Computational scoring system to predict HLA immunogenicity. (25th February 2016)

• Record Type: Journal Article
• Authors: Mallon, Dermot; Kling, Christiane; Bradley, J Andrew; Taylor, Craig; Kabelitz, Dieter; Kosmoliaptsis, Vasilis
• Is Part Of: Lancet. Volume 387(2016)Supplement 1
• Page Start: S68
• Publication Date: 2016-02-25
• Subjects: Medicine -- Periodicals; Medicine -- Periodicals; Medicine; Medicine; Electronic journals; Periodicals; 610.5
• DOI: 10.1016/S0140-6736(16)00455-4
• Languages: English
• ISSNs: 0140-6736

Abstract

Abstract: Background: Our preliminary work has indicated that donor HLA immunogenicity in kidney and haemopoietic stem-cell transplantation might be predicted by assessment of their aminoacid sequence and physicochemical properties. We have now created a novel computational algorithm to quantify structural and surface electrostatic potential differences between donor and recipient HLA and applied it to predict alloantibody responses in a unique patient cohort. Methods: We examined 141 patients undergoing treatment for infertility with lymphocyte immunotherapy. Patients were injected subcutaneously with partners' lymphocytes, and serum samples were collected before and after immunotherapy to assess sensitisation. After four-digit HLA typing and HLA structural modelling, donor–recipient HLA comparisons were performed to determine electrostatic mismatch score (EMS-3D) and assess its ability to predict development of donor-specific alloantibody and overall sensitisation to HLA after immunotherapy. Overall sensitisation was expressed as calculated reaction frequency (cRF), a measure that denotes incompatibility with a UK pool of 10 000 organ donors. Findings: Donor EMS-3D was a strong predictor of overall HLA class I and class II sensitisation risk (cRF >15%) (odds ratio [OR] 5·68 per unit increase, 95% CI 1·84–17·49; p=0·002) with a receiver operating characteristic area under curve score of 0·68 for cRF of more than 15% (p=0·0046) and 0·71 for cRF more than 85% (p<0·0001). Donor HLAs with the highest EMS-3D (fourth quartile) were associated with a higher risk of development of donor-specific alloantibody both for HLA class I (OR 2·90, 95% CI 1·62–5·23; p=0·031) and for HLA class II (16·07, 9·11–142·68; p<0·0001). Physicochemical differences between donor and recipient HLA-DQ were strong predictors of developing HLA-DQ donor-specific alloantibody (30·00, 6·33–87·73; p<0·0001). Interpretation: Our study provides compelling evidence that donor HLA immunogenicity can be predicted by assessment of their structural and physicochemical disparities to recipient HLA. If confirmed in larger studies, our findings might enable better assessment of transplant immunological risk before transplantation and could inform future policies on allocation of deceased-donor kidney and haemopoietic stem-cell transplants to maximise the benefits of transplantation. Funding: Royal College of Surgeons of England Research Fellowship, Kiel University Starter Grant, Evelyn Trust, Academy of Medical Sciences.