This study was approved by the Ethics Committee of our hospital. All patients involved in this study provided informed consent prior to the initiation of the study. From September 2015 to February 2019, 21 patients (male, n = 11; female, n = 10) with VHL, confirmed by genetic diagnosis, underwent surgery for CNS HGBs in our hospital. Thirty CNS HGBs were surgically removed, all of which were verified by postoperative pathology. These tumor tissues were embedded in paraffin or frozen directly and stored at −80°C.

All 21 patients underwent contrast-enhanced magnetic resonance imaging (MRI) examinations (T1-weighted imaging [T1WI], T2-weighted imaging, diffusion-weighted imaging, and contrast-enhanced T1WI; slice thickness, 1 mm) to determine the lesion location. The volume of the tumor and associated cysts was calculated by using a Picture Archiving and Communication System containing a DICOM viewer software (CARESTREAM PACS, CarestreamHealth Inc, Toronto, Canada). All enrolled patients were additionally examined with abdominal computed tomography and fundoscopy to identify related lesions in other target organs. The Karnofsky Performance Scale (KPS) score was used to evaluate neurological function. Postoperative follow-up examinations were performed every 6 months. Patients underwent follow-up contrast-enhanced MRI examinations to determine the presence of unresected tumor or recurrence.

Genomic DNA from peripheral blood was extracted using a QIAamp DNA Blood Mini Kit (QIAGEN, Germany). Three coding exons and their flanking intronic regions were amplified by polymerase chain reaction (PCR) using primers described previously (17). Missense and splicing mutations were detected by direct sequencing. A multiplex ligation-dependent probe amplification kit (MLPA, P016-C2, MRC-Holland, Amsterdam) was used to detect the large exon deletions, which were confirmed by Real-time (RT) quantitative PCR with primers described by Ebenazer et al. (18). All patients were divided into two groups: the missense group and the partial deletion group.

The paraffin-embedded tissue from 30 VHL-related CNS HGBs (medulla oblongata, n = 5, cerebellum, n = 25) and three control brains from cerebral hemorrhage patients was cut at a thickness of 4 μm Paraffin-embedded sections were deparaffinized in xylene and rehydrated through a graded series of alcohols. Heat-induced antigen retrieval was performed in 10 mM citrate buffer, pH 6.0 for 10 min in an 800-Watt microwave. After cooling, sections were incubated in 3% H₂O₂ for 10 min at room temperature to quench endogenous peroxidases. Next, sections were incubated in 5% bovine serum albumin (BSA) for 30 min to block non-specific binding. Sections were then incubated with a rabbit anti-human phospho EGFR (1:200, Abcam, ab40815) and a rabbit anti-human TGFα (1:200, Abcam, ab208156) at 4°C overnight. After rinsing, a rabbit-specific horseradish peroxidase (HRP) polymer kit (Beijing Zhongshan Jinqiao Biotechnology, PV-9001) was used at room temperature for 50 min and the antibody reaction was visualized with chromogen 3,3-diaminobenzidine tetrahydrochloride (DAB). Sections were counterstained with hematoxylin and mounted with DPX after dehydration.

The expression of EGFR and TGFα among these sections was evaluated by measuring integral optical density (IOD) and mean optical density (MOD) using the Image-Pro Plus 6.0 (Media Cybernetics, Rockville, MD, USA) (19). The following scoring method for the assessment of EGFR and TGFα was used: score 0 ≤ 10% stained tumor cells; score 1 ≥ 10% but <25% stained tumor cells; score 2 ≥ 25% but <50% stained tumor cells; and score 3 ≥ 50% stained tumor cells (20).

Total RNA was extracted from 7 snap-frozen CNS HGBs and 3 control brains from cerebral hemorrhage patients using a RNeasy kit (Invitrogen, St. Louis, MO, USA). The RNA was reverse transcribed with a first strand cDNA synthesis kit (MBI Fermentas, Lithuania). RT PCR was used to quantitatively analyze RNA expression using the Premix Taq™ (Ex Taq™ Version 2.0). Related data acquisition and processing was completed with a Roche LightCycler® 480II PCR machine (Roche, Switzerland). The values of cycle threshold (CT) for theses samples was normalized to the corresponding GAPDH CT values. The relative expression of RNA was assessed with the ΔΔCT method.

Patients and tumor characteristics were analyzed using descriptive statistics. The difference between HGBs and controls for the MOD and ΔΔCT values was compared using the Mann–Whitney U test. The immunohistochemical score and the age at first operation between the partial deletion group and missense groups were also compared using the Mann–Whitney U test. The Chi-squared test was used to compare the sex distribution between groups. The relation between immunohistochemical score and cyst size was evaluated by linear regression. Significance was assumed for two-sided p-values < 0.05. The data were analyzed with the Statistical Package for Social Science (SPSS) 21.0 (IBM, Armonk, NY, USA).

A total of 21 VHL patients confirmed by genetic testing (missense group, n = 9; partial deletion group, n = 12) from different pedigrees of 12 Chinese provinces were enrolled in this study. The mean age (SD) was 33.3 ± 10.2 years (range, 21–58; median, 31). The male-to-female ratio was 1.1 (males, n = 11; females, n = 10). All patients had received 44 operations for CNS HGBs, with an average of 2.1 ± 1.4 operations (range, 1–5; median, 1). The mean age at first operation in these patients was 28.7 ± 11.8 years (range, 13–58; median, 27). Our results showed no significant difference in sex (p = 0.806), age at first operation (p = 0.24), or number of operations (p = 0.937) between the missense group and the partial deletion group. However, the age at first operation of men was found significantly lower than that of women (p = 0.005).

The preoperative MRI showed that the 21 patients harbored 116 CNS HGBs on admission, with an average of 5.5 ± 3.1 tumors/patient (range, 2–13; median, 4). Because of obvious occupying effects, 18 cystic tumors and three solid tumors were resected. Because of their proximity to the larger tumors, another nine solid tumors were also removed at the same time to prevent an increase in postoperative volume. The average volume of a resected cyst from HGBs was 16.9 ± 10.7 cm³ (range, 1.4–41.3; median, 18.7). The cyst volumes showed no difference in sex (p= 0.161) and mutation type (p = 0.436). Among the 21 patients, 90% had pancreatic cysts, 14% had pancreatic neuroendocrine tumors, 57% had renal cysts, 38% had renal cancer, and 29% had pheochromocytomas. Retinal HGBs occurred in 43% of patients.

Twenty patients recovered well postoperatively and were discharged ~1 week later. One patient died from aspiration after surgery. Because of the high recurrence and low growth rate, our follow-up interval was 6 months. The mean follow-up duration in these 21 patients was 2.05 ± 1.04 years (range, 0.5–4; median, 2). The KPS score at 6 months postoperatively was 95 ± 7.6 (range, 80–100; median, 100).

The expression of EGFR and TGFα was analyzed in tissue from 18 cystic and 12 solid HGBs with immunohistochemistry. Twenty-five tumors were located in the cerebellum and five in the medulla oblongata, while tissue from three normal brains was used as negative control. The results of Image-Pro Plus for EGFR and TGFα showed that the MOD in VHL-related HGBs was significantly higher than that in the control (EGFR, p = 0.007; TGFα, p = 0.021).

The immunoreactivity for pEGFR was strong in stromal cells in all of the VHL-related CNS HGBs. Immunoreactive products were mainly located in the cytoplasm and along the cell membrane of the stromal cells. TGFα expression was also observed in the cytoplasm. However, local differences in the intensity of immunostaining are observed in most tumors. The staining intensity for TGFα appeared to be weaker than that of EGFR (Figure 1). However, tissue from two of the tumors appeared to be negative in TGFα expression compared to normal brain tissue. The mean positive cell percentage of EGFR in 30 tumor tissues was (68.6 ± 14.7)%. And the TGFα was (61.9 ± 26.8)%. The immunohistochemical score revealed that most of the HGBs had strong positive expression for EGFR and TGFα (Table 1), while there was no difference in expression between the partial deletion group and the missense group (TGFα, p = 0.435; EGFR, p = 0.539). Our analysis also showed that the expression intensity of EGFR and TGFα in relation to the solid tumor size had no statistical difference. The mean immunohistochemical positive cell percentage of EGFR had a positive correlation with the cyst volume (r = 0.782, p < 0.01). However, the combination of EGFR and TGFα expression levels or the TGFα immunohistochemical score showed no correlation with the cyst volume.

Due to the deficiency of fresh-frozen tissue, only tissue from seven VHL-related HGBs and three controls were used in RT PCR to detect expression of EGFR and TGFα mRNAs. The results were similar to those obtained with the immunohistochemistry analysis. EGFR and TGFα mRNAs were upregulated in all the tumors compared to the control group. ΔΔCT analysis revealed that the expression of EGFR (p = 0.017) and TGFα (p = 0.017) mRNAs in VHL-related HGBs was significantly higher than that of the control group (Figure 2).