Timothy W. Randolph

Prevalidation of Salivary Biomarkers for Oral Cancer Detection

Background: Oral cancer is the sixth most common cancer with a 5-year survival rate of approximately 60%. Presently, there are no scientifically credible early detection techniques beyond conventional clinical oral examination. The goal of this study is to validate whether the seven mRNAs and three proteins previously reported as biomarkers are capable of discriminating patients with oral squamous cell carcinomas (OSCC) from healthy subjects in independent cohorts and by a National Cancer Institute (NCI)-Early Detection Research Network (EDRN)-Biomarker Reference Laboratory (BRL). Methods: Three hundred and ninety-five subjects from five independent cohorts based on case controlled design were investigated by two independent laboratories, University of California, Los Angeles (Los Angeles, CA) discovery laboratory and NCI-EDRN-BRL. Results: Expression of all seven mRNA and three protein markers was increased in OSCC versus controls in all five cohorts. With respect to individual marker performance across the five cohorts, the increase in interleukin (IL)-8 and subcutaneous adipose tissue (SAT) was statistically significant and they remained top performers across different cohorts in terms of sensitivity and specificity. A previously identified multiple marker model showed an area under the receiver operating characteristic (ROC) curve for prediction of OSCC status ranging from 0.74 to 0.86 across the cohorts. Conclusions: The validation of these biomarkers showed their feasibility in the discrimination of OSCCs from healthy controls. Established assay technologies are robust enough to perform independently. Individual cutoff values for each of these markers and for the combined predictive model need to be further defined in large clinical studies. Impact: Salivary proteomic and transcriptomic biomarkers can discriminate oral cancer from control subjects. Cancer Epidemiol Biomarkers Prev; 21(4); 664–72. ©2012 AACR.

The clathrin adaptor Dab2 recruits EH domain scaffold proteins to regulate integrin β1 endocytosis

Dab2 binds EH domain proteins. This interaction is required for integrin β1 but not TfnR endocytosis. β1 and TfnR do not colocalize, even though their adaptors sort to the same pits. The data suggest that Dab2 selectively drives β1 endocytosis. It is proposed that specific cargo–adaptor–EH domain protein complexes are needed for efficient endocytosis.

Quantifying Peptide Signal in MALDI-TOF Mass Spectrometry Data

This study addressed the question of which properties in MALDI-TOF spectra are relevant to the task of identifying mass and abundance of a peptide species in human serum. Data of this type are common to biomarker studies, but significant within- and between-spectrum variabilities make quantifying biologically induced features difficult. We investigated this signal content and quantified the existence, or lack, of peptide-induced signal (as manifest in a multiresolution decomposition) by generating spectra from human serum in which the abundance of peptides of specific masses is controlled by a sequence of dilutions. The intensities of the corresponding features were directly proportional to peptide concentration. The primary goal was to exhibit some quantifiable properties of raw spectra from this application of MALDI-TOF mass spectrometry. Although no recommendations are given regarding the best method for processing these data, the results confirm the utility of a simple method, based on wavelets, for defining and quantifying features related to low abundance peptide species in a heterogeneous set of complex spectra. Estimates on lower limits of detectable peptide abundance (in the 20-nmol range) and on the number of features present in a spectrum are made possible by the controlled experimental design, the use of a large external reference data set, and dependence on relatively few modeling assumptions.

Enterolignan Producing Phenotypes are Associated with Increased Gut Microbial Diversity and Altered Composition in Premenopausal Women in the United States

Background: Lignans in plant foods are metabolized by gut bacteria to the enterolignans, enterodiol (END) and enterolactone (ENL). Enterolignans have biologic activities important to the prevention of cancer and chronic diseases. We examined the composition of the gut microbial community (GMC) as a contributor to human enterolignan exposure. Methods: We evaluated the association between the GMC in stool, urinary enterolignan excretion, and diet from a 3-day food record in 115 premenopausal (ages 40–45 years) women in the United States. Urinary enterolignans were measured using gas chromatography–mass spectroscopy. The GMC was evaluated using 454 pyrosequencing of the 16S rRNA gene. Sequences were aligned in SILVA (www.arb-silva.de). Operational taxonomic units were identified at 97% sequence similarity. Taxonomic classification was performed and alpha and beta diversity in relationship to ENL production were assessed. Multivariate analysis and regression were used to model the association between enterolignan excretion and the GMC. Bacteria associated with ENL production were identified using univariate analysis and ridge regression. Results: After adjusting for dietary fiber intake and adiposity, we found a significant positive association between ENL excretion and either the GMC (P = 0.0007), or the diversity of the GMC (P = 0.01). The GMC associated with high ENL production was distinct (UNIFRAC, P < 0.003, MRPP) and enriched in Moryella spp., Acetanaerobacterium spp., Fastidiosipila spp., and Streptobacillus spp. Conclusion: Diversity and composition of the GMC are associated with increased human exposure to enterolignans. Impact: Differences in gut microbial diversity and composition explain variation in gut metabolic processes that affect environmental exposures and influence human health. Cancer Epidemiol Biomarkers Prev; 24(3); 546–54. ©2014 AACR.

Characterization of the gut microbiome in epidemiologic studies: the multiethnic cohort experience.

PURPOSE The development of next-generation sequencing and accompanying bioinformatics tools has revolutionized characterization of microbial communities. As interest grows in the role of the human microbiome in health and disease, so does the need for well-powered, robustly designed epidemiologic studies. Here, we discuss sources of bias that can arise in gut microbiome research. METHODS Research comparing methods of specimen collection, preservation, processing, and analysis of gut microbiome samples is reviewed. Although selected studies are primarily based on the gut, many of the same principles are applicable to samples derived from other anatomical sites. Methods for participant recruitment and sampling of the gut microbiome implemented in an ongoing population-based study, the Multiethnic Cohort (MEC), are also described. RESULTS Variation in methodologies can influence the results of human microbiome studies. To help minimize bias, techniques such as sample homogenization, addition of internal standards, and quality filtering should be adopted in protocols. Within the MEC, participant response rates to stool sample collection were comparable to other studies, and in-home stool sample collection yields sufficient high-quality DNA for gut microbiome analysis. CONCLUSIONS Application of standardized and quality controlled methods in human microbiome studies is necessary to ensure data quality and comparability among studies.

Signal Detection in High-Resolution Mass Spectrometry Data

Mass spectrometry data from high-resolution time-of-flight instruments often contain a vast number of noninformative background-ion peaks whose signal is similar to that of peptide peaks. Consequently, seeking peptide signal in these spectra based on a signal-to-noise ratio will remove signal peaks as well as noise. This work characterizes the background as a precursor to seeking peptide-related features. Robust-regression methods are used to estimate distributions for null (background) peak intensities and locations. Defining signal peaks as outliers with respect to these distributions leads to more precision in detecting the isotopic envelope of peaks from low-abundance peptides in high-resolution spectra.

Comparative Study of Computational Methods for Reconstructing Genetic Networks of Cancer-Related Pathways

Network reconstruction is an important yet challenging task in systems biology. While many methods have been recently proposed for reconstructing biological networks from diverse data types, properties of estimated networks and differences between reconstruction methods are not well understood. In this paper, we conduct a comprehensive empirical evaluation of seven existing network reconstruction methods, by comparing the estimated networks with different sparsity levels for both normal and tumor samples. The results suggest substantial heterogeneity in networks reconstructed using different reconstruction methods. Our findings also provide evidence for significant differences between networks of normal and tumor samples, even after accounting for the considerable variability in structures of networks estimated using different reconstruction methods. These differences can offer new insight into changes in mechanisms of genetic interaction associated with cancer initiation and progression.

Kernel-penalized regression for analysis of microbiome data

The analysis of human microbiome data is often based on dimension-reduced graphical displays and clusterings derived from vectors of microbial abundances in each sample. Common to these ordination methods is the use of biologically motivated definitions of similarity. Principal coordinate analysis, in particular, is often performed using ecologically defined distances, allowing analyses to incorporate context-dependent, non-Euclidean structure. In this paper, we go beyond dimension-reduced ordination methods and describe a framework of high-dimensional regression models that extends these distance-based methods. In particular, we use kernel-based methods to show how to incorporate a variety of extrinsic information, such as phylogeny, into penalized regression models that estimate taxonspecific associations with a phenotype or clinical outcome. Further, we show how this regression framework can be used to address the compositional nature of multivariate predictors comprised of relative abundances; that is, vectors whose entries sum to a constant. We illustrate this approach with several simulations using data from two recent studies on gut and vaginal microbiomes. We conclude with an application to our own data, where we also incorporate a significance test for the estimated coefficients that represent associations between microbial abundance and a percent fat.

Oxidative DNA damage during night shift work

Objectives We previously reported that compared with night sleep, day sleep among shift workers was associated with reduced urinary excretion of 8-hydroxydeoxyguanosine (8-OH-dG), potentially reflecting a reduced ability to repair 8-OH-dG lesions in DNA. We identified the absence of melatonin during day sleep as the likely causative factor. We now investigate whether night work is also associated with reduced urinary excretion of 8-OH-dG. Methods For this cross-sectional study, 50 shift workers with the largest negative differences in night work versus night sleep circulating melatonin levels (measured as 6-sulfatoxymelatonin in urine) were selected from among the 223 shift workers included in our previous study. 8-OH-dG concentrations were measured in stored urine samples using high performance liquid chromatography with electrochemical detection. Mixed effects models were used to compare night work versus night sleep 8-OH-dG levels. Results Circulating melatonin levels during night work (mean=17.1 ng/mg creatinine/mg creatinine) were much lower than during night sleep (mean=51.7 ng/mg creatinine). In adjusted analyses, average urinary 8-OH-dG levels during the night work period were only 20% of those observed during the night sleep period (95% CI 10% to 30%; p<0.001). Conclusions This study suggests that night work, relative to night sleep, is associated with reduced repair of 8-OH-dG lesions in DNA and that the effect is likely driven by melatonin suppression occurring during night work relative to night sleep. If confirmed, future studies should evaluate melatonin supplementation as a means to restore oxidative DNA damage repair capacity among shift workers.

Pharmacometabonomic Prediction of Busulfan Clearance in Hematopoietic Cell Transplant Recipients

Intravenous (IV) busulfan doses are often personalized to a concentration at steady state (Css) using the patients clearance, which is estimated with therapeutic drug monitoring. We sought to identify biomarkers of IV busulfan clearance using a targeted pharmacometabonomics approach. A total of 200 metabolites were quantitated in 106 plasma samples, each obtained before IV busulfan administration in hematopoietic cell transplant (HCT) recipients. Both univariate linear regression with false discovery rate (FDR) and pathway enrichment analyses using the Global test were performed. In the univariate analysis, glycine, N-acetylglycine, 2-hydroxyisovaleric acid, creatine, serine, and tyrosine were statistically significantly associated with IV busulfan clearance at P < 0.05, with the first three satisfying the FDR of q < 0.1. Using pathway enrichment analysis, the glycine, serine, and threonine metabolism pathway was statistically significantly associated with IV busulfan clearance at P < 0.05 and q < 0.1, and a pathway impact >0.1. Glycine is a component of glutathione, which is conjugated with busulfan via glutathione transferase enzymes. These results demonstrate the potential utility of pharmacometabonomics to inform IV busulfan dosing. Future studies are required to validate these findings.