Johan de Rooi

Intrinsic Gene Expression Profiles of Gliomas Are a Better Predictor of Survival than Histology

Gliomas are the most common primary brain tumors with heterogeneous morphology and variable prognosis. Treatment decisions in patients rely mainly on histologic classification and clinical parameters. However, differences between histologic subclasses and grades are subtle, and classifying gliomas is subject to a large interobserver variability. To improve current classification standards, we have performed gene expression profiling on a large cohort of glioma samples of all histologic subtypes and grades. We identified seven distinct molecular subgroups that correlate with survival. These include two favorable prognostic subgroups (median survival, >4.7 years), two with intermediate prognosis (median survival, 1-4 years), two with poor prognosis (median survival, <1 year), and one control group. The intrinsic molecular subtypes of glioma are different from histologic subgroups and correlate better to patient survival. The prognostic value of molecular subgroups was validated on five independent sample cohorts (The Cancer Genome Atlas, Repository for Molecular Brain Neoplasia Data, GSE12907, GSE4271, and Li and colleagues). The power of intrinsic subtyping is shown by its ability to identify a subset of prognostically favorable tumors within an external data set that contains only histologically confirmed glioblastomas (GBM). Specific genetic changes (epidermal growth factor receptor amplification, IDH1 mutation, and 1p/19q loss of heterozygosity) segregate in distinct molecular subgroups. We identified a subgroup with molecular features associated with secondary GBM, suggesting that different genetic changes drive gene expression profiles. Finally, we assessed response to treatment in molecular subgroups. Our data provide compelling evidence that expression profiling is a more accurate and objective method to classify gliomas than histologic classification. Molecular classification therefore may aid diagnosis and can guide clinical decision making.

Integrated Transcript and Genome Analyses Reveal NKX2-1 and MEF2C as Potential Oncogenes in T Cell Acute Lymphoblastic Leukemia

To identify oncogenic pathways in T cell acute lymphoblastic leukemia (T-ALL), we combined expression profiling of 117 pediatric patient samples and detailed molecular-cytogenetic analyses including the Chromosome Conformation Capture on Chip (4C) method. Two T-ALL subtypes were identified that lacked rearrangements of known oncogenes. One subtype associated with cortical arrest, expression of cell cycle genes, and ectopic NKX2-1 or NKX2-2 expression for which rearrangements were identified. The second subtype associated with immature T cell development and high expression of the MEF2C transcription factor as consequence of rearrangements of MEF2C, transcription factors that target MEF2C, or MEF2C-associated cofactors. We propose NKX2-1, NKX2-2, and MEF2C as T-ALL oncogenes that are activated by various rearrangements.

Intrinsic Molecular Subtypes of Glioma Are Prognostic and Predict Benefit From Adjuvant Procarbazine, Lomustine, and Vincristine Chemotherapy in Combination With Other Prognostic Factors in Anaplastic Oligodendroglial Brain Tumors: A Report From EORTC Study 26951

PURPOSE Intrinsic glioma subtypes (IGSs) are molecularly similar tumors that can be identified based on unsupervised gene expression analysis. Here, we have evaluated the clinical relevance of these subtypes within European Organisation for Research and Treatment of Cancer (EORTC) 26951, a randomized phase III clinical trial investigating adjuvant procarbazine, lomustine, and vincristine (PCV) chemotherapy in anaplastic oligodendroglial tumors. Our study includes gene expression profiles of formalin-fixed, paraffin-embedded (FFPE) clinical trial samples. PATIENTS AND METHODS Gene expression profiling was performed in 140 samples, 47 fresh frozen samples and 93 FFPE samples, on HU133_Plus_2.0 and HuEx_1.0_st arrays, respectively. RESULTS All previously identified six IGSs are present in EORTC 26951. This confirms that different molecular subtypes are present within a well-defined histologic subtype. Intrinsic subtypes are highly prognostic for overall survival (OS) and progression-free survival (PFS). They are prognostic for PFS independent of clinical (age, performance status, and tumor location), molecular (1p/19q loss of heterozygosity [LOH], IDH1 mutation, and MGMT methylation), and histologic parameters. Combining known molecular (1p/19q LOH, IDH1) prognostic parameters with intrinsic subtypes improves outcome prediction (proportion of explained variation, 30% v 23% for each individual group of factors). Specific genetic changes (IDH1, 1p/19q LOH, and EGFR amplification) segregate into different subtypes. We identified one subtype, IGS-9 (characterized by a high percentage of 1p/19q LOH and IDH1 mutations), that especially benefits from PCV chemotherapy. Median OS in this subtype was 5.5 years after radiotherapy (RT) alone versus 12.8 years after RT/PCV (P = .0349; hazard ratio, 2.18; 95% CI, 1.06 to 4.50). CONCLUSION Intrinsic subtypes are highly prognostic in EORTC 26951 and improve outcome prediction when combined with other prognostic factors. Tumors assigned to IGS-9 benefit from adjuvant PCV.

MGMT-STP27 methylation status as predictive marker for response to PCV in anaplastic oligodendrogliomas and oligoastrocytomas. A report from EORTC study 26951

Purpose: The long-term follow-up results from the EORTC-26951 trial showed that the addition of procarbazine, CCNU, and vincristine (PCV) after radiotherapy increases survival in anaplastic oligodendrogliomas/oligoastrocytomas (AOD/AOA). However, some patients appeared to benefit more from PCV treatment than others. Experimental Design: We conducted genome-wide methylation profiling of 115 samples included in the EORTC-26951 trial and extracted the CpG island hypermethylated phenotype (CIMP) and MGMT promoter methylation (MGMT-STP27) status. Results: We first show that methylation profiling can be conducted on archival tissues with a performance that is similar to snap-frozen tissue samples. We then conducted methylation profiling on EORTC-26951 clinical trial samples. Univariate analysis indicated that CIMP+ or MGMT-STP27 methylated tumors had an improved survival compared with CIMP− and/or MGMT-STP27 unmethylated tumors [median overall survival (OS), 1.05 vs. 6.46 years and 1.06 vs. 3.8 years, both P < 0.0001 for CIMP and MGMT-STP27 status, respectively]. Multivariable analysis indicates that CIMP and MGMT-STP27 are significant prognostic factors for survival in presence of age, sex, performance score, and review diagnosis in the model. CIMP+ and MGMT-STP27 methylated tumors showed a clear benefit from adjuvant PCV chemotherapy: the median OS of CIMP+ samples in the RT and RT-PCV arms was 3.27 and 9.51 years, respectively (P = 0.0033); for MGMT-STP27 methylated samples, it was 1.98 and 8.65 years. There was no such benefit for CIMP- or for MGMT-STP27 unmethylated tumors. MGMT-STP27 status remained significant in an interaction test (P = 0.003). Statistical analysis of microarray (SAM) identified 259 novel CpGs associated with treatment response. Conclusions: MGMT-STP27 may be used to guide treatment decisions in this tumor type. Clin Cancer Res; 19(19); 5513–22. ©2013 AACR.

Immature MEF2C-dysregulated T-cell leukemia patients have an early T-cell precursor acute lymphoblastic leukemia gene signature and typically have non-rearranged T-cell receptors

Three distinct immature T-cell acute lymphoblastic leukemia entities have been described including cases that express an early T-cell precursor immunophenotype or expression profile, immature MEF2C-dysregulated T-cell acute lymphoblastic leukemia cluster cases based on gene expression analysis (immature cluster) and cases that retain non-rearranged TRG@ loci. Early T-cell precursor acute lymphoblastic leukemia cases exclusively overlap with immature cluster samples based on the expression of early T-cell precursor acute lymphoblastic leukemia signature genes, indicating that both are featuring a single disease entity. Patients lacking TRG@ rearrangements represent only 40% of immature cluster cases, but no further evidence was found to suggest that cases with absence of bi-allelic TRG@ deletions reflect a distinct and even more immature disease entity. Immature cluster/early T-cell precursor acute lymphoblastic leukemia cases are strongly enriched for genes expressed in hematopoietic stem cells as well as genes expressed in normal early thymocyte progenitor or double negative-2A T-cell subsets. Identification of early T-cell precursor acute lymphoblastic leukemia cases solely by defined immunophenotypic criteria strongly underestimates the number of cases that have a corresponding gene signature. However, early T-cell precursor acute lymphoblastic leukemia samples correlate best with a CD1 negative, CD4 and CD8 double negative immunophenotype with expression of CD34 and/or myeloid markers CD13 or CD33. Unlike various other studies, immature cluster/early T-cell precursor acute lymphoblastic leukemia patients treated on the COALL-97 protocol did not have an overall inferior outcome, and demonstrated equal sensitivity levels to most conventional therapeutic drugs compared to other pediatric T-cell acute lymphoblastic leukemia patients.

Deconvolution of pulse trains with the L0 penalty

The output of many instruments can be modeled as a convolution of an impulse response and a series of sharp spikes. Deconvolution considers the inverse problem: estimate the input spike train from an observed (noisy) output signal. We approach this task as a linear inverse problem, solved using penalized regression. We propose the use of an L(0) penalty and compare it with the more common L(2) and L(1) penalties. In all cases a simple and iterative weighted regression procedure can be used. The model is extended with a smooth component to handle drifting baselines. Application to three different data sets shows excellent results.

Molecular subtypes of glioma identified by genome‐wide methylation profiling

Recent studies have indicated a prognostic role for genome‐wide methylation in gliomas: Tumors that show an overall increase in DNA methylation at CpG sites (CIMP+; CpG island methylator phenotype) have a more favorable prognosis than CIMP− gliomas. Here, we have determined whether methylation profiling can identify more and clinically relevant molecular subtypes of glioma by performing genome‐wide methylation profiling on 138 glial brain tumors of all histological diagnosis. Hopach (Hierarchical ordered partitioning and collapsing hybrid) clustering using the 1,000 most variable CpGs identified three distinct glioma subtypes (C+1p19q, C+wt, and C−) and one adult brain subtype. All “C+1p19q” and “C+wt” tumors were CIMP+ whereas most (50/54) “C−” tumors were CIMP−. The C− subtype gliomas contained many glioblastomas and all pilocytic astrocytomas. 1p19q LOH was frequent in the C+1p19q subtype. Other genetic changes (IDH1 mutation and EGFR amplification) and gene‐expression based molecular subtypes also segregated in distinct methylation subtypes, demonstrating that these subtypes are also genetically distinct. Each subtype was associated with its own prognosis: median survival for C−, C+1p19q, and C+wt tumors was 1.18, 5.00, and 2.62 years, respectively. The prognostic value of these methylation subtypes was validated on an external dataset from the TCGA. Analysis of recurrences of 14 primary tumors samples indicates that shifts between some C+wt and C+1p/19q tumors can occur between the primary and recurrent tumor, but CIMP status remained stable. Our data demonstrate that methylation profiling identifies at least three prognostically relevant subtypes of glioma that can aid diagnosis and potentially guide treatment for patients.

Sparse deconvolution in one and two dimensions: applications in endocrinology and single-molecule fluorescence imaging.

Deconvolution of noisy signals is an important task in analytical chemistry, examples being spectral deconvolution or deconvolution in microscopy. When the number of spectral peaks or single emitters in imaging is limited, the solution of the deconvolution is required to be sparse, and desirable results are obtained using a penalized estimation techniques. We impose sparseness by using penalized regression with a penalty based on the L0-norm, as discussed in earlier work. Several extensions to this approach are presented. Results are demonstrated on pulse identification in endocrine data where the aim is to model the secretion pattern as a sparse series of spikes. An application in single-molecule fluorescence imaging demonstrates the algorithm when applied to two-dimensional data.

Mixture models for two-dimensional baseline correction, applied to artifact elimination in time-resolved spectroscopy.

Baseline correction and artifact removal are important pre-processing steps in analytical chemistry. We propose a correction algorithm using a mixture model in combination with penalized regression. The model is an extension of a method recently introduced for baseline estimation in the case of one-dimensional data. The data are modeled as a smooth surface using tensor product P-splines. The weights of the P-splines regression model are computed from a mixture model where a datapoint is either allocated to the noise around the baseline, or to the artifact component. The method is broadly applicable for anisotropic smoothing of two-way data such as two-dimensional gel electrophoresis and two-dimensional chromatography data. We focus here on the application of the approach in femtosecond time-resolved spectroscopy, to eliminate strong artifact signals from the solvent.

Sparse deconvolution of high-density super-resolution images

In wide-field super-resolution microscopy, investigating the nanoscale structure of cellular processes, and resolving fast dynamics and morphological changes in cells requires algorithms capable of working with a high-density of emissive fluorophores. Current deconvolution algorithms estimate fluorophore density by using representations of the signal that promote sparsity of the super-resolution images via an L1-norm penalty. This penalty imposes a restriction on the sum of absolute values of the estimates of emitter brightness. By implementing an L0-norm penalty – on the number of fluorophores rather than on their overall brightness – we present a penalized regression approach that can work at high-density and allows fast super-resolution imaging. We validated our approach on simulated images with densities up to 15 emitters per μm-2 and investigated total internal reflection fluorescence (TIRF) data of mitochondria in a HEK293-T cell labeled with DAKAP-Dronpa. We demonstrated super-resolution imaging of the dynamics with a resolution down to 55 nm and a 0.5 s time sampling.