Immune Surveillance By Autoreactive Helper But Not Cytolytic T Cells Is a Common Phenomenon In Patients With Acute Myeloid Leukemia

The presence of immune surveillance as a mechanism to prevent the development of malignancies and/or to eradicate small numbers of developing tumor cells has been described by many researchers in the past. The development of an overt malignancy may therefore be seen as failure of immune surveillance. In previous studies we demonstrated the existence of autoreactive CD4 T cells showing HLA-restricted reactivity against myeloid cells as a common phenomenon in healthy individuals. These autoreactive T cells show a profound cytokine response, but do not exert significant cytotoxic effects. To investigate whether such autoreactive T cell responses could play a role in immune surveillance against myeloid malignancies, we investigated the functional reactivity and cell lineage specificity of autologous T cell responses mounted against mature acute myeloid leukemia (AML) blasts. T cells were isolated using the pan T isolation kit and magnetic bead separation (MACS) from peripheral blood of patients with AML in remission after induction chemotherapy and labeled with the fluorescent dye PKH26. These T cells were stimulated with autologous AML blasts at a 1/2 responder to stimulator ratio and cultured for 14 days in medium containing heat-inactivated human serum and 1 ng/mL IL7 and 0.01 ng/mL IL-15 (Miltenyi). At day 14, proliferating CD4 and CD8 T cells comprising 7.3% +/- 2.5% and 0.8 +/- 0.5% of the total T cell populations, respectively, were isolated single cell/well using flowcytometric cell sorting based on PKH dilution. T cell clones were expanded and tested against a panel of autologous and HLA-matched or HLA-mismatched allogeneic target cells, comprising AML blasts, EBV-LCL, monocytes, monocyte-derived dendritic cells (DCs), B cells and primary skin fibroblasts. The isolated CD4 clones produced interferon-gamma (IFNg) and/or interleukin 4 (IL-4) in response to stimulation with autologous AML blasts. This cytokine production could be blocked using pan-HLA-class-II and allele-specific blocking antibodies. HLA-DR, -DP, as well as –DQ restricted clones were found and these clones displayed an oligoclonal T cell receptor V-beta (TCR-VB) usage. Interestingly, these clones exerted also reactivity against autologous EBV-LCL, monocytes, DCs, HLA-class II expressing (IFNg pretreated) fibroblasts and to a lesser extend against autologous B cells, as well as to the same target cell populations (including AML blasts) obtained from allogeneic third party individuals that were matched for the HLA-molecules presenting the T cell epitopes. These results indicate recognition of common antigens, not restricted to the malignant cell populations. No reactivity was observed against HLA-mismatched target cells. In addition, a limited number of CD8 clones was isolated that showed a similar HLA-restricted cytokine production profile. Similar experiments were performed in serumfree X-vivo15 medium to exclude recognition of serum components. In contrast to the profound cytokine response, none of the isolated clones exerted substantial cytotoxicity against one of the targets. Some CD8 clones exerted a maximum of 17% lysis at a 10/1 effector to target ratio against AML blasts. Since no direct cytotoxicity by the autoreactive T cells could be demonstrated, we investigated whether crosstalk between the autoreactive T cells and the AML blasts may render the AML cells more sensitive to subsequent immune attack. The expression of costimulatory markers (CD40, CD80 and CD86) and adhesion (ICAM-1 /CD54) on the AML blasts was significantly increased after co-incubation with the autoreactive T cells. Similar AML responsive autoreactive T cell clones were obtained using T cells from HLA-matched healthy donors as responder cells, illustrating that these autoreactive T cells are part of the normal T cell repertoire and were not induced by the high dose chemotherapy that the patients had been subjected to. In conclusion, we here demonstrate that the presence of autoreactive helper T cells is a common phenomenon in patients with AML. We hypothesize that the large burden of myeloid cells at presentation of AML may result in the amplification of an autoreactive AML directed CD4 T cell response. This response does not result in a direct effective anti-AML immune surveillance, but the immune-modulatory effect on the AML phenotype upon crosstalk may pave the way for other immunological interventions.