Genome-wide promoter methylation analysis in neuroblastoma identifies prognostic methylation biomarkers

Background

Accurate outcome prediction in neuroblastoma, which is necessary to enable the optimal choice of risk-related therapy, remains a challenge. To improve neuroblastoma patient stratification, this study aimed to identify prognostic tumor DNA methylation biomarkers.

Results

To identify genes silenced by promoter methylation, we first applied two independent genome-wide methylation screening methodologies to eight neuroblastoma cell lines. Specifically, we used re-expression profiling upon 5-aza-2''-deoxycytidine (DAC) treatment and massively parallel sequencing after capturing with a methyl-CpG-binding domain (MBD-seq). Putative methylation markers were selected from DAC-upregulated genes through a literature search and an upfront methylation-specific PCR on 20 primary neuroblastoma tumors, as well as through MBD- seq in combination with publicly available neuroblastoma tumor gene expression data. This yielded 43 candidate biomarkers that were subsequently tested by high-throughput methylation-specific PCR on an independent cohort of 89 primary neuroblastoma tumors that had been selected for risk classification and survival. Based on this analysis, methylation of KRT19, FAS, PRPH, CNR1, QPCT, HIST1H3C, ACSS3 and GRB10 was found to be associated with at least one of the classical risk factors, namely age, stage or MYCN status. Importantly, HIST1H3C and GNAS methylation was associated with overall and/or event-free survival.

Conclusions

This study combines two genome-wide methylation discovery methodologies and is the most extensive validation study in neuroblastoma performed thus far. We identified several novel prognostic DNA methylation markers and provide a basis for the development of a DNA methylation-based prognostic classifier in neuroblastoma.     

COG5-CDG: expanding the clinical spectrum

Background

The Conserved Oligomeric Golgi (COG) complex is involved in the retrograde trafficking of Golgi components, thereby affecting the localization of Golgi glycosyltransferases. Deficiency of a COG-subunit leads to defective protein glycosylation, and thus Congenital Disorders of Glycosylation (CDG). Mutations in subunits 1, 4, 5, 6, 7 and 8 have been associated with CDG-II. The first patient with COG5-CDG was recently described (Paesold-Burda et al. Hum Mol Genet 2009; 18:4350–6). Contrary to most other COG-CDG cases, the patient presented a mild/moderate phenotype, i.e. moderate psychomotor retardation with language delay, truncal ataxia and slight hypotonia.

Methods

CDG-IIx patients from our database were screened for mutations in COG5. Clinical data were compared. Brefeldin A treatment of fibroblasts and immunoblotting experiments were performed to support the diagnosis.

Results and conclusion

We identified five new patients with proven COG5 deficiency. We conclude that the clinical picture is not always as mild as previously described. It rather comprises a broad spectrum with phenotypes ranging from mild to very severe. Interestingly, on a clinical basis some of the patients present a significant overlap with COG7-CDG, a finding which can probably be explained by subunit interactions at the protein level.

     

Dynamics of Vaginal Bacterial Communities in Women Developing Bacterial Vaginosis, Candidiasis, or No Infection, Analyzed by PCR-Denaturing Gradient Gel Electrophoresis and Real-Time PCR

The microbial flora of the vagina plays a major role in preventing genital infections, including bacterial vaginosis (BV) and candidiasis (CA). An integrated approach based on PCR-denaturing gradient gel electrophoresis (PCR-DGGE) and real-time PCR was used to study the structure and dynamics of bacterial communities in vaginal fluids of healthy women and patients developing BV and CA. Universal eubacterial primers and Lactobacillus genus-specific primers, both targeted at 16S rRNA genes, were used in DGGE and real-time PCR analysis, respectively. The DGGE profiles revealed that the vaginal flora was dominated by Lactobacillus species under healthy conditions, whereas several potentially pathogenic bacteria were present in the flora of women with BV. Lactobacilli were the predominant bacterial population in the vagina for patients affected by CA, but changes in the composition of Lactobacillus species were observed. Real-time PCR analysis allowed the quantitative estimation of variations in lactobacilli associated with BV and CA diseases. A statistically significant decrease in the relative abundance of lactobacilli was found in vaginal fluids of patients with BV compared to the relative abundance of lactobacilli in the vaginal fluids of healthy women and patients with CA.     

Consistent probabilistic outputs for protein function prediction

In predicting hierarchical protein function annotations, such as terms in the Gene Ontology (GO), the simplest approach makes predictions for each term independently. However, this approach has the unfortunate consequence that the predictor may assign to a single protein a set of terms that are inconsistent with one another; for example, the predictor may assign a specific GO term to a given protein ('purine nucleotide binding') but not assign the parent term ('nucleotide binding'). Such predictions are difficult to interpret. In this work, we focus on methods for calibrating and combining independent predictions to obtain a set of probabilistic predictions that are consistent with the topology of the ontology. We call this procedure 'reconciliation'. We begin with a baseline method for predicting GO terms from a collection of data types using an ensemble of discriminative classifiers. We apply the method to a previously described benchmark data set, and we demonstrate that the resulting predictions are frequently inconsistent with the topology of the GO. We then consider 11 distinct reconciliation methods: three heuristic methods; four variants of a Bayesian network; an extension of logistic regression to the structured case; and three novel projection methods - isotonic regression and two variants of a Kullback-Leibler projection method. We evaluate each method in three different modes - per term, per protein and joint - corresponding to three types of prediction tasks. Although the principal goal of reconciliation is interpretability, it is important to assess whether interpretability comes at a cost in terms of precision and recall. Indeed, we find that many apparently reasonable reconciliation methods yield reconciled probabilities with significantly lower precision than the original, unreconciled estimates. On the other hand, we find that isotonic regression usually performs better than the underlying, unreconciled method, and almost never performs worse; isotonic regression appears to be able to use the constraints from the GO network to its advantage. An exception to this rule is the high precision regime for joint evaluation, where Kullback-Leibler projection yields the best performance.     

"Nonclassical" secretion of annexin A2 to the lumenal side of the enterocyte brush border membrane

Annexin A2 is a member of the annexin family of Ca(2+)-dependent lipid binding proteins and believed to be engaged in membrane transport processes in a number of cell types. In small intestinal enterocytes, we localized annexin A2 to the brush border region, where it was found mainly on the lumenal side of the microvilli, showing an apical secretion by a "nonclassical" mechanism. In addition, annexin A2 was associated with surface-connected, deep apical tubules in the apical terminal web region and with an underlying pleiomorphic, tubulo-vesicular compartment (subapical compartment/multivesicular bodies). By subcellular fractionation, the 36 kDa full-length form of annexin A2 was approximately equally distributed between the Mg(2+)-precipitated fraction (containing intracellular and basolateral membranes) and the microvillar membrane fraction. In addition, a 33 kDa molecular form of annexin A2 was seen in the latter fraction that could be generated from the full-length annexin A2 by digestion with trypsin. Taken together, the results suggest that annexin A2 acts in exocytic apical membrane trafficking and is proteolytically cleaved in situ by pancreatic proteinases once it has become externalized to the lumenal side of the brush border membrane. On the basis of its well-known membrane fusogenic properties, we propose a model for the nonclassical membrane translocation of annexin A2.     

Functional stabilization of an RNA recognition motif by a noncanonical N-terminal expansion

RNA recognition motifs (RRMs) constitute versatile macromolecular interaction platforms. They are found in many components of spliceosomes, in which they mediate RNA and protein interactions by diverse molecular strategies. The human U11/U12-65K protein of the minor spliceosome employs a C-terminal RRM to bind hairpin III of the U12 small nuclear RNA (snRNA). This interaction comprises one side of a molecular bridge between the U11 and U12 small nuclear ribonucleoprotein particles (snRNPs) and is reminiscent of the binding of the N-terminal RRMs in the major spliceosomal U1A and U2B″ proteins to hairpins in their cognate snRNAs. Here we show by mutagenesis and electrophoretic mobility shift assays that the β-sheet surface and a neighboring loop of 65K C-terminal RRM are involved in RNA binding, as previously seen in canonical RRMs like the N-terminal RRMs of the U1A and U2B″ proteins. However, unlike U1A and U2B″, some 30 residues N-terminal of the 65K C-terminal RRM core are additionally required for stable U12 snRNA binding. The crystal structure of the expanded 65K C-terminal RRM revealed that the N-terminal tail adopts an α-helical conformation and wraps around the protein toward the face opposite the RNA-binding platform. Point mutations in this part of the protein had only minor effects on RNA affinity. Removal of the N-terminal extension significantly decreased the thermal stability of the 65K C-terminal RRM. These results demonstrate that the 65K C-terminal RRM is augmented by an N-terminal element that confers stability to the domain, and thereby facilitates stable RNA binding.