A statistical framework for integrating two microarray data sets in differential expression analysis.

BACKGROUND: Different microarray data sets can be collected for studying the same or similar diseases. We expect to achieve a more efficient analysis of differential expression if an efficient statistical method can be developed for integrating different microarray data sets. Although many statistical methods have been proposed for data integration, the genome-wide concordance of different data sets has not been well considered in the analysis. RESULTS: Before considering data integration, it is necessary to evaluate the genome-wide concordance so that misleading results can be avoided. Based on the test results, different subsequent actions are suggested. The evaluation of genome-wide concordance and the data integration can be achieved based on the normal distribution based mixture models. CONCLUSION: The results from our simulation study suggest that misleading results can be generated if the genome-wide concordance issue is not appropriately considered. Our method provides a rigorous parametric solution. The results also show that our method is robust to certain model misspecification and is practically useful for the integrative analysis of differential expression.

The autoimmune regulator directly controls the expression of genes critical for thymic epithelial function.

The autoimmune regulator (Aire) gene plays an essential role in negative selection of T cells and deletion of autoreactive T cells in the thymus. The defect in thymic selection in Aire(-/-) mice was attributed to the repressed expression of tissue-specific Ags in the thymic epithelial cells and defective Ag presentation; however, the molecular mechanism underlying these functions has been elusive. Using the chromatin immunoprecipitation technique, we demonstrate here that Aire binds in vivo to specific DNA sequence motifs and directly regulates thymic expression of genes important for thymic functions including expression of autoantigens, cytokines, transcription factors, and posttranslational modifiers. These results unambiguously established Aire as a key transcriptional regulator of the immune system.

Early pathogenic events associated with Sjögren's syndrome (SjS)-like disease of the NOD mouse using microarray analysis.

Recently, we reported development of the C57BL/6.NOD-Aec1Aec2 mouse carrying two genetic intervals derived from the NOD mouse. These two genetic regions confer full Sjögren's syndrome (SjS)-like disease in SjS-non-susceptible C57BL/6 mice. The current study was undertaken to apply microarray technology to define the molecular basis underlying onset of SjS-disease in C57BL/6.NOD-Aec1Aec2 mice. Using oligonucleotide microarrays, gene expression profiles of submandibular glands derived from 8- to 12-week-old C57BL/6.NOD-Aec1Aec2 mice and 8-week-old C57BL/6 mice were performed for comparison. Significant differential expressions were determined using the Mann-Whitney U test. Hybridizations using submandibular cDNA probes revealed 75 differentially expressed genes at 8 weeks and 105 differentially expressed genes at 12 weeks of age in C57BL/6.NOD-Aec1Aec2 mice compared to 8-week-old C57BL/6 mice. These genes were related generally to basic cellular activities such as transcription, translation, DNA replication, and protein folding. During the predisease phase, genes upregulated encode proteins associated with the IFN-gamma signal-transduction-pathway (Jak/Stat1), TLR-3 (Irf3 and Traf6) and apoptosis (casp11 and casp3), indicative of chronic proinflammatory stimuli, especially IL-1. Between 8 and 12 weeks of age, sets of clustered genes were upregulated that are associated with adaptive immune responses, especially B cell activation, proliferation and differentiation (Baffr, Taci, Bcma, Blys, April, CD70, CD40L, Traf1, Traf3, Pax5, c-Jun, Elk1 and Nf-kB), and neural receptors (Taj/Troy). Altered gene expressions of TLR3 and TNF-superfamily-receptors and ligands during this early phase of SjS suggest a possible viral etiology in the altered glandular homeostasis with an upregulated, possibly overstimulated, B-lymphocyte activation in the early autoimmune response present in the submandibular glands. The importance of NF-kappaB as a critical signal transduction pathway is also suggested but its link is not yet clear.

The autoimmune regulator (Aire) controls iNKT cell development and maturation.

The mechanism underlying the autoimmune polyglandular syndrome type-1 (APS1) has been attributed to defective T-cell negative selection resulting from reduced expression and presentation of autoantigens in thymic medullary epithelial cells (MECs). It has also been postulated that Aire is involved in development of regulatory T cells, although supporting evidence is lacking. Here we show that expression of Aire in MECs is required for development of iNKT cells, suggesting a role for iNKT cells in APS1.

Alterations of renal phenotype and gene expression profiles due to protein overload in NOD-related mouse strains.

BACKGROUND: Despite multiple causes, Chronic Kidney Disease is commonly associated with proteinuria. A previous study on Non Obese Diabetic mice (NOD), which spontaneously develop type 1 diabetes, described histological and gene expression changes incurred by diabetes in the kidney. Because proteinuria is coincident to diabetes, the effects of proteinuria are difficult to distinguish from those of other factors such as hyperglycemia. Proteinuria can nevertheless be induced in mice by peritoneal injection of Bovine Serum Albumin (BSA). To gain more information on the specific effects of proteinuria, this study addresses renal changes in diabetes resistant NOD-related mouse strains (NON and NOD.B10) that were made to develop proteinuria by BSA overload. METHODS: Proteinuria was induced by protein overload on NON and NOD.B10 mouse strains and histology and microarray technology were used to follow the kidney response. The effects of proteinuria were assessed and subsequently compared to changes that were observed in a prior study on NOD diabetic nephropathy. RESULTS: Overload treatment significantly modified the renal phenotype and out of 5760 clones screened, 21 and 7 kidney transcripts were respectively altered in the NON and NOD.B10. Upregulated transcripts encoded signal transduction genes, as well as markers for inflammation (Calmodulin kinase beta). Down-regulated transcripts included FKBP52 which was also down-regulated in diabetic NOD kidney. Comparison of transcripts altered by proteinuria to those altered by diabetes identified mannosidase 2 alpha 1 as being more specifically induced by proteinuria. CONCLUSION: By simulating a component of diabetes, and looking at the global response on mice resistant to the disease, by virtue of a small genetic difference, we were able to identify key factors in disease progression. This suggests the power of this approach in unraveling multifactorial disease processes.

A functional variant of SUMO4, a new I kappa B alpha modifier, is associated with type 1 diabetes.

Previous studies have suggested more than 20 genetic intervals that are associated with susceptibility to type 1 diabetes (T1D), but identification of specific genes has been challenging and largely limited to known candidate genes. Here, we report evidence for an association between T1D and multiple single-nucleotide polymorphisms in 197 kb of genomic DNA in the IDDM5 interval. We cloned a new gene (SUMO4), encoding small ubiquitin-like modifier 4 protein, in the interval. A substitution (M55V) at an evolutionarily conserved residue of the crucial CUE domain of SUMO4 was strongly associated with T1D (P = 1.9 x 10(-7)). SUMO4 conjugates to I kappa B alpha and negatively regulates NF kappa B transcriptional activity. The M55V substitution resulted in 5.5 times greater NF kappa B transcriptional activity and approximately 2 times greater expression of IL12B, an NF kappa B-dependent gene. These findings suggest a new pathway that may be implicated in the pathogenesis of T1D.

Gene expression profiles define a key checkpoint for type 1 diabetes in NOD mice.

cDNA microarrays with >11,000 cDNA clones from an NOD spleen cDNA library were used to identify temporal gene expression changes in NOD mice (1-10 weeks), which spontaneously develop type 1 diabetes, and changes between NOD and NOD congenic mice (NOD.Idd3/Idd10 and NOD.B10Sn-H2(b)), which have near zero incidence of insulitis and diabetes. The expression profiles identified two distinct groups of mice corresponding to an immature (1-4 weeks) and mature (6-10 weeks) state. The rapid switch of gene expression occurring around 5 weeks of age defines a key immunological checkpoint. Sixty-two known genes are upregulated, and 18 are downregulated at this checkpoint in the NOD. The expression profiles are consistent with increased antibody production, antigen presentation, and cell proliferation associated with an active autoimmune response. Seven of these genes map to confirmed diabetes susceptibility regions. Of these seven, three are excellent candidate genes not previously implicated in type 1 diabetes. Ten genes are differentially expressed between the NOD and congenic NOD at the immature stage (Hspa8, Hif1a, and several involved in cellular functions), while the other 70 genes exhibit expression differences during the mature (6-10 week) stage, suggesting that the expression differences of a small number of genes before onset of insulitis determine the disease progression.

Molecular pathways altered by insulin b9-23 immunization.

Type 1 diabetes (T1D) in the nonobese diabetic (NOD) mouse can be delayed by administration of insulin or specific insulin peptides. To better understand how insulin treatment delays diabetes development, NOD mice treated with an insulin peptide (B9-23) were compared with age-matched NOD and NOD congenic mice for gene expression changes in spleen using cDNA microarray. Fifty genes were identified that were significantly altered by B9-23 treatment. Thirty-three of these genes are downregulated by the treatment while they are upregulated during the natural disease progression in NOD from immature (3-4 weeks) to mature (10 weeks) stages. Taken together, our data suggest that the B9-23 treatment, like the protective genes in NOD congenic strains, reduces pro-inflammatory activation of lymphocytes that normally occurs in NOD mice. Furthermore, our studies discovered two genes (Irf4 and Tra1) with increased expression in B9-23-treated mice that promote the Th2 response, providing a molecular basis for the B9-23-protective therapy.

Gene expression profiling during all-trans retinoic acid-induced cell differentiation of acute promyelocytic leukemia cells.

Using cDNA microarrays we determined the gene expression patterns in the human acute promyelocytic leukemia (APL) cell line NB4 during all-trans retinoic acid (ATRA)-induced differentiation. We analyzed the expression of 12,288 genes in the NB4 cells after 12 hours, 24 hours, 48 hours, 72 hours, and 96 hours of ATRA exposure. During this time course, we found 168 up-regulated and more than 179 down-regulated genes, most of which have not been reported before. Many of the altered genes encode products that participate in signaling pathways, cell differentiation, programmed cell death, transcription regulation, and production of cytokines and chemokines. Of interest, the CD52 and protein kinase A regulatory subunit alpha (PKA-Rlalpha) genes, whose products are being used as therapeutic targets for certain human neoplasias in currently ongoing clinical trials, were among the genes observed to be markedly up-regulated after ATRA treatment. The present study provides valuable data to further understand the mechanism of ATRA-induced APL cell differentiation and suggests potential therapeutic alternatives for this leukemia.

Microarray analysis of gene expression in the kidneys of new- and post-onset diabetic NOD mice.

We profiled the expression of 5,760 clones from a kidney subtraction library in the kidneys of three groups of NOD mice: nondiabetic, new-onset, and long-term diabetic. A total of 27 genes had lower expression and 1 gene (Gpx3) had higher expression in the new-onset diabetic mice compared with nondiabetic control NOD mice (P < 0.001). Similarly, 19 of the above 27 genes and 7 additional genes had higher expression and the Gpx3 gene had lower expression in long-term diabetic mice compared with controls (P < 0.001). Interestingly, only three genes may be different between new-onset and long-term diabetic mice (P < 0.0004). These genes are from diverse functional groups, including oxidative phosphorylation, free radical neutralization, channels, pumps, lipid processing, transcription and translation machinery, protein trafficking, constitutive protein processing, and immune function. The majority of these genes fall into four signaling pathways: insulin, transforming growth factor-beta, tumor necrosis factor-alpha, and peroxisome proliferator-activated receptor. The most significant expression change was found for the stearoyl-coenzyme A desaturase 1 (SCD1) gene (P < 10(-7)). The lower expression levels of the SCD1 gene in both diabetic groups compared with controls were further confirmed by Northern blot analysis and immunohistochemistry.

Gene expression associated with interferon alfa antiviral activity in an HCV replicon cell line.

Interferon alfa (IFN-alpha)-based treatment is the only therapeutic option for chronic hepatitis C viral infection. However, the molecular mechanisms of IFN-alpha antiviral activity are not completely understood. The recent development of an HCV replicon cell culture system provides a feasible experimental model to investigate the molecular details of IFN-induced direct antiviral activity in hepatocytes. In this report, we show that IFN-alpha can effectively inhibit HCV subgenomic RNA replication and suppress viral nonstructural protein synthesis. Using cDNA microarray analysis, we also show that the replicon cells have different gene expression profile compared with the parental hepatoma cells (Huh7). IFN-alpha can induce a number of responsive genes in the replicon cells. One of the genes, 6-16 (G1P3), can enhance IFN-alpha antiviral efficacy. In addition, we demonstrate that IFN-alpha can significantly activate STAT3 in hepatoma cells, suggesting that this pathway plays a role in IFN-alpha signaling. In conclusion, our results indicate that IFN-alpha antiviral activity is associated with activation of STAT3-signaling pathway and intracellular gene activation. Our results also suggest that IFN-alpha-induced target genes may play an important role in IFN-alpha anti-HCV activity.

Protein microarrays to detect protein-protein interactions using red and green fluorescent proteins.

Proteomics, the study of protein function on a global scale, will play an important role in furthering our understanding of gene functions, complex biological pathways, and discovery of novel drug targets. A number of techniques have been developed for proteomic studies to identify and analyze proteins, compare protein expression levels, and study protein-protein interactions. Recent developments have applied a DNA array-type approach to immobilize proteins on a surface for high-throughput analysis. Here we report the development and construction of protein chips using derivatized glass and nitrocellulose-coated slides and the employment of recombinant proteins fused with green and red fluorescent proteins for detection. Fluorescent signals were found to be proportional to the amount of arrayed proteins and could be readily detected with a conventional fluorescence slide scanner. This technique allows the investigation of protein-protein interactions without the need for additional labeling steps of probe proteins.