Immune repertoire diversity in allogeneic stem cell transplantation and its implications for infections and the graft versus leukemia effect

Sammanfattning: The beneficial graft versus leukemia effect (GVL) and its detrimental counterparts, graft versus host disease (GVHD) and susceptibility to infections, are all coupled to a multitude of events during the immune reconstitution (IR) after hematopoietic stem cell transplantation (HSCT). The general aim of this thesis has been to learn more about the IR in HSCT with a particular focus on the impact of infections, natural killer (NK) cell mediated GVL effects and the possibility to apply GVL effects in adoptive cell therapy. In paper I, we identified factors interfering with the IR, thereby making the patients susceptible to late lethal infections. We found that cytomegalovirus (CMV) was an independent risk factor for death in late infections. NK cells are important for controlling CMV infections and patients lacking NK cells suffer from cyclic herpes virus reactivations.1-3 NK cells are also known to mediate GVL-effects and have been coupled to reduced relapse rates after HSCT. The results of paper I thus prompted us to study NK cell-mediated GVL effects and the interaction between CMV and NK cell repertoire dynamics. In paper II we examined NK cell-mediated alloreactivityin 105 patients with myeloid malignancies undergoing human leukocyte antigen (HLA )-identicalsibling transplantation. A longitudinal analysis revealed maintained NK cell tolerance at all timepoints during the IR. In agreement with these experimental data, no clinically evident GVL effect was observed based on stratification of missing ligands to killer cell immunoglobulin-like receptors (KIRs ) in the recipients. In paper III we determined the size of the alloreactive subset and graded the ability of different donors to deliver GVL effects in HLA -mismatched transplantation. The educated subsets expressing KIRs in presence of a corresponding HLA - receptor ligand varied between 12-68% (mean 33%) resulting in 0-62% (mean 8%) alloreactive NK cells depending on recipient HLA -ligands. This algorithm served as a template for studies conducted in paper IV, where we further dissected the role of pre-transplant NK cell repertoires in the donor and post-transplant repertoires developing after 9-12 months. Unsupervised hierarchical clustering was used to group donors and recipients based on their NK cell receptor repertoires. The result showed that donors with naïve receptor repertoires had less relapse and recipients with a tendency to reset their repertoires towards naivety had less relapse and better overall survival. In summary this thesis shed new light on the relationships between early and late infections and the recovery of the immune system after HSCT, linking specific NK cell repertoires to protection from relapse and increased overall survival. This knowledge may be useful for the development of new strategies utilizing NK cells in cellular therapies against hematological malignancies.