Better bioinformatics through usability analysis.

MOTIVATION: Improving the usability of bioinformatics resources enables researchers to find, interact with, share, compare and manipulate important information more effectively and efficiently. It thus enables researchers to gain improved insights into biological processes with the potential, ultimately, of yielding new scientific results. Usability 'barriers' can pose significant obstacles to a satisfactory user experience and force researchers to spend unnecessary time and effort to complete their tasks. The number of online biological databases available is growing and there is an expanding community of diverse users. In this context there is an increasing need to ensure the highest standards of usability. RESULTS: Using 'state-of-the-art' usability evaluation methods, we have identified and characterized a sample of usability issues potentially relevant to web bioinformatics resources, in general. These specifically concern the design of the navigation and search mechanisms available to the user. The usability issues we have discovered in our substantial case studies are undermining the ability of users to find the information they need in their daily research activities. In addition to characterizing these issues, specific recommendations for improvements are proposed leveraging proven practices from web and usability engineering. The methods and approach we exemplify can be readily adopted by the developers of bioinformatics resources.

Beta-2-glycoprotein specificity of human anti-phospholipid antibody resides on the light chain: a novel mechanism for acquisition of cross-reactivity by an autoantibody.

We have recently shown that the anti-cardiolipin activity of human anti-phospholipid antibody UK4 (lambda) resides on its heavy chain. We now show that UK4 possesses strong reactivity to the plasma-protein beta2-Glycoprotein I (beta2-GPI) also. Utilizing chain shuffling experiments involving an unrelated anti-p185 antibody 4D5 (kappa) with no reactivity to beta2-GPI, we now demonstrate that both the constructs possessing the auto-antibody-derived light chain exhibited significant binding to beta2-GPI. However, the construct possessing UK4 heavy chain in association with 4D5 light chain, exhibited no anti-beta2-GPI activity. Furthermore, there was a low increase (approximately 10%) in the binding of UK4 to cardiolipin in the presence of beta2-GPI. The results demonstrate that anti-beta2-GPI activity resides on UK4 light chain and, importantly, this activity could be transferred to a novel antibody construct via the light chain alone. Computer-generated models of the three-dimensional structures of UK4 and its hybrids, suggest predominant interaction of UK4 light chain with domain IV of beta2-GPI. Molecular docking experiments highlight a number of potential sites on beta2-GPI for interaction of UK4 and indicate as to how beta2-GPI recognition may occur primarily via the autoantibody light chain. The study provides first demonstration of the occurrence of anti-phospholipid and anti-beta2-GPI activities separately on heavy and light chains of an autoantibody. The possible mechanisms that such antibodies may employ to recognise their antigens, are discussed.

Progression of atherosclerosis is associated with variation in the alpha1-antitrypsin gene.

OBJECTIVE: Alpha1-Antitrypsin (AAT) protects elastic tissue and may play a role in atherogenesis. The association of atherosclerosis progression with common AAT variants was considered in 2 clinical trials. METHODS AND RESULTS: We examined the association of AAT V213A, S and Z deficiency alleles, and the functional 3' UTR 11478G>A with change in minimal luminal diameter, a measure of focal disease, in the Lopid Coronary Angiography Trial gemfibrozil study of post-bypass men. S or Z carriers (n=14) showed strong progression of disease on placebo (11.5%) but responded well to treatment (3% regression). 11478A carriers treated with placebo or gemfibrozil showed significantly more disease progression (n=8, -14.5% and n=16, -4.0%, respectively) than 11478GG men (n=107, -7.0% and n=108, -1.4%, respectively; overall, P=0.003). VV213 men treated with gemfibrozil (n=68) showed -4.8% progression, whereas A213 carriers (n=55) showed +1.4% regression of disease (P=0.001). No V213A effect was seen on placebo (P=0.11). In the Diabetes Atherosclerosis Intervention Study fenofibrate trial of angiographic progression in type 2 diabetes, the association of 11478A with increased disease progression was confirmed in the treatment group, but not the gemfibrozil-treated A213 association with regression, suggesting a pharmacogenetic difference. CONCLUSIONS: Disease progression is associated with variation in AAT, and low AAT levels promote atherogenesis.

Anti-cardiolipin/beta-2 glycoprotein activities co-exist on human anti-DNA antibody light chains.

We have recently shown that the human anti-DNA antibodies B3 and 33H11 also bind cardiolipin and that the anti-autoantigen activity resides predominantly on their lambda light chains. We now show that the two auto-antibodies possess strong reactivity to the plasma-protein 2-Glycoprotein I (beta2-GPI) also. Utilizing chain shuffling experiments involving an unrelated anti-p185 antibody 4D5 with insignificant reactivity to cardiolipin or to beta2-GPI, we now demonstrate that hybrid Fabs with constituent light chain, but not the heavy chain, of B3 or 33H11, exhibit anti-cardiolipin activity. Furthermore, the constructs possessing the auto-antibody-derived light chain also exhibited significant reactivity to beta2-GPI. The results suggest that anti-DNA, anti-cardiolipin and anti-beta2-GPI activities co-exist on the light chains of the antibodies studied and, importantly, these activities could be transferred to antibody constructs by their light chains alone. Computer-generated models of the three-dimensional structures of the auto-antibodies and their hybrids, suggest predominant interaction of their light chains with domain IV of beta2-GPI.

Somatic mutations to arginine residues affect the binding of human monoclonal antibodies to DNA, histones, SmD and Ro antigen.

Autoantibodies to a wide variety of antigens are associated with systemic lupus erythematosus (SLE). Antibodies to double-stranded DNA (anti-dsDNA) are thought to be particularly closely related to tissue damage and disease activity in SLE. Autoantibodies to histones, Sm and Ro are found in patients with SLE, but their role in pathogenesis is unclear. Using a transient expression system, we previously showed that particular sequence motifs in CDRs of light chains derived from the human Vlambda gene 2a2 are very important in determining their ability to form a DNA-binding site, when paired with the heavy chain of the human monoclonal anti-dsDNA antibody B3. These motifs are often sites of somatic mutation and/or contain arginine residues. In the experiments reported in this paper, the same expression system was used to show that these CDR motifs also affect binding to histones, Ro antigen and Sm antigen, but that binding to different antigens is affected in diverse ways by particular changes in the sequence of the CDRs. The heavy chain also plays a role in binding to these antigens. Pairing of the same range of 11 2a2 derived light chains with the heavy chain of a different anti-DNA antibody, 33.H11, gave reduced ability to bind DNA in comparison with the results obtained using the B3 heavy chain. Computer-generated models of the three-dimensional structures of these heavy/light chain combinations were used to define the positions occupied by the important sequence motifs at the binding sites of these antibodies, and to explain the different effects exerted by arginine residues at different positions in the light chains.

Relative importance of different human aPL derived heavy and light chains in the binding of aPL to cardiolipin.

INTRODUCTION: Previous studies have suggested the importance of somatic mutations and arginine, asparagine and lysine residues in the complementarity determining regions (CDRs) of antiphospholipid antibodies (aPL) implicated in the pathogenesis of the antiphospholipid antibody syndrome. The relative contributions of the heavy and light chains of aPL in binding to cardiolipin (CL) were assessed by pairing the heavy and light chains of two IgG, beta(2)GPI dependent aPL (IS4 and CL24) with different partner chains from other IgG, beta(2)GPI independent aPL (UK4) and anti-DNA antibodies (B3 and 33H11). METHODS: Four heavy (V(H)) and five light (V(L)) chain variable sequences from three aPL and two anti-DNA antibodies were cloned into expression vectors containing appropriate gamma(1), lambda or kappa constant region cDNA. Paired combinations of heavy and light chain expression plasmids were transfected into COS-7 cells allowing transient expression of whole IgG molecules, which were harvested and tested for the ability to bind CL and DNA by enzyme-linked immunosorbant assay (ELISA). RESULTS: Whole IgG was produced from 19 heavy/light chain combinations. IS4V(H) was dominant in conferring the ability to bind CL with four of the five V(L) tested. The identity of the V(L) region paired with IS4V(H) was important in determining the strength of binding to CL. IS4V(H) contains multiple arginine residues in CDR3, which may have accumulated due to antigen driven selection. It is likely that these arginine residues may interact with CL. The combination B3V(H)/B3V(L) also bound CL, but none of the other 14 combinations showed any binding in this assay. CONCLUSION: Whole IgG molecules capable of binding CL were produced by in vitro expression in COS-7 cells. Arginine residues play important roles in binding to CL and double-stranded DNA. However, different patterns of mutation to arginine are associated with binding to each of these antigens.

Novel sites in the p65 subunit of NF-kappaB interact with TFIIB to facilitate NF-kappaB induced transcription.

Nuclear factor kappaB (NF-kappaB) transcription factors regulate a large number of genes in response to inflammation, infection and stressful conditions. In this study, we investigated whether NF-kappaB p65 regulates the transcription of target genes by interacting with components of the basal transcription machinery. We examined the interaction of p65 with the basal transcription factor IIB (TFIIB). Glutathione S-transferase pull down assays showed that the Rel homology domain of p65 is important for binding to TFIIB. Molecular modelling, together with the generation of specific point mutants, revealed that residues 41 R and 42 S in the Rel homology domain of p65 facilitate the interaction with TFIIB. Mutation of these residues showed a decrease in p65 induced transcription, suggesting that they are involved in a functional interaction with TFIIB.

A systematic analysis of sequences of human antiphospholipid and anti-beta2-glycoprotein I antibodies: the importance of somatic mutations and certain sequence motifs.

OBJECTIVE: Previous studies have suggested the importance of somatic mutations and certain residues in the complementarity determining regions (CDRs) of antiphospholipid antibodies (aPL) implicated in the pathogenesis of antiphospholipid antibody syndrome (APS). The authors tested this hypothesis by carrying out a systematic analysis of all published aPL sequences. METHODS: Each aPL variable region sequence was aligned to the closest germline counterpart in the VBASE Sequence Directory by using DNAPLOT software, allowing analysis of nucleotide homology and distribution of somatic mutations. The probability that this distribution arose as a result of antigen-driven accumulation of replacement mutations in the CDRs was tested statistically. RESULTS: There was no preferential gene or family use in the 36 aPL sequences identified. Immunoglobulin (Ig) M aPL had few somatic mutations compared with IgG. Of the IgG aPL, 9 of 14 showed evidence of antigen-driven accumulation of replacement mutations in the CDRs. Multinomial analysis allowed a clearer statistical identification of sequences that had been subject to antigen drive. The more specific IgM aPL and some IgG aPL displayed an accumulation of arginine, asparagine, and lysine residues in CDRs. CONCLUSIONS: High-specificity binding in IgG aPL, but not in more specific IgM aPL, is conferred by antigen-driven somatic mutation. This may in part be caused by an accumulation of arginine, asparagine, and lysine residues in the CDRs, which are germlines encoded in the more specific IgM aPL, but often arise because of somatic mutation in IgG aPL. RELEVANCE: An understanding of the role of arginine, asparagine, and lysine residues in the binding of pathogenic aPL to phospholipids, and to beta(2)-glycoprotein I, may eventually help in the development of drugs to interfere with those interactions, and thereby improve the treatment of antiphospholipid antibody syndrome.

Predicting response to radioimmunotherapy from the tumor microenvironment of colorectal carcinomas.

Solid tumors have a heterogeneous pathophysiology, which directly affects antibody-targeted therapies. Here, we consider the influence of selected tumor parameters on radioimmunotherapy, by comparing the gross biodistribution, microdistribution, and therapeutic efficacy of either radiolabeled or fluorescently labeled antibodies (A5B7 anti-carcinoembryonic antigen antibody and a nonspecific control) after i.v. injection in two contrasting human colorectal xenografts in MF1 nude mice. The LS174T is moderately/poorly differentiated, whereas SW1222 has a well-differentiated glandular structure. Biodistribution studies (1.8 MBq (131)I-labeled A5B7, four mice per group) showed similar gross tumor uptake at 48 h in the two models (25.1% and 24.0% injected dose per gram, respectively). However, in therapy studies (six mice per group), LS174T required a 3-fold increase in dose (18 versus 6 MBq) to equal SW1222 growth inhibition ( approximately 55 versus approximately 60 days, respectively). To investigate the basis of this discrepancy, high-resolution multifluorescence microscopy was used to study antibody localization in relation to tumor parameters (5 min, 1 and 24 h, four mice per time point). Three-dimensional microvascular corrosion casting and transmission electron microscopy showed further structural differences between xenografts. Vascular supply, overall antigen distribution, and tumor structure varied greatly between models, and were principally responsible for major differences in antibody localization and subsequent therapeutic efficacy. The study shows that multiparameter, high-resolution imaging of both therapeutic and tumor microenvironment is required to comprehend complex antibody-tumor interactions, and to determine which tumor regions are being successfully treated. This will inform the design of optimized clinical trials of single and combined agents, and aid individual patient selection for antibody-targeted therapies.

MicroCore: mapping genome expression to cell pathways and networks.

The MicroCore toolkit is a suite of analysis programs for microarray and proteomics data that is open source and programmed exclusively in Java. MicroCore provides a flexible and extensible environment for the interpretation of functional genomics data through visualization. The first version of the application (downloadable from the MicroCore website: http://www.ucl.ac.uk/oncology/MicroCore/microcore.htm), implements two programs-PIMs (protein interaction maps) and MicroExpress-and is soon to be followed by an extended version which will also feature a fuzzy k-means clustering application and a Java-based R plug-in for microarray analysis. PIMs and MicroExpress provide a simple yet powerful way of graphically relating large quantities of expression data from multiple experiments to cellular pathways and biological processes in a statistically meaningful way.

Engineering antibodies for clinical applications in cancer.

The 'magic bullet' concept predicted over a century ago that antibodies would be used to target cancer therapy. Since then initial problems that were related to specificity, purity and immungenicity of antibody-based reagents have slowly been overcome due to developments in technology and increased knowledge. As a result, antibodies are in use for many clinical applications and now comprise the second largest category of medicines in clinical development after vaccines. For antibody-based cancer therapeutics the last 20 years have met with an explosion of knowledge about the biology of the disease and potential targets as well as new technology which allows cloning and manipulation of multifunctional antibody-based molecules. However, the focus still remains on developing therapeutics that will have potential for treating cancer in people and this is efficiently assessed in mechanistic clinical trials that feed back to the laboratory for further development. This review illustrates the mechanistic approach to making new molecules for antibody imaging and therapy of cancer. It is illustrated by examples of radioimmunotherapy and antibody-directed enzyme prodrug therapy developed by the authors.

Key residues of a major cytochrome P4502D6 epitope are located on the surface of the molecule.

Eukaryotically expressed CYP2D6 is the universal target of liver kidney microsomal Ab type 1 (LKM1) in both type 2 autoimmune hepatitis (AIH) and chronic hepatitis C virus (HCV) infection. In contrast, reactivity to prokaryotically expressed CYP2D6 protein and synthetic peptides is significantly lower in HCV infection than in AIH. The aim of the present study was to characterize LKM1 reactivity against a panel of eukaryotically expressed CYP2D6 constructs in the two conditions. LKM1-positive sera obtained from 16 patients with AIH and 16 with HCV infection were used as probes to perform a complete epitope mapping of CYP2D6. Reactivity to the full-length protein and 16 constructs thereof was determined by radioligand assay. We found that antigenicity is confined to the portion of the molecule C-terminal of aa 193, no reactivity being detectable against the aa sequence 1-193. Reactivity increases stepwise toward the C-terminal in both AIH and HCV, but the frequency of reactivity in the two conditions differs significantly between aa 267-337. To further characterize this region, we introduced a five and a three amino acid swap mutation selected from the homologous regions of CYP2C9 and HCV. This maneuver resulted in a substantial loss of LKM1 binding in both conditions, suggesting that this region contains a major epitope. Molecular modeling revealed that CYP2D6(316-327) is exposed on the surface of the protein, and may represent a key target for the autoantibody. These findings provide an initial characterization of the antigenic constitution of the target of LKM1 in AIH and HCV infection.