Background

Recognition of pathogens by dendritic cells (DCs) through interaction with pattern recognition receptors, including Toll like receptors (TLR), is crucial for the initiation of appropriate polarized T helper (Th) cell responses. Yet, the characteristics and differences in molecular profiles of DCs with different T cell polarizing capacities are still poorly defined. To address this issue, the molecular profile of human monocyte derived DCs was characterized after exposure to TLR4 ligand LPS in combination with the Th1 promoting bacterial extracts from Listeria monocytogenes and Escherichia coli or the Th2 promoting helminth derived phospholipids from Schistosoma mansoni and Ascaris lumbricoides, all with TLR2 activating capacity.

Results

With regard to the signalling pathways activated upon exposure to LPS and the TLR2 activating compounds, we find that the ratio of activated Mitogen Activated Protein Kinases (MAPK) p-ERK/p-p38 is lower in DCs stimulated with the bacterial products compared to DCs stimulated with the helminth products, which correlates with the Th1 and Th2 polarizing capacity of these compounds. Furthermore, analysis of the mRNA expression levels of a set of 25 carefully selected genes potentially involved in modulation of T cell polarization revealed that the mRNA expression of notch ligand delta-4 and transcription factor c-fos are differentially regulated and show a strong correlation with Th1 and Th2 polarization, respectively.

Conclusion

This study shows that combined TLR2 and TLR4 activation in the context of different antigen sources can induce very distinct molecular profiles in DCs and suggests that the Th1/Th2 polarizing capacity of compounds can be predicted with the molecular signature they induce in DCs.

Background

The statistical analysis of immunological data may be complicated because precise quantitative levels cannot always be determined. Values below a given detection limit may not be observed (nondetects), and data with nondetects are called left-censored. Since nondetects cannot be considered as missing at random, a statistician faced with data containing these nondetects must decide how to combine nondetects with detects. Till now, the common practice is to impute each nondetect with a single value such as a half of the detection limit, and to conduct ordinary regression analysis. The first aim of this paper is to give an overview of methods to analyze, and to provide new methods handling censored data other than an (ordinary) linear regression. The second aim is to compare these methods by simulation studies based on real data.

Results

We compared six new and existing methods: deletion of nondetects, single substitution, extrapolation by regression on order statistics, multiple imputation using maximum likelihood estimation, tobit regression, and logistic regression. The deletion and extrapolation by regression on order statistics methods gave biased parameter estimates. The single substitution method underestimated variances, and logistic regression suffered loss of power. Based on simulation studies, we found that tobit regression performed well when the proportion of nondetects was less than 30%, and that taken together the multiple imputation method performed best.

Conclusion

Based on simulation studies, the newly developed multiple imputation method performed consistently well under different scenarios of various proportion of nondetects, sample sizes and even in the presence of heteroscedastic errors.

Background

The number of subjects that can be recruited in immunological studies and the number of immunological parameters that can be measured has increased rapidly over the past decade and is likely to continue to expand. Large and complex immunological datasets can now be used to investigate complex scientific questions, but to make the most of the potential in such data and to get the right answers sophisticated statistical approaches are necessary. Such approaches are used in many other scientific disciplines, but immunological studies on the whole still use simple statistical techniques for data analysis.

Results

The paper provides an overview of the range of statistical methods that can be used to answer different immunological study questions. We discuss specific aspects of immunological studies and give examples of typical scientific questions related to immunological data. We review classical bivariate and multivariate statistical techniques (factor analysis, cluster analysis, discriminant analysis) and more advanced methods aimed to explore causal relationships (path analysis/structural equation modelling) and illustrate their application to immunological data. We show the main features of each method, the type of study question they can answer, the type of data they can be applied to, the assumptions required for each method and the software that can be used.

Conclusion

This paper will help the immunologist to choose the correct statistical approach for a particular research question.