Year: 2014
Type: article
Abstract: Chemical exchange transfer (CEST) is an MR technology that provides a form of contrast which is distinct from standard MRI sequences. The method is based on a transfer of spins from bound proton sites to bulk water via chemical exchange. In one form of CEST imaging, amide proton transfer imaging (APT), this exchange occurs between peptide/protein amide (-NH) groups and free tissue water to create negative contrast. Malignant gliomas are thought to have a higher concentration of these mobile peptides/proteins than normal brain. APT has shown promise in differentiating normal brain and edematous brain from tumor, in differentiating true from pseudoprogression, and in the non-invasive grading of gliomas. In a recent study, Sagiyama et al employed APT imaging to assess the effects of temozolomide treatment on gliomas. They used a model in which tumors removed from glioblastoma (GBM) patients were injected into mouse brains and passaged without exposure to cell culture conditions. The advantage of this model is that it appears to more faithfully represent histologic and imaging features of the original human tumor than cell-culture derived xenografts. In the first experiment, 11 mice were generated from a patient tumor that was responsive to TMZ treatment. Mice were given one course of TMZ (80 mg/kg) for 3 days followed by 4 days of no drug treatment. TMZ treatment resulted in a significant reduction in APT signal compared to non-treated controls, and this correlated with a reduction in the tumor's Ki67 index (but not tumor volume, cell density, or apoptosis levels). In the next experiment the authors generated mice from a patient tumor both at initial presentation and at recurrence. The initial tumor, but not the recurrent tumor, was responsive to TMZ clinically, even though the tumors were genomically very similar (and thus offered a well-controlled experimental model despite the heterogeneity of GBM). As expected, the implanted TMZ-sensitive tumors shrank and the TMZ-resistant tumors grew following TMZ treatment. The authors again showed reduction in APT contrast in the TMZ-sensitive tumors, but there was an increase in APT contrast in TMZ-resistant tumors. This change was independent of tumor size. Following treatment the TMZ-sensitive tumors again had a significantly lower Ki-67 index than the resistant tumors. The authors conclude that resistance was acquired following patient treatment with radiation and TMZ, and that the APT signal decrease was correlated with treatment response. These findings suggest that APT could be developed into a clinically useful technique that could serve as an early response marker for TMZ response (and potentially other treatments) and which may be independent of (or precede) changes in tumor size. Many physiological imaging techniques such as perfusion MRI have been proposed as methods to develop early response parameters in GBM; none of the resultant markers have been successfully validated in multicenter trials. APT offers a promising new avenue of investigation that may help achieve this important goal. The development of a link between MRI features and underlying molecular data holds the potential to identify surrogate biomarkers that accurately predict natural history and response to therapy in glioblastoma (GBM). The purpose of the study of Gevaert et al was to derive quantitative image features from MR images that characterize the radiographic phenotype of GBM lesions and to create radiogenomic maps associating these features with various molecular data. The study included 55 patients with GBM whose clinical, molecular, and MR imaging data were obtained from TCGA and Cancer Imaging Archive. Quantitative image features by using the electronic Physician Annotation Device (ePAD) were derived from three distinct regions of interest (ROIs): necrosis and enhancement ROIs on T1-weighted post-contrast images and edema ROI on T2-weighted fluid-attenuated inversion recovery (FLAIR) images. In parallel, a comprehensive integration strategy, called Amaretto, was used to summarize the different molecular data (gene expression, DNA methylation, and copy number) into gene expression modules, thereby creating 100 co-expressed gene expression modules. Next, those modules that significantly correlated with survival, ie 35 modules, were associated with the quantitative image features by using the Spearman correlation test, thus creating a prognostic radiogenomic map for necrosis, enhancement, and edema. Some interesting correlations between TCGA subtypes and important molecular aberrations and quantitative image features were found. The mesenchymal subtype was correlated with one edema feature, i.e. the minimum intensity, while the classic subtype was correlated with one necrosis and two edema image features related to edge sharpness and intensity. More specifically, edema in classic tumors has lower intensity than in mesenchymal tumors. As for the final radiogenomic maps, the authors reported preliminary data and hypotheses: for instance, GAP43 and WWTR1 genes that define module 20 could represent driving necrosis processes. In conclusion, this paper extends to GBM a complex methodology that has been already applied successfully in non-small cell and hepatocellular carcinoma. The possibility to provide the biological correlates of imaging characteristics of GBM is extremely attractive. One of the limitations of the study is represented by the fact that ROIs were manually annotated in a single section containing the largest cross-section of the lesion; in the future, semiautomatic approaches for outlining the different ROIs with probably improve the sensitivity. All of these preliminary data need validation in larger cohorts from different institutions. Numerous studies have now documented the extensive molecular heterogeneity characterizing glioblastoma, both on an intertumoral and intratumoral basis. Regarding the latter, recent work has shown that a variety of distinct pathogenic alterations in receptor tyrosine kinases (RTK) such as EGFR are frequently co-existent in a single tumor. Such findings are reminiscent of other studies documenting distinct RTK amplification events characterizing non-overlapping cellular subclones within an individual neoplasm. This sort of heterogeneity has obvious clinical implications, particularly with regard to targeted therapeutics and latent resistance mechanisms. In their recent Cancer Discovery paper, Francis et al examined how heterogeneous abnormalities in EGFR distribute within cellular subclones in glioblastoma. Using two tumor samples, they performed both bulk sequencing and single nucleus sequencing (SNS) in parallel and inferred subclonal relationships with a novel algorithm based on haplotype ratios of somatic copy number alterations. Clonal hierarchies were further established by comparing SNS and bulk sequencing data. In both tumors, multiple EGFR abnormalities were identified, including two deletion mutations—EGFR vIII and the less common EGFR vII. In either case, distinct deletion events were found to segregate to non-overlapping cellular subclones. A variety of additional molecular alterations were also identified in subclonal populations, including chromothripsis—a complex chromosomal rearrangement thought to occur by way of chromosome shattering—and a unique translocation event that likely results in enhanced TERT expression. Finally, functional analyses demonstrated that EGFR vII was both oncogenic and responsive to a variety of EGFR inhibitors. Taken together, these findings definitively demonstrate that distinct genetic profiles evolve in non-overlapping cell populations within glioblastoma. As such, this study provides compelling support for the notion that that intratumoral heterogeneity involving actionable molecular targets must be addressed in the design of effective therapeutic regimens moving forward. Medulloblastoma, now known as a broad category of tumor types, is thought to exist as 4 major molecular subtypes based on several types of molecular criteria. As such, it stands to reason that each subtype requires independent investigation and characterization. Accordingly, this paper by Vanner et al addresses the biology of progenitor cells in sonic hedgehog (SHH) subtype medulloblastoma. The authors focus on the stem cell marker Sox2 and first show in a mouse model of SHH-driven medulloblastoma that Sox2 was present in only a small proportion of the tumor population. Using an elegant set of thymidine analog label-chase experiments, they show that Sox2-positive cells were relatively quiescent compared with the Sox2-negative tumor bulk. The progeny of Sox2-expressing cells are shown to be rapidly cycling doublecortin+ progenitors and post-mitotic progeny expressing NeuN-positive cells, to comprise the bulk of the tumor. Slowly growing cells were preferentially Sox2 positive compared to fast-growing cells. Next, they show that cells with tumor propagating potential were enriched for Sox2 expression. Lineage tracing in Ptc-deficient mice prone to medulloblastoma showed that over time, a substantial proportion of the tumor cell were derived from Sox2-expressing cells. Clinically relevant therapeutic regimens (Ara-C, an anti-mitotic, and vismodegib, an inhibitor of SHH signaling) were shown to decrease overall tumor proliferation but were relatively ineffective against the tumor propagating subpopulation, since residual tumor cells were enriched for Sox2-positivity. A comparison of gene expression profiles of Sox2-positive versus -negative cells revealed genes indicative of SHH signaling as common to both, but genes known to be upregulated in multiple quiescent stem cell populations, including neural stem cells, were highly enriched in Sox2-expressing cells. Extending these data to the clinic, a Sox2-gene signature was correlated with poor prognosis in patients with medulloblastoma. With respect to potential therapeutics, the authors utilize the NCI oncology drug set and find that aureolic acids, including dactinomycin and mithramycin (MM), were potentially effective against cell self-renewal in SHH-driven medulloblastoma cells. Importantly, MM, known to cross the blood-brain barrier, was effective in nanomolar concentrations. Overall, this study outlines the role of tumor propagating cells in a major medulloblastoma subtype (SHH-driven medulloblastoma) and identifies a marker that defines a pool of cells resistant to current therapies that can cause clinical relapse. Future treatment strategies could be considered that control both the tumor bulk as well as the minority Sox2-positive cells that can result in tumor regrowth.
Source: Neuro-Oncology
Authors: Whitney Pope, Riccardo Soffietti, Jason Huse, Kenneth Aldape
Language: English
Subfield: Genetics (clinical)
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W4251621882