Ly108, a glycoprotein of the signaling lymphocytic activation molecule family of cell surface receptors expressed by T, B, NK, and APCs has been shown to have a role in NK cell cytotoxicity and T cell cytokine responses. In this study, we describe that CD4+ T cells from mice with a targeted disruption of exons 2 and 3 of Ly108 (Ly108ΔE2+3) produce significantly less IL-4 than wild-type CD4+ cells, as judged by in vitro assays and by in vivo responses to cutaneous infection with Leishmania mexicana. Surprisingly, neutrophil functions are controlled by Ly108. Ly108ΔE2+3 mice are highly susceptible to infection with Salmonella typhimurium, bactericidal activity of Ly108ΔE2+3 neutrophils is defective, and their production of IL-6, IL-12, and TNF-α is increased. The aberrant bactericidal activity by Ly108ΔE2+3 neutrophils is a consequence of severely reduced production of reactive oxygen species following phagocytosis of bacteria. Thus, Ly108 serves as a regulator of both innate and adaptive immune responses.
The signaling lymphocytic activation molecule (SLAM) 4 family of immune receptors, which includes the SLAM-associated protein (SAP)-binding receptors SLAM, Ly108, CD84, CS1, Ly-9, 2B4, CD48, BLAME, and SF2001, are thought to play a role in innate and adaptive immunity (1, 2). Ly108 (NTB-A, SF2000, KALI, SF-3) is a membrane glycoprotein of the SLAM family expressed on T cells, B cells, macrophages, dendritic cells, and granulocytes (3, 4, 5). Ly108 has been shown to function on NK cells by augmenting cytotoxicity (4); this function is impaired in NK cells derived from X-linked lymphoproliferative disease patients who lack expression of the adapter SAP. A recent report suggests that anti-Ly108 Ab cross-linking induces IFN-γ production by T cells (6). Both Ly108 and SLAM are homotypic adhesion receptors with two cytoplasmic immunoreceptor tyrosine-based switch motif domains which are tyrosine-phosphorylated upon receptor cross-linking (6, 7). The cell and tissue distribution of SLAM and Ly108 are very similar and T cell IFN-γ responses are augmented by Abs against either receptor (8, 9). T cell signals mediated by SLAM are partially regulated by the adapter SAP (SH2D1A), which binds to the immunoreceptor tyrosine-based switch motif domains in the receptors’ cytoplasmic tail, inducing activation of Fyn and downstream phosphorylation of Dok1/2, SH2-containing protein, Ras-GTPase-activating protein, and SHIP in T cells (7, 10, 11). EWS/FL11-associated transcript 2 is structurally related to SAP and thought to have a similar function in APCs (1, 12).
We have recently demonstrated that mice deficient in SLAM have impaired macrophage responses to LPS stimulation and diminished Th2 cytokine production (2). Because an IL-4 defect is also observed in CD4+ cells from mice deficient in SAP and because this defect appears to be more robust in SAP-deficient animals than in SLAM-deficient mice, we hypothesize that other SLAM-related receptors might have a similar phenotype (13, 14). This prompted us to investigate the role of Ly108 in adaptive and innate immune responses. In this study, we report that in a mouse with a targeted disruption of the Ly108 gene CD4+ T cell and innate responses are defective. The results of these studies demonstrate a surprising role for Ly108 in the control of responses to bacteria by neutrophils while macrophage functions are intact.
Materials and Methods
Generation of Ly108-deficient (Ly108ΔE2+3) mice
A targeting construct was generated from a 129/Sv mouse pBAC clone (CD84.361) and was cloned into the plasmid vector pPNT. The second and third exons of the Ly108 gene, encoding the complete ectodomain of Ly108, were replaced with the neomycin resistance gene.
The targeting vector was linearized and electroporated into embryonic stem (ES) cells. G418-resistant ES cell colonies were screened by Southern blot. SpeI digestion of genomic DNA generated a 12-kb band from the endogenous wild-type (WT) Ly108 allele, while the correctly targeted Ly108 allele generated an additional 6.1-kb band (Fig. 1, A and B). The single integration site was confirmed with the internal 5′ probe upon SpeI digestion of the DNA (our unpublished data). A Ly108+/− ES cell clone was injected into C57BL/6 blastocysts. F1 mice with germline transmission of Ly108+/− (C57BL/6 × 129/Sv) were bred to homozygosity. Ly108ΔE2+3 mice were kept under specific pathogen-free conditions.
Mice were genotyped by genomic PCR. Primers P1–P4 were used for typing, P1 and P2 amplify the second exon, giving a 540-bp band, while P3 and P4 amplify the neomycin gene to produce a 700-bp band (Fig. 1 B). The sequences are: P1, 5′-GAGACCATAAGTTAGGATCATC-3′; P2, 5′-CAGTGTATGATCCTGTGTCTG-3′; P3, 5′-GCAGCGCATCGCCTTCTATC-3′; and P4, 5′-CACCTAGATCTCTTACTCCTC-3′. All mice in this study were of the C57BL/6 × 129sv background; control mice (C57BL/6 × 129sv F1) were purchased from The Jackson Laboratory.
RT-PCR was performed as previously described (2, 8) Ly108 fragments spanning exons 2 and 3 were amplified using a 5′ primer in exon 2 which was 5′-GGGAAGATAGCCAATATCATCAT-3′ and 3′ primer in exon 3 which was 5′-GCAGAGACTCTGGGTCGAAA-3′. Fragments spanning exons 1–8 were amplified with a 5′ primer TCAGAGGATGGTCTGGCTCT in exon 1 and a 3′ primer AGCGTGTGGATGAGTTACCC in exon 8.
T cell stimulation proliferation assays and Th1/Th2 polarization
ELISA for T cell, polymorphonuclear neutrophils (PMNs), and macrophage cytokine secretion
Infection with Leishmania mexicana
L. mexicana infections were performed as previously described (15). Lesion diameter was measured at 1-wk intervals for up to 8 wk.
Infection with Salmonella typhimurium
WT and Ly108ΔE2+3 mice were challenged i.p with 1 × 105 CFU of S. typhimurium 14028s or sseB, an attenuated isogenic mutant of the 14028s strain of S. typhimurium. Mice were injected i.p. with 1 × 105 bacteria in 2 ml of PBS. Blood samples were taken from the tail vein at 24 h for cytokine analysis. Time to death was recorded.
Isolation of macrophages and PMNs
Thioglycolate-elicited macrophages (TEPM) were obtained as previously described (2). PMNs were isolated from bone marrow or alternatively from the peritoneum 4 h after injection with 2 ml of 5% Brewer’s thioglycolate medium. Bone marrow or thioglycolate peritoneal lavage was washed three times in HBSS/5% FCS. PMNs were then isolated by discontinuous Percoll gradient centrifugation. Using this technique, >95% purity was routinely obtained as assessed by Wright-Giemsa staining.
Gentamicin-protection bacterial killing assay
Macrophage bactericidal activity was measured using a gentamicin protection assay as previously described (2).
PMN opsono-phagocytic killing assay
Bone marrow-derived or peritoneal-derived PMNs were washed three times in HBSS/5% FCS before re-suspension at 1 × 106 in HBSS supplemented with 50% fresh autologous mouse serum. Bacteria opsonized in 20% fresh normal mouse serum at 37°C for 30 min were added to the PMNs at ratios of 3:1, 2:1, or 1:1 PMNs:bacteria and incubated at 37°C with end-over-end mixing. Fifty-microliter aliquots were extracted at 0, 30, 60, 90, and 120 min and lysed in 10 ml of sterile water for 15 min at 25°C. Twenty microliters was then plated directly onto Luria-Bertani agar plates, and bacterial colonies were counted after an 18-h incubation at 37°C.
Flow cytometric measurement of PMN phagocytosis
Bone marrow-derived PMNs (4 × 106/ml in HBSS/5% FCS) were incubated for various periods with 4 × 108 paraformaldehyde-fixed and opsonized GFP-expressing Escherichia coli strain MS589 (a kind gift from Dr. P. Klemm, Technical University of Denmark, Lyngby, Denmark). Cells were washed three times in ice-cold PBS followed by a 60-s wash in 0.4% trypan blue to quench extracellular GFP and a final wash in PBS before flow cytometry. As a negative control for nonspecific bacterial-PMN adhesion, a portion of the PMNs was fixed for 10 min in 2% paraformaldehyde before the assay.
Measurement of superoxide generation
Superoxide production was measured with lucigenin. PMNs and macrophages resuspended in HBSS/5% FCS at 2.5 × 105 and 1 × 106/ml, respectively, were stimulated for 3 h with 8 × 107 heat-killed, opsonized E. coli strain F18 or PMA at 1 μg/ml for 15 min. Luminescence was measured with a TD2020 luminometer (Turner Designs).
Results and Discussion
Generation of Ly108ΔE2+3 mice
A mouse with a targeted disruption of the second and third exons of Ly108, encoding its entire ectodomain, was generated by homologous recombination in ES cells (Fig. 1, A and B). Ly108ΔE2+3 mice were fertile, morphologically indistinguishable from WT littermates, and no differences in T, B, or NK development were detected by cell surface marker analysis (our unpublished data). No transcripts encoding Ly108 exons 2 and 3 mRNA was detected by RT-PCR in knockout thymus (Fig. 1 C). A 650-bp transcript was detectable with primers spanning the signal peptide to the 3′ untranslated region, indicating that a short residual transcript encoding the transmembrane and cytoplasmic domain remained in these mice. Since Ly108 is a self-ligand, it was anticipated that removal of the entire extracellular portion of the receptor would result in a total loss of Ly108 function. It is possible however that the Ly108ΔE2+3 mutation results in a different phenotype from a Ly108null mouse.
Impaired IL-4 production by Ly108ΔE2+3 CD4+ T cells
To determine whether Ly108ΔE2+3 T cells deviated in their cytokine production in a way similar to those derived from SLAM−/− and SAP−/− mice, we performed various in vitro and in vivo analyses. First, splenic CD4+ T cells were stimulated in vitro with anti-CD3 and anti-CD28 or PMA and ionomycin, followed by analysis of cell supernatant cytokines using ELISA. Ly108ΔE2+3 CD4+ T cells produce significantly less IL-4 than WT T cells even after PMA/ionomycin stimulation, whereas production of IFN-γ was normal (Fig. 2,A). Ly108ΔE2+3 CD4+ T cells stimulated with anti-CD3 Abs also produced less IL-13 than WT T cells as assessed by semiquantitative cytokine array analysis (our unpublished data). Ly108ΔE2+3 CD8+ T cell production of IFN-γ did not differ from WT CD8+ T cells (our unpublished data). We next determined whether the observed defect in IL-4 production by Ly108ΔE2+3 CD4+ T cells could be “rescued” by a secondary stimulation or by Th2 polarization. Following secondary stimulation, IL-4 production by Ly108ΔE2+3 CD4+ T cells was still lower and IL-4 production following Th2 polarization was ∼50% of that of WT CD4+ T cells (Fig. 2,B). Conversely, polarization of CD4+ T cells toward a Th1 phenotype resulted in equivalent IFN-γ production by CD4+ T cells from WT and Ly108ΔE2+3 mice (Fig. 2 B). The proliferative response of Ly108ΔE2+3 CD4+ T cells to anti-CD3/CD28 stimulation was normal (our unpublished observations).
To confirm the defect in IL-4 production in vivo, we analyzed the ability of Ly108ΔE2+3 mice to mount an inflammatory response to infection with L. mexicana. Th2 responses upon infection with L. mexicana are a prerequisite for controlling the progression of lesions caused by the parasite and consequently serves as a useful indicator for the correct balance of Th1 and Th2 responses (16). Whereas IL-4 is necessary for lesion formation, IFN-γ production is required for protective host immunity after L. mexicana infection (17, 18, 19). In IL-4−/− mice, Th1 responses are predominant, which results in healing of the lesions (17). Ly108ΔE2+3 mice infected with L. mexicana exhibited delayed formation of lesions compared with WT mice (5 wk in Ly108−/− animals, 3 wk in WT) and developed significantly smaller lesions (Fig. 2 D). In vitro Ag restimulation of lymph node CD4+ T cells from the L. mexicana-infected mice revealed lower IL-4 production by the Ly108ΔE2+3 T cells (536 ± 124 vs 229 ± 35 pg/ml). This result was consistent with the in vitro observation of impaired IL-4 production by CD4+ T cells. Thus, in Ly108ΔE2+3, SLAM−/−, and SAP−/− mice Th2 functions are impaired; the Ly108ΔE2+3 phenotype is more robust than observed in the SLAM−/− mouse, since IL-4 production is impaired even upon stimulation with PMA and ionomycin (13).
Increased susceptibility to bacterial infection in the absence of Ly108
Because we had observed altered innate immune responses and, in particular, a macrophage defect in SLAM−/− mice (2), we next examined a role for Ly108 in innate immunity. To this end, WT and Ly108ΔE2+3 mice were infected i.p with 1 × 105 of S. typhimurium 14028s or a congenic sseB mutant of S. typhimurium, deficient in the SPI2-encoded type III secretory system. S. typhimurium sseB is attenuated for virulence and is therefore cleared efficiently by WT mice. Twenty-four hours after infection with 14028s, Ly108ΔE2+3 mice displayed signs of severe salmonellosis including hunching, pilial erection, and lethargy. WT controls appeared to be healthy. Ly108ΔE2+3 mice suffered from accelerated morbidity in response to the WT S. typhimurium strain 14028s (100% succumbed to infection in 3 days vs 5 days for WT mice; Fig. 3,A) and displayed unusual sensitivity to the attenuated sseB strain (40% of Ly108ΔE2+3 mice succumbed to infection vs no WT animals). Analysis of serum cytokines in S. typhimurium sseB-infected Ly108ΔE2+3 mice showed a 4- to 7-fold increase over WT mice in the amounts of circulating IL-12p40, TNF-α, and IL-6 (Fig. 3 B).
We next investigated the possibility that increased susceptibility to bacterial infection in Ly108ΔE2+3 mice might be due to aberrant neutrophil or macrophage responses. Bone marrow PMNs were tested for in vitro cytokine production in response to stimulation with bacterial (E. coli) LPS and peptidoglycan (PGN) from Staphylococcus aureus (Fig. 3,C). Surprisingly, PMNs from Ly108ΔE2+3 mice produced 5-fold more IL-12p40 than WT PMNs in response to LPS and twice as much TNF-α. IL-6 production was also moderately elevated by PMNs from Ly108ΔE2+3 mice. Cytokine production by macrophages was, however, not significantly different between WT and Ly108ΔE2+3 mice (our unpublished observations). Cytokine production by neutrophils in response to PGN did not increase significantly above constitutive levels, and no differences were observed between WT and Ly108ΔE2+3 neutrophils in this respect, despite PGN inducing robust TNF-α responses in both WT and Ly108ΔE2+3 macrophages (our unpublished data). Both bone marrow neutrophils and peritoneal macrophages expressed Ly108 mRNA (Fig. 3 D).
We then tested the Ly108ΔE2+3 PMNs ability to phagocytose and kill bacteria. As shown in Fig. 4,A, PMNs from Ly108ΔE2+3 mice were impaired in their bactericidal activity, displaying a significant lag in time to clear bacteria in vitro. Killing of S. aureus was also diminished in PMNs from Ly108ΔE2+3 mice, as was killing of E. coli by thioglycolate-elicited peritoneal PMNs from Ly108ΔE2+3 mice (our unpublished data). To assess whether the defect in PMN killing in the absence of Ly108 was attributable to impaired uptake of bacteria, a flow cytometric analysis of phagocytosis was used. Ly108ΔE2+3 PMNs were efficient in phagocytosis of paraformaldehyde-fixed GFP expressing E. coli (Fig. 4,B). In contrast to PMNs, Ly108ΔE2+3 peritoneal macrophages were competent in both phagocytosis (2-h time point) and killing of bacteria after 6 and 24 h (Fig. 4 C). Thus, Ly108ΔE2+3 PMNs are defective in their responses to bacteria, while macrophage functions appear normal.
Reduced PMN oxidative burst in the absence of Ly108
Following phagocytosis of bacteria, both PMNs and macrophages elicit a respiratory burst of reactive oxygen species (ROS) and NO into the bacteria-containing phagolysosome. To explain the significantly delayed bacterial killing by Ly108ΔE2+3 PMNs, we examined both their NO and ROS production. No difference in NO production in response to LPS and IFN-γ was observed between WT and Ly108ΔE2+3 PMNs (our unpublished data). However, a dramatic reduction in ROS production by Ly108ΔE2+3 PMNs in response to heat-killed E. coli was observed (Fig. 4 D). As predicted by the bacterial killing experiments, production of ROS by Ly108ΔE2+3 macrophages in response to bacterial phagocytosis was normal. Analysis of ROS generation in response to PMA, a stimulus which bypasses receptor involvement, indicated that both PMNs and macrophages from Ly108ΔE2+3 mice made robust ROS responses equal to or, in the case of Ly108ΔE2+3 macrophages, exceeding that of WT cells.
In conclusion, we report here for the first time a critical role for the SLAM family receptor Ly108 in CD4+ T cell responses and innate immunity to bacteria and parasites. This is the first report of the involvement of such a cell surface receptor in bacterial phagosomal killing. It will be of great interest to elucidate the biochemical mechanisms involved in Ly108 induction of cytokines in T cells and oxidative burst in PMNs.
We thank Ana Aabadia Molina, Tanya Mayadas, Xavier Cullere, and Ahmad Utomo for experimental advice and review of this manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
D.H. was supported by a Fellowship from the Leukemia and Lymphoma Society of America and C.T. was supported by a grant from the March of Dimes.
Abbreviations used in this paper: SLAM. signaling lymphocytic activation molecule; SAP, SLAM-associated protein; ROS, reactive oxygen species; TEPM, thioglycolate-elicited peritoneal macrophage; ES, embryonic stem; WT, wild type; PMN, polymorphonuclear neutrophil; PGN, peptidoglycan.