The reactivity of Bw4+ HLA-B and HLA-A alleles with KIR3DL1: implications for patient and donor suitability for haploidentical stem cell transplantations.

Natural killer (NK)-cell alloreactivity can be exploited in haploidentical hematopoietic stem cell transplantation (HSCT). NK cells from donors whose HLA type includes Bw4, a public epitope present on a subset of HLA-B alleles, can be alloreactive toward recipients whose cells lack Bw4. Serologically detectable epitopes related to Bw4 also exist on a subset of HLA-A alleles, but the interaction of these alleles with KIR3DL1 is controversial. We therefore undertook a systematic analysis of the ability of most common HLA-B alleles and HLA-A alleles with Bw4 serologic reactivity to protect target cells from lysis by KIR3DL1-dependent NK cells. All Bw4(-) HLA-B alleles failed to protect target cells from lysis. All Bw4(+) HLA-B alleles with the exception of HLA-B*1301 and -B*1302 protected targets from lysis. HLA-A*2402 and HLA-A*3201 unequivocally protected target cells from lysis, whereas HLA-A*2501 and HLA-A*2301 provided only weak protection from lysis. KIR3DL1-dependent alloreactive NK clones were identified in donors with HLA-A*2402 but not in donors with HLA-B*1301 or -B*1302. These findings clarify the HLA types that donors and recipients need in haploidentical HSCT and other NK allotherapies in order to benefit from NK alloreactivity.

KIR2DS1-mediated activation overrides NKG2A-mediated inhibition in HLA-C C2-negative individuals.

NK cell cytotoxicity is controlled through a balance of both activating and inhibitory signals. The HLA specificity of alloreactive NK cells has been previously shown to be controlled by inhibitory killer immunoglobulin-like receptors (KIRs). Alloreactive NK cells lyse targets that lack the HLA ligand for their inhibitory KIR. We have characterized in detail an alloreactive NK clone in which the specificity is controlled by an activating receptor, KIR2DS1. Only target cells expressing the HLA-C group 2 (C2) epitope were lysed by this clone and homozygous C2 targets were lysed more strongly than heterozygous C1/C2 targets. Anti-CD158a (KIR2DS1) blocked lysis of targets confirming KIR2DS1 was responsible. Although this NK clone expressed NKG2A, an inhibitory receptor whose ligand is HLA-E, targets with ligands for both KIR2DS1 and NKG2A were lysed by this clone indicating that the KIR2DS1-mediated activation signal overrides the NKG2A-mediated inhibitory signal. KIR2DS1 activated NK clones in polyclonally expanded NK cultures from a donor that lacked the C2 epitope accounted for approximately 1% of all NK cells. This study highlights a potential role for NK cells controlled by activating KIR in mediating NK alloreactivity.

Killer Ig-like receptor expression in uterine NK cells is biased toward recognition of HLA-C and alters with gestational age.

Immunogenetic studies suggest that interactions between maternal killer Ig-like receptor (KIR) expressed by uterine NK (uNK) cells, and fetal HLA-C molecules on trophoblast, influence the success of human placentation. However, the exact functional response of fresh uNK cells to trophoblast HLA-C molecules is unknown. In this study, we show by quantitative RT-PCR and FACS that both activating and inhibitory KIR specific for HLA-C are expressed at higher levels and on an increased proportion of NK cells in the human decidua compared with blood. In contrast, expression of KIR3DL1/S1, which is specific for HLA-B, is similar in both NK cell populations. Remarkably, there is also a temporal change in the expression pattern of HLA-C-specific KIR, with a decline in both intensity of expression and frequency on uNK cells throughout the first trimester of pregnancy. This selective up-regulation of KIR has functional consequences because uNK cells show increased binding of HLA-C tetramers compared with blood NK cells. Ab cross-linking shows that these KIR are functional and results in increased cytokine secretion. uNK cells, therefore, exhibit a unique KIR profile that enhances their ability to recognize trophoblast cells expressing HLA-C at the materno-fetal interface. This is the first report to demonstrate selective regulation of KIR expression over time in vivo in a normal physiological situation and suggests that KIR expression by uNK cells is regulated by the tissue microenvironment in the decidua.

Three common alleles of KIR2DL4 (CD158d) encode constitutively expressed, inducible and secreted receptors in NK cells.

Genetic polymorphism of KIR2DL4 results in alleles with either 9 or 10 consecutive adenines in exon 6, which encodes the transmembrane domain. "10A" alleles encode a membrane-expressed receptor that is constitutively expressed on resting CD56bright NK cells and on CD56dim cells after culture. However, in some individuals with the 10A allele, KIR2DL4 cannot be detected on their resting CD56bright NK cells. "9A" alleles have been predicted to encode a secreted receptor due to the splicing out of the transmembrane region. In this publication, we show that those individuals with a 10A allele who lack detectable KIR2DL4 on CD56bright NK cells express a KIR2DL4 receptor in which the D0-domain is excised. This Delta-D0 receptor cannot be detected by the available anti-KIR2DL4 monoclonal antibodies. In such individuals, KIR2DL4 becomes detectable on cultured NK cells due to up-regulation of the full-length KIR2DL4 transcript. In all individuals with 10A alleles, KIR2DL4 ceases to be expressed at the cell surface 16 days after activation, despite the maintenance of maximal levels of KIR2DL4 mRNA transcription, suggesting the existence of a negative regulator of cell surface expression. Finally, we show that the 9A allele can produce a secreted KIR2DL4 receptor.

The murine cytomegalovirus M73.5 gene, a member of a 3' co-terminal alternatively spliced gene family, encodes the gp24 virion glycoprotein.

We have identified a novel family of five 3' co-terminal transcripts in murine cytomegalovirus (MCMV) arranged in a tail-to-tail orientation with respect to the MCMV glycoprotein H (gH) gene M75. They share the same exon 2 sequence but possess different exon 1 sequences. Two of these spliced transcripts (M73) encode the MCMV homolog of glycoprotein N (gN) entirely within exon 1. Two other transcripts designated M73.5 encode a previously described virion glycoprotein gp24 that shares its first 20 amino acids with gN, but which has another 64 amino acids encoded within exon 2. The fifth transcript, designated m60, has an 80-bp exon 1 near the MCMV oriLyt region 10.8 kb upstream of exon 2. Both MCMV M73.5 and m60 encode type II glycoproteins expressed at the cell surface. Their shared exon 2 coding sequences likely represent the highly conserved region of an as yet unidentified betaherpesvirus-specific glycoprotein species.