH for aromatase reaction. Aliquots (100 μl) of microsome samples and ovarian microsome that were treated as positive controls, were then added for an incubation in a Dubnoff shaking water bath for 1 h at 37°C. The reaction was terminated by addition of 0.4 ml ice-cold 15% trichloroacetic acid containing 40 mg charcoal/ml, and then the labelled aqueous phases were separated from the unmetabolized substrate, and from the steroid products formed during the incubation, by chloroform extraction and 5% charcoal/0.5% dextran precipitation. After a rapid passage through a cotton filter, the amount of the tritiated water formed during the aromatization process was measured in a liquid scintillation spectrometer. A calibration standard curve helped obtain quench corrected dpm of the isotope. The amount of protein in each tube was evaluated using the Bradford method (Bradford, 1976). To validate the ³H₂O assay, aliquots of microsomes were incubated with increasing amounts of (1β-³H) androstenedione (e.g. 10–200 nM), and the amount of ³H₂O generated was measured and corrected for counts in buffer blanks. The data were plotted as a saturation plot; the maximal velocity (Vmax) and Michaelis constant (Km), derived by non-linear regression analysis.

A 50 μg sample of the microsomal protein was loaded into each lane along with a prestained protein size marker (Bio-Rad Laboratories, Inc., Hercules, CA), electrophoresed on a 10% SDS-polyacrylamide gel at 18 V/cm, and electroblotted onto a polyvinylidene difluoride membrane (Micron Separations, Westboro, MA) using a wet electroblotter. After blocking in fat-free milk, incubation was conducted with the antiaromatase antibody (1:200; Boster Biological Technology LTD., China) and β-actin antibody (1:3000) at room temperature for 4 h in TBS-T solution (20 mM Tris, 137 mM NaCl, and 0.1% Tween-20, pH 7.6). After extensive washing, blots were incubated with AP-labelled goat antirabbit antiserum (Sino-American Biotechnology Co., China) for 60 min at room temperature and developed using NBT/BCIP detection system (Amersham Pharmacia Biotech). The intensities of the bands were evaluated using the Image Master Software (SYDR-1990, SYNGENE, U.S.A.), and values were normalized to β-actin immunoreactivity in each sample and expressed as percent of the control. Specificity of the aromatase immuno-staining was determined by preincubation of antiserum for 24 h at 4°C with varying concentrations of aromatase, with the primary antibody omitted to identify non-specific staining as well.

Total RNA (2 μg) was transcribed in reverse, in a final volume of 20 μl, using 200 IU M-MLV reverse transcriptase in the presence of 25 pmol downstream primer (Sangon Inc), 0.5 mM deoxy-NTP and 20 IU Rnasin (from Promega) for 60 min at 42°C before heat denatured for 5 min at 95°C. The cDNAs obtained were further amplified by PCR using 25 pmol of upstream primer (Sangon Inc) designed to amplify P450 arom highly conserved sequence (289 bp length) including helical and aromatic region, (upstream Aro-Ex8, 5'GCT TCT CAT CGC AGA GTA TCC GG-3' located in the exon 8; downstream Aro-Ex9, 5'-CAA GGG TAA ATT CAT TGG GCT TGG-3' located in the exon 9; amplified product 289 bp) 12. We first determined the linear range of amplification of cDNA using each of the primer sets, and then chose an appropriate amplification cycle within this range for each cDNA species. For P450 arom, we used 35 PCR amplification cycles, and 20 cycles for β-actin gene expression. P450 arom PCR was performed (1 min at 94°C, 1 min at 60°C, 2 min at 72°C) with Taq DNA polymerase (3 U per tube) and 2.2 mM magnesium chloride (from Promega) in a final volume of 50 μl. To check the presence of DNA contamination RT-PCR was performed on 2 μg of total RNA without M-MLV reverse transcriptase (negative control). An internal control (water instead of RNA) for each RT-PCR was performed to investigate RNA contamination of the mixture. For each sample 5 μl of the PCR amplification products were analysed on 2% agarose gels and stained with ethidium bromide. The intensities of the bands were evaluated using the Image Master Software (SYDR-1990, SYNGENE, U.S.A.). The RT-PCR products were extracted and purified from agarose gel by Golden Beads Gel Extraction kit (Sangon Inc., China) and sequenced using radioactive dideoxychain terminating method (Sangon Inc., China).

At the time of sacrifice, the blood samples (0.8 ml) of the OVX, OVX+C, OVX+EA and INT+EA rats were collected (6 hours after the last EA) from tail veins respectively, and those of the INT animals, during the period of proestrus. The plasma was separated by centrifugation and stored at -70#176;C until assayed. Concentration of blood hormones were determined by double-antibody RIA kits purchased from the National Atomic Energy Research Institute (Beijing, China.). The samples were assayed in duplicate, and all the subjects' samples were assayed together. The sensitivity of the kit was 0.8 pg/ml (testosterone); 1.4 pg/ml (estrogen) and 0.05 pg/ml (corticosterone), the intra- and interassay coefficients of variation, 3.7–8.0% and 4.74–7.7%.

All data are presented as means ± S.E.M. Statistical analysis was performed on raw data using one-way analysis of variance (ANOVA), with the significance concentrations of p < 0.05 and p < 0.01 in two-tailed testing chosen.

The epithelial cells were stained by haematoxylin-eosin (HE). The INT and INT+EA rats showed regular 4-day estrus cycle change, and the cyclic change disappeared in the OVX, OVX+EA and OVX+C rats. Few mature vaginal epithelia were observed in the smears of the OVX and OVX+C rats, and the percent of mature epithelia increased significantly in the OVX+EA rats (p < 0.05) (Table 1).

The blood E₂ concentration decreased significantly in the OVX (p < 0.01) compared with that in the INT and INT+EA groups and was higher in the OVX+EA (p < 0.05) than in the OVX. The blood testosterone concentration decreased significantly in the OVX (p < 0.01) compared with that in the INT and INT+EA, but there is no statistical disparity between the OVX and OVX+EA (Fig. 2). The corticosterone concentration in the OVX+EA increased significantly compared with that in the INT, INT+EA and OVX groups (p < 0.05) (Fig. 2). Moreover, there were no disparities between the OVX and OVX+C, the INT and the INT+EA as well.

The tissues were incubated with (1β-³H) androstenedione for various periods of time, and the production of ³H₂O (aromatase activity) was linear with time up to 1 h of incubation. The aromatase activity was also linear with increasing amounts of tissue up to 5 mg/incubation. The optimal pH for the aromatase was 7.4. The maximum rates were obtained with a substrate concentration of 0.2 μM, and the enzyme preparation exhibited an apparent Michaelis-Menten constant (Km) of 0.04 μM. Therefore, incubation conditions were optimised that tissue homogenates ~3 mg tissue/ incubation; and incubation for 1 h in 0.2 ml potassium phosphate buffer containing 0.13 μM substrate (pH 7.4). The aromatase activity of the SA adipose and liver tissues presented no changes among the INT, INT+EA and OVX, and the activity of the SA adipose and liver tissue in the OVX+EA increased significantly (p < 0.01) compared with that in the INT, INT+EA and OVX (Fig. 3).

Densitometric analysis of the protein concentration using aromatase/β-actin expressed as the mean with SEM. The ratio of liver aromatase to β-actin in the OVX decreased significantly compared with that in the INT (p < 0.05). And the ratio of the SA adipose produced a higher level in the OVX than in the INT (p < 0.05). The ratio of liver tissues in the OVX+EA increased significantly compared with that in the OVX (p < 0.05). And the ratio of SA adipose tissues in the OVX+EA was as much as that in the OVX, but was higher than in the INT (p < 0.05). No disparity was detected between the INT and INT+EA (Fig. 4). No immunoreactive bands detected in the samples when using antiserum after preabsorption with excessive antigens and omission of the primary antibody.

Comparison of the amplified PCR fragment with rat ovary aromatase sequence revealed 100% homology (data not shown). Densitometric analysis of the mRNA concentration using aromatase/β-actin expressed as the mean with SEM. The ratio of liver aromatase to β-actin in the OVX decreased significantly compared with that in the INT (p < 0.05), and the ratio of liver tissue in the OVX+EA was higher than in the OVX (p < 0.05). No changes occurred on the ratios of SA adipose tissue between the OVX and INT, but a significant increase of the ratio in the OVX+EA (p < 0.05) compared with that in the INT and OVX was detected. However, no disparity was found between the INT and INT+EA (Fig. 5).