Selective allodepletion is a strategy to eliminate host-reactive donor T cells from hematopoietic stem cell allografts to prevent graft-versus-host disease while conserving useful donor immune functions. To overcome fluctuations in activation-based surface marker expression and achieve a more consistent and effective allodepletion, we investigated a photodepletion process targeting activation-based changes in p-glycoprotein that result in an altered efflux of the photosensitizer TH9402. Expanded lymphocytes, generated using anti-CD3 and IL-2, were cocultured with responder cells from HLA-matched or -mismatched donors. Optimal results were achieved when cocultured cells were incubated with 7.5 muM TH9402, followed by dye extrusion and exposure to 5 Joule/cm(2) light energy at 5 x 10(6) cells/mL. In mismatched stimulator-responder pairs, the median reduction of alloreactivity was 474-fold (range, 43-fold to 864-fold) compared with the unmanipulated responder. Third-party responses were maintained with a median 1.4-fold (range, 0.9-fold to 3.3-fold) reduction. In matched pairs, alloreactive helper T-lymphocyte precursors were reduced to lower than 1:100 000, while third-party responses remained higher than 1:10 000. This establishes a clinical-scale process capable of highly efficient, reproducible, selective removal of alloreactive lymphocytes from lymphocyte transplant products performed under current Good Manufacturing Practice. This procedure is currently being investigated in a clinical trial of allotransplantation.