Regulatory landscape and clinical implication of MBD3 in human malignant glioma

In this article we inspect the roles and functions of the methyl-CpG-binding domain protein 3 (MBD3) in human malignant glioma, to assess its potential as an epigenetic biomarker for prognosis. The regulatory effects of MBD3 on glioma transcriptome were first profiled by high-throughput microarray. Our results indicate that MBD3 is involved in both transcriptional activation and repression. Furthermore, MBD3 fine-controls a spectrum of proteins critical for cellular metabolism and proliferation, thereby contributing to an exquisite anti-glioma network. Specifically, the expression of MHC class II molecules was found to positively correlate with MBD3, which provides new insight into the immune escape of gliomagenesis. In addition, MBD3 participates in constraining a number of oncogenic non-coding RNAs whose over-activation could drive cells into excessive growth and higher malignancy. Having followed up a pilot cohort, we noted that the survival of malignant glioma patients was proportional to the content of MBD3 and 5-hydroxymethylcytosine (5hmC) in their tumor cells. The progression-free survival (PFS) and overall survival (OS) were relatively poor for patients with lower amount of MBD3 and 5hmC in the tissue biopsies. Taken together, this work enriches our understanding of the mechanistic involvement of MBD3 in malignant glioma.


Fluorescence correlation spectroscopy (FCS)
FCS was conducted with a scanning confocal timeresolved microscope (Microtime, PicoQuant GmbH) to test the chromatin binding states of MBD3 in SF767 cells. Using autocorrelation function, the diffusion profile of fluorescent molecules can be calculated by: where <N> is the average number of diffusing molecules inside the detection volume, τ is the lag time between two measured fluorescence intensities, and τ Di is the diffusion time for the ith component of the molecules. For most proteins in living cells, i can be set to 2 to differentiate the bound and unbound status of the molecules, as well as the respective proportion f i . The MBD3-GFP plasmid was transfected following the previously optimized protocol and the phototoxicity was carefully controlled in FCS measurements [1].
In FCS data collection and analysis, only cells with a MBD3-GFP expression level at 80-100 nM were selected. In parallel to be fitted with the least-square algorithm provided in the SymPhoTime software (PicoQuant GmbH), autocorrelation functions were fitted by a Maximum Entropy Method (MEMFCS) as validation [2,3]. In a general comparison, the percentage of chromatinbound MBD3-GFP in glioma cells was found to be 50-70% while the bound component in noncancerous cells was less than 30%.

Colorimetric immunoassay for 5mC
DNA from control and siRNA-treated SF767 cells was extracted and purified with DNeasy Blood & Tissue Kit (Qiagen), according to the manufacturer's instruction. Global DNA methylation was quantified by MethylFlash Methylated DNA Quantification Kit (Epigentek). In brief, 100 ng of DNA was deposited to the bottom of the assay well followed by incubation with capture antibody and detection antibody, sequentially. The developed colorimetric signals were normalized over the standard input DNA to give out the global methylation percentage in samples.

MSP for CIITA promoter IV
The methylation status for the promoter IV region of CIITA gene was determined using MSP. After DNA extraction, bisulfite conversion was performed with EZ DNA Methylation-Lightning Kit (Zymo Research). PCR primer design and reaction condition followed previous research [4]. The amplicons by unmethylated primers and methylated primers were run through 3% agarose gel prestained with GelRed dye (Biotium). The gel band intensity was analyzed with the ImageJ software (http://imagej.nih. gov/ij).

Flow cytometry
Flow cytometry was applied to determine cell cycle progression. In general, cells were fixed with 70% ethanol and then stained with 20 μg/ml propidium iodide (PI) solution containing 0.1% Triton X-100 and 200 μg/ ml RNase A, referring to the established protocol [5]. A FC500 MPL system (Beckman Coulter) was used to quantify the single-cell PI intensity from at least 15,000 cells. The generated cell cycle histograms were analyzed with Flowjo software.

Immunofluorescence staining
Cells were first fixed with 4% paraformaldehyde for 15 minutes and then permeabilized with 0.4% Triton X-100 for 30 minutes. For 5mC staining, cells were additionally treated with 1 N HCl at 37°C for 30 minutes. Blocking was conducted with fresh PBS buffer containing 5% goat serum and 0.3% Triton X-100. Primary antibodies used in this study include mouse anti-5mC antibody (Eurogentec) and rabbit anti-MBD3 antibody (Pierce). Cells were incubated with primary antibody at 4°C overnight and then with secondary antibodies: AlexaFluor-546 F(ab') 2 fragment of Goat anti-Mouse IgG and AlexaFluor-488 F(ab') 2 fragment of Goat anti-Rabbit IgG, respectively. Confocal images were taken with an Olympus IX71 inverted microscope fitted with diode lasers.

Therapeutic strategy for enrolled patients
All patients received surgical resection to the extent feasible, followed by radiotherapy and chemotherapy which were given around 2~4 weeks after the surgical removal. Radiotherapy was given with a dose of 1.8 Gy/d and a total dose of 50-56 Gy which includes 30-36 Gy for brain and spine, and 20 Gy for the original tumor site as local boost. Chemotherapy mainly involved oral administration of temozolomide. One treatment cycle of chemotherapy was a successive dosage of 150 mg/m 2 /d for 5 days within every 4 weeks. 8 cycles were performed in total. All the surgeries were performed in the Department of Neurosurgery, Xiangya Hospital, and the radiotherapy and chemotherapy were conducted in the Department of Oncology at Xiangya Hospital, Changsha City, China.