Katja Ickstadt

Role of thioredoxin reductase 1 and thioredoxin interacting protein in prognosis of breast cancer

IntroductionThe purpose of this work was to study the prognostic influence in breast cancer of thioredoxin reductase 1 (TXNRD1) and thioredoxin interacting protein (TXNIP), key players in oxidative stress control that are currently evaluated as possible therapeutic targets.MethodsAnalysis of the association of TXNRD1 and TXNIP RNA expression with the metastasis-free interval (MFI) was performed in 788 patients with node-negative breast cancer, consisting of three individual cohorts (Mainz, Rotterdam and Transbig). Correlation with metagenes and conventional clinical parameters (age, pT stage, grading, hormone and ERBB2 status) was explored. MCF-7 cells with a doxycycline-inducible expression of an oncogenic ERBB2 were used to investigate the influence of ERBB2 on TXNRD1 and TXNIP transcription.ResultsTXNRD1 was associated with worse MFI in the combined cohort (hazard ratio = 1.955; P < 0.001) as well as in all three individual cohorts. In contrast, TXNIP was associated with better prognosis (hazard ratio = 0.642; P < 0.001) and similar results were obtained in all three subcohorts. Interestingly, patients with ERBB2-status-positive tumors expressed higher levels of TXNRD1. Induction of ERBB2 in MCF-7 cells caused not only an immediate increase in TXNRD1 but also a strong decrease in TXNIP. A subsequent upregulation of TXNIP as cells undergo senescence was accompanied by a strong increase in levels of reactive oxygen species.ConclusionsTXNRD1 and TXNIP are associated with prognosis in breast cancer, and ERBB2 seems to be one of the factors shifting balances of both factors of the redox control system in a prognostic unfavorable manner.

Retention time alignment algorithms for LC/MS data must consider non-linear shifts

MOTIVATION Proteomics has particularly evolved to become of high interest for the field of biomarker discovery and drug development. Especially the combination of liquid chromatography and mass spectrometry (LC/MS) has proven to be a powerful technique for analyzing protein mixtures. Clinically orientated proteomic studies will have to compare hundreds of LC/MS runs at a time. In order to compare different runs, sophisticated preprocessing steps have to be performed. An important step is the retention time (rt) alignment of LC/MS runs. Especially non-linear shifts in the rt between pairs of LC/MS runs make this a crucial and non-trivial problem. RESULTS For the purpose of demonstrating the particular importance of correcting non-linear rt shifts, we evaluate and compare different alignment algorithms. We present and analyze two versions of a new algorithm that is based on regression techniques, once assuming and estimating only linear shifts and once also allowing for the estimation of non-linear shifts. As an example for another type of alignment method we use an established alignment algorithm based on shifting vectors that we adapted to allow for correcting non-linear shifts also. In a simulation study, we show that rt alignment procedures that can estimate non-linear shifts yield clearly better alignments. This is even true under mild non-linear deviations. AVAILABILITY R code for the regression-based alignment methods and simulated datasets are available at http://www.statistik.tu-dortmund.de/genetik-publikationen-alignment.html. SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.

Improved Criteria for Clustering Based on the Posterior Similarity Matrix

In this paper we address the problem of obtaining a single clustering estimate bc based on an MCMC sample of clusterings c (1) ;c (2) :::;c (M) from the posterior distribution of a Bayesian cluster model. Methods to derive b when the number of groups K varies between the clusterings are reviewed and discussed. These include the maximum a posteriori (MAP) estimate and methods based on the posterior similarity matrix, a matrix containing the posterior probabilities that the observations i and j are in the same cluster. The posterior similarity matrix is related to a commonly used loss function by Binder (1978). Minimization of the loss is shown to be equivalent to maximizing the Rand index between esti- mated and true clustering. We propose new criteria for estimating a clustering, which are based on the posterior expected adjusted Rand index. The criteria are shown to possess a shrinkage property and outperform Binders loss in a simulation study and in an application to gene expression data. They also perform favorably compared to other clustering procedures.

Genetic variants in urinary bladder cancer: collective power of the “wimp SNPs”

In recent years, genome-wide association studies (GWAS) have identified more than 300 validated associations between genetic variants and risk of approximately 70 common diseases. A small number of rare variants with a frequency of usually less than 1% are associated with a strongly enhanced risk, such as genetic variants of TP53, RB1, BRCA1, and BRCA2. Only a very small number of SNPs (with a frequency of more that 1% of the rare allele) have effects of a factor of two or higher. Examples include APOE4 in Alzheimer’s disease, LOXL1 in exfoliative glaucoma, and CFH in age-related macular degeneration. However, the majority of all identified SNPs have odds ratios between 1.1 and 1.5. In the case of urinary bladder cancer, all known SNPs that have been validated in sufficiently large populations are associated with odds ratios smaller than 1.5. These SNPs are located next to the following genes: MYC, TP63, PSCA, the TERT-CLPTM1L locus, FGFR3, TACC3, NAT2, CBX6, APOBEC3A, CCNE1, and UGT1A. It is likely that these moderate risk or “wimp SNPs” interact, and because of their high number, collectively have a strong influence on whether an individual will develop cancer or not. It should be considered that variants identified so far explain only approximately 5–10% of the overall inherited risk. Possibly, the remaining variance is due to an even higher number of SNPs with odds ratios smaller than 1.1. Recent studies have provided the following information: (1) The functions of genes identified as relevant for bladder cancer focus on detoxification of carcinogens, control of the cell cycle and apoptosis, as well as maintenance of DNA integrity. (2) Many novel SNPs are far away from the protein coding regions, suggesting that these SNPs are located on distant-acting transcriptional enhancers. (3) The low odds ratio of each individual bladder cancer-associated SNP is too low to justify reasonable preventive measures. However, if the recently identified SNPs interact, they may collectively result in a substantial risk that is of preventive relevance. In addition to the “novel SNPs” identified by the recent GWAS, at least 163 further variants have been reported in relation to bladder cancer, although they have not been consistently validated in independent case–control series. Moreover, given that only 60 of these 163 “old SNPs” are covered by the SNP chips used in the recent GWAS, there are in principle 103 published variants still awaiting validation or disproval. In future, besides identifying novel disease-associated rare variants by deep sequencing, it will also be important to understand how the already identified variants interact.

Susceptibility to urinary bladder cancer: relevance of rs9642880[T], GSTM1 0/0 and occupational exposure.

Recently, a genome-wide single nucleotide polymorphism association study has identified a sequence variant 30 kb upstream of the c-Myc gene (allele T of rs9642880) that confers susceptibility to bladder cancer. However, the role of exposure to bladder carcinogens has not been considered. This prompted us to analyse the relevance of this polymorphism in 515 bladder cancer cases and 893 controls where the quality and quantity of occupational exposure to bladder carcinogens has been documented. When we analysed a hospital-based case-control series not selected for occupational exposure, rs9642880[T] was influential, in contrast to GSTM1 0/0. However, in a case-control series of patients that have been occupationally exposed to aromatic amines and polycyclic aromatic hydrocarbons, rs9642880[T] was not influential but GSTM1 0/0 was significantly associated with bladder cancer risk. Therefore, the degree to which rs9642880[T] and GSTM1 0/0 confer susceptibility to urinary bladder cancer seems to depend on the extent of exposure to urinary bladder carcinogens.

Detecting high-order interactions of single nucleotide polymorphisms using genetic programming

MOTIVATION Not individual single nucleotide polymorphisms (SNPs), but high-order interactions of SNPs are assumed to be responsible for complex diseases such as cancer. Therefore, one of the major goals of genetic association studies concerned with such genotype data is the identification of these high-order interactions. This search is additionally impeded by the fact that these interactions often are only explanatory for a relatively small subgroup of patients. Most of the feature selection methods proposed in the literature, unfortunately, fail at this task, since they can either only identify individual variables or interactions of a low order, or try to find rules that are explanatory for a high percentage of the observations. In this article, we present a procedure based on genetic programming and multi-valued logic that enables the identification of high-order interactions of categorical variables such as SNPs. This method called GPAS cannot only be used for feature selection, but can also be employed for discrimination. RESULTS In an application to the genotype data from the GENICA study, an association study concerned with sporadic breast cancer, GPAS is able to identify high-order interactions of SNPs leading to a considerably increased breast cancer risk for different subsets of patients that are not found by other feature selection methods. As an application to a subset of the HapMap data shows, GPAS is not restricted to association studies comprising several 10 SNPs, but can also be employed to analyze whole-genome data. AVAILABILITY Software can be downloaded from http://ls2-www.cs.uni-dortmund.de/~nunkesser/#Software

Genotyping NAT2 with only two SNPs (rs1041983 and rs1801280) outperforms the tagging SNP rs1495741 and is equivalent to the conventional 7-SNP NAT2 genotype.

Genotyping N-acetyltransferase 2 (NAT2) is of high relevance for individualized dosing of antituberculosis drugs and bladder cancer epidemiology. In this study we compared a recently published tagging single nucleotide polymorphism (SNP) (rs1495741) to the conventional 7-SNP genotype (G191A, C282T, T341C, C481T, G590A, A803G and G857A haplotype pairs) and systematically analysed if novel SNP combinations outperform the latter. For this purpose, we studied 3177 individuals by PCR and phenotyped 344 individuals by the caffeine test. Although the tagSNP and the 7-SNP genotype showed a high degree of correlation (R=0.933, P<0.0001) the 7-SNP genotype nevertheless outperformed the tagging SNP with respect to specificity (1.0 vs. 0.9444, P=0.0065). Considering all possible SNP combinations in a receiver operating characteristic analysis we identified a 2-SNP genotype (C282T, T341C) that outperformed the tagging SNP and was equivalent to the 7-SNP genotype. The 2-SNP genotype predicted the correct phenotype with a sensitivity of 0.8643 and a specificity of 1.0. In addition, it predicted the 7-SNP genotype with sensitivity and specificity of 0.9993 and 0.9880, respectively. The prediction of the NAT2 genotype by the 2-SNP genotype performed similar in populations of Caucasian, Venezuelan and Pakistani background. A 2-SNP genotype predicts NAT2 phenotypes with similar sensitivity and specificity as the conventional 7-SNP genotype. This procedure represents a facilitation in individualized dosing of NAT2 substrates without losing sensitivity or specificity.

Simulation of Lévy Random Fields

An efficient recently developed method, the Inverse Levy Measure (ILM) algorithm, is presented for drawing random samples from gamma, skewed stable and other nonnegative independent-increment random fields, which we call Levy random fields. The method is useful for computing posterior distributions in nonparametric hierarchical Bayesian statistical analysis. Algorithms are illustrated through prototype implementations in S-PLUS.

Bayesian Nonparametric Estimation of Continuous Monotone Functions with Applications to Dose–Response Analysis

In this article, we consider monotone nonparametric regression in a Bayesian framework. The monotone function is modeled as a mixture of shifted and scaled parametric probability distribution functions, and a general random probability measure is assumed as the prior for the mixing distribution. We investigate the choice of the underlying parametric distribution function and find that the two-sided power distribution function is well suited both from a computational and mathematical point of view. The model is motivated by traditional nonlinear models for dose-response analysis, and provides possibilities to elicitate informative prior distributions on different aspects of the curve. The method is compared with other recent approaches to monotone nonparametric regression in a simulation study and is illustrated on a data set from dose-response analysis.

ERBB2 induces an antiapoptotic expression pattern of Bcl-2 family members in node-negative breast cancer.

Purpose: Members of the Bcl-2 family act as master regulators of mitochondrial homeostasis and apoptosis. We analyzed whether ERBB2 influences the prognosis of breast cancer by influencing the proapoptotic versus antiapoptotic balance of Bcl-2 family members. Experimental Design: ERBB2-regulated Bcl-2 family members were identified by inducible expression of ERBB2 in MCF-7 breast cancer cells and by correlation analysis with ERBB2 expression in breast carcinomas. The prognostic relevance of ERBB2-regulated and all additional Bcl-2 family members was determined in 782 patients with untreated node-negative breast cancer. The biological relevance of ERBB2-induced inhibition of apoptosis was validated in a murine tumor model allowing conditional ERBB2 expression. Results: ERBB2 caused an antiapoptotic phenotype by upregulation of MCL-1, TEGT, BAG1, BNIP1, and BECN1 as well as downregulation of BAX, BMF, BNIPL, CLU, and BCL2L13. Upregulation of the antiapoptotic MCL-1 [P = 0.001, hazard ratio (HR) 1.5] and BNIP3 (P = 0.024; HR, 1.4) was associated with worse prognosis considering metastasis-free interval, whereas clusterin (P = 0.008; HR, 0.88) and the proapoptotic BCL2L13 (P = 0.019; HR, 0.45) were associated with better prognosis. This indicates that ERBB2 alters the expression of Bcl-2 family members in a way that leads to adverse prognosis. Analysis of apoptosis and tumor remission in a murine tumor model confirmed that the prototypic Bcl-2 family member Bcl-xL could partially substitute for ERBB2 to antagonize tumor remission. Conclusions: Our results support the concept that ERBB2 influences the expression of Bcl-2 family members to induce an antiapoptotic phenotype. Antagonization of antiapoptotic Bcl-2 family members might improve breast cancer therapy, whereby MCL-1 and BNIP3 represent promising targets. Clin Cancer Res; 16(2); 451–60