Research
Molecular evolution: Affinity maturation of antibodies
In the immune system, affinity maturation describes the improvement of antibody affinity over the course of a B cell response. Affinity maturation is linked to the locales of several hundreds of germinal centers (GC) which transiently form in secondary lymphoid organs like spleen and lymph nodes upon antigenic challenge. Within the unique multi-cell type environment of GCs, mature B lymphocytes undergo clonal expansion and pass through a micro-evolutionary process including diversification of their B cell receptors (antibodies) by somatic hypermutation (SHM) and subsequent receptor-mediated affinity selection. Furthermore, class switch recombination and commitment of B lymphocytes to either B cell memory or plasma cell fate occur in the context of GCs.
Kinetics of germinal center growth and cellular composition
Since the GC response spans a period of about three weeks and more, GC growth cannot be followed by real-time imaging and research has to resort to acquisition of time-resolved cross-sectional data. A drawback to the latter is that it allows for determining the average kinetics of the overall GC ensemble, but not necessarily the growth kinetics of individual GCs. Most importantly, as we have shown with the aid of mathematical simulations, the empirically observed cross-sectional profile of GC growth is compatible with very different GC ensemble kinetics. A scenario where all GCs emerge almost simultaneously and then grow synchronized, for instance, can result in the same cross-sectional profile than a scenario that allows for asynchronous GC development. Whereas in the first scenario one GC is like the other in terms of size at a given time point, there is no such thing like a typical GC in the second scenario. Which is it then? And what about the cellular composition of GCs – typical or not? These questions are crucial, for mathematical models of affinity maturation often rely on fitting of GC population kinetics to assess the efficiency of presumed selection mechanisms.
By means of computer-aided three-dimensional reconstructions of GCs from histological sections we could show that GC growth is indeed nonsynchronized and that there is no typical GC in terms of size. Acquisition of large-scale confocal data on the cellular composition of GCs in a second study let us further reveal that established GCs have a typical cellular composition, independent of their size and age.
Publications:
[1] Or-Guil M, Wittenbrink N, Weiser AA, Schuchhardt J. Recirculation of germinal center B cells: a multilevel selection strategy for antibody maturation, Immunol Rev. 2007 Apr;216:130-41.
[2] Wittenbrink N, Weber TS, Klein A, Weiser AA, Zuschratter W, Sibila M, Schuchhardt J, Or-Guil M. Broad volume distributions indicate nonsynchronized growth and suggest sudden collapses of germinal center B cell populations, J Immunol. 2010 Feb 1;184(3):1339-47. Epub 2010 Jan 6.
[3] Wittenbrink N, Klein A, Weiser AA, Schuchhardt J, Or-Guil M. Is there a typical germinal center? A large-scale immunohistological study on the cellular composition of germinal centers during the hapten- carrier-driven primary immune response in mice, J Immunol. 2011 Dec 15;187(12):6185-96. Epub 2011 Nov 18.
Tools:
GCImagePresenter - A Database on Histological Images of Murine Splenic Germinal Centers
Spectrum of mutations relevant to maturation of antibodies
Despite advances in revealing the dynamics and molecular base of affinity maturation (and SHM in particular), the fundamentals of receptor-mediated B cell selection within GCs are as yet little understood. For instance, it is often not clear how and how many mutations lead to improvement of a given B cell receptor over time. This is because both selection and intrinsic biases of SHM contribute to experimentally obtained BCR mutation patterns. By following an approach centered on statistical comparison of empirical mutation frequencies and mutation frequencies expected under the null hypothesis of no selection we challenged the concept of single/a few key mutations. Indeed, we found that the spectrum of mutations relevant to maturation of canonical anti‑(4‑hydroxy‑3-nitrophenyl)acetyl (NP) antibodies is much broader than previously acknowledged. Furthermore, our results imply that next to affinity also expression and stability of BCRs drive maturation of B lymphocytes in GCs.
Publications:
[4] Weiser AA, Wittenbrink N, Zhang L, Schmelzer AI, Valai A, Or-Guil M. Affinity maturation of B cells involves not only a few but a whole spectrum of relevant mutations. Int Immunol. 2011 May;23(5):345-56.
Molecular recognition: Antibody profiling
Peptide arrays are frequently applied to study protein-peptide interactions, for example to identify biologically active peptides, to map epitopes or to perform disease diagnostics. We study the use antibody binding profiles measured with peptide arrays to determine biomarkers for diagnostics. This is accomplished within the joint projects SYSTHER (BMBF), NephroFIT (IBB) and SeroPep (BMWi).
Antibody affinity profiling with peptide microarrays
Little is known about the reliability of high-throughput antibody-peptide binding measurements. To assess the technological precision of peptide microarray probing, we performed a study on the qualitative and quantitative reliability of peptide arrays.
By using a model system, we systematically compared dissociation constants of protein-peptide complexes with obtained signal intensities. We chose a system of binding partners composed of the CB4-1 monoclonal antibody (directed against the p24 element of the HIV-1 envelope) and an array of 26 different peptides for which the CB4-1 binding affinity has been measured independently by SPR binding assays. We characterized technological precision in terms of intra- and inter support experimental error as well as the measurement accuracy in terms of SI/Kd correlations using a phenomenological model derived from the mass-and-action law. ROC-curves generated for peptide classification emphasize that the peptide array technology is an accurate method for assigning the measured spot's signal intensity to one of two binding affinity classes.
Publications:
[5] Tapia V, Bongartz J, Schutkowski M, Bruni N, Weiser A, Ay B, Volkmer R, Or-Guil M. Affinity profiling using the peptide microarray technology: a case study. Anal Biochem. 2007 Apr 1;363(1):108-18. Epub 2007 Jan 4.
Classification of serum antibody profiles with random peptide microarrays
Serum antibody reactivity to antigen is the basis of many diagnostics tests for diseases; prominent examples include HIV ELISA and HIV dipstick as well as Hashimoto's Thyroiditis diagnostics tests. The essential premise of such tests is the discovery of at least one disease-specific antigen. However, there are a great many diseases, where – despite extensive research – the identity of specific antigens remains obscure. By investigating an experimental model system of helminth infection in mice we demonstrated that synthetic libraries of random peptides are a well-suited alternative to classical disease-specific antigens in serological diagnostics. Probing of sera on random peptide microarrays led us to classify different strains of mice and to predict H. bakeri infection with high accuracy.
Publications:
[6] Bongartz J, Bruni N, Or-Guil M. Epitope mapping using randomly generated peptide libraries, Methods Mol Biol. 2009;524:237-46.
Prediction of ensemble properties of serum antibodies
Although serum probing on peptide microarrays is steadily gaining in diagnostic importance, understanding of the technology itself is very limited. To date, little is known about the origin/genesis and composition of antibody-peptide binding signals, in particular where probing with antibody mixtures of unknown complexity, such as sera, is concerned. By relating measured signals to the amino acid sequences of peptides in a multivariate regression model, we defined amino acid-associated weights that – although they do not contain information on amino acid position – predict up to 50% of variation in serum antibody binding profiles. To further assess this result, we formulated a minimal model of peptide binding. Our model emphasized that position-independent amino acid-associated weights predict linear epitope binding of antibody mixtures only if the mixture is random, highly diverse, and contains no dominant antibodies.
Publications:
[7] Greiff V, Redestig H, Lück J, Bruni N, Valai A, Hartmann S, Rausch S, Schuchhardt J, Or-Guil M. A minimal model of peptide binding predicts ensemble properties of serum antibodies. BMC Genomics. 21;13:79.