David Tritchler

Fish consumption and breast cancer risk: an ecological study

There is experimental evidence that fish oils protect against mammary carcinogens in animals. However, there has been little investigation of the possible relevance of this finding to breast cancer in humans. We compared breast cancer incidence and mortality rates with estimates of the consumption of fish and other foods and nutrients in the countries for which reliable data are available. The results showed an inverse association between percent calories from fish and breast cancer rates that was consistent with a protective effect. This analysis confirmed the finding of others that dietary fat is strongly associated with international variation in breast cancer rates. It also showed that of the dietary components considered, percent calories from fish was the factor most strongly correlated with breast cancer rates after statistical adjustment for dietary fat intake. This result is therefore in accord with animal experimental data and suggests that the omega-3 fatty acids contained in certain fish may protect against breast cancer.

Sparse canonical correlation analysis with application to genomic data integration.

Large scale genomic studies with multiple phenotypic or genotypic measures may require the identification of complex multivariate relationships. In multivariate analysis a common way to inspect the relationship between two sets of variables based on their correlation is canonical correlation analysis, which determines linear combinations of all variables of each type with maximal correlation between the two linear combinations. However, in high dimensional data analysis, when the number of variables under consideration exceeds tens of thousands, linear combinations of the entire sets of features may lack biological plausibility and interpretability. In addition, insufficient sample size may lead to computational problems, inaccurate estimates of parameters and non-generalizable results. These problems may be solved by selecting sparse subsets of variables, i.e. obtaining sparse loadings in the linear combinations of variables of each type. In this paper we present Sparse Canonical Correlation Analysis (SCCA) which examines the relationships between two types of variables and provides sparse solutions that include only small subsets of variables of each type by maximizing the correlation between the subsets of variables of different types while performing variable selection. We also present an extension of SCCA - adaptive SCCA. We evaluate their properties using simulated data and illustrate practical use by applying both methods to the study of natural variation in human gene expression.

Automated analysis of mammographic densities and breast carcinoma risk

There is considerable evidence that one of the strongest risk factors for breast carcinoma can be assessed from the mammographic appearance of the breast. However, the magnitude of the risk factor and the reliability of the prediction depend on the method of classification. Subjective classification requires specialized observer training and suffers from inter‐ and intraobserver variability. Furthermore, the categoric scales make it difficult to distinguish small differences in mammographic appearance. To address these limitations, automated analysis techniques that characterize mammographic density on a continuous scale have been considered, but as yet, these have been evaluated only for their ability to reproduce subjective classifications of mammographic parenchyma.

Mammographic Densities and Breast Cancer Risk

The radiological appearance of the female breast varies among individuals because of differences in the relative amounts and X-ray attenuation characteristics of fat and epithelial and stromal tissues. Fat is radiolucent and appears dark on a mammogram, and epithelium and stroma are radiodense and appear light. We review here the evidence that these variations, known as mammographic parenchymal patterns, are related to risk of breast cancer. Studies that used quantitative measurement to classify mammographic patterns have consistently found that women with dense tissue in more than 60-75% of the breast are at four to six times greater risk of breast cancer than those with no densities. These risk estimates are independent of the effects of other risk factors and have been shown to persist over at least 10 years of follow up. Estimates of attributable risk suggest that this risk factor may account for as many as 30% of breast cancer cases. Mammographically dense breast tissue is associated both with epithelial proliferation and with stromal fibrosis. The relationship between these histological features and risk of breast cancer may by explained by the known actions of growth factors that are thought to play important roles in breast development and carcinogenesis. Mammographically dense tissue differs from most other breast cancer risk factors in the strength of the associated relative and attributable risks for breast cancer, and because it can be changed by hormonal and dietary interventions. This risk factor may be most useful as a means of investigating the etiology of breast cancer and of testing hypotheses about potential preventive strategies.

Breast cancer risk and measured mammographic density.

It has been well established that there is a positive correlation between the dense appearance of breast stroma and parenchyma on a mammogram and the risk of breast cancer. Subjective assessment by radiologists indicated relative risks on the order of 4 to 6 for the group of women whose mammograms showed a density of over 75% or more of the projected area compared to those with an absence of density. In order to obtain a more quantitative, continuous and reproducible means of estimating breast density, which is sensitive to small changes, we have developed quantitative methods for the analysis of mammographic density, which can be applied to digitized mammograms. These techniques have been validated in a nested case-control study on 708 women aged 40–59 years(on entry) who participated in a national mammographic screening study. An interactive image segmentation method and two completely automated techniques based on image texture and grey scale histogram measures have been developed and evaluated. While our methods all show statistically significant risk factors for dense breasts, the interactive method currently provides the highest risk values (relative risk 4.0, 95% confidence interval (CI) = 2.12–7.56) compared to a measure based on the shape of the image histogram (relative risk 3.35, 95% CI = 1.57–7.12) or the fractal dimension of the mammogram (relative risk 2.54, 95% CI = 1.14–5.68), All methods were highly consistent between images of the left and right breast and between the two standard views (cranio-caudal and medio-lateral oblique) of each breast, so that studies can be done by sampling only one of the four views per examination. There is a large number of factors in addition to breast density which affect the appearance of the mammogram. In particular, the assessment of density is made difficult where the breast is not uniformly compressed, e.g. at the periphery. We have designed and are currently evaluating an image processing algorithm that effectively corrects for this problem and have considered methods for controlling some of the variables of image acquisition in prospective studies. Measurements of breast density may be helpful in assigning risk groups to women. Such measurements might guide the frequency of mammographic screening, aid the study of breast cancer aetiology, and be useful in monitoring possible risk-modifying interventions. Using our techniques, we have been able to show that reduction of the proportion of fat in the diet can result in reductions of breast density, although the direct connection to risk has not yet been made. The relationship between breast density and hormone-related and genetic factors is also of great interest. It is often not possible or ethical to obtain mammograms on some groups of women for whom information on density would be very useful. This includes younger women as well as groups in which it would be desirable to obtain such information at frequent intervals. For this reason, we are exploring the use of imaging approaches such as ultrasound and magnetic resonance imaging, which do not require ionizing radiation, to make measurements analogous to those now being performed by using X-ray mammograms.

Albumin Activates ERK Via EGF Receptor in Human Renal Epithelial Cells

Emerging clinical and experimental evidence strongly implicates proteinuria in the progression of kidney disease. One pathway involves the activation of NFkappaB by albumin, and it has been demonstrated that the activation of NFkappaB induced by albumin is dependent on mitogen-activated protein kinase ERK1/ERK2. To study the effect of albumin on gene expression, primary human renal tubular cells were exposed in vitro to albumin (1%) for 6 h, and gene expression profiling was performed with the human oligonucleotide microarray, U133A Affymetrix Gene Chip. In all, 223 genes were differentially regulated by albumin, including marked upregulation of the EGF receptor (EGFR) and IL-8. Accordingly, the authors sought to delineate the signaling pathway linking albumin to the EGFR and activation of ERK1/ERK2. It was found that albumin led to a dose- and time-dependent activation of ERK1/ERK2. Treatment with albumin led to EGFR phosphorylation, but the activation of ERK1/ERK2 was prevented by pretreatment of the cells with AG-1478, the EGFR kinase inhibitor, at a dose that inhibited EGF-induced ERK1/ERK2 activation. Exogenously administered reactive oxygen species (ROS) were found to activate ERK1/ERK2 via the EGFR and src tyrosine kinase activity and pretreatment of cells with the antioxidant N-acetylcysteine (NAC) and the NADPH oxidase inhibitor DPI abrogated albumin-induced activation of ERK1/ERK2. The src tyrosine kinase inhibitor, PP2, also inhibited the albumin-induced activation of ERK1/ERK2. Finally, pretreatment with AG-1478, the MEK inhibitor UO126, and NAC prevented the albumin-induced increase in IL-8 expression. The authors conclude that the EGF receptor plays a central role in the signaling pathway that links albumin to the activation of ERK1/ERK2 and increased expression of IL-8. Gene profiling studies suggest that there may be a positive feedback loop through the EGFR that amplifies the response of the proximal tubule cell to albumin. Taken together, these results suggest that the EGFR may be an important treatment target for kidney disease associated with proteinuria.

An Algorithm for Exact Logistic Regression

Abstract An efficient algorithm is given for the exact logistic analysis of a single parameter. Hypothesis tests and confidence intervals for a logistic regression parameter are obtained. The method provides for the control of the effect of a stratification factor when making inferences concerning the regression parameter.

Genome-wide sparse canonical correlation of gene expression with genotypes

There is a growing interest in studying natural variation in human gene expression. Studies mapping genetic determinants of expression profiles are often carried out considering the expression of one gene at a time, an approach that is computationally intensive and may be prone to high false-discovery rate because the number of genes under consideration often exceeds tens of thousands. We present an exploratory method for investigating such data and apply it to the data provided as Problem 1 of Genetic Analysis Workshop 15 (GAW15). In multivariate analysis, canonical correlation analysis is a common way to inspect the relationship between two sets of variables based on their correlation. It determines linear combinations of all variables from each data set such that the correlation between the two linear combinations is maximized. However, due to the large number of genes, linear combinations involving all single-nucleotide polymorphism (SNP) loci and gene expression phenotypes lack biological plausibility and interpretability. We introduce sparse canonical correlation analysis, which examines the relationships of many genetic loci and gene expression phenotypes by providing sparse linear combinations that include only a small subset of loci and gene expression phenotypes. These correlated sets of variables are sufficiently small for biological interpretability and further investigation. Applying this method to the GAW15 Problem 1 data, we identified groups of 41 loci and 150 gene expressions with the highest between-group correlation of 43%.

Association between the T27C polymorphism in the cytochrome P450 c17α (CYP17) gene and risk factors for breast cancer

Mammographic density is associated with increased breast cancer risk and is influenced by sex hormones. A T27C polymorphism (alleles A1 and A2, respectively) in the 5′ promoter region of CYP17 may be associated with elevated sex hormone levels. In a cross-sectional study of 181 pre- and 173 postmenopausal women, we examined the relationship of this polymorphism with mammographic density and other risk factors for breast cancer. Subjects were recruited across five categories of density. Risk factor and dietary information, anthropometric measures, and blood samples were obtained. Sex hormone, lipid, growth factor levels, and CYP17 genotypes were determined. CYP17 genotype was not associated with mammographic density levels before or after adjusting for risk factors for breast cancer. In premenopausal women, the A2 allele was associated with higher levels of dehydroepiandrosterone sulfate, and in postmenopausal women, with higher levels of total estradiol and lower levels of follicle stimulating hormone. Among premenopausal women, interactions were observed between CYP17 genotype and endogenous insulin levels as well as dietary variables associated with mammographic density. Our findings suggest that the CYP17 A2 allele is associated with hormone levels, and interacts with insulin levels and diet to affect breast density levels and potentially breast cancer risk.

A molecular signature of proteinuria in glomerulonephritis

Proteinuria is the most important predictor of outcome in glomerulonephritis and experimental data suggest that the tubular cell response to proteinuria is an important determinant of progressive fibrosis in the kidney. However, it is unclear whether proteinuria is a marker of disease severity or has a direct effect on tubular cells in the kidneys of patients with glomerulonephritis. Accordingly we studied an in vitro model of proteinuria, and identified 231 “albumin-regulated genes” differentially expressed by primary human kidney tubular epithelial cells exposed to albumin. We translated these findings to human disease by studying mRNA levels of these genes in the tubulo-interstitial compartment of kidney biopsies from patients with IgA nephropathy using microarrays. Biopsies from patients with IgAN (n = 25) could be distinguished from those of control subjects (n = 6) based solely upon the expression of these 231 “albumin-regulated genes.” The expression of an 11-transcript subset related to the degree of proteinuria, and this 11-mRNA subset was also sufficient to distinguish biopsies of subjects with IgAN from control biopsies. We tested if these findings could be extrapolated to other proteinuric diseases beyond IgAN and found that all forms of primary glomerulonephritis (n = 33) can be distinguished from controls (n = 21) based solely on the expression levels of these 11 genes derived from our in vitro proteinuria model. Pathway analysis suggests common regulatory elements shared by these 11 transcripts. In conclusion, we have identified an albumin-regulated 11-gene signature shared between all forms of primary glomerulonephritis. Our findings support the hypothesis that albuminuria may directly promote injury in the tubulo-interstitial compartment of the kidney in patients with glomerulonephritis.