Storage-induced changes in erythrocyte membrane proteins promote recognition by autoantibodies

Abstract

Physiological erythrocyte removal is associated with a selective increase in expression of neoantigens on erythrocytes and their vesicles, and subsequent autologous antibody binding and phagocytosis. Chronic erythrocyte transfusion often leads to immunization and the formation of alloantibodies and autoantibodies. We investigated whether erythrocyte storage leads to the increased expression of non-physiological antigens. Immunoprecipitations were performed with erythrocytes and vesicles from blood bank erythrocyte concentrates of increasing storage periods, using patient plasma containing erythrocyte autoantibodies. Immunoprecipitate composition was identified using proteomics. Patient plasma antibody binding increased with erythrocyte storage time, while the opposite was observed for healthy volunteer plasma, showing that pathology-associated antigenicity changes during erythrocyte storage. Several membrane proteins were identified as candidate antigens. The protein complexes that were precipitated by the patient antibodies in erythrocytes were different from the ones in the vesicles formed during erythrocyte storage, indicating that the storage-associated vesicles have a different immunization potential. Soluble immune mediators including complement factors were present in the patient plasma immunoprecipitates, but not in the allogeneic control immunoprecipitates. The results support the theory that disturbed erythrocyte aging during storage of erythrocyte concentrates contributes to transfusion-induced alloantibody and autoantibody formation.

Figure 1. Erythrocyte autoantibody immunoprecipitation of erythrocytes sampled at regular time intervals during storage.

Analysis was performed by SDS-PAGE, followed by silver staining. (A) Protein patterns of precipitates obtained using Ringer, autologous plasma, and a representative example from one out of three allogeneic plasmas, and one out of six autoantibody-containing plasmas (patient No. 2). For the allogeneic controls, day 14 is missing. (B) Mean optical density (OD) of patient (•, solid line, N = 6 patients) and allogeneic control plasma (▴, dotted line, N = 3 volunteers) precipitations. Numbers indicate approximate molecular weight (kDa). Error bars represent standard error, *p<0.05.

Figure 2. Erythrocyte autoantibody immunoprecipitation of stored erythrocytes.

(A) Immunoprecipitation of 35 day old erythrocytes with erythrocyte autoantibody-containing patient plasma and allogeneic control plasma, using TX100 or RIPA extraction buffer and analyzed by SDS-PAGE under reducing or non-reducing conditions, followed by silver staining. A representative result (patient No. 2) from one out of three patient plasmas is shown. (B) Example of a silver stained gel of an immunoprecipitation of 35 day stored erythrocytes with plasma of patient No. 1. The same sample was used for Coomassie blue gel staining and subsequent proteomics analysis (Table 2). Gel slices which were excised for proteomic analyses are indicated as slices I and II (see also Table S1). Numbers indicate molecular weight (kDa). Heavy [H] and light [L] antibody chains are indicated by arrows.

Figure 3. Erythrocyte autoantibody immunoprecipitation of biotinylated erythrocyte vesicles from a 35 day old transfusion unit.

(A) Immunoprecipitation with either plasma from patient No. 2, or a monoclonal antibody against band 3 (see Materials and Methods). Analysis was performed by SDS-PAGE, followed by detection of biotinylated membrane proteins (red, streptavidin) and band 3 (green, polyclonal rabbit antibody). A protein G bead control was included. (B) Example immunoprecipitation of biotinylated erythrocyte vesicles from a 35 day old transfusion unit using plasma from patient No. 1. Analysis was performed by SDS-PAGE, followed by detection of biotinylated membrane proteins using fluorochrome conjugated streptavidin. The same sample was used for Coomassie blue gel staining and subsequent proteomics analysis (Table 2). The gel slice which was excised for proteomic analysis is indicated as slice III (see also Table S1). Numbers indicate approximate molecular weight (kDa). Blots were analyzed using the Odyssey Infrared Imaging System.