Graft-Versus-Host Disease or Infection: Rapid Detection of HLA Mismatch-Reactive T Cells Ex Vivo Can Facilitate Diagnosis and Guide Therapy after Allogeneic Transplantation.

Diagnosis of graft-versus-host disease (GVHD) is mainly based on clinical features and on tissue biopsies. However, clinicians and pathologists are well aware of cases, in which GVHD cannot be distinguished from infections arising from severe immunodeficiency after allogeneic stem-cell transplantation (SCT). This may pose a deep therapeutic dilemma of whether to modify immunosuppressive treatment or to use donor lymphocyte infusion (DLI) for promoting anti-microbial immunity. We observed a 68-year-old patient with myelodysplastic syndrome who developed acute GVHD grade II of skin and gut at d+16 after T-cell depleted reduced-intensity SCT (Fig. 1). GVHD was confirmed by histology and responded to prednisolone therapy. From d+90 to d+240, the patient suffered from massive diarrhea (>2L per day) and recurrent episodes of lower gastrointestinal bleeding. Histopathology analysis on gut biopsies showed a heterogeneous picture with signs of GVHD and ulcerative inflammation. In stool screening, we isolated norovirus type 2 (ELISA, PCR) between d+111 and d+229, thereby confirming the longest infection with this virus ever reported. Tapering immunosuppression did not improve diarrhea, and the patient required intensive care due to serious fluid imbalance. Because of severe lymphopenia (<100 per μL CD3 T cells), we considered DLI therapy to promote T cell reconstitution and norovirus clearance. To balance the risk for DLI-induced exacerbation of GVHD, we first analyzed peripheral blood CD8 T cells for anti-recipient reactivity ex vivo by IFN-γ ELISPOT assay. Due to the limited availability of recipient cells and a single HLA disparity between donor (B*3508) and patient (B*3503), we used HLA-deficient K562 cells transfected with mismatched (B*3503) or matched (A*0201) HLA alleles as antigen-presenting cells. Post-transplant CD8 T cells specifically recognized disparate HLA-B*3503 (Fig. 1), but not shared HLA-A*0201. Mismatch-reactive CD8 T cells were detectable at significant numbers (71/10⁵) during the first episode of GVHD and correlated closely with the intensity of diarrhea beyond d+110. Maximum anti-HLA-B*3503 reactivity (439/10⁵) on d+205 was in the range of IFN-γ spot production obtained in mitogen-stimulated controls. Considering this vigorous alloreactivity, we were concerned that scheduled DLI would boost GVHD rather than facilitating norovirus clearance. Therefore, we decided to omit DLI. Patient’s clinical condition improved spontaneously around d+240, and diarrhea did not recur thereafter. In conclusion, we have established a new T cell assay for the rapid detection of anti-HLA mismatch reactivity using K562-HLA transfectants as substitutes for patient cells. In the meantime, we have validated this assay in two further HLA-incompatible donor-patient pairs and generated off-the-shelf K562 transfectants for more than 15 HLA alleles. Ex vivo alloreactivity toward mismatch HLA might be a surrogate marker to facilitate GVHD diagnosis and guide therapy in ambiguous clinical situations, such as the coincident viral infection described herein.

Fig. 1:

View largeDownload slide

Clinical course and monitoring of all-HLA-B*3503 reactive CD8 T cells ex vivo

Fig. 1:

View largeDownload slide

Clinical course and monitoring of all-HLA-B*3503 reactive CD8 T cells ex vivo

Close modal