Project

Background and purpose: Gliomas are the most common brain tumors, comprising approximately 80% of all malignant brain tumors. Aberrant epigenomes happened in many adult brain cancers, as demonstrated by widespread changes to DNA methylation patterns, redistribution of histone marks and disruption of chromatin structure. In adult glioblastoma, the most aggressive and prevalent adult primary intrinsic brain cancer, nearly 46% of patients harbor at least one mutation of an epigenetic regulator amid a diversity of oncogenic pathway mutations. Therefore, predictions of tumor behavior, treatment response, and patient survival remain challenging. The uncertainty of prognosis can be attributed to the interobserver variability in histological diagnosis and the inherent tumor heterogeneity, especially in higher grade gliomas. Recently, large cohort studies based on the Cancer Genome Atlas (TCGA) database identified molecular profiles of gliomas and proposed that the IDH mutations, mutations of TERT promotor, and codeletion of 1p/19q chromosome arms, can be reliable biomarkers to categorized gliomas into five principal groups with different patient survivals and distinct mechanisms of pathogenesis. However, the invasive tissue biopsy, high costs of DNA assay, and tumor heterogeneity prohibit its applications in diagnosis and patient follow-up. Our purpose is to develop the MRI-based radiomics of gliomas and to determine the relations between image features and the underlying genotypes. We hypothesize that the tumor MR phenotypes can reflect the genotypes pattern and resistance related epigenomic modification status at a certain level, and therefore can be used to differentiate subtypes, predict the patient survivals, thus guide therapy (1-3: Development of diagnosis, prognosis and novel therapeutics). Study design and methods: Two datasets will be included in this study, namely TCGA database cohort and recruited study cohort. TCGA dataset will be used to develop MR radiomics/radiogenomics platform and examine the feasibility of imaging-based differentiation of glioma subtypes. The study cohort will be employed to evaluate the tumor heterogeneity, recurrence, and epigenomic alteration; and will be recruited from three hospitals (50 patients in each year) with the application of advanced MR techniques (diffusion kurtosis imaging, MR spectroscopy, and dynamic susceptibility contrast imaging) and DNA mutation, methylation, gene/protein expression and histone modification analyses for locus-specific tumor samples. Significance and specific aims: This study will provide insights into the connections between the genetic profiles and MR imaging phenotypes of gliomas. Novel imaging platform combining multi-modality MR techniques and multi-feature analyses, such as histogram, geometry, and texture analyses, will be constructed to decode the tumor phenotypes into quantifiable features. A mixture model of imaging traits estimated by the multivariate linear regression will provide optimal formulae in predicting tumor behavior and differentiating tumor heterogeneity. This is a multi-site (Taipei Medical University Hospital, Taipei Veteran General Hospital, Tri-Service General Hospital, Linkou Chang Gung Memorial Hospital, and Mackay Memorial Hospital). Clinical data, tissue samples, and advanced MRI for patients with glioma will be collected from the five hospitals in Taiwan. The analyses of TCGA database, MR techniques, and radiomic/radiogenomic/radioproteomic approaches will be developed by our research center with supports from different lab collaborations. Five specific aims will be achieved in this four-year project. Aim 1: Extract locus-specific tumor tissues and corresponding DNA mutation, gene expressions and histone modification to investigate tumor heterogeneity. Aim 2: Connect MR phenotypes to the underlying genotypes (radiogenomics) and validated by the protein expressions (radioproteomics). Aim 3: Predict treatment response, tumor recurrence, and patient survivals by the established imaging predicted model. Aim 4: Developing MR epigenomic biomarker for treatment response and tumor recurrent prediction. Aim 5: Establish combination treatment strategy to sensitize brain tumor for chemotherapeutics through ‘epigenetic priming’ on tumor bearing animal model.

Post-stroke trans-neuronal degeneration (TND) refers to secondary neuron deaths following the disruption of input from or output to other synapsed neurons sustaining ischemic insults after few days to several weeks of stroke onset. Secondary TND in the thalamus following ischemic infarct of the middle cerebral arterial (MCA) territory has been reported associated with the degeneration of the thalamocortical fibers. Given the integral role of the thalamus in the sensorimotor and other neurocognitive functions, damage to the thalamus or its projections is likely to have detrimental consequences. Although global morphological changes of thalamus have been reported after MCA infarcts, in vivo studies of TND on the intrinsic integrity of specific nuclei and their effects on distinct cortical functions have yet to be fully described. The questions whether the TND is nucleus-specific, how the nucleic degeneration may affect clinical recovery, and can the trans-synapse apoptosis be prevented remain unclear. To effectively translate the results in animal models into human stroke, mechanisms underlying the ischemic injury of TND need to be elucidated with more comparative studies across human and animals. We will first implement thalamic parcellation to determine the major nuclear topography in rat and man using structural MRI. Based on the thalamic parcellation, quantitative measurements of the microstructural and functional alterations in selective thalamic nuclei will be implemented using advanced MRI technology. TND will be conducted on a photothrombotic rat stroke model for specific thalamo-cortical pathways and translated into human stroke by correlating the imaging findings with pathological data from the animal model. The relevant pathophysiological mechanisms underpinning the imaging findings can then be assessed. By the novel imaging design and a longitudinal follow-through study using high Tesla MRI (7T and 3T), we will integrate these measurements and the neurological function tests of sensorimotor, memory, language and other neurocognitive functions as well as immunohistopathological results to explore the relations between the neuroimaging and clinical findings. We will test the hypothesis that the effects of TND on the thalamus would be nucleus-specific and the functional modifications following neuronal insults would be interrelated with the distinct connection of thalamic nuclei to the cerebral cortex associated with sensorimotor and other neurocognitive functions. We will further evaluate the longitudinal changes of TND by characterizing the time course of TND-related brain injury to provide the optimal time window for possible therapeutic strategies using integrated measurements. A specific rehabilitation training of robot-assisted gait training (RAGT) system designed to enhance neuroplasticity and facilitate functional recovery will be enrolled in this study to investigate the changes of thalamic connectivity for patients with chronic stroke after therapeutic intervention. We will use MRI to identify how the functional and structural remodeling of the thalamocortical pathway will be correlated with locomotor changes after RAGT therapeutic intervention. We aim to develop imaging biomarkers that can predict the risk of specific neurological function deterioration and thereby identify subjects for targeted intervention to provide the potential recovery biomarkers and assist therapeutic approaches.

The relation between thalamocortical dysrhythmia (TCD) and post-concussive syndrome (PCS) in mild traumatic brain injury (mTBI) patients has recently been recognized in several studies [1-3]; however the potential imaging biomarkers to evaluate neurological functional deficits and to localize TCD, in particular the thalamocortical circuits, have not been established. Based on the current understanding of the mTBI, network dysfunction of thalamocortical circuits is the one of main points of investigation which is characterized by microstructural injuries [4-7], disrupted functional connectivity [1, 8, 9], and the altered interplay between excitatory and inhibitory neurotransmitters [10]. The elevated synaptic glutamate levels after mTBI can further exacerbate the post-traumatic cellular injury [11]. Recent studies have indicated that mTBI patients can benefit from N-acetylcysteine (NAC) treatment by effectively improving the behavioral deficits which had a significant impact on neuropsychological test results [12-14]. Though the cognitive tests outcome after administration of NAC is promising, the neuroimaging evidence of NAC treatment efficacy on mTBI and its prevention of subsequent brain atrophy had not been well explored. The purpose of this study is to propose a TCD model in mTBI by simultaneous measurements of MRI and MR-compatible EEG on human participants followed by a pre-clinical experiment using animal mTBI model to evaluate the therapeutic effect of NAC on mTBI. Through the characterization of the neural network dysfunction from the perspectives of neural electrophysiology, microstructure integrity, and hemodynamic synchrony, we will be able to have a complete vision on the pathophysiological mechanism in mTBI. In addition, this study can help provide the neuroimaging evidence of the NAC efficacy in treating mTBI. The specific aims for this three-year project comprise: 1. Characterize the TCD in patients with mTBI by measuring the theta shift and gamma activity on cortical surface (EEG), microstructural injuries in white matter (MR fractional anisotropy), disrupted thalamocortical connectivity (MR diffusion tensor tractography and fMRI connectivity), and the altered GABA concentrations (MR spectroscopy). 2. Construct a norm of thalamocortical oscillation including the above-mentioned EEG and MRI features from 40 healthy volunteers. 3. Localize TCD at specific cortical regions based on the alterations of structural and functional thalamocortical connectivity compared to the constructed norm. 4. Correlate the impairments of thalamocortical circuits with PCS and the impaired neurocognitive performance by comparing between the sub-categorized mTBI cohorts. 5. Provide neuroimaging evidence of the treatment efficacy of NAC in mTBI and investigate the restoration or reconnection of thalamocortical circuits at follow-up using an animal model. 6. Identify the potential image biomarkers for predicting the cortical regions of brain atrophy 3 months after brain injury.

Glioblastoma is the most common malignant tumor of the central nervous system. Over 90% of diagnosed glioblastoma are primary cases, originating from glial cells and making up 80% of malignant brain tumors through multi-step oncogenesis. Despite the advances of treatment, the cure of malignant GBM remains elusive and the prognosis appears poor, partly due to under-sampling and mis-grading that influence the therapeutic strategy. Moreover, intra-tumoral and individual genetic heterogeneity remains another important issue. Therefore, an accurate grading of glioma by advanced imaging biomarkers that can address the heterogeneity of glioblastoma cannot be overemphasized. In addition to the heterogeneity at the histology level, primary glioblastoma are also characterized by heterogeneous alterations of gene expression, which consist of EGFR amplification, loss of PTEN, and loss of cyclin-dependent kinase inhibitors. The incorporation of oncogenomics to the imaging biomarkers, the radiogenomic, has recently emerged for cancer management. The key pathological features of glioblastoma, including cellularity, invasiveness, mitotic activity, angiogenesis, and necrosis can be evaluated with advanced MR imaging which can be used as glioma genomic signatures. By linking specific imaging biomarkers with specific gene expression profile could allow for more accurate diagnosis, prognosis prediction and better therapeutic guidance. In this three-year translational glioma project, we aim to investigate the advanced MR imaging biomarkers, the phonotypical expression, of gliomas with respect to their gene expression in preclinical animal model at 7T MRI and human subjects at 3T. To study intra-tumoral and inter-tumoral heterogeneity of glioblastomas, advanced MR imaging biomarkers such as perfusion-weighted imaging (PWI), CBF measurements by arterial spin labeling (ASL), high resolution diffusion tensor imaging (DTI), MR spectroscopy (MRS), Amide proton transfer (APT) image and Vessel size image will be used to depict the physiological as well as molecular events within the tumor. Genomic information will be acquired from microarray analysis. In addition to the frequently affected pathway including proliferation and angiogenesis, we will also focus on the DNA damage/repair genes, as radiation response of glioblastomas in Taiwanese is uniquely different from that Caucasian glioblastoma population. With the combined application of genomic information and multiparametric MRI, the tumor extent can be accurately defined, the therapeutic response can be better monitored and the prognosis can be well predicted. In summary, with the establishment of tumor radiogenomic platform, we will be able to unveil the genomic modification of glioblastoma from the perspective of molecular MRI through a bedside-to-bench translational research.

1234

Address: No.250, Wu Hsing Street, Taipei City 110, Taiwan (R.O.C.)
Telephone: +886 02-27361661 ext.8902
E-mail: tmutirc@tmu.edu.tw

© Copyright – TMU Translational Imaging Research Center – design by Morcept