Novel computational approaches to improve transplantation outcomes
EBPOD 2017: Project 2
This is one of 11 joint postdoctoral fellowships offered by EMBL-EBI, the NIHR Cambridge Biomedical Research Centre and the University of Cambridge’s School of the Biological Sciences in 2017.
Human organ transplantation is one of the most significant medical advances over the past 50 years. The key challenge of organ rejection is tackled through the use of immunosuppressant drugs and techniques that aim to identify the most suitable tissue match between donor and recipient. Nevertheless, despite the undoubted positive impact of organ transplantation on the lives of thousands of patients worldwide, there remain areas where further progress is needed. In this project, we aim to combine our complementary expertise in transplant surgery and computational modelling techniques to capitalise on the greater understanding of the molecular basis of the immunological response and the ever-growing volumes of genetic and clinical outcomes data. Our overall goals are to develop new and improved methods for identifying the most suitable donor-recipient pairs, so leading to improved clinical outcomes and enhanced patient benefit.
Introduction and background
Central to the regulation of the immune system, and thus the response to foreign pathogens, the immune regulation of neoplastic cells, autoimmunity and the transplantation of a donor organ, is the Human Leukocyte Antigen (HLA) gene complex that encodes the major histocompatibility complex (MHC) proteins. MHC gene products bind peptides derived from potential antigens for subsequent presentation to receptors on T-cells, thus initiating the immune response. Exposure to an ever-evolving environment has resulted in the human MHC being one of the most polymorphic regions of the human genome.
The most significant MHC proteins in histocompatibility are the products of the HLA class I (A, B, C) and class II (DR, DQ, DP) loci. Current national and international organ allocation algorithms seek to minimise the number of mismatches between donor and recipient. However, the extensive polymorphism of the HLA system and the need to balance often competing clinical and societal factors makes exact matches difficult to achieve and so most recipients receive HLA mismatched grafts. In kidney transplantation, current matching algorithms aim to minimise incompatibilities only at the HLA-A, -B and -DR loci. As a consequence, HLA incompatibilities are common, often resulting in heavier immunosuppression drug therapy (increasing morbidity and mortality), the development of antibody-mediated organ rejection (the major cause of long-term graft loss) and an increased risk of sensitisation that can severely limit options for repeat transplantation should the primary graft fail. In haematopoietic stem cell transplantation (HSCT) the requirements for donor-recipient HLA matching are more stringent (considered at the HLA-A, -B, -C, -DR, -DQ and -DP loci). The best results are obtained using HLA-identical sibling transplantation, but this is often not feasible, necessitating the selection of unrelated donor-recipient pairs that are not fully matched, leading to increased morbidity and mortality.
The major limitation of existing matching algorithms is that they assign an equal weight to all HLA mismatches despite potential differences in immunogenic potential. Detailed amino acid sequence analysis of HLA proteins and structural information obtained from X-ray crystallography now offer the exciting prospect of comparing donor-recipient HLA types at the molecular level, thereby providing opportunities for more rational approaches to the study of HLA immunogenicity. Preliminary studies by our group have demonstrated that immunogenicity can be assessed more accurately by evaluating differences in the number, location and physicochemical properties of amino acid polymorphisms between donor and recipient HLA proteins1. We have recently extended this approach using 3D electrostatic similarity calculations based on structural data2 which showed significant promise in predicting HLA-specific alloantibody responses3 and in identifying donor-recipient HLA combinations associated with better graft outcomes4.
1. Kosmoliaptsis V, et al. (2016) Alloantibody Responses After Renal Transplant Failure Can Be Better Predicted by Donor-Recipient HLA Amino Acid Sequence and Physicochemical Disparities Than Conventional HLA Matching. Am J Transplantation 16:2139
2. Mallon DH, et al., (2015) Three-dimensional structural modelling and calculation of electrostatic potentials of HLA Bw4 and Bw6 epitopes to explain the molecular basis for alloantibody binding: toward predicting HLA antigenicity and immunogenicity. Transplantation 99:385
3. Mallon DH, et al. (2016) Computational scoring system to predict HLA immunogenicity. Lancet 387 (S1):68
4. Kosmoliaptsis V, et al. (2015) Physiochemical disparity of mismatched HLA class I alloantigens and risk of acute GVHD following HSCT. Bone Marrow Transplant 50:540
We now wish to build upon our exciting early results to explore the full potential of applying advanced computational and structural biology methods in organ transplantation research. In addition to extending the models to incorporate different mechanistic aspects of the immune system we will analyse large clinical datasets to both guide and validate our new methods. The proposed research plan involves three related work packages (WP) that provide ever deeper insights and expand the computational toolbox.
WP1: Optimise donor-recipient HLA matching in kidney transplantation The objective of this WP is to further develop and optimise our existing HLA immunogenicity algorithm for kidney transplantation. We will extend our whole-molecule HLA molecular electrostatic similarity method to incorporate immunogenic ‘hot-spots’ on the molecular surface (B-cell epitopes). In addition, we will use computational docking and in silico epitope prediction to examine the potential of recipient HLA class II molecules to present donor HLA-derived peptides, thus accounting for the cellular (T-cell) and humoral (B-cell) alloimmune responses. The computational algorithm will be validated using large clinical datasets (outcomes will include HLA-specific alloantibody responses and graft survival) via our established collaborations with NHS Blood and Transplant and with Eurotransplant. We anticipate that this WP will inform a new national deceased-donor kidney allocation scheme and will lead to proposals for randomised controlled trials based on our HLA immunogenicity algorithms.
WP2: Optimise donor-recipient HLA matching in haematopoietic stem cell transplantation The principal immunological mechanisms leading to graft-versus-host disease and to graft failure in HSCT differ to those in kidney transplantation, with T-cell responses thought to be dominant. We will expand our previous approach and investigate the role of polymorphisms in the HLA peptide binding groove and in HLA/T-cell receptor contact areas at the structural level (including quantification of differences in properties such as the electrostatic potential) to identify donor-recipient HLA combinations likely to lead to better clinical outcomes. The potential of donor T-cells to recognise recipient HLA-derived peptides will be examined (informed by methods developed in WP1) and the computational predictions will be validated based on in vitro assays with prospectively collected patient samples. The bioinformatics approach to donor-recipient matching will be optimised and validated using large clinical datasets (existing collaboration with the British Society of Blood and Marrow Transplantation and the Center for International Blood and Marrow Transplant Research). We anticipate that WP2 will lead to a step change in the field by enhancing options for donor selection and by improving outcomes for patients in need of allogeneic HSCT.
WP3: Investigation of the molecular basis of HLA-alloantibody interactions
The development of alloantibodies against donor HLA has emerged as a crucial determinant of adverse outcomes after kidney, heart and lung transplantation. However, little is known about the molecular basis of alloantibody-HLA interactions, the structural characteristics of HLA B-cell epitopes and the epitope-paratope interactions mediating high affinity binding. We propose to use a combination of experimental structure determination methods including X-ray crystallography (collaboration with the Dept. of Biochemistry, University of Cambridge) and computational techniques (e.g. structure prediction and molecular dynamics simulations) to study specific alloantibody-HLA interactions that mediate different effector functions. The results of this work will provide mechanistic insights into the physicochemical determinants of HLA B-cell epitopes (will also inform WP1) and into the clinical significance of alloantibody profiles from patient sera.
This project will provide an exciting opportunity for a talented scientist to develop novel computational methods to address important clinical problems with significant potential for patient benefit. The collaboration also has the potential to drive the future research in this area (e.g. in donor specific T-cell alloreactivity), including clinical studies, and thus to open new multidisciplinary areas of enquiry across the BRC.