Antibodies targeting surface membrane antigens in patients with chronic graft-versus-host disease

Abstract

Chronic graft-versus-host disease (cGVHD) after allogeneic hematopoietic stem cell transplant reflects a complex immune response resulting in chronic damage to multiple tissues. Previous studies indicated that donor B cells and the antibodies they produce play an important role in the development of cGVHD. To understand the pathogenic role of antibodies in cGVHD, we focused our studies on posttransplant production of immunoglobulin G antibodies targeting cell surface antigens expressed in multiple cGVHD affected tissues, due to their potential functional impact on living cells in vivo. Using plate-bound cell membrane proteins as targets, we detected a significantly higher level of antibodies reactive with these membrane antigens in patients who developed cGVHD, compared with those who did not and healthy donors. Plasma-reactive antibody levels increased significantly prior to the clinical diagnosis of cGVHD and were reduced following cGVHD therapies including prednisone, interleukin-2, or extracorporeal photophoresis. Using cell-based immunoprecipitation with plasma from cGVHD patients and mass spectrometry, we identified 43 membrane proteins targeted by these antibodies. The presence of antibodies in cGVHD patients' plasma that specifically target 6 of these proteins was validated. Antibodies reactive with these 6 antigens were more frequently detected in patients with cGVHD compared with patients without cGVHD and healthy donors. These results indicate that antibodies that target membrane antigens of living cells frequently develop in cGVHD patients and further support a role for B cells and antibodies in the development of cGVHD.

Graphical abstract

Figure 1.

Comparison of antibody reactivities to plate-bound antigens detected by FLISA with antigens on live cells by FACS. (A) Antibody detection by FLISA in 4-patient samples against plate-bound membrane antigens from the AML cell line (Kasumi-1). (B) Antibody reactivities of the same patient samples against live cells (Kasumi-1) detected by FACS. APC, allophycocyanin; FACS, fluorescence-activated cell sorter; MFI, mean fluorescence intensity.

Figure 2.

Antibody reactivities against skin cell membrane antigens. Healthy donor (n = 90). No cGVHD indicates patients with no cGVHD (n = 85). cGVHD indicates patients with active cGVHD (n = 106) as described in “Materials and methods” cGVHD-T indicates patients with active cGVHD and receiving intensive treatments as described in “Materials and methods” (n = 93). *P < .0001. (A) Antibody reactivity was tested with FLISA using membrane proteins from the skin keratinocyte cell line (CCD1106) as antigens. (B) Antibody reactivity was tested with FLISA using membrane proteins from the skin fibroblast cell line (Detroit 551) as antigens. HD, healthy donor.

Figure 3.

Antibody reactivities in HSCT patient plasma against skin keratinocyte antigens. Antibody reactivities in patient plasma were tested against membrane antigens from the skin keratinocyte cell line (CCD1106). (A) Antibody reactivities from individual patients who did not develop cGVHD after HSCT (n = 19). (B) Antibody reactivities from individual patients who developed cGVHD during the 3 years after HSCT (n = 30). (C) The mean fluorescence intensity from cGVHD group and No cGVHD group. (D) The mean fluorescence intensity of the cGVHD group after the time points of all patients within the group were synchronized to the time of cGVHD clinical onset. P = .03.

Figure 4.

Identification of targeted antigens by cell-based immunoprecipitation. (A) Analysis of the immunoprecipitation product with cGVHD patient plasma and keratinocyte cells by capillary protein gel electrophoresis. (B) Summary of the unique antigens identified when cell lines indicated were used. cyto, cytoplasmic fraction; mem, membrane fraction; MWM, molecular weight marker.

Figure 5.

Tissue protein expression of identified targeted antigens. Proteins are listed by their gene name. (A) The antigen source from which they were identified is highlighted by the green color. (B) The level of protein expression in human tissues for each gene is indicated by the height of the bars as high, medium, or low, based on data from Human Protein Atlas (www.proteinatlas.org). *Antigens selected for confirmation study. N/A, data not available.

Figure 6.

Expression of targeted antigens on different cell types. The expression of the 6 antigens targeted by antibodies was examined by flow cytometry. The expression on 2 skin keratinocyte cell lines (A-B), 1 AML cell line (C), and CD34⁺ circulation AML cells from 2 patients (D-E) is shown. Red indicates isotype control–PE; blue, anti-antigen–PE. Pt., patient.

Figure 7.

Antibody reactivity against 6 antigens. (A) Antibody reactivity in plasma of HSCT patients and healthy donors (HD) against the 6 antigens listed were tested with the FACS-based antigen-trapping bead assay. Fluorescence intensities are displayed as heatmap with Excel conditional formatting by equalizing the highest intensity of each antigen. (B) The percentages of high producers for each antigen. High producers were defined as having florescence intensity values for the specific antigen greater than the 95th percentile of HD. (C) The P values between groups in panel B for each antigen. Ab, antibody; Ag, antigen.