Search

Biomarker development for prediction of patient response to therapy is one of the goals of molecular profiling of human tissues. Due to the large number of transcripts, relatively limited number of samples, and high variability of data, identification of predictive biomarkers is a challenge for data analysis. Furthermore, many genes may be responsible for drug response differences, but often only a few are sufficient for accurate prediction. Here we present an analysis approach, the Convergent Random Forest (CRF) method, for the identification of highly predictive biomarkers. The aim is to select from genome-wide expression data a small number of non-redundant biomarkers that could be developed into a simple and robust diagnostic tool. Our method combines the Random Forest classifier and gene expression clustering to rank and select a small number of predictive genes. We evaluated the CRF approach by analyzing four different data sets. The first set contains transcript profiles of whole blood from rheumatoid arthritis patients, collected before anti-TNF treatment, and their subsequent response to the therapy. In this set, CRF identified 8 transcripts predicting response to therapy with 89% accuracy. We also applied the CRF to the analysis of three previously published expression data sets. For all sets, we have compared the CRF and recursive support vector machines (RSVM) approaches to feature selection and classification. In all cases the CRF selects much smaller number of features, five to eight genes, while achieving similar or better performance on both training and independent testing sets of data. For both methods performance estimates using cross-validation is similar to performance on independent samples. The method has been implemented in R and is available from the authors upon request: Jadwiga.Bienkowska@biogenidec.com.

The 620W allelic variant of the intracellular tyrosine phosphatase, PTPN22, is associated with a number of different autoimmune disorders, and this provides direct evidence for common mechanisms underlying many of these diseases. The associated allele appears to influence thresholds for T cell receptor signaling, and a variety of disease models involving both central and peripheral tolerance can be proposed. However, given the fact that PTPN22 is expressed in a variety of immunologically relevant cell types, the precise mechanisms for these associations remain unclear. In general, the PTPN22 620W allele appears to play a role in autoimmune disorders that have a prominent humoral component, suggesting that further investigation of PTPN22 activity in B cells will be useful. From a genetic perspective, the data highlights the genetic heterogeneity underlying autoimmunity in different ethnic groups.

We have completed a genome wide linkage scan using >5700 informative single-nucleotide polymorphism (SNP) markers (Illumina IV SNP linkage panel) in 642 Caucasian families containing affected sibling pairs with rheumatoid arthritis (RA), ascertained by the North American Rheumatoid Arthritis Consortium. The results show striking new evidence of linkage at chromosomes 2q33 and 11p12 with logarithm of odds (LOD) scores of 3.52 and 3.09, respectively. In addition to a strong and broad linkage interval surrounding the major histocompatibility complex (LOD>16), regions with LOD>2.5 were observed on chromosomes 5 and 10. Additional linkage evidence (LOD scores between 1.46 and 2.35) was also observed on chromosomes 4, 7, 12, 16 and 18. This new evidence for multiple regions of genetic linkage is partly explained by the significantly increased information content of the Illumina IV SNP linkage panel (75.6%) compared with a standard microsatellite linkage panel utilized previously (mean 52.6%). Stratified analyses according to whether or not the sibling pair members showed elevated anticyclic citrullinated peptide titers indicates significant variation in evidence for linkage among strata on chromosomes 4, 5, 6 and 7. Overall, these new linkage data should reinvigorate efforts to utilize positional information to identify susceptibility genes for RA.

We carried out gene expression profiling of peripheral blood mononuclear cells (PBMCs) in 29 patients with active rheumatoid arthritis (RA) and 21 control subjects using Affymetrix U95Av2 arrays. Using cluster analysis, we observed a significant alteration in the expression pattern of 81 genes (P

Monitoring of gene and protein expression in peripheral blood cells has significant potential for improving the diagnosis and therapy of many human diseases. As genomic-scale microarray and proteomic technologies are applied to peripheral blood, it is important to consider the variables that may affect interpretation of data. Here we report experiments performed to identify genes that are particularly sensitive to ex vivo handling prior to RNA extraction for gene expression microarrays or quantitative real-time RT-PCR assays. We examined Affymetrix gene expression in samples from eight normal individuals where blood was processed for RNA either immediately after blood draw or the next day following overnight incubation. These studies identified hundreds of genes that are sensitive to ex vivo handling of blood, and suggest that this is an important variable to consider when designing and interpreting human PBMC experiments.

The development of effective adjuvant therapies for the treatment of high-risk melanoma patients is critical for the prevention of metastatic disease and improvement of patient survival. Active specific immunotherapy has been tested as an adjuvant treatment in numerous clinical trials with overall limited, but occasionally promising, success rates. Newcastle disease virus (NDV) oncolysate has been utilized as an adjunctive immunotherapeutic agent in the postsurgical management of these patients. A phase II study initiated in 1975 using adjuvant vaccine therapy composed of allogeneic and autologous human melanoma cells infected with live NDV (NDV oncolysate) in patients with AJCC stage III melanoma following therapeutic lymph node dissection has shown >60% survival rate at 10 years with no adverse effects. Continued long-term analysis of trials with promising early results as well as assessment of immunologic responses generated in these patients may result in improved therapeutic decisions for clinical trials in the future. We analyzed the 15-year survival of patients treated postsurgically with NDV oncolysate in the phase II study described above. In an attempt to understand the immunological effects of this treatment, we have also carried out a comprehensive analysis of the peripheral blood T cell repertoire in these patients. The overall 15-year survival of this group of patients is 55%. Previous studies have suggested that improved outcome in patients undergoing immunotherapy is correlated with increased numbers of CD8+CD57+ cells. In surviving patients, we observed a striking oligoclonality in the CD8+ T cell population in peripheral blood, which reflects clonal expansions in the CD8+CD57+ subset. The data suggest that adjuvant vaccination with NDV oncolysates is associated with prolonged survival of patients with lymph node-positive malignant melanoma and that CD8+ T cells may be an important component of therapeutic efficacy.

Griffin et al., respond: In January 2011, we reported the unique CD20+CD27+CD43+ phenotype of human B1 cells (Griffin et al., 2011a). In December 2011, we reported that the human B1 cell pool consists of functionally distinct CD11b− and CD11b+ populations, with the former particularly proficient at antibody secretion and the latter particularly efficient in stimulating T cells (Griffin and Rothstein, 2011). We also noted that the frequency of CD11b+ B1 cells is increased in the circulation of patients with lupus. As part of this work, we included microarray data that pointed to transcriptional differences between CD11b− and CD11b+ B1 cells. We have read with great interest the letter by Reynaud and Weill, who have mined the microarray data we deposited in the National Center for Biotechnology Information (NCBI) repository and drawn inferences, incorrectly as it turns out, regarding the nature and origin of CD11b− and CD11b+ B1 cells. Reynaud and Weill speculated that CD11b− B1 cells represent B cell–T cell doublets because this population, in aggregate, expressed genes characteristic of T cells along with genes characteristic of B cells. The potential issue of B cell–T cell doublets, raised previously by these same authors (Descatoire et al., 2011), is a topic we have already addressed by demonstrating that our sort-purified B1 cells exist as singlets (Griffin and Rothstein, 2011; Griffin et al., 2011b). At the same time, we identified procedures that can be used to circumvent the potential problem of doublet events (Griffin et al., 2011b). In our published work, we reported that B1 cells sort-purified without CD3− gating contained an estimated 1–3% CD3+ events (Griffin and Rothstein, 2011; Griffin et al., 2011b). We have now sort-purified CD11b− B1 cells again, exactly as was done to generate our published microarray results, and have analyzed three independent postsort samples for CD3 expression, along with purified T cells (Fig. 1 A). We again find a very small proportion of CD3+ events (1.9 ± 0.66%; mean ± SEM) in the CD11b− B1 cell population. This very low level of CD3 positivity is likely attributable to a minor degree of T cell contamination that could account for the appearance of very low levels of some characteristically T cell transcripts among CD11b− B1 cells. In keeping with this, flow cytometric analysis of unsorted CD19-enriched PBMCs demonstrates a small degree of CD2, CD7, CD8, and CD3 staining among CD11b− B1 cells that is eliminated by CD3 exclusion and subsequent analysis of CD3−CD20+CD27+CD43+CD11b− B1 cells (Fig. 1, B and C). Importantly, when CD3 exclusion is combined with sort purification, expression of T cell genes such as CD7, CD8, and granzyme A is essentially eliminated (to

Genotype-expression association analysis using linear regression may produce different test results depending on whether founders only or all pedigreed members are used. This difference is not due to the correlation of samples within a pedigree, because linear mixed models have been applied to account for that correlation. We investigated the possibility that the difference is due to a dependence of expression levels on, among other things, the generation number in the pedigree. Indeed, of the 30 or so studied expression quantitative traits, several of them show significant dependence on the generation number. We propose to use all pedigree members in genotype-expression association analyses whenever the complete genotyping information is available.

Contains fulltext : 70190.pdf (Publisher’s version ) (Open Access) The prediction of response (or non-response) to anti-TNF treatment for rheumatoid arthritis (RA) is a pressing clinical problem. We conducted a genome-wide association study using the Illumina HapMap300 SNP chip on 89 RA patients prospectively followed after beginning anti-TNF therapy as part of Autoimmune Biomarkers Collaborative Network (ABCoN [Autoimmune Bio-markers Collaborative Network]) patient cohort. Response to therapy was determined by the change in Disease Activity Score (DAS28) observed after 14 wks. We used a two-part analysis that treated the change in DAS28 as a continuous trait and then incorporated it into a dichotomous trait of "good responder" and "nonresponder" by European League Against Rheumatism (EULAR) criteria. We corrected for multiple tests by permutation, and adjusted for potential population stratification using EIGENSTRAT. Multiple single nucleotide polymorphism (SNP) markers showed significant associations near or within loci including: the v-maf musculoaponeurotic fibrosarcoma oncogene homolog B (MAFB) gene on chromosome 20; the type I interferon gene IFNk on chromosome 9; and in a locus on chromosome 7 that includes the paraoxonase I (PON1) gene. An SNP in the IL10 promoter (rs1800896) that was previously reported as associated with anti-TNF response was weakly associated with response in this cohort. Replications of these results in independent and larger data sets clearly are required. We provide a reference list of candidate SNPs (P < 0.01) that can be investigated in future pharmacogenomic studies.

The current methods for molecular typing of HLA-DR alleles incur a substantial financial burden when performing large population studies. In the current study, we aimed to provide much less expensive typing approach with high predictability for DRB1 genotype. We have used a panel of three microsatellite markers in the class II region (D6S2666, D6S2665 and D6S2446) for genotyping and haplotype reconstruction in a total of 1687 Caucasian (1313 RA patients and 374 controls) and 1364 Korean individuals (744 RA patients and 620 controls), all of whom were previously genotyped for DRB1. We found that a total of 88.4 and 87.4% of all observed three-marker haplotypes could determine the DR type with a positive predictive value >0.8 with high sensitivity and specificity. There was a high degree of haplotype conservation when comparing Caucasian and Asian populations. Interestingly, we found that the majority of DRB1*09 and DRB1*10 alleles share a common three-marker haplotype in both Caucasian and Asian populations. This is unexpected, since these two alleles are found on very different haplotype families. In addition, these two alleles are both associated with rheumatoid arthritis, making the elucidation of these haplotype relationships potentially important for understanding disease susceptibility.

Psoriatic arthritis (PsA) is a chronic and erosive form of arthritis of unknown cause. We aimed to characterize the PsA phenotype using gene expression profiling and comparing it with healthy control subjects and patients rheumatoid arthritis (RA). Peripheral blood cells (PBCs) of 19 patients with active PsA and 19 age- and sex-matched control subjects were used in the analyses of PsA, with blood samples collected in PaxGene tubes. A significant alteration in the pattern of expression of 313 genes was noted in the PBCs of PsA patients on Affymetrix U133A arrays: 257 genes were expressed at reduced levels in PsA, and 56 genes were expressed at increased levels, compared with controls. Downregulated genes tended to cluster to certain chromosomal regions, including those containing the psoriasis susceptibility loci PSORS1 and PSORS2. Among the genes with the most significantly reduced expression were those involved in downregulation or suppression of innate and acquired immune responses, such as SIGIRR, STAT3, SHP1, IKBKB, IL-11RA, and TCF7, suggesting inappropriate control that favors proin-flammatory responses. Several members of the MAPK signaling pathway and tumor suppressor genes showed reduced expression. Three proinflammatory genes--S100A8, S100A12, and thioredoxin--showed increased expression. Logistic regression and recursive partitioning analysis determined that one gene, nucleoporin 62 kDa, could correctly classify all controls and 94.7% of the PsA patients. Using a dataset of 48 RA samples for comparison, the combination of two genes, MAP3K3 followed by CACNA1S, was enough to correctly classify all RA and PsA patients. Thus, PBC gene expression profiling identified a gene expression signature that differentiated PsA from RA, and PsA from controls. Several novel genes were differentially expressed in PsA and may prove to be diagnostic biomarkers or serve as new targets for the development of therapies.

We genotyped 525 independent North American white individuals with systemic lupus erythematosus (SLE) for the PTPN22 R620W polymorphism and compared the results with data generated from 1,961 white control individuals. The R620W SNP was associated with SLE (genotypic P=.00009), with estimated minor (T) allele frequencies of 12.67% in SLE cases and 8.64% in controls. A single copy of the T allele (W620) increases risk of SLE (odds ratio [OR]=1.37; 95% confidence interval [CI] 1.07–1.75), and two copies of the allele more than double this risk (OR=4.37; 95% CI 1.98–9.65). Together with recent evidence showing association of this SNP with type 1 diabetes and rheumatoid arthritis, these data provide compelling evidence that PTPN22 plays a fundamental role in regulating the immune system and the development of autoimmunity.

Rheumatoid arthritis (RA) is the most common systemic autoimmune disease, affecting approximately 1% of the adult population worldwide, with an estimated heritability of 60%. To identify genes involved in RA susceptibility, we investigated the association between putative functional single-nucleotide polymorphisms (SNPs) and RA among white individuals by use of a case-control study design; a second sample was tested for replication. Here we report the association of RA susceptibility with the minor allele of a missense SNP in PTPN22 (discovery-study allelic P=6.6 x 10(-4); replication-study allelic P=5.6 x 10(-8)), which encodes a hematopoietic-specific protein tyrosine phosphatase also known as "Lyp." We show that the risk allele, which is present in approximately 17% of white individuals from the general population and in approximately 28% of white individuals with RA, disrupts the P1 proline-rich motif that is important for interaction with Csk, potentially altering these proteins' normal function as negative regulators of T-cell activation. The minor allele of this SNP recently was implicated in type 1 diabetes, suggesting that the variant phosphatase may increase overall reactivity of the immune system and may heighten an individual carrier's risk for autoimmune disease.

To the editor: We recently reported that human B-1 cells express the phenotype CD20+CD27+CD43+CD70−.[1][1],[2][2] Covens et al[3][3] now assert that B cells with this phenotype resemble plasmablasts and thus represent preplasmablasts. Their conclusion rests to a large extent on microarray data