Anti Leukemic T Cells in Myeloid Neoplasia: Abnormal Cytokine Secretion Patterns May Prevent Immunological Detection and Possibly Reflect Functional Impairment.

T cells recognising leukemia-associated antigens are present in the T cell repertoire of a subgroup of patients with AML and CML. Difficulties in the detection of such leukemia-reactive T cells in peripheral blood samples of patients are mainly attributed to the low frequency of leukemia-specific T cells.

Previously we have reported that simultaneous cytokine analysis by cytokine bead array (CBA) is a useful tool to detect low frequency leukemia-reactive T cells with an aberrant cytokine secretion profile in CML patients in stable hematologic/cytogenetic remission (Westermann et al. Br J Haematol 2005).

This study includes patients with AML (n=10) and CML (n=30). AML patients were in CR after allogeneic stem cell transplantation, CML patients had achieved a cytogenetic response under imatinib treatment or were in CR after allogeneic stem cell transplantation. Blood samples were studied by different structural and functional T cell assays (ELISpot, CBA, flowcytometric intracellular cytokine staining and tetramer staining) in order to detect T cells specifically recognising leukemia-associated antigens such as proteinase-3, WT-1, SOCS-2, bcr/abl and c-abl.

In 2/10 HLA-A2+ AML patients after allogeneic transplantation, functional T cell assays showed a peptide- specific g-interferon secretion upon stimulation with a HLA-A2-restricted proteinase-3 peptide while no peptide-specific TNF-a or IL-4 secretion could be found. Another 2/10 HLA-A2+ AML patients only showed peptide-specific TNF-a release upon stimulation with HLA-A2-restricted peptides from proteinase-3 or WT-1 respectively. No peptide-specific g-interferon release was detected in these latter patients. In one HLA-A2+ AML patient, WT-1-specific T cells could be demonstrated by tetramer staining whereas all functional assays were negative.

In CML, 5/30 patients showed a peptide-specific g-interferon secretion upon stimulation with HLA-matched peptides from proteinase-3, c-abl and SOCS-2. However, in 19/30 CML patients TNF-a was the only cytokine which was released specifically upon stimulation with HLA-matched peptides from proteinase-3,bcr/abl, c-abl and SOCS-2. In these latter patients no peptide-specific g-interferon response could be detected. Bcr/abl-specific T cells recognising the GFKQSSKAL fusion peptide were detected in 2/9 HLA-B8+ CML patients by tetramer staining.

Our results show that the cytokine secretion profile of leukemia-reactive T cells after in vitro stimulation may substantially differ from a standard TH1/TH2 profile and that TNF-a may be the only cytokine which is released in an antigen-specific manner. Antigen-specific TNF-a release by T cells seems to be a common feature both in AML and CML. The functional status of these TNF-a secreting T cells remains to be determined since they might represent T cells which have been functionally impaired, i.e. by tolerogenic antigen-presentation by leukemia cells.

We conclude that T cell monitoring in leukemia patients should contain both structural and functional assays which may detect several cytokines simultaneously. Otherwise functionally abnormal leukemia-reactive T cells may be missed.