Differential expression of p42.3 in low- and high-grade gliomas

All patients provided their written informed consent to participate in the study, and the clinical study was approved by the Institutional Review Board of Beijing Tiantan Hospital.

In this study, we used 162 paraffin-embedded samples, 51 nitrogen-frozen samples of human brain gliomas which that had been resected surgically and 13 normal tissues derived from traumatic decompression operations. Histopathologic analyses were independently performed by two pathologists.

The RT-PCR fragment of the p42.3 gene was cloned into the pGEM-T easy vector system (Promega) that contained SP6 and T7 promoters. After sequencing, to confirm the orientation of the insert, the plasmids were linearized by either ApaI or SacI enzyme digestion. Digoxigenin (DIG)-labeled, anti-sense- or sense-strand RNA probes were prepared by in vitro transcription with T7 RNA polymerase and SP6 RNA polymerase according to the manufacturer’s instruction manual (Roche). A tissue array was incubated overnight at 43°C in prehybridization buffer containing 5 ng/μl of a sense or antisense DIG-labeled RNA probe. Detection of hybridization was undertaken using a DIG nucleic acid detection kit (Roche Applied Science, Indianapolis, IN, USA) according to the manufacturer’s instructions. Hybridization with sense-strand RNA probes was used as a negative control.

For protein analysis, tissues were lysed in protein lysis buffer. Protein (100 μg) was separated by 12% Tris-glycine polyacrylamide gel and transferred to a polyvinylidene fluoride membrane by electrophoretic blotting at 90 V for 2 hours. The membrane was blocked with 5% nonfat dry milk in 0.1% phosphate-buffered saline with Tween 20 (PBST) and was incubated with a 1:100 dilution of polyclonal antibody at 4°C overnight. The blot was then washed in PBST and incubated for 2 hours with horseradish peroxidase–conjugated secondary antibody goat anti-rabbit immunoglobulin G (IgG) (1:1,000) and washed in PBST. The membrane was then developed using an electrochemiluminescence Western blotting detection reagent system (Amersham, Little Chalfont, UK) according to the manufacturer’s instructions.

Immunohistochemistry was performed as follows. Briefly, sections were deparaffinized by incubation at 70°C for 2 hours. Next, the slides were subjected to antigen retrieval by autoclave heating in an ethylenediaminetetraacetic acid buffer for 2.5 minutes and left to cool down to room temperature. After that, slides were then incubated in 3% H₂O₂ to block endogenous peroxidase after rehydration in ethanol gradient, and then they were incubated in 10% nonfat dry milk to reduce nonspecific binding. Afterward, the slides were sequentially incubated with purified anti-p42.3 antibody (1:50), biotinylated goat anti-mouse IgG and avidin-biotin complex were added for 30 minutes at room temperature and the slides were developed with diaminobenzidine (Dako, Carpinteria, CA, USA).

A χ² test (SPSS 10.0 for Windows, Chicago, IL, USA) was performed to analyze the expression of p42.3 in grade 2 gliomas, and grade 4 GBMs were examined by RT-PCR (Table 1). Logistic regression analysis (SAS 8.01 software; SAS Institute, Cary, NC, USA) was carried out to assess the frequency of p42.3 gene expression in gliomas identified by mRNA in situ hybridization (ISH) (Tables 2 and3). Statistical significance was set at P < 0.05.

We knew little about the novel gene p42.3 after cloning its full-length cDNA. However, bioinformatics proved to be a very useful tool in helping us to find clues for a successive study; therefore, we predicted the expression profile by digital Northern analysis of the National Center for Biotechnology Information (NCBI) reference sequence. The results showed that the p42.3 gene is highly expressed in several cancer tissues, such as prostate adenocarcinoma, brain GBM, breast carcinoma and colon adenocarcinoma, as well as in embryonic kidney tissue (Figure 1).

Digital Northern analysis showed that the p42.3 gene is expressed in brain GBMs and confirmed that p42.3 is expressed in human fetal brain tissue as well. Therefore, we performed RT-PCR to examine its expression in gliomas, including astrocytomas (grade 2), OAs (grade 2), anaplastic OAs (grade 3), GBMs (grade 4) and corresponding normal tissues. The results showed that the positive frequency of this gene expression was similar between astrocytomas (1 of 11, 9.1%) and OAs (1 of 10, 10%); however, we found that p42.3 gene was expressed in 5 (45.5%) of 11 GBMs (Figure 2A and B) and was not expressed in 6 anaplastic OA samples. Moreover, this gene was undetectable in nine normal brain tissue samples (Figure 2C and Table 1). The prostate cell line PC3 served as a positive control. Using mRNA in situ hybridization with a variety of tissues, we examined p42.3 gene expression in brain gliomas. Table 2 shows that the expression of this gene was positive in 10 (15.9%) of 63 astrocytomas, 2 (12.5%) of 16 OAs, 3 (30%) of 10 anaplastic OAs, 2 (9.52%) of 21 ODs and 19 (36.5%) of 52 GBMs. In contrast, the p42.3 gene was not expressed in four normal brain tissues (Figure 2C and Table 2). More importantly, the statistical significance of expression between grade 2 and grade 4 gliomas was 0.038 (P < 0.05) (Table 1), and the expression of the p42.3 gene was significantly different between low-grade (grade 2) gliomas and high-grade (grades 3 and 4) gliomas (P < 0.01), but there was no statistically significant difference associated with age or sex (Table 3).

We produced polyclonal antibody to the p42.3 protein to analyze the corresponding protein expression and generated six sets of rabbit antiserum against three synthetic peptide fragments of the p42.3 protein. The p42.3 protein was expressed in E. coli and purified by Ni-NTA agarose (Figure 3A). Western blot analysis showed that two of the six sets could specifically bind intact protein expressed in E. coli (Figure 3B).

Western blot analysis was carried out to examine the p42.3 protein expression in gliomas and normal brain tissues. The glioma specimens included eight GBM samples (grade 4), four OA samples (grade 2), three anaplastic OA samples (grade 3), and three normal tissue samples. The results showed that p42.3 protein was frequently expressed in all eight GBM samples and detectable in one OA sample and one anaplastic OA sample. p42.3 protein was not expressed in the three normal tissue samples. β-actin served as an internal control (Figure 4).

We further performed immunohistochemistry on ten astrocytoma samples and ten GBM samples to test p42.3 expression. Our results show that this protein was expressed in both astrocytomas (tumor grades 5 ≥ ++, 5+) and GBMs (tumor grades 4 ≥ ++, 6+). Moreover, the location of p42.3 expression is different in these two glioma types. p42.3 protein was tested in both the cytoplasm and nucleus in nine of ten astrocytoma samples (Figure 5A and B), especially the protein highly expressed in the nucleus. In contrast, this protein was strongly expressed in the cytoplasm in only nine of ten GBM samples, and there was little expression of p42.3 protein in the nucleus (Figure 5C and D).