Innate lymphocyte function in malaria

National Institute Of Allergy And Infectious Diseases

Investigators

Linked publications, trials & patents

Abstract

We have shown that IgG isolated from adults living in a malaria-endemic region induced ADCC by primary human NK cells toward infected erythrocytes in vitro. NK cells lysed P. falciparum-infected erythrocytes and inhibited parasite growth in an in vitro assay for ADCC-dependent growth inhibition. In the absence of antibodies, NK cells exhibited negligible natural cytotoxicity towards infected erythrocytes. Granzyme B activity in infected erythrocytes was detected during co-incubation with NK cells after loading erythrocytes with a fluorescent probe that is activated by proteolytic cleavage by granzyme B. Here, we addressed two major questions. First, what are the requirements for tight conjugate formation between NK cells and infected erythrocytes, and for the delivery of granzyme B into target cells? Second, what is the fate of infected erythrocytes and of parasitophorous vacuoles during incubation with NK cells. Lysis of target cells by NK cells depends on two processes, which are triggered by different signals. First, degranulation of NK cells, which releases perforin and granzymes, is triggered by either Fc receptor CD16 or a combination of signals from synergistic co-activation receptors. In the case of infected erythrocytes, degranulation is induced by antibodies that trigger CD16 on NK cells. Second, tight adhesion to target cells is required. Signaling by integrin alphaLbeta2 (LFA-1) is sufficient to induce conjugation and polarization of NK cells upon binding to integrin ligands ICAM-1 or ICAM-2 on target cells. As erythrocytes do not express ICAM-1 or ICAM-2, it is not known how NK cells bind and what receptor(s) might be involved. Erythrocytes do not have ligands for the main NK co-activation receptors, such as Ncr1, Ncr2, Ncr3, NKG2D, 2B4, DNAM-1 and CD28H. Degranulation alone, in the absence of a polarized immunological synapse, is not sufficient for lysis of target cells. What then are the receptor-ligand interactions that bring NK cells and infected erythrocytes into conjugates? Among several candidate ligands on erythrocytes, the two that emerged as likely to be functional in binding NK cells were CD58 (LFA-3) and ICAM-4. CD58 is the ligand for human CD2, a co-receptor that enhances immunological synapse formation and signaling by the T-cell receptor. Most NK cells express CD2. ICAM-4 (blood group LW) is a non-canonical member of the ICAM family and lacks the key glutamate residue required for high affinity binding of beta2 integrins. Conflicting reports have suggested that ICAM-4 binds either integrin alpha4beta1 (VLA-4) or alphaMbeta2 (Mac-1). We used blocking antibodies to integrin subunits to test their effect on NK interactions with infected erythrocytes. Assays for conjugate formation, granzyme B delivery into target cells, and hemoglobin release were performed. Consistently, a CD2 antibody reduced the NK response. A CD11b (alphaM) antibody had a smaller effect, while several antibodies to other molecules had no effect, such as CD49d (alpha4) and CD29 (beta1). To validate these results and obtain genetic evidence, we initiated a collaboration with Dr. Duraisingh (Harvard U.). Using CRISPR/Cas9, his lab has knocked out individual genes in erythrocytic precursors, which were then differentiated into late-stage erythrocytes. A CD58-KO clone consistently showed reduced interaction with and sensitivity to NK cells. Furthermore, blocking experiments with the CD11b (alphaM) antibody showed greater inhibition with the CD58-KO erythrocytes than it did with the Cas9+ wild-type erythrocyte control. These results suggest some redundancy of CD2 and Mac-1 for conjugation of NK cells with erythrocytes, but that their combined engagement with their respective ligands provides an optimal response. ICAM-4-KO erythrocytic cells supported this interpretation, as there was a reduced NK cell response as well as a stronger inhibitory effect of the blocking CD2 antibody. Transmission electron and scanning electron microscopy showed tight interactions between NK cells and infected erythrocytes in the presence of plasma from adults living in a malaria endemic region of Mali. Such interactions were not observed with uninfected erythrocytes. In some of the NK-infected erythrocyte conjugates, loss of erythrocyte membrane integrity was evident. Furthermore, smaller, round bodies were observed, which had a size consistent with that of parasitophorous vacuoles (PV). Flow cytometry analysis showed that PVs reacted very strongly with plasma antibodies of clinically immune Mali individuals. The 3D7 P. falciparum line was engineered to express GFP-tagged Pf-EXP2 in order to have a fluorescent marker in the PV membrane. EXP2 is a transmembrane subunit of a transporter in the PV membrane, which is exposed to the cytosol of erythrocytes. GFP+ PVs were detected inside erythrocyte ghosts after incubation with NK cells and Mali plasma. Free PVs accumulated after longer incubations with NK cells. In the presence of Mali plasma, PVs were efficiently phagocytosed by primary monocytes. This work has shown that PVs are released from infected erythrocytes in the presence of NK cells and immune plasma and are phagocytosed by monocytes. It suggests that the release of relatively intact PVs from lysed erythrocytes and their removal by monocytes serves to limit the inflammatory response that occurs when hemozoin is released into the blood stream.