Abstract

Abstract 647

PNH is caused by somatic mutations in the PIG-A gene in hematopoietic stem cells leading to disruption of the synthesis of GPI, a glycolipid used for post-translational modification of proteins. PNH is classically characterized by the triad of intravascular hemolysis, severe acquired thrombophilia and multiple cytopenias. Whereas the first two components are the consequence of increased susceptibility to activated complement, due to the deficiency on red cells of GPI-linked CD55 and CD59, cytopenias are an expression of bone marrow failure. This fact, together with strong clinical evidence associating PNH with aplastic anemia, has led to the notion that a T cell-mediated attack underlies both conditions whereby in PNH the PIG-A mutant clone is able to escape the attack. Consistent with this, we previously found skewing of the spectratyping pattern of T cells, and recurrent TCR beta chain sequences. Here, using complementary assays, we tested the hypothesis that PIG-A mutant cells escape an auto-immune T cell attack targeted to the GPI molecule itself, which would be presented by CD1d.

First, CD8+ T cells from PNH patients and from normal donors were co-cultured with C1R-CD1d B cells pulsed with either trypanosomal GPI (t-GPI) or with solvent. Reactive T cells producing interferon gamma (IFNG) were then enumerated by ELISPOT technology. We found that in the presence of t-GPI-pulsed C1R-CD1d cells, the frequency of reactive T cells was significantly increased (up to 5-fold) in patients on each of days 1, 7, 14 but not in controls. To confirm that T cell reactivity to GPI depended on the presence of CD1d, when similar experiments were carried out with the untransfected, CD1d-negative C1R cells, no such increase was seen. Similar results were obtained using as antigen presenting cells (APC) K562 cells and appropriate derivatives.

Next, we used for antigen presentation, in addition to the C1R-CD1d cell line, a derivative that lacks the ability to synthesize GPI (GPI- C1R-CD1d). In these experiments, reactive (IFNG-producing) T cells were counted by intracellular staining and flow-cytometry:in PNH patients, but not in controls, we found a 3-fold increase in the frequency of reactive CD8+ T cells when the APC were GPI+ C1R-CD1d compared to when they were GPI- C1R-CD1d. This increase was not seen when APC were the C1R cells. By contrast, with GPI- C1R-CD1d the increase was restored by the addition of exogenous human GPI (h-GPI).

Next, we tested the reactivity of T cells from 11 PNH patients using as APC mature dendritic cells (DC) obtained from autologous peripheral blood monocytes. By selecting CD14-neg cells – that are the majority in PNH patients - we took advantage of the fact that in this way we had APC that were naturally GPI negative. In the absence of added GPI we could not demonstrate reactive T cells; however, in 7 out of 11 patients there was a significant increment when the DC were loaded with h-GPI (p =0,0149; Mann-Whitney test).

Since CD1d is required for development and function of iNKT cells, immunoregulatory, glycolipid-specific T cells characterized by an invariant TCRValpha24Jalpha18 chain, we tested the hypothesis that CD1d-restricted, GPI-specific T cells might also possess an invariant TCR alpha chain. Global TCR alpha chain repertoire analysis of flow-sorted dimer CD1d/GPI+ T cells identified a novel TCRValpha21Jalpha31 invariant chain (9/9 clones) in 1 out of 3 patients tested. Further analysis of the TCRValpha21 repertoire of patients and controls by next generation sequencing (GS Junior, Roche) revealed that the invariant TCRValpha21Jalpha31 chain was present but not expanded in 5/8 controls (frequency<0.5% of TCRValpha21-Calpha T cell repertoire); by contrast, it was expanded in 6/11 patients (median frequency: 1.86%, range: 1.05–75% of TCRValpha21-Calpha T cell repertoire).

In conclusion, we provide the first direct evidence that GPI-specific, CD1d-restricted T cells, enriched in a novel invariant TCR alpha chain sequence, are increased in patients with PNH but not normal controls.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.