Tumor-infiltrating lymphocytes (TILs) have prognostic significance in many cancers, yet their roles in glioblastoma (GBM) have not been fully defined. We hypothesized TILs in GBM are associated with molecular alterations, histologies and survival.
We used data from The Cancer Genome Atlas (TCGA) to investigate molecular, histologic and clinical correlates of TILs in GBMs. Lymphocytes were categorized as absent, present or abundant in histopathologic images from 171 TCGA GBMs. Associations were examined between lymphocytes and histologic features, mutations, copy number alterations, CpG island methylator phenotype, transcriptional class and survival. We validated histologic findings using CD3G gene expression.
We found a positive correlation between TILs and GBMs with gemistocytes, sarcomatous cells, epithelioid cells and giant cells. Lymphocytes were enriched in the mesenchymal transcriptional class and strongly associated with mutations in NF1 and RB1. These mutations are frequent in the mesenchymal class and characteristic of gemistocytic, sarcomatous, epithelioid and giant cell histologies. Conversely, TILs were rare in GBMs with small cells and oligodendroglioma components. Lymphocytes were depleted in the classical transcriptional class and in EGFR-amplified and homozygous PTEN-deleted GBMs. These alterations are characteristic of GBMs with small cells and GBMs of the classical transcriptional class. No association with survival was demonstrated.
TILs were enriched in GBMs of the mesenchymal class, strongly associated with mutations in NF1 and RB1 and typical of histologies characterized by these mutations. Conversely, TILs were depleted in the classical class, EGFR-amplified and homozygous PTEN-deleted tumors and rare in histologies characterized by these alterations.
Glioblastoma; Lymphocytes; The Cancer Genome Atlas; Transcriptional class; Mesenchymal; Classical; Neurofibromatosis 1 (NF1); Retinoblastoma 1 (RB1); Epidermal growth factor receptor (EGFR); Tumor protein 53 (TP53); Phosphatase and tensin homolog (PTEN)
BACKGROUND: The mechanisms driving glioblastoma (GBM) relapse remain elusive. To investigate evolutionary patterns in recurrence and therapy-resistance of GBM, we analyzed the genomic profiles from 252 primary GBM and 60 biopsy samples taken from 23 pairs of pre- and post-treatment GBM tumors. METHODS: We integrated mutation allele fraction, DNA copy number and genotype information to determine the cellular frequencies of all mutations and found that 70.5% of mutations were classified as clonal and 30.5% as subclonal. We validated our classification approach through multi-sector sequencing of 13 GBM tumors. RESULTS: Separating patients into discrete age groups by an interval of 10 years, we observed a significant linear correlation between clonal mutations and age (P < 3.69x10−7), but no correlation between age and subclonal mutations (P = 0.62). This result suggested that clonal mutations predominantly accumulated over the life span of the cell population that gave rise to the cell of origin before neoplastic onset, whereas_ENREF_22 subclonal mutations were acquired during tumorigenesis and may reflect intratumor heterogeneity. Interestingly, mutation or aneuploidy of p53 pathway members strongly correlated with increased fraction of subclonal mutations (P = 6 × 10−5, Wilcoxon rank test), unrelated to patient age. To further evaluate the association between TP53 mutation status, intratumoral heterogeneity and clonal evolution, we analyzed exome sequencing and DNA copy number data from 23 pairs of first and recurrent GBM tumors. We found that TP53 mutant GBM showed a significant increase in subclonal mutation frequency relative to the matched primary tumor (P = 0.0015). The clonal mutation frequency was unaffected (P = 0.23). These data suggested that TP53 mutant GBM became increasingly clonally complex at time of recurrence, whereas TP53 wildtype GBM showed a reduced level of intratumoral heterogeneity. To comprehensively investigate tumor recurrence we performed whole genome sequencing of seven primary-recurrent tumor pairs. We found that major clones in the tumor recurrence resembled major clones in the matching primary, and were both derived from an ancestral cell population. The same applied to minor clones, suggesting that multiple types of ancestral cells existed. CONCLUSIONS: We showed that mutations in the p53 pathway resulted in increased intratumoral heterogeneity at time of diagnosis and after tumor recurrence. Disease relapse was fueled by ancestral cell populations. Further research is needed to determine the impact of intratumoral heterogeneity on treatment response. Our study provides a molecular snapshot of tumor evolution in glioblastoma. SECONDARY CATEGORY: Neuropathology and Tumor Biomarkers.
BACKGROUND: Lower grade gliomas (LGG) are infiltrative brain tumors that include astrocytomas, oligodendrogliomas and oligoastrocytomas, grades II and III. LGGs almost always progress, typically to glioblastoma (GBM, grade IV), and are uniformly fatal. The Cancer Genome Atlas (TCGA) is conducting a comprehensive molecular analysis of LGG, incorporating genetic and genomic alterations, DNA methylation profiles, and RNA and proteomic signatures. The rich, integrated clinical data of TCGA provides an outstanding platform to uncover biomarkers of therapy response and outcome. METHODS: We analyzed 289 lower grade gliomas for exome sequence to uncover somatic mutations; DNA copy number alterations; RNA sequencing for expression and gene fusion; DNA methlyation; microRNA expression; and protein level and phosphorylation. Supervised and unsupervised clustering was performed to segregate LGGs into robust molecular categories. RESULTS: Among 289 LGGs, exome sequencing identified 19 significantly mutated genes, including IDH1, TP53, ATRX, CIC, FUBP1, NOTCH1, PIK3CA, NF1, PIK3R1, ARID1A, PTEN, SMARCA4 and EGFR. Frequent arm-level deletions were 1p, 19q, 13q, 9p, 10q, 10p, 4q and 4p; arm-level gains included 7q, 7p, 19p and 11q. There were 15 significant genomic amplifications, among which were 7p11.2 (EGFR), 12q14.1 (CDK4), 1q32.1 (MDM4), 8q24.21 (MYC), 12p13.32 (CCND2) and 4q12 (PDGFR). Among 28 deletions were 9p21.3 (CDKN2A, CDKN2B), 19q13.42 (TFPT, ZNF331), 2q37.3, 10q26.2 (FGFR2, DUX4) and 14q24.3 (TSHR, GPHN).Global analysis of DNA methylation identified 5 stable clusters, with one showing substantial hypomethylation. Gene expression profiling identified 4 stable clusters. On integrative analysis, IDH1/2 wt LGGs had features of “pre-GBM”: they formed distinct hypomethylation and gene expression clusters; had molecular alterations typical of GBM, such as EGFR amplifications, PTEN mutations, CDKN2A loss and RTK gene fusions; were mostly grade III astrocytomas; and had short survivals. IDH1/2 mutant LGGs were hypermethylated. One subgroup was enriched for 1p/19q co-deletion, CIC and FUBP mutations, and oligodendrogliomas. Another subset was enriched for TP53 and ATRX mutations and astrocytomas. These groups clustered separately on gene expression analysis. A fourth gene expression cluster had elements of these two IDH1/2 mutant subgroups and the longest survival. A cluster-of-cluster analysis of mRNA, copy number, miRNA and methylation analyses reinforced the clear separation of three molecular classes of LGG based on the status of IDH1/2 and 1p/19q. CONCLUSIONS: These data illustrate a potential stratification of LGGs that emphasizes molecular characteristics. IDH wt LGG have molecular alterations and clinical behavior similar to GBMs, whereas IDH mutant LGGs segregate based on 1p/19q status. SECONDARY CATEGORY: Tumor Biology.
To evaluate the cellular nature of and diagnostic terminology used in connection with acquired retinal “vasoproliferative tumors.”
Retrospective clinicopathologic study.
Clinical records and microscopic slides of 4 enucleated globes were reviewed. Special stains and immunohistochemical probes for CD31, CD34, p53, glial fibrillary acidic protein (GFAP), CD163, and Ki67 (cell replication) were employed; ultrastructural and fluorescence in situ hybridization (FISH) analyses were performed.
Tumors were located inferotemporally in middle-aged patients. They were uniformly composed of compacted elongated, GFAP-positive spindle cells (due to intermediate filaments identified ultrastructurally) with a Ki67 index of less than 1%. Rosenthal fibers and eosinophilic granular bodies were observed. Hyalinized periodic acid–Schiff-positive vessels were widely separated. CD31 and CD34 revealed a sparse microvasculature. Tumor-associated exudate spread predominantly subretinally. The retinal pigment epithelium had undergone extensive placoid fibrous metaplasia with focal ossification. P53 upregulation, BRAF-KIAA gene rearrangement, and IDH1 R132H mutation typically associated with low-grade astrocytic neoplasms were absent.
Retinal “vasoproliferative” tumors have been mischaracterized, because they actually display a paucity of microvessels. Proliferating fibrous astrocytes with a very low proliferation index predominate, without immunohistochemical or genetic evidence favoring a neoplasm. Subretinal exudate appeared capable of provoking widespread fibrous metaplasia of the pigment epithelium that was mainly responsible for secondary retinal damage. The term “reactive retinal astrocytic tumor” is proposed as more appropriate for this entity. In carefully selected progressive lesions, consideration should be given to earlier surgical intervention before extensive subretinal exudate accumulates and pigment epithelial proliferation with fibrous metaplasia ensues.
Open surgery effectively treats mesial temporal lobe epilepsy (MTLE), but carries risks of neurocognitive deficits, which may be reduced with minimally invasive alternatives.
To describe technical and clinical outcomes of stereotactic laser amygdalohippocampotomy (SLAH) with real-time magnetic resonance thermal imaging (MRTI) guidance.
Under general anesthesia and utilizing standard stereotactic methods, 13 adult patients with intractable MTLE (with and without mesial temporal sclerosis, MTS) prospectively underwent insertion of a saline-cooled fiber-optic laser applicator into amygdalohippocampal structures from an occipital trajectory. Computer-controlled laser ablation was performed during continuous MRTI followed by confirmatory contrast-enhanced anatomic imaging and volumetric reconstruction. Clinical outcomes were determined from seizure diaries.
A mean 60% volume of the amygdalohippocampal complex was ablated in 13 patients (9 with MTS) undergoing 15 procedures. Median hospitalization was one day. With follow-up ranging from 5-26 (median 14) months, 77% (10/13) of patients achieved meaningful seizure reduction, of which 54% (7/13) were free of disabling seizures. Of patients with preoperative MTS, 67% (6/9) achieved seizure freedom. All recurrences were observed by<6 months. Variances in ablation volume and length did not account for individual clinical outcomes. Whereas no complications of laser therapy itself were observed, one significant complication, a visual field defect, resulted from deviated insertion of a stereotactic aligning rod, which was corrected prior to ablation.
Real-time MR-guided SLAH is a technically novel, safe, and effective alternative to open surgery. Further evaluation with larger cohorts over time is warranted.
Epilepsy; laser therapy; magnetic resonance imaging; minimally invasive surgical procedures; stereotactic techniques; temporal lobe; thermometry
How necrotic areas develop in tumors is incompletely understood but can impact progression. Recent findings suggest that formation of vascular microthrombi contributes to tumor necrosis, prompting investigation of coagulation cascades. Here we report that loss of tumor suppressor p14ARF can contribute to activating the clotting cascade in glioblastoma (GBM). p14ARF transcriptionally upregulated TFPI2, a Kunitz-type serine protease in the tissue factor pathway that inhibits the initiation of thrombosis reactions. p14ARF activation in tumor cells delayed their ability to activate plasma clotting. Mechanistically, p14ARF activated the TFPI2 promoter in a p53-independent manner that relied upon c-JUN, SP1 and JNK activity. Taken together, our results identify the critical signaling pathways activated by p14ARF to prevent vascular microthrombosis triggered by glioma cells. Stimulation of this pathway might be used as a therapeutic strategy to reduce aggressive phenotypes associated with necrotic tumors including glioblastoma.
Diffuse gliomas consist of both low- and high-grade varieties, each with distinct morphological and biological features. The often extended periods of relative indolence exhibited by low-grade gliomas (LGG; WHO grade II) differ sharply from the aggressive, rapidly fatal clinical course of primary glioblastoma (GBM; WHO grade IV). Nevertheless, until recently, the molecular foundations underlying this stark biological contrast between glioma variants remained largely unknown. The discoveries of distinctive and highly recurrent genomic and epigenomic abnormalities in LGG have both informed a more accurate classification scheme and pointed to viable avenues for therapeutic development. As such, the field of neuro-oncology now seems poised to capitalize on these gains to achieve significant benefit for LGG patients. This report will briefly recount the proceedings of a workshop held in January 2013 and hosted by Accelerate Brain Cancer Cure (ABC2) on the subject of LGG. While much of the meeting covered recent insights into LGG biology, its focus remained on how best to advance the clinical management, whether by improved preclinical modeling, more effective targeted therapeutics and clinical trial design, or innovative imaging technology.
clinical trials; genomics; low-grade glioma; personalized medicine
Breast cancer aberrantly expresses tissue factor (TF) in cancer tissues and cancer vascular endothelial cells (VECs). TF plays a central role in cancer angiogenesis, growth, and metastasis and, as such, is a target for therapy and drug delivery. TF is the cognate receptor of factor VIIa (fVIIa). We have coupled PTX (paclitaxel, also named Taxol) with a tripeptide, phenylalanine-phenylalanine-arginine chloromethyl ketone (FFRck) and conjugated it with fVIIa. The key aim of the work is to evaluate the antiangiogenic effects of PTX-FFRck-fVIIa against a PTX-resistant breast cancer cell line. Matrigel mixed with VEGF and MDA-231 was injected subcutaneously into the flank of athymic nude mice. Animals were treated by tail vein injection of the PTX-FFRck-fVIIa conjugate, unconjugated PTX, or PBS. The PTX-FFRck-fVIIa conjugate significantly reduces microvessel density in matrigel (p < 0.01–0.05) compared to PBS and unconjugated PTX. The breast cancer lung metastasis model in athymic nude mice was developed by intravenous injection of MDA-231 cells expressing luciferase. Animals were similarly treated intravenously with the PTX-FFRck-fVIIa conjugate or PBS. The conjugate significantly inhibits lung metastasis as compared to the control, highlighting its potential to antagonize angiogenesis in metastatic carcinoma. In conclusion, PTX conjugated to fVIIa is a promising therapeutic approach for improving selective drug delivery and inhibiting angiogenesis.
The integration and visualization of multimodal datasets is a common challenge in biomedical informatics. Several recent studies of The Cancer Genome Atlas (TCGA) data have illustrated important relationships between morphology observed in whole-slide images, outcome, and genetic events. The pairing of genomics and rich clinical descriptions with whole-slide imaging provided by TCGA presents a unique opportunity to perform these correlative studies. However, better tools are needed to integrate the vast and disparate data types.
To build an integrated web-based platform supporting whole-slide pathology image visualization and data integration.
Materials and methods
All images and genomic data were directly obtained from the TCGA and National Cancer Institute (NCI) websites.
The Cancer Digital Slide Archive (CDSA) produced is accessible to the public (http://cancer.digitalslidearchive.net) and currently hosts more than 20 000 whole-slide images from 22 cancer types.
The capabilities of CDSA are demonstrated using TCGA datasets to integrate pathology imaging with associated clinical, genomic and MRI measurements in glioblastomas and can be extended to other tumor types. CDSA also allows URL-based sharing of whole-slide images, and has preliminary support for directly sharing regions of interest and other annotations. Images can also be selected on the basis of other metadata, such as mutational profile, patient age, and other relevant characteristics.
With the increasing availability of whole-slide scanners, analysis of digitized pathology images will become increasingly important in linking morphologic observations with genomic and clinical endpoints.
Digital Pathology; Computer-Assisted Image Analysis; Cancer; Cell Morphology; Image Cytometry; TCGA
The first US Food and Drug Administration–approved clinical trial to treat amyotrophic lateral sclerosis (ALS) with neural stem cell–based therapy is in progress. The goal of the current study was to identify and assess the survival of human spinal cord–derived neural stem cells (HSSCs) transplanted into the spinal cord in patients with ALS.
Spinal cords transplanted with HSSCs were examined from six autopsy cases. Homogenized tissues were interrogated for the presence of donor versus recipient DNA using real-time PCR methods (qPCR). Fluorescence in situ hybridization (FISH) was performed using DNA probes for XY chromosomes to identify male donor HSSCs in one female case, and immunohistochemistry (IHC) was used to characterize the identified donor cells.
Genomic DNA from donor HSSCs was identified in all cases, comprising 0.67–5.4% of total tissue DNA in patients surviving 196 to 921 days after transplantation. In the one female patient a “nest” of cells identified on H&E staining were XY-positive by FISH, confirming donor origin. A subset of XY-positive cells labeled for the neuronal marker NeuN and stem cell marker SOX2.
This is the first study to identify human neural stem cells transplanted into a human spinal cord. Transplanted HSSCs survived up to 2.5 years posttransplant. Some cells differentiated into neurons, while others maintained their stem cell phenotype. This work is a proof of concept of the survival and differentiation of human stems cell transplanted into the spinal cord of ALS patients.
Pathology is a medical subspecialty that practices the diagnosis of disease. Microscopic examination of tissue reveals information enabling the pathologist to render accurate diagnoses and to guide therapy. The basic process by which anatomic pathologists render diagnoses has remained relatively unchanged over the last century, yet advances in information technology now offer significant opportunities in image-based diagnostic and research applications. Pathology has lagged behind other healthcare practices such as radiology where digital adoption is widespread. As devices that generate whole slide images become more practical and affordable, practices will increasingly adopt this technology and eventually produce an explosion of data that will quickly eclipse the already vast quantities of radiology imaging data. These advances are accompanied by significant challenges for data management and storage, but they also introduce new opportunities to improve patient care by streamlining and standardizing diagnostic approaches and uncovering disease mechanisms. Computer-based image analysis is already available in commercial diagnostic systems, but further advances in image analysis algorithms are warranted in order to fully realize the benefits of digital pathology in medical discovery and patient care. In coming decades, pathology image analysis will extend beyond the streamlining of diagnostic workflows and minimizing interobserver variability and will begin to provide diagnostic assistance, identify therapeutic targets, and predict patient outcomes and therapeutic responses.
Biomedical imaging; biomedical informatics; digital pathology; image analysis; virtual microscopy
High-grade astrocytomas (HGAs), corresponding to WHO grades III (AA) and IV (GBM), are biologically aggressive and their molecular classification is increasingly relevant to clinical management. PDGFRA amplification is common in HGAs, although its prognostic significance remains unclear. Using fluorescence in situ hybridization (FISH), the most sensitive technique for detecting PDGFRA copy number gains, we determined PDGFRA amplification status in 123 pediatric and 263 adult HGAs. A range of PDGFRA FISH patterns were identified and cases were scored as non-amplified (normal and polysomy) or amplified (low-level and high-level). PDGFRA amplification was frequent in pediatric (29.3%) and adult (20.9%) tumors. Amplification was not prognostic in pediatric HGAs. In adult tumors diagnosed initially as GBM, the presence of combined PDGFRA amplification and IDH1R132H mutation was a significant independent prognostic factor (p=0.01). In HGAs, PDGFRA amplification is common and can manifest as high-level and focal or low-level amplifications. Our data indicate that the latter is more prevalent than previously reported with copy number averaging techniques. To our knowledge, this is the largest survey of PDGFRA status in adult and pediatric HGAs and suggests PDGFRA amplification increases with grade and is associated with a less favorable prognosis in IDH1 mutant de novo GBMs.
PDGFRA; astrocytoma; FISH; IDH1; isocitrate dehydrogenase 1; prognosis
Hypoxia and necrosis are fundamental features of glioblastoma (GBM) and their emergence is critical for the rapid biological progression of this fatal tumor; yet, underlying mechanisms are poorly understood. We have suggested that vaso-occlusion following intravascular thrombosis could initiate or propagate hypoxia and necrosis in GBM. Tissue factor (TF), the main cellular initiator of coagulation, is overexpressed in GBMs and likely favors a thrombotic microenvironment. Epidermal growth factor receptor (EGFR) amplification and PTEN loss are two common genetic alterations seen in GBM but not in lower-grade astrocytomas that could be responsible for TF up-regulation. The most frequent EGFR mutation in GBM involves deletion of exons 2 to 7, resulting in the expression of a constitutively active receptor, EGFRvIII. Here, we show that overexpression of EGFR or EGFRvIII in human glioma cells causes increased basal TF expression and that stimulation of EGFR by its ligand, EGF, leads to a marked dose-dependent up-regulation of TF. In all cases, increased TF expression led to accelerated plasma coagulation in vitro. EGFR-mediated TF expression depended most strongly on activator protein-1 (AP-1) transcriptional activity and was associated with c-Jun NH2-terminal kinase (JNK) and JunD activation. Restoration of PTEN expression in PTEN-deficient GBM cells diminished EGFR-induced TF expression by inhibiting JunD/AP-1 transcriptional activity. PTEN mediated this effect by antagonizing phosphatidylinositol 3-kinase activity, which in turn attenuated both Akt and JNK activities. These mechanisms are likely at work in vivo, as EGFR expression was highly correlated with TF expression in human high-grade astrocytoma specimens.
Gliomas are the most common primary malignant brain tumors in adults with great heterogeneity in histopathology and clinical course. The intent was to evaluate the relevance of known glioblastoma (GBM) expression and methylation based subtypes to grade II and III gliomas (ie. lower grade gliomas).
Gene expression array, single nucleotide polymorphism (SNP) array and clinical data were obtained for 228 GBMs and 176 grade II/II gliomas (GII/III) from the publically available Rembrandt dataset. Two additional datasets with IDH1 mutation status were utilized as validation datasets (one publicly available dataset and one newly generated dataset from MD Anderson). Unsupervised clustering was performed and compared to gene expression subtypes assigned using the Verhaak et al 840-gene classifier. The glioma-CpG Island Methylator Phenotype (G-CIMP) was assigned using prediction models by Fine et al.
Unsupervised clustering by gene expression aligned with the Verhaak 840-gene subtype group assignments. GII/IIIs were preferentially assigned to the proneural subtype with IDH1 mutation and G-CIMP. GBMs were evenly distributed among the four subtypes. Proneural, IDH1 mutant, G-CIMP GII/III s had significantly better survival than other molecular subtypes. Only 6% of GBMs were proneural and had either IDH1 mutation or G-CIMP but these tumors had significantly better survival than other GBMs. Copy number changes in chromosomes 1p and 19q were associated with GII/IIIs, while these changes in CDKN2A, PTEN and EGFR were more commonly associated with GBMs.
GBM gene-expression and methylation based subtypes are relevant for GII/III s and associate with overall survival differences. A better understanding of the association between these subtypes and GII/IIIs could further knowledge regarding prognosis and mechanisms of glioma progression.
Hypoxia strongly up-regulates tissue factor and promotes plasma clotting by glioblastoma multiforme, but transcriptional mechanisms remain undefined. Here, we investigated the potential roles of early growth response gene-1 (Egr-1), Sp1, nuclear factor-κB (NF-κB), activator protein-1 (AP-1), and hypoxia-inducible factor-1 (HIF-1) in the hypoxic regulation of tissue factor by glioblastoma multiforme cells in vitro. Hypoxia (1% O2) strongly induced Egr-1 mRNA within 1 hour and led to nuclear localization of Egr-1 protein. Using luciferase reporter plasmids in glioma cells, we found that hypoxia dramatically increased luciferase activity in cells with constructs containing Egr-1-binding sites but not in cells with constructs containing AP-1- or NF-κB-binding sites. Electrophoretic mobility shift assays revealed hypoxia-induced Egr-1, but not Sp1, binding to oligonucleotides containing the Egr-1/Sp1 motif of tissue factor gene promoter. Using an expression vector containing the minimal tissue factor promoter (−111 to +14 bp) and small interfering RNA (siRNA) directed at Egr-1 and Sp1 mRNAs, we found that Egr-1 was required for maximal hypoxic induction of promoter activity. Forced overexpression of Egr-1 but not Sp1 by cDNA transfection caused up-regulation of tissue factor in glioma cells under normoxia (21% O2), whereas siRNA directed at Egr-1 strongly attenuated hypoxia-induced tissue factor expression. To examine the effects of HIF-1α on tissue factor expression, we used glioma cells stably transfected with a HIF-1α siRNA expression vector and found that HIF-1α mRNA silencing did not affect tissue factor expression under hypoxia. We conclude that hypoxic up-regulation of tissue factor in glioblastoma multiforme cells depends largely on Egr-1 and is independent of HIF-1.
The presence of necrosis within a diffuse glioma is a powerful predictor of poor prognosis, yet little is known of its origins. Intravascular thrombosis is a frequent finding in glioblastoma [GBM; World Health Organization (WHO) grade IV] specimens and could potentially be involved in astrocytoma progression to GBM or represent a surrogate marker of GBM histology. We investigated whether intravascular thrombosis was more frequent or prominent in GBM than other central nervous system (CNS) malignancies and considered its prognostic significance in anaplastic astrocytoma (AA; WHO grade III), which lacks necrosis. Histologic sections were examined for thrombosis, necrosis and microvascular hyperplasia from each of 297 CNS tumors, including 103 GBMs, 46 AAs, 20 diffuse astrocytoma (DAs; WHO grade II), eight anaplastic oligodendrogliomas (AOs; WHO grade III), 20 oligodendrogliomas (ODs; WHO grade II), 49 metastatic carcinomas (METs), 31 primary central nervous system lymphomas (PCNSLs) and 20 medulloblastomas (MBs). Among newly diagnosed tumors, thrombosis was present in 92% of GBM resections, significantly greater than other types of CNS malignancies. Of tumors with thrombosis, GBMs had a higher frequency of affected vessels than AAs, DAs, AOs, ODs and MBs, but had a frequency similar to METs and PCNSLs. The sensitivity of thrombosis for the diagnosis of GBM in this set of tumors was 92% and the specificity was 91%. Intravascular thrombosis was uncommon in AAs and was only noted in stereotactic biopsies. This subset of patients had shorter survivals than those AAs without thrombosis. Thus, intravascular thrombosis is more frequent in GBM than other CNS malignancies. When present in AAs, it appears to indicate aggressive clinical behavior.
astrocytoma; brain tumor; glioblastoma; hypoxia; necrosis; thrombosis
Pathologic review of tumor morphology in histologic sections is the traditional method for cancer classification and grading, yet human review has limitations that can result in low reproducibility and inter-observer agreement. Computerized image analysis can partially overcome these shortcomings due to its capacity to quantitatively and reproducibly measure histologic structures on a large-scale. In this paper, we present an end-to-end image analysis and data integration pipeline for large-scale morphologic analysis of pathology images and demonstrate the ability to correlate phenotypic groups with molecular data and clinical outcomes. We demonstrate our method in the context of glioblastoma (GBM), with specific focus on the degree of the oligodendroglioma component. Over 200 million nuclei in digitized pathology slides from 117 GBMs in the Cancer Genome Atlas were quantitatively analyzed, followed by multiplatform correlation of nuclear features with molecular and clinical data. For each nucleus, a Nuclear Score (NS) was calculated based on the degree of oligodendroglioma appearance, using a regression model trained from the optimal feature set. Using the frequencies of neoplastic nuclei in low and high NS intervals, we were able to cluster patients into three well-separated disease groups that contained low, medium, or high Oligodendroglioma Component (OC). We showed that machine-based classification of GBMs with high oligodendroglioma component uncovered a set of tumors with strong associations with PDGFRA amplification, proneural transcriptional class, and expression of the oligodendrocyte signature genes MBP, HOXD1, PLP1, MOBP and PDGFRA. Quantitative morphologic features within the GBMs that correlated most strongly with oligodendrocyte gene expression were high nuclear circularity and low eccentricity. These findings highlight the potential of high throughput morphologic analysis to complement and inform human-based pathologic review.
Glioblastoma (GBM) is a highly aggressive primary brain tumor with a poor prognosis. Despite aggressive therapy with surgery, radiotherapy, and chemotherapy, nearly all patients succumb to disease within 2 years. Several studies have supported the presence of stem-like cells in brain tumor cultures that are CD133-positive, are capable of self-renewal, and give rise to all cell types found within the tumor, potentially perpetuating growth. CD133 is a widely accepted marker for glioma-derived cancer stem cells; however, its reliability has been questioned, creating a need for other identifiers of this biologically important subpopulation. We used a panel of 20 lectins to identify differences in glycan expression found in the glycocalyx of undifferentiated glioma-derived stem cells and differentiated cells that arise from them. Fluorescently labeled lectins that specifically recognize α-N-acetylgalactosamine (GalNAc) and α-N-acetylglucosamine (GlcNAc) differentially bound to the cell surface based on the state of cellular differentiation. GalNAc and GlcNAc were highly expressed on the surface of undifferentiated cells and showed markedly reduced expression over a 12-day duration of differentiation. Additionally, the GalNAc-recognizing lectin Dolichos biflorus agglutinin was capable of specifically selecting and sorting glioma-derived stem cell populations from an unsorted tumor stock and this subpopulation had proliferative properties similar to CD133+ cells in vitro and also had tumor-forming capability in vivo. Our preliminary results on a single cerebellar GBM suggest that GalNAc and GlcNAc are novel biomarkers for identifying glioma-derived stem cells and can be used to isolate cancer stem cells from unsorted cell populations, thereby creating new cell lines for research or clinical testing.
Survival periods vary considerably for patients with high-grade astrocytomas, and reliable prognostic markers are not currently available. We therefore investigated whether genetic losses from chromosomes 1p, 19q, 9p, or 10q were associated with survival in 89 high-grade astrocytomas using tissue microarrays (TMAs) derived from Radiation Therapy Oncology Group clinical trials. Cases included 15 anaplastic astrocytomas (AAs) and 74 glioblastomas (GBMs) selected on the basis of survival times significantly shorter or longer than the expected median. Genetic analysis was performed by TMA-fluorescence in situ hybridization (FISH) on array sections using 8 DNA probes, including those directed at 1p32, 19q13.4, 9p21 (p16/CDKN2A), and 10q (PTEN and DMBT1). Genetic status for each locus was correlated with patient survival group, and data were analyzed by using Fisher’s exact test of association (adjusted P = 0.025). Losses of chromosome 1p, either alone or in combination with 19q, were encountered in only 2 cases, both AAs. This contrasts with oligodendrogliomas, in which combined 1p and 19q losses are frequent and predictive of prolonged survival. Solitary 19q loss was noted in 3/15 AAs and in 7/70 GBMs and was more frequent in the long-term survival group (P = 0.041, AA and GBM combined). Chromosome 9p loss was seen in 5/8 AAs and 39/57 GBMs, whereas chromosome 10q loss was detected in 4/15 AAs and 48/68 GBMs. The 9p and 10q deletions were slightly more frequent in short-term survivors, though none of the comparisons achieved statistical significance. Long-term and short-term survival groups of high-grade astrocytomas appear to have dissimilar frequencies of 19q, 9p, and 10q deletions. TMA-FISH is a rapid and efficient way of evaluating genetic alterations in such tumors.
The brain tumor glioblastoma multiforme (GBM) is among the most lethal forms of human cancer. Here, we report that a small subset of GBMs (3.1%; 3 of 97 tumors examined) harbors oncogenic chromosomal translocations that fuse in-frame the tyrosine kinase coding domains of fibroblast growth factor receptor (FGFR) genes (FGFR1 or FGFR3) to the transforming acidic coiled-coil (TACC) coding domains of TACC1 or TACC3, respectively. The FGFR-TACC fusion protein displays oncogenic activity when introduced into astrocytes or stereotactically transduced in the mouse brain. The fusion protein, which localizes to mitotic spindle poles, has constitutive kinase activity and induces mitotic and chromosomal segregation defects and triggers aneuploidy. Inhibition of FGFR kinase corrects the aneuploidy, and oral administration of an FGFR inhibitor prolongs survival of mice harboring intracranial FGFR3-TACC3–initiated glioma. FGFR-TACC fusions could potentially identify a subset of GBM patients who would benefit from targeted FGFR kinase inhibition.
Algorithm evaluation provides a means to characterize variability across image analysis algorithms, validate algorithms by comparison with human annotations, combine results from multiple algorithms for performance improvement, and facilitate algorithm sensitivity studies. The sizes of images and image analysis results in pathology image analysis pose significant challenges in algorithm evaluation. We present an efficient parallel spatial database approach to model, normalize, manage, and query large volumes of analytical image result data. This provides an efficient platform for algorithm evaluation. Our experiments with a set of brain tumor images demonstrate the application, scalability, and effectiveness of the platform.
The paper describes an approach and platform for evaluation of pathology image analysis algorithms. The platform facilitates algorithm evaluation through a high-performance database built on the Pathology Analytic Imaging Standards (PAIS) data model.
(1) Develop a framework to support algorithm evaluation by modeling and managing analytical results and human annotations from pathology images; (2) Create a robust data normalization tool for converting, validating, and fixing spatial data from algorithm or human annotations; (3) Develop a set of queries to support data sampling and result comparisons; (4) Achieve high performance computation capacity via a parallel data management infrastructure, parallel data loading and spatial indexing optimizations in this infrastructure.
Materials and Methods:
We have considered two scenarios for algorithm evaluation: (1) algorithm comparison where multiple result sets from different methods are compared and consolidated; and (2) algorithm validation where algorithm results are compared with human annotations. We have developed a spatial normalization toolkit to validate and normalize spatial boundaries produced by image analysis algorithms or human annotations. The validated data were formatted based on the PAIS data model and loaded into a spatial database. To support efficient data loading, we have implemented a parallel data loading tool that takes advantage of multi-core CPUs to accelerate data injection. The spatial database manages both geometric shapes and image features or classifications, and enables spatial sampling, result comparison, and result aggregation through expressive structured query language (SQL) queries with spatial extensions. To provide scalable and efficient query support, we have employed a shared nothing parallel database architecture, which distributes data homogenously across multiple database partitions to take advantage of parallel computation power and implements spatial indexing to achieve high I/O throughput.
Our work proposes a high performance, parallel spatial database platform for algorithm validation and comparison. This platform was evaluated by storing, managing, and comparing analysis results from a set of brain tumor whole slide images. The tools we develop are open source and available to download.
Pathology image algorithm validation and comparison are essential to iterative algorithm development and refinement. One critical component is the support for queries involving spatial predicates and comparisons. In our work, we develop an efficient data model and parallel database approach to model, normalize, manage and query large volumes of analytical image result data. Our experiments demonstrate that the data partitioning strategy and the grid-based indexing result in good data distribution across database nodes and reduce I/O overhead in spatial join queries through parallel retrieval of relevant data and quick subsetting of datasets. The set of tools in the framework provide a full pipeline to normalize, load, manage and query analytical results for algorithm evaluation.
Algorithm validation; parallel database; pathology imaging; spatial database
While alterations in xenobiotic metabolism are considered causal in the development of bladder cancer (BCa), the precise mechanisms involved are poorly understood. In this study, we used high-throughput mass spectrometry to measure over 2,000 compounds in 58 clinical specimens, identifying 35 metabolites which exhibited significant changes in BCa. This metabolic signature distinguished both normal and benign bladder from BCa. Exploratory analyses of this metabolomic signature in urine showed promise in distinguishing BCa from controls, and also non-muscle from muscle-invasive BCa. Subsequent enrichment-based bioprocess mapping revealed alterations in phase I/II metabolism and suggested a possible role for DNA methylation in perturbing xenobiotic metabolism in BCa. In particular, we validated tumor-associated hypermethylation in the CYP1A1 and CYP1B1 promoters of BCa tissues by bisulfite sequence analysis and methylation-specific PCR, and also by in vitro treatment of T-24 BCa cell line with the DNA demethylating agent 5-aza-2′-deoxycytidine. Further, we showed that expression of CYP1A1 and CYP1B1 was reduced significantly in an independent cohort of BCa specimens compared to matched benign adjacent tissues. In summary, our findings identified candidate diagnostic and prognostic markers and highlighted mechanisms associated with the silencing of xenobiotic metabolism. The metabolomic signature we describe offers potential as a urinary biomarker for early detection and staging of BCa, highlighting the utility of evaluating metabolomic profiles of cancer to gain insights into bioprocesses perturbed during tumor development and progression.
Multi-modal, multi-scale data synthesis is becoming increasingly critical for successful translational biomedical research. In this paper, we present a large-scale investigative initiative on glioblastoma, a high-grade brain tumor, with complementary data types using in silico approaches. We integrate and analyze data from The Cancer Genome Atlas Project on glioblastoma that includes novel nuclear phenotypic data derived from microscopic slides, genotypic signatures described by transcriptional class and genetic alterations, and clinical outcomes defined by response to therapy and patient survival. Our preliminary results demonstrate numerous clinically and biologically significant correlations across multiple data types, revealing the power of in silico multi-modal data integration for cancer research.
Glioblastoma; multi-modal data process; in silico; cluster analysis; translational integration
To determine whether temozolomide is an active agent in the treatment of children with high-grade astrocytomas and whether survival is influenced by the expression of the O6-methylguanine-methyltransferase gene (MGMT) in these patients. In the Children's Oncology Group study ACNS0126, 107 patients with a diagnosis of anaplastic astrocytoma (AA), glioblastoma multiforme (GBM), or gliosarcoma were enrolled. All patients underwent concomitant chemoradiotherapy with temozolomide, followed by adjuvant chemotherapy with temozolomide. The outcomes were compared with those of children treated in Children's Cancer Group (CCG) study CCG-945. Formalin-fixed, paraffin-embedded tumor tissue was available in 71 cases for immunohistochemical analysis of MGMT expression. Ninety patients were deemed eligible, 31 with AA, 55 with GBM, and 4 with other eligible diagnoses. The 3-year event-free survival (EFS) and overall survival (OS) rates were 11 ± 3% and 22 ± 5%, respectively. There was no evidence that temozolomide given during radiation therapy and as adjuvant therapy resulted in improved EFS compared with that found in CCG-945 (p = 0.98). The 3-year EFS rate for AA was 13 ± 6% in ACNS0126 compared with 22 ± 5.5% in CCG-945 (p = 0.95). The 3-year EFS rate for GBM was 7 ± 4% in ACNS0126 compared with 15 ± 5% in CCG-945 (p = 0.77). The 2-year EFS rate was 17 ± 5% among patients without MGMT overexpression and 5 ± 4% among those with MGMT overexpression (p = 0.045). Temozolomide failed to improve outcome in children with high-grade astrocytomas. MGMT overexpression was adversely associated with survival.
High-grade glioma; temozolomide; pediatric brain tumors
Malignant gliomas are the most common and the most lethal primary brain tumors in adults. Among malignant gliomas, 60%–80% show loss of P14ARF tumor suppressor activity due to somatic alterations of the INK4A/ARF genetic locus. The tumor suppressor activity of P14ARF is in part a result of its ability to prevent the degradation of P53 by binding to and sequestering HDM2. However, the subsequent finding of P14ARF loss in conjunction with TP53 gene loss in some tumors suggests the protein may have other P53-independent tumor suppressor functions. Here, we report what we believe to be a novel tumor suppressor function for P14ARF as an inhibitor of tumor-induced angiogenesis. We found that P14ARF mediates antiangiogenic effects by upregulating expression of tissue inhibitor of metalloproteinase–3 (TIMP3) in a P53-independent fashion. Mechanistically, this regulation occurred at the gene transcription level and was controlled by HDM2-SP1 interplay, where P14ARF relieved a dominant negative interaction of HDM2 with SP1. P14ARF-induced expression of TIMP3 inhibited endothelial cell migration and vessel formation in response to angiogenic stimuli produced by cancer cells. The discovery of this angiogenesis regulatory pathway may provide new insights into P53-independent P14ARF tumor-suppressive mechanisms that have implications for the development of novel therapies directed at tumors and other diseases characterized by vascular pathology.