The application of genomic and proteomic technologies in predictive, preventive and personalized medicine.

The long asymptomatic period before the onset of chronic diseases offers good opportunities for disease prevention. Indeed, many chronic diseases may be preventable by avoiding those factors that trigger the disease process (primary prevention) or by use of therapy that modulates the disease process before the onset of clinical symptoms (secondary prevention). Accurate prediction is vital for disease prevention so that therapy can be given to those individuals who are most likely to develop the disease. The utility of predictive markers is dependent on three parameters, which must be carefully assessed: sensitivity, specificity and positive predictive value. Specificity is important if a biomarker is to be used to identify individuals either for counseling or for preventive therapy. However, a reciprocal relationship exists between sensitivity and specificity. Thus, successful biomarkers will be highly specific without sacrificing sensitivity. Unfortunately, biomarkers with ideal specificity and sensitivity are difficult to find for many diseases. One potential solution is to use the combinatorial power of a large number of biomarkers, each of which alone may not offer satisfactory specificity and sensitivity. Recent technological advances in genetics, genomics, proteomics, and bioinformatics offer a great opportunity for biomarker discovery. The newly identified biomarkers have the potential to bring increased accuracy in disease diagnosis and classification, as well as therapeutic monitoring. In this review, we will use type 1 diabetes (T1D) as an example, when appropriate, to discuss pertinent issues related to high throughput biomarker discovery.

Microarray experimental design: power and sample size considerations.

Gene expression analysis using high-throughput microarray technology has become a powerful approach to study systems biology. The exponential growth in microarray experiments has spawned a number of investigations into the reliability and reproducibility of this type of data. However, the sample size requirements necessary to obtain statistically significant results has not had as much attention. We report here statistical methods for the determination of the sufficient number of subjects necessary to minimize the false discovery rate while maintaining high power to detect differentially expressed genes. Two experimental designs were considered: 1) a comparison between two groups at a single time point, and 2) a comparison of two experimental groups with sequential time points. Computer programs are available for the methods discussed in this paper and are adaptable to more complicated situations.

A statistical method for flagging weak spots improves normalization and ratio estimates in microarrays.

Over the last few years, there has been a dramatic increase in the use of cDNA microarrays to monitor gene expression changes in biological systems. Data from these experiments are usually transformed into expression ratios between experimental samples and a common reference sample for subsequent data analysis. The accuracy of this critical transformation depends on two major parameters: the signal intensities and the normalization of the experiment vs. reference signal intensities. Here we describe and validate a new model for microarray signal intensity that has one multiplicative variation and one additive background variation. Using replicative experiments and simulated data, we found that the signal intensity is the most critical parameter that influences the performance of normalization, accuracy of ratio estimates, reproducibility, specificity, and sensitivity of microarray experiments. Therefore, we developed a statistical procedure to flag spots with weak signal intensity based on the standard deviation (delta(ij)) of background differences between a spot and the neighboring spots, i.e., a spot is considered as too weak if the signal is weaker than cdelta(ij). Our studies suggest that normalization and ratio estimates were unacceptable when this threshold (c) is small. We further showed that when a reasonable compromise of c (c = 6) is applied, normalization using trimmed mean of log ratios performed slightly better than global intensity and mean of ratios. These studies suggest that decreasing the background noise is critical to improve the quality of microarray experiments.

Rapid decrease of RNA level of a novel mouse mitochondria solute carrier protein (Mscp) gene at 4-5 weeks of age.

We cloned a novel mouse gene that encodes a protein with homology to the mitochondria solute carrier proteins (Mscp). The major full-length Mscp transcript contains 4112 bp of cDNA and a deduced protein of 338 amino acids. The Mscp protein shares 50%, 40%, and 39% sequence identity with the C. elegans hypothetical protein T26089 and the yeast mitochondria carrier proteins MRS3 and MRS4, respectively. It also showed homology with the uncoupling proteins (UCP1, UCP2, and UCP3; 22%, 24%, and 29% identity, respectively). The protein has six transmembrane domains and three mitochondria energy-transfer protein signature motifs, which are conserved among all the members of mitochondria carrier protein family. Northern analysis indicated that the Mscp gene is highly expressed in the spleen. Using cDNA microarray and Northern analysis, we have shown a significant decrease of the splenic Mscp mRNA levels around 4-5 weeks of age in several mouse strains including C57BL/6J, nonobese diabetic (NOD), and several NOD-congenic mice. These results suggest that the Mscp gene is decreased during splenic lymphocyte maturation in these mice.

WECHE: a novel hematopoietic regulatory factor.

Previously, we described AGM-derived endothelial cell lines that either inhibited or permitted the development of erythroid or B cells. We utilized a differential gene expression method to isolate a chemokine, termed WECHE, from one of these cell lines. WECHE inhibited the formation of erythroid cells but had no effect on either myeloid or B cell formation. WECHE repressed BFU-E development from either mouse fetal liver or bone marrow progenitor cells but had no effect on colony formation induced by IL-3 or IL-7. WECHE reduced HPP-CFC production from fetal liver-derived stem cells. WECHE hindered the growth of yolk sac-derived endothelial cells. WECHE was also chemotactic for bone marrow cells. Thus, WECHE is a novel chemokine that regulates hematopoietic differentiation.

Linkage analysis of 150 high-risk prostate cancer families at 1q24-25.

Confirmation of linkage and estimation of the proportion of families who are linked in large independent datasets is essential to understanding the significance of cancer susceptibility genes. We report here on an analysis of 150 high-risk prostate cancer families (2,176 individuals) for potential linkage to the HPC1 prostate cancer susceptibility locus at 1q24-25. This dataset includes 640 affected men with an average age at prostate cancer diagnosis of 66. 8 years (range, 39-94), representing the largest collection of high-risk families analyzed for linkage in this region to date. Linkage to multiple 1q24-25 markers was strongly rejected for the sample as a whole (lod scores at theta = 0 ranged from -30.83 to -18. 42). Assuming heterogeneity, the estimated proportion of families linked (alpha) at HPC1 in the entire dataset was 2.6%, using multipoint analysis. Because locus heterogeneity may lead to false rejection of linkage, data were stratified based on homogeneous subsets. When restricted to 21 Caucasian families with five or more affected family members and mean age at diagnosis < = 65 years, the lod scores at theta = 0 remained less than -4.0. These results indicate that the overall portion of hereditary prostate cancer families whose disease is due to inherited variation in HPC1 may be less than originally estimated.

Evidence for a rare prostate cancer-susceptibility locus at chromosome 1p36.

Combining data from a genomic screen in 70 families with a high risk for prostate cancer (PC) with data from candidate-region mapping in these families and an additional 71 families, we have localized a potential hereditary PC-susceptibility locus to chromosome 1p36. Because an excess of cases of primary brain cancer (BC) have been observed in some studies of families with a high risk for PC, and because loss of heterozygosity at 1p36 is frequently observed in BC, we further evaluated 12 families with both a history of PC and a blood relative with primary BC. The overall LOD score in these 12 families was 3.22 at a recombination fraction (theta) of .06, with marker D1S507. On the basis of an a priori hypothesis, this group was stratified by age at diagnosis of PC. In the younger age group (mean age at diagnosis <66 years), a maximum two-point LOD score of 3.65 at straight theta = .0 was observed, with D1S407. This linkage was rejected in both early- and late-onset families without a history of BC (LOD scores -7.12 and -6.03, respectively, at straight theta = .0). After exclusion of 3 of the 12 families that had better evidence of linkage to previously described PC-susceptibility loci, linkage to the 1p36 region was suggested by a two-point LOD score of 4.74 at straight theta = .0, with marker D1S407. We conclude that a significant proportion of these families with both a high risk for PC and a family member with BC show linkage to the 1p36 region.

Localization of non-Mhc collagen-induced arthritis susceptibility loci in DBA/1j mice.

One approach to understanding common human diseases is to determine the genetic defects responsible for similar diseases in animal models and place those defective genes in their corresponding biochemical pathways. Our laboratory is working with an animal model for human rheumatoid arthritis called collagen-induced arthritis (CIA). We are particularly interested in determining the location of disease-predisposing loci. To that end, we performed experiments to localize susceptibility loci for CIA in an F2 cross between the highly susceptible mouse strain DBA/1j and the highly resistant mouse strain SWR/j. Specifically, a quantitative trait locus analysis was performed to localize regions of the mouse genome responsible for susceptibility/severity to CIA. One susceptibility locus, Cia1 in the major histocompatibility locus, had been identified previously. Two additional loci were detected in our analysis that contribute to CIA severity (Cia2, Cia3) on chromosomes 2 and 6. A third locus was detected that contributes to the age of onset of the disease. This locus (Cia4) was located on chromosome 2 and was linked to the same region as Cia2. Determining the identity of these loci may provide insights into the etiology of human rheumatoid arthritis.

Linkage analysis of 49 high-risk families does not support a common familial prostate cancer-susceptibility gene at 1q24-25.

Linkage of a putative prostate cancer-susceptibility locus (HPC1) to chromosome 1q24-25 has recently been reported. Confirmation of this linkage in independent data sets is essential because of the complex nature of this disease. Here we report the results of a linkage analysis using 10 polymorphic markers spanning approximately 37 cM in the region of the putative HPC1 locus in 49 high-risk prostate cancer families. Data were analyzed by use of two parametric models and a nonparametric method. For the parametric LOD-score method, the first model was identical to the original report by Smith and coworkers ("Hopkins"), and the second was based on a segregation analysis previously reported by Carter and coworkers ("Seattle"). In both cases, our results do not confirm the linkage reported for this region. Calculated LOD scores from the two-point analysis for each marker were highly negative at small recombination fractions. Multipoint LOD scores for this linkage group were also highly negative. Additionally, we were unable to demonstrate heterogeneity within the data set, using HOMOG. Although these data do not formally exclude linkage of a prostate cancer-susceptibility locus at HPC1, it is likely that other prostate cancer-susceptibility loci play a more critical role in the families that we studied.

Data-adaptive algorithms for calling alleles in repeat polymorphisms.

Data-adaptive algorithms are presented for separating overlapping signatures of heterozygotic allele pairs in electrophoresis data. Application is demonstrated for human microsatellite CA-repeat polymorphisms in LiCor 4000 and ABI 373 data. The algorithms allow overlapping alleles to be called correctly in almost every case where a trained observer could do so, and provide a fast automated objective alternative to human reading of the gels. The algorithm also supplies an indication of confidence level which can be used to flag marginal cases for verification by eye, or as input to later stages of statistical analysis.

An analysis of the dynamic range and linearity of an infrared DNA sequencer.

We are currently developing genotyping software and protocols for use on the Li-Cor model 4000(L) infrared fluorescence DNA sequencers. During the development of the genotyping software it became apparent that the potential dynamic range of the instrument was not being realized when the data collection parameters were optimized to produce a high-contrast image on the computer screen. In particular, we were unable to obtain peaks with signal to noise (S/N) greater than about 30:1 and the stronger peaks often saturated the detector. Because the numerical dynamic range available in 16-bit data collection mode exceeds 65000, our limited dynamic range of about 30 in actual data was somewhat puzzling. Hence, we undertook a study to explore the dynamic range and linearity of the Li-Cor DNA sequencer in order to minimize background fluorescence and noise as well as maximize the available S/N. Data is presented on the background and noise using different polyacrylamide gel mixes with various signal gain and offset values, and the relative contribution of the glass, gel and instrumentation to the background fluorescence is discussed. Based on these results, optimum gain and offset values were determined that maintain the background fluorescence at approximately 1-2% of the maximum dynamic range with a minimum amount of noise. Using these optimized values, we determined the detection system is linear over three orders of magnitude by titrating known quantities of an infrared fluorophore-labeled primer. In addition, we were able to detect approximately 15.2 amoles of labeled primer. The results provided by this study establish a set of guidelines for evaluating how to set the signal gain and offset in order to minimize the background and maximize the S/N and dynamic range of the Li-Cor sequencer.

The relation of blood sodium concentration to blood pressure in the rat.

Rises or falls in blood sodium concentration ([Na]) within a physiological range of +/- 15 mmol/l, sustained for 5 h, were produced in the rat by intraperitoneal dialysis with physiological salt solutions containing variable amounts of Na. In general, systolic and diastolic blood pressure rose and fell in direct relation to the alteration in [Na]. Solutions of equivalent osmolarity produced changes in blood pressure that were the inverse of those induced by Na. These effects could not be explained in terms of changes in blood or extracellular fluid volume, and indicated the need for an exploration of the partition of Na and water across the vascular smooth muscle cell.

The origin of MHC class II gene polymorphism within the genus Mus.

The I region of the major histocompatibility complex (MHC) of the mouse (H-2) contains a tightly-linked cluster of highly polymorphic genes (class II MHC genes) which control immune responsiveness. Speculation on the origin of this polymorphism, which is believed to be essential for the function of the class II proteins in immune responses to disease, has given rise to two hypotheses. The first is that hypermutational mechanisms (gene conversion or segmental exchange) promote the rapid generation of diversity in MHC genes. The alternative is that polymorphism has arisen from the steady accumulation of mutations over long evolutionary periods, and multiple specific alleles have survived speciation (trans-species evolution). We have looked for evidence of 'segmental exchange' and/or 'trans-species evolution' in the class II genes of the genus Mus by molecular genetic analysis of I-A beta alleles. The results indicate that greater than 90% (28 out of 31) of the alleles examined can be organized into two evolutionary groups both on the basis of restriction site polymorphisms and by the presence or absence of a short interspersed nucleotide element (SINE). Using this SINE sequence as an evolutionary tag, we demonstrate that I-A beta alleles in these two evolutionary groups diverged at least three million years ago and have survived the speciation events leading to several modern Mus species. Nucleotide sequence comparisons of eight Mus m. domesticus I-A beta alleles representing all three evolutionary groups indicate that most of the divergence in exon sequences is due to the steady accumulation of mutations that are maintained independently in the different alleles. But segmental exchanges between alleles from different evolutionary groups have also played a role in the diversification of beta 1 exons.

The relation of cellular sodium to the onset of hypertension induced by DOCA-saline in the rat.

Groups of 15 rats were injected subcutaneously with a microcrystalline suspension of DOCA and given 1% saline as drinking water for periods ranging from 2 to 16 days. The in vivo transmembrane distribution of Na and K in relation to blood pressure was assessed in terms of plasma [Na] and [K] measured with ion-specific electrodes; smooth muscle cell Na, K and water were measured in the rapidly excised tail artery. A small increase in blood pressure was observed in week 1 of treatment and was followed by an abrupt increase to higher levels at about day 10. Plasma [Na] was elevated and [K] lowered throughout in a new steady state with a reciprocal change of about 1 mmol/l. There were minor changes in intracellular [K] in rapidly excised fresh tail artery samples so that the transmembrane K gradient (EK) was at all times increased by greater than 10 mV, indicating enhanced Na-K transport. In general, intracellular [Na] was directly and Na gradient inversely related to blood pressure (P less than 0.01).

Prehypertensive changes in sodium transport induced by deoxycorticosterone acetate in incubated rat tail artery.

The transport and distribution of sodium, potassium, and water were examined in tail arteries of rats treated with deoxycorticosterone acetate (DOCA)-saline for 10 days, a time that marks the earliest onset of a rise of blood pressure in the strain (Wistar) used. The arteries were incubated for more than 20 hours to ensure that any change observed was sufficiently built in so that it could not readily be washed out. Three distinct changes were observed. First, the steady state transmembrane sodium gradient (operationally [Na]o/[Na]i) was increased. Second, the amount of sodium excluded from participation in the sodium gradient, and hence probably bound, was increased. Third, after prolonged potassium depletion, the ouabain-insensitive loss of cell water and sodium that follows the readmittance of potassium was increased. These results suggest that fundamental embedded changes in sodium transport occur well before the blood pressure rises in response to DOCA-saline.

Na transport and blood pressure in the adrenalectomized rat.

The transmembrane distribution of Na, K, and water in freshly excised rat tail artery was measured 7 days after adrenalectomy, 3 days after withdrawal of support with 1% NaCl as drinking water. Cell Na decreased from 35.8 +/- 0.9 to 31.6 +/- 0.7 mmol/kg dry wt, tissue water increased from 2.93 +/- 0.03 to 3.08 +/- 0.04 l/kg dry wt, and tissue K remained unchanged. As concentration, [Na]i decreased from 22.5 +/- 0.9 to 18.3 +/- 1.0 mM in parallel with [Na]o which fell from 140.6 +/- 0.3 to 133.6 +/- 1.0. Conversely, while [K]i fell from 140 +/- 2 to 128 +/- 3, [K]o rose from 2.15 +/- 0.12 to 5.40 +/- 0.33 so that EK dropped from -109 to -83 mV. Aldosterone (120 micrograms X 100 g-1 X 24 h-1 sc; 3 X 10(-7) M if immediately distributed in the extracellular fluid) partially restored blood pressure, plasma [Na] and [K], and tissue water within 24-26 h but did not increase cell Na. Similar effects were produced with corticosterone at 400 microgram X 100 g-1 X 24 h-1 (3 X 10(-6) M if immediately distributed in the extracellular fluid). We conclude that the restoration of blood pressure in the adrenalectomized rat is independent of [Na]i or [Na]o/[Na]i but is associated with enhanced Na transport activity.