Invariant NKT (iNKT) cells are a potent, CD1d-restricted immunomodulatory subset of T cells that regulate a variety of experimental immune responses including alloreactivity and acute graft versus host disease (aGVHD).However, their role in human immune responses and in particular in the regulation of clinical aGVHD remains unknown. Given that the frequency of iNKT cells in the blood of normal individuals varies up to 1000-fold we surmised that there was likely to be similar variability in their frequency in peripheral blood stem cell (PBSC) grafts and that the graft iNKT cell dose might be important in the development of aGVHD. To address these hypotheses, using multiparameter flow-cytometry, we determined in G-CSF mobilised PBSC grafts of 61 sibling donors the frequency and the cell dose (in × 106 cells/Kg) of CD3+TCRVbeta11+Valpha24+ iNKT cells as well as of effectors (i.e., CD3+ T cells and CD3-CD56+CD16+ NK cells) or regulators (CD3+CD161+ T cells and CD3+CD4+CD25++FoxP3+ Tregs) of alloreactivity and aGVHD.
The median (minimum-maximum) frequency of T cells in the grafts was 62.3% (44.8%–80.7%) of the mononuclear cells and the dose of T cells given with the graft 179.8 (28.7–607.3), while for the NK cells it was 3.3% (0.15%–19.7%) and 9 (0.27–99.6) respectively. The frequency of the CD3+CD161+ cells was 10.5% of the total T cells (1.2%–32.5%) and the dose 17.3 (1.3–120.2). The frequency of Tregs was 5.9% of the CD3+CD4+ population (1.7%–11.9%) and the dose given 6.1 (1.1–36.4). Unlike these cell populations and similar to peripheral blood, iNKT cell content in the grafts varied up to 1000-fold with median frequency 0.045% of the total T cells (0.001%–1.07%) and cell dose 0.075 (0.0014–1.6). There was no correlation between donor age or sex with the frequency or cell dose of any the above cell populations.
To assess the role of graft iNKT cells in the development of aGVHD, we selected 41 of the 61 cases, whose G-CSF mobilised PBSC graft was used for a T cell replete, sibling HLA-identical allograft (myeloablative n=34, RIC n=7) for the treatment of haematological malignancies: CML (n=16), AML (n=16), ALL (n=4), MDS (n=3) and myeloma (n=2). aGvHD prophylaxis was cyclosporine-A plus methotrexate (n=39) or cyclosporine-A only (n=2). Only patients who survived >100 days were included unless aGVHD developed earlier.
Twenty one of the 41 recipients (51%) developed overall grade 0-I aGVHD (group 1) and 20 (49%) grade II-IV aGVHD (group 2). We found no difference in the CD34+ cell dose (median 5.14 vs 5.95, Mann Whitney p = 0.48), sex mismatched allografts (6 cases in each aGVHD group, chi square = 0.92), donor age (42.3yrs vs 46.1yrs, p = 0.4) or recipient age (44.2yrs vs 48yrs, p = 0.4) between the two groups. Similarly, there was no difference in the frequency or cell dose of CD3+ cells (63.3% vs 62.3% p = 0.9 and 179.8 vs 176.5 p = 0.9), CD3+CD161+ cells (14.1% vs 9%, p = 0.17 and 25.8 vs 19.6 p = 0.42), NK cells (3.5% vs 2.9% p = 0.36 and 10.8 vs 9.9 p = 0.35) and Tregs (6.3% vs 5.7% p = 0.68 and 6.1 vs 5.9 p = 0.9). In contrast, there was a significant difference in the graft iNKT cell content between the two aGVHD groups. Specifically, group 1 (grade 0-I) compared to group 2 (grade II-IV) received grafts with an almost 3-fold higher frequency (0.067% vs 0.024% p = 0.026) and dose (0.181 vs 0.07, p = 0.027) of iNKT cells. Further analysis of the relative role of the CD4+ and CD4- iNKT cell subsets revealed that the latter accounted for the protective effect of the high graft iNKT cell content (frequency 0.05% vs 0.014% p = 0.016 and cell dose 0.072 vs 0.023 p = 0.01) but not their CD4+ counterpart (0.038% vs 0.011% p = 0.12 and 0.051 vs 0.029 p = 0.14).
Thus, this study is the first to demonstrate that graft iNKT cells are an important determinant of aGVHD in humans and suggests that enrichment of the graft with iNKT cells might be a useful strategy to prevent clinically significant aGVHD.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.