HLA-E is the ligand for the natural killer cell CD94/NKG2 receptors

Genetic Variants of the NKG2C/HLA-E Receptor-Ligand Axis Are Determinants of Progression-Free Survival and Therapy Outcome in Aggressive B-Cell Lymphoma

Simple Summary

NKG2C and its ligand HLA-E represent key molecules for NK cell-mediated immune responsiveness. However, the impact of genetic variants in NKG2C and HLA-E on clinical outcomes of aggressive B-cell non-Hodgkin lymphoma patients (B-NHL) has not been clarified. In this study, we analyzed the distribution of NKG2C deletion status and HLA-E variants in 441 patients and 192 healthy individuals. Homozygous deletion of NKG2C (NKG2C^−/−) was more often found in high-risk patients compared to patients with a lower risk and consequently was associated with reduced 2-year progression-free survival. The HLA-E*01:01 allele frequency was increased in B-NHL patients and was strongly related with complete remission. Our results show that absence of NKG2C and HLA-E allelic variations is predictive for B-NHL outcome; while carriers of HLA-E*01:01 are characterized by high, complete remission rates, NKG2C^−/− was rare, but associated with poorer outcome. Prospective validation of our results identifies patients that may benefit from risk-adapted therapy.

Abstract

Aggressive B-cell lymphomas account for the majority of non-Hodgkin lymphomas (B-NHL). NK cells govern the responses to anti-CD20 monoclonal antibodies and have emerged as attractive targets for immunotherapy in subtypes of B-NHL. NKG2C and its cognate ligand HLA-E represent key molecules for fine-tuning of NK cell-mediated immune responses. Here, we investigated the impact of genetic variants of NKG2C and HLA-E on clinical outcomes of 441 B-NHL patients. Homozygous deletion of NKG2C (NKG2C^−/−) was three-fold increased in patients compared to 192 healthy controls. Among studied patients, NKG2C^−/− was more abundant in International Prognostic Index (IPI) high-risk patients compared to patients with a lower IPI (p = 0.013). Strikingly, NKG2C^−/− was associated with a significantly reduced 2-year PFS (progression-free survival) (p = 0.0062) and represented an independent risk factor for 2-year PFS in multivariate analysis (p = 0.005). For HLA-E, the cognate ligand of NKG2C, the HLA-E*01:01 allele frequency was increased in B-NHL patients compared to controls (p = 0.033) and was associated with complete remission in univariate (p = 0.034) and multivariate (p = 0.018) analysis. Our data suggest that NKG2C and HLA-E genotyping is a promising tool for both defining risk groups of aggressive B-NHL and predicting response to immune therapeutic approaches.

Utilizing cell-based therapeutics to overcome immune evasion in hematologic malignancies

Hematologic malignancies provide a suitable testing environment for cell-based immunotherapies, which were pioneered by the development of allogeneic hematopoietic stem cell transplant. All types of cell-based therapies, from donor lymphocyte infusion to dendritic cell vaccines, and adoptive transfer of tumor-specific cytotoxic T cells and natural killer cells, have been clinically translated for hematologic malignancies. The recent success of chimeric antigen receptor-modified T lymphocytes in B-cell malignancies has stimulated the development of this approach toward other hematologic tumors. Similarly, the remarkable activity of checkpoint inhibitors as single agents has created enthusiasm for potential combinations with other cell-based immune therapies. However, tumor cells continuously develop various strategies to evade their immune-mediated elimination. Meanwhile, the recruitment of immunosuppressive cells and the release of inhibitory factors contribute to the development of a tumor microenvironment that hampers the initiation of effective immune responses or blocks the functions of immune effector cells. Understanding how tumor cells escape from immune attack and favor immunosuppression is essential for the improvement of immune cell-based therapies and the development of rational combination approaches.

HLA-E: a novel player for histocompatibility

The classical class I human leukocyte antigens (HLA-A, -B, and -C) present allele-specific self- or pathogenic peptides originated by intracellular processing to CD8(+) immune effector cells. Even a single mismatch in the heavy chain (hc) of an HLA class I molecule can impact on the peptide binding profile. Since HLA class I molecules are highly polymorphic and most of their polymorphisms affect the peptide binding region (PBR), it becomes obvious that systematic HLA matching is crucial in determining the outcome of transplantation. The opposite holds true for the nonclassical HLA class I molecule HLA-E. HLA-E polymorphism is restricted to two functional versions and is thought to present a limited set of highly conserved peptides derived from class I leader sequences. However, HLA-E appears to be a ligand for the innate and adaptive immune system, where the immunological response to peptide-HLA-E complexes is dictated through the sequence of the bound peptide. Structural investigations clearly demonstrate how subtle amino acid differences impact the strength and response of the cognate CD94/NKG2 or T cell receptor.

NKG2A is a marker for acquisition of regulatory function by human CD8+ T cells activated with anti-CD3 antibody

Treatment with anti-CD3 mAb modulates immune responses that cause type 1 diabetes and other diseases. CD8+ Tregs can be induced in vitro and in vivo by mAb. However, 1/3 of patients do not respond to drug therapy and in an equal proportion, anti-CD3 mAb does not induce Tregs in vitro. The acquisition of CD8+ Treg activity is a function of the CD8+ cells and not the targets in the assay. To identify markers to differentiate responses of CD8+ Tregs, we analyzed genes differentially expressed in CD8+ T cells of non-responders compared with responders, and found that an inhibitory receptor NKG2A (CD159a) was highly expressed in cells from all non-responders tested. Application of a mAb agonistic to NKG2A during in vitro CD8+ Treg induction by anti-CD3 prevented induction of CD8+ Tregs. CD8+ T cells that are TNFR2+ but NKG2A- are the most potently induced Tregs. The level of NKG2A expression on resting CD8+ T cells inversely correlated with acquisition of regulatory function when activated. We suggest that the induction of human CD8+ Tregs by anti-CD3 mAb is controlled by a negative signaling through NKG2A, and that NKG2A may serve as a negative marker of human CD8+ Tregs.

Human cytomegalovirus immunity and immune evasion

Human cytomegalovirus (HCMV) infection induces both innate immune responses including Natural Killer cells as well as adaptive humoral and cell mediated (CD4+ helper, CD8+ cytotoxic and γδ T cell) responses which lead to the resolution of acute primary infection. Despite such a robust primary immune response, HCMV is still able to establish latency. Long term memory T cell responses are maintained at high frequency and are thought to prevent clinical disease following periodic reactivation of the virus. As such, a balance is established between the immune response and viral reactivation. Loss of this balance in the immunocompromised host can lead to unchecked viral replication following reactivation of latent virus, with consequent disease and mortality. HCMV encodes multiple immune evasion mechanisms that target both the innate and acquired immune system. This article describes the current understanding of Natural killer cell, antibody and T cell mediated immune responses and the mechanisms that the virus utilizes to subvert these responses.

The emerging role of HLA-E-restricted CD8+ T lymphocytes in the adaptive immune response to pathogens and tumors

Human leukocyte antigen (HLA)-E is a nonclassical major histocompatibility complex (MHC) class I molecule of limited sequence variability that is expressed by most tissues albeit at low levels. HLA-E has been first described as the ligand of CD94/NKG2 receptors expressed mainly by natural killer (NK) cells, thus confining its role to the regulation of NK-cell function. However, recent evidences obtained by our and other groups indicate that HLA-E complexed with peptides can interact with alphabeta T-cell receptor (TCR) expressed on CD8(+) T cells. Although, HLA-E displays a selective preference for nonameric peptides, derived from the leader sequence of various HLA class I alleles, several reports indicate that it can present also "noncanonical" peptides derived from both stress-related and pathogen-associated proteins. Because HLA-E displays binding specificity for innate CD94/NKG2 receptors, as well as all the features of an antigen-presenting molecule, its role in both natural and acquired immune responses has recently been re-evaluated.

Intracellular sequestration of the NKG2D ligand ULBP3 by human cytomegalovirus

Human CMV (HCMV) encodes multiple genes that control NK cell activation and cytotoxicity. Some of these HCMV-encoded gene products modulate NK cell activity as ligands expressed at the cell surface that engage inhibitory NK cell receptors, whereas others prevent the infected cell from upregulating ligands that bind to activating NK cell receptors. A major activating NKR is the homodimeric NKG2D receptor, which has eight distinct natural ligands in humans. It was shown that HCMV is able to prevent the surface expression of five of these ligands (MIC A/B and ULBP1, 2, and 6). In this article, we show that the HCMV gene product UL142 can prevent cell surface expression of ULBP3 during infection. We further show that UL142 interacts with ULBP3 and mediates its intracellular retention in a compartment that colocalizes with markers of the cis-Golgi complex. In doing so, UL142 prevents ULBP3 trafficking to the surface and protects transfected cells from NK-mediated cytotoxicity. This is the first description of a viral gene able to mediate downregulation of ULBP3.

NKG2A inhibits TH2 cell effector function in vitro

Background

We previously reported that NKG2A, a key inhibitory ligand for HLA-E, is expressed on activated TH2 but not TH1 cells. Here we measured cytokine expression in human ex vivo TH2 cells upon activation with anti-CD3/28 and challenge with an NKG2A-specific agonist.

Methods

TH2 cells were purified from healthy volunteers and activated with anti-CD3/28 in the presence and absence of NKG2A-specific agonist. IL-4 was used as a marker of TH2 effector function and measured by flow cytometry.

Results

Activation of TH2 cells increased NKG2A positivity from (Mean ± SE) 7.3 ± 2.4% to 13.7 ± 3.8%; (p = 0.03). The presence of NKG2A agonist did not significantly alter NKG2A expression, however, the percentage of activated TH2 cells expressing intracellular IL-4 decreased from 25.5 ± 6.8% to 9.3 ± 4.8% (p = 0.001).

Conclusion

We show that signalling through NKG2A suppresses TH2 effector function. This may provide a means to modulate Th1/Th2 balance in diseases where Th2 cytokines predominate.

Granzyme B and natural killer (NK) cell death

Granzyme B is a unique serine protease, which plays a crucial role for target cell death. Several mechanisms of delivery of granzyme B to target cells have been recently identified. Granzyme B directly activates Bid, a specific substrate for granzyme B, resulting in caspase activation. Granzyme B efficiently cleaves many prominent autoantigens, and the hypothesis that autoantibodies arise when cryptic determinants are revealed to the immune system has been proposed. Some autoantibodies directed against granzyme B-specific neoepitopes are present in serum from patients with autoimmune diseases. In the tissues from autoimmune diseases, granzyme B might play an important role for disease progression (i.e., rheumatoid arthritis synovium) or inhibition (i.e., regulatory T cells). We have identified a novel type of activation-induced cell death (granzyme B leakage-induced cell death). Activation-induced natural killer (NK) cell death is accompanied by the leakage of granzyme B from intracellular granules into the cytoplasm, and it triggers apoptosis by directing Bid to mitochondrial membranes. An excess of "leaked" granzyme B over its inhibitor, serpin proteinase inhibitor 9, is a major determinant of cell death. The role of granzyme B in autoimmunity and its influence on NK cell death are discussed.

NKG2A and CD56 are coexpressed on activated TH2 but not TH1 lymphocytes

NKG2A is commonly expressed on cytotoxic cells but has been found on activated T helper (TH) cells. In identifying novel markers differentiating between TH1 and TH2 lymphocytes, we focused on NKG2A expression. TH1 and TH2 cells were negatively isolated from healthy volunteers for microarray analysis and reverse transcription polymerase chain reaction (RT-PCR). Flow cytometry of quiescent and activated TH1 and TH2 cells was performed. Isolates were >95% pure CD3+CD4+ cells (TH1=90.3% and TH2=84.1%). Microarrays revealed differential expression of NKG2A and NKG2C isoforms between TH1 and TH2 cells. RT-PCR indicated greater expression of NKG2A in TH2 cells (4-fold) and NKG2C in TH1 cells (3-fold). Flow studies revealed tripling of TH2 NKG2A with activation to 10.76+/-4.01% (p=0.05), a 23-fold increase in CD56 to 35+/-14.54% (p=0.03), and an increase in NKG2A+CD56+ double-positive cells to 3.04+/-1.38% (p=0.04). TH1 lymphocytes did not differ with activation. We identified co-induction of NKG2A and CD56 on activation of TH2 cells. These cells would likely bind more HLA-E and exhibit increased effector inhibition. Given that certain viruses are known to decrease MHC class I and thus HLA-E production by antigen-presenting cells, activated TH2 cells would bind less HLA-E in this scenario. This would likely result in less effector inhibition and a relatively robust TH2 response.

Role that each NKG2A immunoreceptor tyrosine-based inhibitory motif plays in mediating the human CD94/NKG2A inhibitory signal

The human NKG2A chain of the CD94/NKG2A receptor contains two immunoreceptor Tyr-based inhibitory motifs (ITIMs) in its cytoplasmic tail. To determine the relative importance of membrane-distal (residues 6-11) and membrane-proximal (residues 38-43) ITIMs in mediating the inhibitory signal, we made site-directed mutants of NKG2A at the Y (Y8F, Y40F, Y8F/Y40F) and the residues two positions N-terminal (Y-2) of Y (V6A, I38A, V6A/I38A) in each motif. Wild-type (wt) and mutated NKG2A were then cotransfected with CD94 into rat basophilic leukemia 2H3 cells. Immunochemical analyses after pervanadate treatment showed that each of the mutant molecules could be phosphorylated to expected levels relative to wt NKG2A and that all the mutations significantly reduced the avidity of SH2 domain-bearing tyrosine phosphatase-1 for NKG2A. Confocal microscopy was used to determine whether SH2 domain-bearing tyrosine phosphatase-1 and CD94/NKG2A colocalized intracellularly after receptor ligation. Only the Y8F/Y40F and Y8F mutant NKG2A molecules failed to show a dramatic colocalization. In agreement with this result, the Y8F/Y40F mutant was unable to inhibit FcepsilonRI-mediated serotonin release and the Y8F mutant was relatively ineffective compared with wt NKG2A. In contrast, the Y40F mutant was 70% as effective as wt in mediating inhibition, and the Y-2 mutations did not remarkably affect inhibitory function. These results show that, like KIR, both NKG2A ITIMs are required for mediating the maximal inhibitory signal, but opposite to KIR, the membrane-distal ITIM is of primary importance rather than the membrane-proximal ITIM. This probably reflects the opposite orientation of the ITIMs in type II vs type I proteins.

Structure and function of natural killer cell receptors: multiple molecular solutions to self, nonself discrimination

In contrast to T cell receptors, signal transducing cell surface membrane molecules involved in the regulation of responses by cells of the innate immune system employ structures that are encoded in the genome rather than generated by somatic recombination and that recognize either classical MHC-I molecules or their structural relatives (such as MICA, RAE-1, or H-60). Considerable progress has recently been made in our understanding of molecular recognition by such molecules based on the determination of their three-dimensional structure, either in isolation or in complex with their MHC-I ligands. Those best studied are the receptors that are expressed on natural killer (NK) cells, but others are found on populations of T cells and other hematopoietic cells. These molecules fall into two major structural classes, those of the immunoglobulin superfamily (KIRs and LIRs) and of the C-type lectin-like family (Ly49, NKG2D, and CD94/NKG2). Here we summarize, in a functional context, the structures of the murine and human molecules that have recently been determined, with emphasis on how they bind different regions of their MHC-I ligands, and how this allows the discrimination of tumor or virus-infected cells from normal cells of the host.

Structure and function of major histocompatibility complex (MHC) class I specific receptors expressed on human natural killer (NK) cells

Natural killer (NK) cells express receptors that are specific for MHC class I molecules. These receptors play a crucial role in regulating the lytic and cytokine expression capabilities of NK cells. In humans, three distinct families of genes have been defined that encode for receptors of HLA class I molecules. The first family identified consists of type I transmembrane molecules belonging to the immunoglobulin (Ig) superfamily and are called killer cell Ig-like receptors (KIR). A second group of receptors belonging to the Ig superfamily, named ILT (for immunoglobulin like transcripts), has more recently been described. ILTs are expressed mainly on B, T and myeloid cells, but some members of this group are also expressed on NK cells. They are also referred to as LIRs (for leukocyte Ig-like receptor) and MIRs (for macrophage Ig-like receptor). The ligands for the KIR and some of the ILT receptors include classical (class Ia) HLA class I molecules, as well as the nonclassical (class Ib) HLA-G molecule. The third family of HLA class I receptors are C-type lectin family members and are composed of heterodimers of CD94 covalently associated with a member of the NKG2 family of molecules. The ligand for most members is the nonclassical class I molecule HLA-E. NKG2D, a member of the NKG2 family, is expressed as a homodimer, along with the adaptor molecule DAP10. The ligands of NKG2D include the human class I like molecules MICA and MICB, and the recently described ULBPs. Each of these three families of receptors has individual members that can recognize identical or similar ligands yet signal for activation or inhibition of cellular functions. This dichotomy correlates with particular structural features present in the transmembrane and intracytoplasmic portions of these molecules. In this review we will discuss the molecular structure, specificity, cellular expression patterns, and function of these HLA class I receptors, as well as the chromosomal location and genetic organization.

Inhibition of CD28-mediated natural cytotoxicity by KIR2DL2 does not require p56(lck) in the NK cell line YT-Indy

CD28 functions as a cytotoxicity activation receptor in the NK cell line YT-Indy. To analyze the requirement of p56(lck) kinase in the function of killer inhibitory receptors, we transfected the p56(lck) negative YT-Indy cell line with the cl43 gene encoding for KIR2DL2. Pervanadate treatment revealed KIR2DL2 phosphorylation in YT-Indy-cl43, as well as SHP1/SHP2 recruitment. YT-Indy-cl43 cells were inhibited in their ability to lyse target cells expressing HLA-Cw3, a ligand for KIR2DL2. This inhibition was blocked by anti-KIR2DL2 or anti-HLA class I mAb. CD28 crosslinking on YT-Indy-cl43 enhanced tyrosine phosphorylation of PLC-gamma1. The simultaneous ligation of KIR2DL2 with mAb resulted in a decrease in CD28-induced tyrosine phosphorylation of PLC-gamma1 confirming that dephosphorylation of this protein is involved in the KIR2DL2-induced inhibition of CD28-mediated cytotoxicity. As YT-Indy-cl43 did not express detectable levels of p56(lck), these results indicate that this kinase is not required for transmitting the negative signals generated by KIR2DL2 ligation.

Immunophenotypic characterization of normal blood CD56+lo versus CD56+hi NK-cell subsets and its impact on the understanding of their tissue distribution and functional properties

In the present study we have compared the immunophenotypic characteristics of the CD56+lo and CD56+hi NK-cell subsets in a group of normal healthy adults. Our results show that CD56+hi NK-cells display greater light-scatter properties than CD56+lo NK-cells at the same time they have higher levels of CD25 and CD122 IL-2 chains, together with a higher reactivity for HLA-DR and CD45RO and lower levels of CD45RA, supporting that, as opposed to the majority of the CD56+lo population, CD56+hi NK-cells might correspond to a subset of activated circulating NK-lymphocytes. Higher expression of the CD2 and CD7 costimulatory molecules found for the CD56+hi NK-cells would support their greater ability to respond to various stimuli. In addition, CD56+hi NK-cells expressed higher levels of several adhesion molecules such as CD2, CD11c, CD44, CD56, and CD62L compared to CD56+lo NK-cells, supporting a particular ability of these cells to migrate from blood to tissues and/or a potential advantage to form conjugates with target cells. Interestingly, CD56+lo and CD56+hi NK-cells showed a different pattern of expression of killer receptors that might determine different activation requirements for each of these NK-cell subsets. For instance, absence or low levels of CD16 expression might explain the lower antibody-dependent cytotoxicity activity of CD56+hi NK-cells. On the other hand, the virtual absence of expression of the CD158a and NKB1 immunoglobulin-like and the greater reactivity for the CD94 lectin-like killer receptors on CD56+hi in comparison to CD56+lo NK-cells might determine different MHC-class I specificities for both NK-cell subsets, a possibility that deserves further studies to be confirmed.

Genomic diversity of natural killer cell receptor genes in three populations

We report the distribution of genes encoding 11 killer cell immunoglobulin-like receptors (KIR) and 2 CD94:NKG2 receptors, in 32 Caucasians, 67 Australian Aborigines and 59 Vietnamese. The inhibitory and the activating KIR genes were found at different frequency in the three populations. No correlation was found between the polymorphism of the KIR genes and the HLA specificities of the tested samples. The most significant KIR associations were 2DL2 with 2DS2; 2DL2 with 2DS3 and 3DL1 with 2DS4 in all three study groups. In Caucasians and Vietnamese 2DS2 was associated with 2DS3 and 2DS1with 3DS1. KIR 2DL1 was strongly associated with three other KIRs: 2DL3, 3DL1 and 2DS4 in Aborigines. The distribution of the KIR phenotypes was different in the three populations. The AA1 phenotype was frequent in Vietnamese (42.4%) and Caucasians (31.2%), but very rare in Aborigines (1.5%). In contrast, the BB7 phenotype was very common for Aborigines (22.4%) and was absent in the two other groups. Our data demonstrate that different associations and putative KIR haplotypes could be distinguished in different populations.

Mapping the ligand of the NK inhibitory receptor Ly49A on living cells

We have used a recombinant, biotinylated form of the mouse NK cell inhibitory receptor, Ly49A, to visualize the expression of MHC class I (MHC-I) ligands on living lymphoid cells. A panel of murine strains, including MHC congenic lines, was examined. We detected binding of Ly49A to cells expressing H-2D(d), H-2D(k), and H-2D(p) but not to those expressing other MHC molecules. Cells of the MHC-recombinant strain B10.PL (H-2(u)) not only bound Ly49A but also inhibited cytolysis by Ly49A(+) effector cells, consistent with the correlation of in vitro binding and NK cell function. Binding of Ly49A to H-2D(d)-bearing cells of different lymphoid tissues was proportional to the level of H-2D(d) expression and was not related to the lineage of the cells examined. These binding results, interpreted in the context of amino acid sequence comparisons and the recently determined three-dimensional structure of the Ly49A/H-2D(d) complex, suggest a role for amino acid residues at the amino-terminal end of the alpha1 helix of the MHC-I molecule for Ly49A interaction. This view is supported by a marked decrease in affinity of an H-2D(d) mutant, I52 M, for Ly49A. Thus, allelic variation of MHC-I molecules controls measurable affinity for the NK inhibitory receptor Ly49A and explains differences in functional recognition in different mouse strains.