Human NKG2A overrides NKG2C effector functions to prevent autoreactivity of NK cells

To the editor:

In April 2009, we saw a 64-year-old woman with persistent indolent proliferation of CD3⁻CD56⁺ large granular lymphocytes.¹  She had no other clinical symptoms and no autoimmune events, and her clinical course has remained benign over 22 months. Neither clinical nor imaging studies have shown signs of lymphoma or acute infection. IgG serology is negative for HIV-1/2, human T cell lymphotrophic virus (HTLV)–1/2, hepatitis B virus, hepatitis C virus, and HHV8, and positive for EBV, CMV, HSV, and varicella zoster virus. Most notably, we have observed an elevated lymphocyte count, ranging from 5.2 to 8.0 × 10⁹/L, extremely unbalanced with CD3⁻CD56^dim natural killer (NK) cells accounting for 62.0%-83.2% of lymphocytes overall (Figure 1A). Intriguingly, more than 95% of the NK cells simultaneously coexpress inhibitory NKG2A and triggering NKG2C receptors (Figure 1B), although coexpression of these markers is generally observed in only 1%-2% of decidual and peripheral blood NK cells from healthy donors.^2-4  However, half of NK cells express HLA-DR, an activation marker, and present an intermediate cell-differentiation phenotype with a high proportion of CD57⁺ cells and an insignificant quantity of killer cell immunoglobulin–like receptors (KIR) (Figure 1C-D),⁴  although her genome includes all KIR variants.

Consistent with recent studies exploring the relation between NK-cell responsiveness and receptors that engage self-MHC class I molecules,⁵  polyfunctional analysis showed that this patient's KIR⁻NKG2A⁺NKG2C⁺ NK cells were fully functional against MHC class I–deficient target cells (K562, and 721.221) and capable of antibody-dependent cellular cytotoxicity (Figure 1E).

Inhibitory NKG2A and activating NKG2C receptors are both specific for HLA-E, a nonclassical MHC class Ib molecule expressing peptides derived from signal sequences of other HLA class I molecules.⁶  Given their opposing functional effect, the coexpression of both markers raises the question of how NK cells can establish self-tolerance under this circumstance, as the lack of symptoms shows they do. A redirected killing assay against FcγR⁺ P815 targets in the presence of anti-NKG2C mAb showed that the NKG2C was functional, but more interestingly, costimulation of NKG2A and NKG2C with their respective mAbs demonstrated that the inhibitory function of NKG2A prevailed (Figure 1F). The NKG2A⁺NKG2C⁺ NK cells were clearly less polyfunctional against targets expressing HLA-E than against wild-type 721.221 cells (Figure 1E), and HLA-E⁺ target interaction inhibited NK-cell proliferation, except in the presence of IL-2 (Figure 1G), as described.⁷  A redirect killing assay confirmed that simultaneous engagement of NKG2C and IL-2 induced a marked proliferative response, abrogated by NKG2A engagement (Figure 1H).

We therefore report that the NKG2A inhibitory receptor plays a central inhibitory role, providing a regulatory feedback mechanism for controlling the activating NKG2C receptor, reminiscent of the activation-dependent expression of CTLA-4 on T lymphocytes that counteracts CD28-mediated costimulation.⁸  These findings are consistent with the higher affinity of NKG2A for HLA-E, which likely favors its competition with NKG2C for ligand engagement.⁹ 

To our knowledge, this case provides the first in vivo evidence that, in the absence of KIR, expression of NKG2A on NKG2C⁺ NK cells establishes an inhibitory threshold that prevents their potential autoreactivity against self-HLA-E⁺ cells and explains this woman's absence of autoimmune disorders or collateral damage.

In accordance with national ethics committee guidelines, the patient provided written informed consent before peripheral blood samples were collected for this study.