MDV3100 (enzalutamide) was purchased from Selleck Chemicals and dihydrotestosterone (DHT) from Sigma-Aldrich. X-treme GENE HP DNA transfection reagent was obtained from Roche and the simpleChIP^® enzymatic chromatin IP kit (Magnetic Beads) from Cell Signaling Technology, Inc. All the reagents were stored and used according to the protocol provided by the manufacturer. The primary antibodies for neurofibromin (ab17963) and VEGFA (ab46154) were purchased from Abcam. The primary antibody for AR (#5153) was purchased from Cell Signaling Technology, Inc.

Murine SW10 Schwann cells and human skin fibroblasts were propagated in DMEM/F12 with 10% fetal bovine serum (FBS). The cells were maintained in a humid atmosphere with 5% CO₂ at 37°C. Recombinant replication-defective lentiviruses harboring Nf1-specific short hairpin RNA (shRNA) or control shRNA were employed to transfect the cells, which were then named shNf1 or shNC cells, respectively. The efficiency of shRNA transfection was detected by western blot analysis. The siRNAs targeting AR were purchased from RiboBio. The cells were infected by lentiviruses at a confluence of 70-80% or transfected with siRNA using X-treme GENE siRNA Transfection Reagent (Roche) at a confluence of 30-50%.

Paraffin-embedded tissues from 29 patients with NF1 were collected from the first Affiliated Hospital of Xi'an Jiaotong University and used for the immunohistochemical (IHC) analysis of AR, CD31 and VEGFA expression. The tissue procurement protocol was approved by the Institutional Review Board and informed consent was provided from each patient. The sections with a thickness of 5 µm were derived from all the paraffin-embedded tissues and were accordingly prepared. The staining for AR (1:200), VEGFA (1:200) and CD31 (1:150) was performed using the DAKO Autostainer Plus system. One pathologist analyzed the sections under a high-power field (x400 magnification) in a double-blind protocol setup. Microvessels were defined as CD31⁺ endothelial cells or a cell cluster detached from any microvessel structures. The average number of microvessels from 10 random fields was defined as the microvessel density (MVD). The AR staining score was calculated by both intensity and percentage. The tendency score was estimated as follows: 0, 1, 2 and 3 indicated no staining, weak positive staining, moderate positive staining and strong positive staining, respectively. The percentage score was estimated as follows: 0, 0%; 1, ≤25%; 2, 25-50%; 3, 50-75%; and 4, ≥75%. The total score was calculated by the multiplication of the intensity score with the percentage score. The clinicopathological characteristics of the patients are presented in Table I.

A total RNA extraction kit (Fastagen Biotech) was used to extract the total RNA, which was subsequently subjected to reverse transcription by SuperScript III Transcriptase (Thermo Fisher Scientific). qPCR was performed using the Bio-Rad CFX96 system with SYBR-Green for analysis of the mRNA levels of each specific gene. Human GAPDH cDNA was used as the internal control. The primer sequences are listed in Table SI. The PCR thermocycling conditions were as follows: Step 1: 95°C, 30 sec, 1×; step 2: 95°C, 0 sec, and 60°C, 30 sec, 39×; step 3: 4°C, +∞, 1×.

RIPA lysis buffer containing protease inhibitor was used for total protein extraction. The Bradford assay was used to detect the protein concentration. Briefly, 12% SDS-polyacrylamide gels were used to separate 30 µg of protein prior to blotting onto the nitrocellulose filter membranes. Non-fat-milk (5%) in Tris-buffered saline and Tween-20 were used to block the non-specific bindings sites of the membranes. The membranes were incubated with specific primary antibodies overnight at 4°C and the following day with secondary antibodies for 1 h at room temperature. Molecular Imager ChemiDoc XRS System (Bio-Rad Laboratories, Inc.) was used to visualize the protein bands. Immunoblotting of GAPDH was used as the internal control. The primary antibodies for neurofibromin (ab17963, 1:1,500) and VEGFA (ab46154, 1:1,500) were purchased from Abcam. Primary antibody for AR (#5153, 1:1,000) and horseradish peroxidase-conjugated secondary antibodies (#7074 for anti-rabbit IgG and #7076 for anti-mouse IgG, 1:200) were purchased from Cell Signaling Technology, Inc.

A total of 5×10⁵ shNf1-SW10 cells or shNf1 fibroblasts were seeded in a 60-mm culture dish. Following 24 h of incubation, the cells were washed with serum-free medium (SFM) 3 times, and an additional 5 ml of SFM were added. The cells were then cultured for an additional 24 h at 37°C. The supernatants were centrifuged (1,000 × g, 5 min) to remove the cell debris and the conditioned medium (CM) was stored at -80°C. The RayBio^® Human VEGF ELISA kit (RayBiotech Inc.) was used to examine the concentration of VEGF in CM. The VEGF concentration was modified by the addition of neutralizing VEGF antibody (MAB293) or the addition of recombinant human VEGF (AF-293-NA) (both from R&D Systems).

Cell viability and growth rate were analyzed by MTT assay. Briefly, 4×10³ cells were seeded in 96-well culture plates. After 24 h, the cells were treated with DHT (5 nM) or DMSO for the indicated time periods (0, 24, 48 and 72 h). The cells were then washed and incubated with MTT solution (M2128; Sigma-Aldrich) (0.5 mg/ml) at 37°C. Following 4 h of incubation, the medium was carefully removed and 150 µl of DMSO was used to solubilize the formazan crystals. The absorbance of each well was detected by a microplate automatic reader (Bio-Tek Instruments, Inc.) at 490 nm.

A total of 12 4-week-old nude mice were randomly allocated to 2 groups with random digits. All mice were injected with 2×10⁶ shNf1-SW10 cells subcutaneously. One week later, the mice in the 2 groups were treated daily with 100 µl of saline and with 0.6 mg/kg of DHT, respectively. The body weight was measured every 3 days. Following 4 weeks of xenograft tumor growth, the mice were sacrificed and the tumors were excised surgically and measured. A solution with 4% paraformaldehyde was used to fix the tumors. The samples were embedded in paraffin, and IHC staining of VEGFA expression was performed. The animal experiments were approved by the Institutional Review Board of the First Affiliated Hospital of Xi'an Jiaotong University.

A total of 1,000 shNf1-SW10 cells or shNf1 fibroblasts were seeded into 1 well of a 6-well plate. The cells were treated with DHT (5 nM) or saline for 2 weeks. Subsequently, the cells were washed, fixed with 4% paraformaldehyde and stained with 0.1% crystal violet (C6158; Sigma-Aldrich) solutions for 15 min at room temperature. The cells were then washed with PBS 3 times at room temperature. Visible cell colonies in each well were counted and the average colony number of each treatment was calculated.

The human umbilical vascular endothelial cell (HUVEC) cell line was purchased from the American Type Culture Collection (ATCC). SFM or SFM added with CM (1:1) were prepared for HUVEC (1×10⁵) culture. A 24-well plate with Matrigel-coated wells was used for the tube formation assay. Imaging was conducted using an optical microscope (IX50-S8F2; Olympus) at the 6-h time period. A tube was identified as 2 branching points that were perfectly connected.

The cells were passaged at a confluence of 60-80% with Trypsin/EDTA. Subsequently, the cells were washed with cold PBS resuspended in ice-cold 70% ethanol and stored for >24 h at −20°C. On the day of the analysis, the ethanol was removed and the cells were washed 2 times with cold PBS. RNAse A (0.5 µg/ml) and propidium iodide (50 µg/ml) were added to the cells that were incubated at room temperature for 30 min in the dark prior to flow cytometric analysis (BD FACSCalibur™ Flow Cytometer; BD Biosciences).

Transwell migration analysis was conducted by 8-µm-pore Transwell inserts (Millipore Corp.). HUVECs (3×10⁵ cells/ml) were mixed with 300 µl of SFM and subsequently seeded into the upper chamber of the Transwell. In the lower chamber, 1 ml of CM or neurofibroma cells treated with reagents was added. Following 16 h of culture, the cells that migrated to the lower surface of the inserts were fixed with 4% paraformaldehyde and stained using 0.1% crystal violet for 15 min at room temperature. The visible cells were counted in 5 random fields (×200 magnification) for each insert with an optical microscope (IX50-S8F2; Olympus).

The promoter region of VEGFA (-1618 to +100) was amplified and inserted into the pGL3-basic plasmids (GenePharma), denoted as pGL3-VEGFA. Nf1-ablated SW10 cells and fibroblasts were transfected with pGL3-basic or pGL3-VEGFA by X-tremeGENE HP DNA transfection reagent (Roche). A control sample was prepared by the addition of the DNA transfection reagent without any plasmids (pGL3-control). A dual luciferase assay kit (Promega) was used for the luciferase assay following the instructions of the manufacturer. The luciferase activity of each well was normalized by comparison with Renilla luciferase activity. The data from 3 wells were collected and the average value was used for analysis of luciferase activity of each sample. The primers used for PCR amplification were the following: forward (5′-ATTCCCATTCTCAGTCCATG-3′) and reverse (5′-CTGACCGGTCCACCTAACCG-3′).

The SimpleChIP^® enzymatic chromatin IP kit was used for ChIP assay in Nf1-ablated SW10 cells and fibroblasts following the indicated protocol. The precipitation of the protein/DNA complex was achieved by antibodies against the AR (#5153, 1:50; Cell Signaling Technology, Inc.) or against normal rabbit IgG (from the kit). The DNA of interest was detected with genetic region-specific primers (Table SII) and amplified by PCR.

Biotin was added to the oligonucleotides of the specific sets of the VEGFA promoter (biotin-5′-CTTCCCCTGCCCCCTTCAATATTCCTAGCAAA GAGGGAACGGCTCT-3′, synthesized by GENEWIZ). The lysis buffer contained NaCl (150 mM), Tris HCl (50 mM, pH 7.4), EDTA (1 mM) and Triton X-100 (1%). Nf1-ablated SW10 cells and fibroblasts were treated with 10 ng/ml of DHT for the indicated time periods (0, 12, 24 and 36 h). The cells were lysed in lysis buffer with protease and phosphatase inhibitors. Following centrifugation (15,000 × g, 4°C, 15 min), 20 µl of ImmunoPure streptavidin-agarose beads were added to each cell extract at 4°C. Following 1 h of incubation at 4°C, the bead-bound cell extracts were centrifuged at 5,000 × g for 1 min at 4°C. The supernatant was collected and incubated with 100 pmol of biotinylated oligonucleotides at 4°C for 24 h. The immobilized streptavidin-agarose beads (30 µl) were used to precipitate DNA-bound proteins at 4°C. Following 1 h of incubation, centrifugation (5,000 × g, 4°C, 1.5 min) was performed and the supernatant was carefully discarded. The precipitate was washed 3 times with lysis buffer prior to western blot analysis.

Statistical analyses were performed by GraphPad Prism (version 5.0; GraphPad Software, Inc.) software, and the Student's t-test was used for 2-group comparisons. For comparisons of >2 groups, we used one-way ANOVA and Fisher's Least Significant Difference test (LSD test) with statistical software SPSS for Windows 10.0. For the correlation analysis, we employed the Spearman's correlation test with SPSS. A P-value <0.05 (P<0.05) was considered to indicate statistically significant differences.

To explore the potential association of AR expression and cNF angiogenesis, we performed IHC assays with antibodies against AR and CD31 (marker of vascular endothelial cells) of 29 cNF and 29 adjacent normal dermal tissues from male patients. The results revealed significantly elevated AR expression levels in cNF tissues compared with those noted in adjacent tissues (Fig. 1A). CD31 staining of the cNF tissues indicated an increased MVD (Fig. 1B) and an enhanced AR expression that was associated with MVD (Fig. 1D). Moreover, a positive linear correlation was noted between MVD and AR expression in the human cNF tissues (r=0.862, P<0.001, Fig. 1C). Taken collectively, the results indicated that AR expression was positively associated with angiogenesis in human cNF samples (Fig. 1).

The loss-of-function mutation of the Nf1 gene is considered a major genetic change responsible for the development of the NF1 type disease. Therefore, we constructed Nf1 knockout (k/o) SW10 cells and Nf1 k/o fibroblasts using lentivirus-delivered specific shRNA (Fig. 2A). To further identify the role of AR signaling in neurofibroma, we monitored the activation of the AR with DHT binding and examined the enhanced cellular proliferation by MTT assay (Fig. 2B). In addition, we performed s colony formation assay and noted an increased colony number in the aforementioned groups (Fig. 2C). Moreover, DHT treatment resulted in an increased number of cells entering the S phase, as demonstrated by flow cytometric assay (Fig. 2D).

Furthermore, the subcutaneous tumorigenesis potential of the shNf1-SW10 cells was significantly increased following DHT intraperitoneal injection. Following 4 weeks of xenograft growth, tumor weight was considerably higher in the DHT-treated nude mice compared with that noted in the control animals (1.468±0.1434 mg, n=7 vs. 1.098±0.08905 mg, n=7) (Fig. 2E).

Taken collectively, the results demonstrated that active AR signaling promoted neurofibroma cell viability (Fig. 2), which is consistent with the clinical observations regarding the increased number and size of neurofibroma during puberty.

Since AR expression was increased with MVD in human cNF tissues, we employed HUVEC recruitment assays to further confirm the effects of AR signaling on angiogenesis. Briefly, DHT- or saline-treated shNf1-SW10 cells and shNf1 fibroblasts were seeded to the lower chamber or CM from Nf1 ablated cNF cells with or without treatment with DHT were added to the lower chamber of the Transwell and the HUVECs were seeded in the upper chamber of the Transwell. Following incubation at 37°C for 16 h, the CM from DHT-treated cells attracted additional HUVECs to the lower surface of the upper chamber compared with those of the control group (Fig. 3A and B, quantitative data). As regards the co-culture system, neurofibroma cells were seeded in the upper chamber and HUVECs were added to the lower chamber. We detected multiple viable cells in the HUVECs co-cultured with neurofibroma cells and in HUVECs treated with DHT (Fig. 3C). Moreover, the capillary tube formation on Matrigel was significantly higher in the DHT-treated shNf1-SW10 cells compared with that of the control cells (Fig. 3D).

On the whole, the results shown in Fig. 3 suggest that the activated AR signaling may modulate the paracrine function of neurofibroma cells and may alter the interaction of neurofibroma cells with HUVECs. These processes can promote angiogenesis.

Since AR binds to DNA and promotes the transcription of target genes, we hypothesized that the activation of AR signaling may cause the release of angiogenic factors and promote angiogenesis in neurofibroma. Therefore, we performed RT-qPCR assay and detected the expression levels of angiogenic cytokines in shNf1-SW10 cells with or without DHT treatment. Statistical analysis indicated that VEGFA and IGF-1 were key players involved in angiogenesis of neurofibroma cells (Fig. 4A). Considering the predominant roles of VEGFA in tumor angiogenesis, we focused on the regulation of VEGFA by AR. The mRNA and protein expression levels of VEGFA were significantly increased following DHT treatment (Fig. 4B and C). In addition, the expression levels of VEGFA were monitored in DHT-treated tumors (Fig. 4D). The transfection of shNf1-SW10 cells and shNf1 fibroblasts with AR-specific siRNA effectively impaired AR expression and reduced VEGFA expression (Fig. 4E). Furthermore, ELISA was performed to detect the reduction in VEGFA secretion caused by siAR transfection and the increase in VEGFA expression following DHT treatment in the medium (Fig. 4F and G). The data indicated that the inhibition of the AR signaling could be considered a novel strategy to abolish HUVEC recruitment. Taken together, these data demonstrated that AR signaling regulated neurofibroma angiogenesis by promoting VEGFA expression.

AR has been shown to bind to the promoters of target genes at the androgen receptor binding element (ARE). We found 7 putative AREs approximately 1,700 bp from the transcription initiation site (Fig. 5A). Subsequently, we cloned the DNA fragment, which was inserted into the pGL3-basic luciferase reporter plasmid to construct the pGL3-VEGFA plasmid. The dual luciferase assay in shNf1-SW10 cells and shNf1 fibroblasts indicated that DHT significantly enhanced the activity of pGL3-VEGFA plasmids (Fig. 5B), which confirmed the increase in VEGFA mRNA levels in DHT-treated cells. Four primer pairs specific for amplifying different regions of the VEGFA promoter were employed in the ChIP assay. A total of 2 out of 4 regions maintained a higher binding affinity with the AR (Fig. 5C). To further clarify whether AR binds directly to the −400/−50 region, which is close to the transcription initiation site of the VEGFA promoter, oligonucleotides with the same sequences were synthesized with biotin bound at the 5' terminus to pull-down the AR proteins. The cells that had been treated with DHT indicated the amplification of AR binding to the VEGFA promoter, which explained the upregulation in the expression of VEGFA following DHT treatment (Fig. 5D). Taken collectively, these data suggested that AR acted as a transcription factor bound directly to the VEGFA promoter that could initiate its transcription (Fig. 5).

The activation of the AR signaling can increase VEGFA expression and secretion, by promoting neurofibroma angiogenesis. Therefore, we further explored the latent therapeutic action of AR targeting in neurofibroma. Enzalutamide or MDV3100 is an androgen receptor antagonist that was selected to reduce the activation of the AR signaling. MDV3100 reduced AR and VEGFA protein levels, which was consistent with the results obtained from the cells transfected with siRNA for AR (Fig. 6A and B). To examine the effects of MDV3100 on HUVEC recruitment, CM from shNf1-SW10 cells and shNf1 fibroblasts was used to attract HUVECs. The results indicated a weakened HUVEC migration in the aforementioned groups (Fig. 6C). In addition, diminished tube formation was noted in the presence of CM from cells with MDV3100 treatment. The activation of the AR signaling following DHT treatment promoted HUVEC recruitment and tube formation, which was consistent with the results noted in Fig. 3.

The results shown in Fig. 6 demonstrated that the inhibition of AR signaling suppressed HUVEC recruitment and impaired tube formation of these cells, which further suggested that the inhibition of AR signaling could partially disrupt the angiogenesis of neurofibroma.