Research ArticleANTIBODIES

Enhancing FcγR-mediated antibody effector function during persistent viral infection

See allHide authors and affiliations

Science Immunology  21 Sep 2018:
Vol. 3, Issue 27, eaao3125
DOI: 10.1126/sciimmunol.aao3125

Giving antibodies a boost

Persistent immune activation during chronic infections is often associated with increased generation and deposition of immune complexes. The actions of antibody-based drugs can therefore be severely impaired in individuals with chronic infections. Using lymphocytic choriomeningitis virus (LCMV) as a model of chronic infection, Wieland et al. have examined how to enhance antibody functions in this setting. They found that the ability of antibodies to deplete target cells is dependent on antigen expression levels. Further, they also found that afucosylation of antibodies directed against CD4 and CD8α enhanced the ability of these antibodies to deplete CD4+ and CD8+ T cells in mice persistently infected with LCMV. Whether afucosylation can be universally used to enhance antibody functions during chronic infections remains to be seen.

Abstract

Persistent viral infections can interfere with FcγR-mediated antibody effector functions by excessive immune complex (IC) formation, resulting in resistance to therapeutic FcγR-dependent antibodies. We and others have previously demonstrated that mice persistently infected with lymphocytic choriomeningitis virus (LCMV) are resistant to a wide range of depleting antibodies due to excessive IC formation. Here, we dissect the mechanisms by which two depleting antibodies overcome the obstacle of endogenous ICs and achieve efficient target cell depletion in persistently infected mice. Efficient antibody-mediated depletion during persistent LCMV infection required increased levels of antibody bound to target cells or use of afucosylated antibodies with increased affinity for FcγRs. Antibodies targeting the highly expressed CD90 antigen or overexpressed human CD20 efficiently depleted their target cells in naïve and persistently infected mice, whereas antibodies directed against less abundant antigens failed to deplete their target cells during persistent LCMV infection. In addition, we demonstrate the superior activity of afucosylated antibodies in the presence of endogenous ICs. We generated afucosylated antibodies directed against CD4 and CD8α, which, in contrast to their parental fucosylated versions, efficiently depleted their respective target cells in persistently infected mice. Efficient antibody-mediated depletion can thus be achieved if therapeutic antibodies can outcompete endogenous ICs for access to FcγRs either by targeting highly expressed antigens or by increased affinity for FcγRs. Our findings have implications for the optimization of therapeutic antibodies and provide strategies to allow efficient FcγR engagement in the presence of competing endogenous ICs in persistent viral infections, autoimmune diseases, and cancer.

INTRODUCTION

Various immune cell dysfunctions have been documented during persistent viral infections such as hepatitis B virus (HBV), hepatitis C virus (HCV), and human immunodeficiency virus (HIV). Whereas the mechanisms and consequences of T cell dysfunction during persistent viral infections are well documented, little is known about the impact of antigen persistence on the humoral immune response and antibody-dependent effector functions. We and others have recently shown that excessive immune complex (IC) formation in mice persistently infected with lymphocytic choriomeningitis virus clone-13 (LCMV clone-13) interferes with various Fcγ receptor (FcγR)–mediated antibody effector functions (1, 2). IC formation during persistent LCMV infection was shown to interfere with FcγR-mediated antibody effector functions without negatively affecting the overall expression pattern of FcγRs (1, 2). Furthermore, the inhibition of FcγR-mediated effector functions during persistent LCMV infection was due to IC formation and not the persistent infection per se, as persistent LCMV infection in the absence of an antiviral B cell response did not interfere with FcγR-mediated effector function.

The efficient engagement of FcγRs on immune effector cells, such as macrophages and natural killer cells, by therapeutic immunoglobulin Gs (IgGs) such as rituximab, an anti-CD20 monoclonal antibody (mAb) used for the treatment of B cell non-Hodgkin lymphoma (NHL), is crucial for their in vivo activity (3, 4). Furthermore, broadly neutralizing antibodies against persistent viral infections such as HIV require Fc-FcγR interactions for optimal therapeutic activity in vivo (5, 6). During persistent LCMV infection, ICs inhibited FcγR-mediated effector functions involved in controlling viral infections, including antibody-dependent cytotoxicity (ADCC) of infected cells and antigen cross-presentation to CD8+ T cells, and required for the therapeutic activity of depleting and agonistic antibodies (1, 2).

Elevated levels of circulating ICs and deposition of ICs in the kidney resulting in glomerulonephritis have been reported to occur in patients persistently infected with viruses such as HBV, HCV, and HIV (79). Furthermore, circulating ICs have been detected in the sera of cancer patients, and epidemiological data show an association between membranous nephropathy and cancer (10, 11). In addition, excessive IC formation due to the presence of autoantibodies is a hallmark of systemic lupus erythematosus (SLE), an autoimmune disease in which, upon rituximab treatment, considerable variation in B cell depletion efficacy is observed (12). It was subsequently shown in the murine MRL/lpr lupus model that increased levels of circulating ICs could reduce or even completely abrogate B cell depletion by rituximab (13, 14). Together, these studies demonstrate that excessive IC formation and subsequent interference with FcγR-mediated effector functions occur in many diseases, whereas the underlying mechanisms and therapeutic consequences are still ill-defined.

Whereas persistent LCMV infection abrogated the therapeutic activity of a wide range of depleting antibodies (Table 1), the depletion of T cells using an anti-CD90.2 antibody was not impaired, demonstrating the general feasibility to deplete cells in the presence of high levels of ICs (1). However, the exact mechanism underlying the depletion efficiency of the anti-CD90.2 antibody during persistent LCMV infection remained unknown.

Table 1 Antibodies with severely impaired activity during persistent LCMV infection and murine SLE model. tg, transgenic; Treg cells, regulatory T cells; DCs, dendritic cells.
View this table:

In this study, we thus wanted to explore why some antibodies efficiently deplete their target cells in the presence of excessive IC formation. We analyzed two antibodies that efficiently depleted their respective target cells in an FcγR-dependent manner during persistent LCMV infection. These antibodies revealed two distinct approaches enabling antibody-mediated depletion in the presence of competing ICs: (i) targeting highly abundant surface antigens such as CD90 or (ii) using antibodies with altered Fc glycosylation known to increase affinity for FcγRs. Our results demonstrate that therapeutic antibodies can outcompete endogenous ICs for FcγRs, resulting in efficient depletion if their target antigen is highly expressed or if they have increased affinity for FcγRs.

RESULTS

FcγR-dependent anti-CD90.2 efficiently depletes T cells during persistent LCMV infection

We and others have previously shown that persistent LCMV infection impairs antibody-mediated depletion and other FcγR-mediated antibody effector functions due to excessive IC formation (1, 2). This defect was observed with a wide range of therapeutic antibodies differing in antigen specificity, isotype, or species of origin (Table 1). Injection of 500 μg of anti-CD4 (clone GK1.5) and anti-CD8α (clone 2.43) on two consecutive days resulted in efficient depletion of CD4+ and CD8+ T cells, respectively, from the peripheral blood of naïve mice within 2 days (Fig. 1A). In contrast, the same treatment regimen failed to deplete CD4+ or CD8+ T cells from the peripheral blood of persistently infected mice. However, two injections of 250 μg of anti-CD90.2 (clone 30H12) were sufficient to efficiently deplete T cells (>94%) from the peripheral blood of naïve and persistently infected animals within 2 days (Fig. 1, B and C). Similar results were obtained when we analyzed the T cell depletion efficiency in the spleen (Fig. 1D). The superior depletion activity of the anti-CD90.2 antibody in persistently infected mice could not simply be explained by its isotype or species of origin, because all antibodies (anti-CD4, anti-CD8α, and anti-CD90.2) were confirmed to be rat IgG2b isotype.

Fig. 1 FcγR-dependent anti-CD90.2 efficiently depletes T cells during persistent LCMV infection.

(A) Naïve and LCMV clone-13–infected mice (28 dpi) were injected twice intraperitoneally with PBS, 500 μg of anti-CD4, or 500 μg of anti-CD8α. Representative flow plots show the frequency of CD8β+ and CD4+ T cells among peripheral blood mononuclear cells (PBMCs) 2 days after treatment start. Graphs show the depletion efficiency of anti-CD4 and anti-CD8α. (B to D) Naïve and LCMV clone-13–infected mice (28 dpi) were injected twice intraperitoneally with PBS or 250 μg of anti-CD90.2. (B) Representative flow plots showing the frequency of CD3+ T cells among PBMCs 2 days after treatment start. (C and D) T cell depletion efficiency in peripheral blood and spleen. (E) Wild-type (wt) and FcRγ−/− mice were injected twice intraperitoneally with PBS or 250 μg of anti-CD90.2, and the number of splenic CD3+ T cells was assessed after 2 days. (F) Wt mice were injected with 200 μl of clodronate liposomes or PBS, followed by anti-CD90.2 treatment and analysis as described in (E). Data (mean and SEM) from one representative experiment of at least two are shown (n = 3 to 4 mice per group). Unpaired two-sided Student’s t test was used for analyses shown in (A) and (E). ***P < 0.001. Ordinary one-way ANOVA with post hoc Tukey’s test for multiple comparisons was used for analysis shown in (F). ns, not significant. ****P < 0.0001.

FcγR-mediated phagocytosis by macrophages is the main mechanism of depleting antibodies targeting circulating cells in vivo (3, 15). However, FcγR-independent mechanisms of action, such as complement-dependent cytotoxicity and direct induction of cell death, have also been proposed (16, 17). We thus sought to determine whether the anti-CD90 antibody was potentially circumventing the obstacle of competing ICs during persistent infection by depleting T cells via FcγR-independent mechanisms. To address this question, we used mice deficient for the common γ chain subunit required for the surface expression and function of activating FcγR I, III, and IV (FcRγ−/−). In contrast to wild-type mice, treatment of FcRγ−/− mice with anti-CD90.2 did not result in a measurable reduction in T cell numbers, suggesting that anti-CD90.2 mediates its depletion activity in vivo solely through the engagement of activating FcγRs (Fig. 1E). In addition, the depletion activity of anti-CD90.2 was dependent on the presence of phagocytes because mice depleted of phagocytes using clodronate liposomes were resistant to T cell depletion mediated by anti-CD90.2 (Fig. 1F). Injection of clodronate liposomes resulted in the efficient depletion of phagocytic cells, such as red pulp macrophages (>85%), without affecting T cell numbers (fig. S1, A and B). These data demonstrate that the CD90.2-specific antibody used in this study can efficiently deplete T cells during persistent LCMV infection despite its dependence on FcγRs, raising the question about the mechanism allowing it to bypass the obstacle of competing endogenous ICs.

High expression of CD90 allows efficient T cell depletion during persistent LCMV infection

We next sought to determine how the FcγR-dependent anti-CD90.2 antibody could efficiently mediate T cell depletion during persistent LCMV infection. CD90 is one of the most abundantly expressed proteins on the surface of T cells with about 200,000 copies per cell (18). Using quantitative flow cytometry, we confirmed that CD90.2 is highly expressed on splenic T cells of persistently infected mice and compared its expression with CD4 and CD8α (Fig. 2A), both targeted by rat IgG2b antibodies unable to deplete their targets during persistent LCMV infection (Fig. 1A). Using calibration beads with known numbers of fluorophores and antibodies with known fluorophore-to-protein ratios under saturating staining conditions, we determined the number of surface molecules per cell assuming a 1:1 binding interaction between antibodies and surface antigens. The number of CD90.2 molecules on splenic T cells was about four- to sixfold higher compared with the number of CD4 and CD8α molecules on the respective splenic T cell subset (Fig. 2A). We thus hypothesized that the high abundance of CD90 on the surface of T cells and thus increased amount of bound antibody surpass the threshold required for efficient FcγR engagement in the presence of competing ICs.

Fig. 2 High expression of CD90 allows efficient T cell depletion during persistent LCMV infection.

(A) Surface expression of CD90.2, CD4, and CD8α on CD3+ T cells, CD4+ T cells, and CD8+ T cells, respectively, in the spleen of LCMV clone-13–infected mice (28 dpi). Data (mean and SEM) from one representative experiment (n = 5 mice) of two are shown. Ordinary one-way ANOVA with post hoc Tukey’s test for multiple comparisons was used for analysis. (B) Experimental design. iv, intravenous. (C) Representative flow plots of transferred cell populations (CD3+ TCRβ+ CellTrace Violet/CFSE+) before and 3 to 4 hours after transfer to naïve and LCMV clone-13–infected (28 dpi) recipients. Numbers indicate the percentages of the individual populations among transferred cells. (D) Graph shows dose-dependent depletion efficiency of T cells in naïve and LCMV clone-13–infected mice. (E) EC50 values (anti-CD90.2 coating concentration required to achieve 50% depletion) were calculated using data shown in (D). Data (mean and SEM) from one representative experiment (n = 5 recipients per group) of three are shown. Unpaired two-sided Student’s t test was used. ****P < 0.0001.

To determine whether increased amounts of bound antibody are required to promote efficient depletion during persistent LCMV infection, we performed in vivo depletion assays using fluorescently labeled T cells coated with antibody (1). We cotransferred differentially labeled T cells coated with titrated amounts of anti-CD90.2 antibody into naïve or infected hosts (Fig. 2B). T cells coated with high amounts of anti-CD90.2 (0.8 to 4 μg/107 T cells) were efficiently depleted in naïve and persistently infected mice, whereas low to intermediate amounts of anti-CD90.2 (3 to 50 ng/107 T cells) only resulted in measurable depletion in naïve but not infected hosts (Fig. 2, C and D). Achieving 50% depletion of transferred cells in persistently infected recipients required about 10-fold more anti-CD90.2 compared with naïve controls (Fig. 2E). Similar results were obtained upon transfer of T cells coated with saturating amounts of a pan-CD90 specific antibody (clone T24/31, rat IgG2b) into naïve or persistently infected mice (fig. S2, A and B). Together, these data demonstrate that, in the presence of competing ICs, increased levels of antibody are required to bind to the target cells for their efficient depletion.

Increased surface antigen expression facilitates efficient antibody-mediated depletion during persistent LCMV infection

We next sought to confirm our findings and exclude that antigen-specific features of CD90 such as clustering patterns, cell surface half-life, or reinternalization rate contributed to the high in vivo depletion efficiency of anti-CD90 antibodies. We thus developed a model allowing us to investigate the impact of differential expression of an additional model antigen, human CD20 (hCD20), on a standardized target cell (EL4) on antibody-mediated depletion in vivo. We generated three hCD20-expressing EL4 cell lines differing in their surface expression of hCD20, ranging from ~2700 molecules (low) to ~80,000 molecules per cell (high; Fig. 3A). To assess in vivo depletion, we cotransferred differentially labeled EL4 hCD20 cell lines together with the parental EL4 cell line to naïve or persistently infected hosts that received the hCD20-specific murine mAb 2H7 before cell transfer (Fig. 3B). In naïve mice, the depletion efficiency coincided with the expression level of hCD20 on EL4 cells—with low hCD20 expression resulting in ~50%, intermediate expression in ~70%, and high expression in ~90% depletion (Fig. 3, B and C). In persistently infected mice, EL4 cells expressing high levels of hCD20 were efficiently depleted to an extent that was slightly lower but overall comparable with naïve mice, whereas low and intermediate expression levels failed to efficiently deplete cells (Fig. 3, B and C). Despite expressing ~7-fold less antigen, EL4 hCD20 low cells were more efficiently depleted in naïve mice than EL4 hCD20 medium cells in infected mice. Overall, these data demonstrate that, although in naïve mice depletion efficiency is also dependent on the surface antigen expression level, increased antigen expression on the cell surface is required to promote efficient depletion in the presence of competing ICs.

Fig. 3 Increased surface antigen expression facilitates efficient antibody-mediated depletion.

(A) Histogram shows surface expression of hCD20 on parental EL4 and hCD20 transgenic EL4 cell lines. Graph shows calculated number of hCD20 molecules per cell. (B) Representative, concatenated flow plots of transferred EL4 cells 3 to 4 hours after transfer to naïve and LCMV clone-13–infected (31 dpi) recipients that received 250 μg of anti-hCD20 (2H7) 3 hours before cell transfer or untreated naïve mice. Numbers indicate the percentages of the individual populations among transferred cells. (C) Graph shows depletion efficiency of EL4 hCD20 cells in naïve and LCMV clone-13–infected mice. Data (mean and SEM) from one representative experiment (n = 4 to 8 recipients per group) of two are shown. Ordinary one-way ANOVA with Sidak’s test for multiple comparisons was used. ns, not significant. *P < 0.05, ****P < 0.0001.

FcγR-dependent anti-CD8β efficiently depletes CD8+ T cells during persistent LCMV infection

During our study, we also tested an anti-CD8β antibody (clone H35-17.2, rat IgG2b) for its ability to deplete CD8+ T cells in naïve and LCMV-infected mice. Unexpectedly, two injections of 250 μg of this antibody were sufficient to achieve efficient depletion in the peripheral blood and spleen of naïve and persistently infected animals (Fig. 4, A to C). The depletion activity of this anti-CD8β antibody, which was confirmed to be of the rat IgG2b isotype, crucially depended on the engagement of activating FcγRs, because FcγR−/− mice showed no reduction in splenic CD8+ T cells upon treatment with anti-CD8β (Fig. 4D). We also show that CD8+ T cell depletion by anti-CD8β requires phagocytosis, because mice lacking phagocytes due to injection of clodronate liposomes failed to deplete CD8+ T cells upon anti-CD8β treatment (Fig. 4E). Our above experiments with CD90-specific antibodies showed that high surface expression allows efficient target cell depletion during persistent infection. We thus quantified the expression of CD8β on splenic CD8+ T cells and compared its expression with CD90.2 and CD8α, an antigen also targeted by rat IgG2b antibodies, which failed to deplete during persistent LCMV infection (Table 1). The expression level of CD8β was about five- to sixfold lower compared with that of CD90.2 and was even slightly lower, although not statistically significant, compared with that of CD8α (Fig. 4E). These data demonstrate that the FcγR-dependent anti-CD8β antibody can efficiently deplete CD8+ T cells during persistent LCMV infection despite antigen expression levels being similar to other “unresponsive” target antigens such as CD8α and CD4.

The anti-CD8β antibody exhibits increased glycosylation heterogeneity

The above data suggested that the anti-CD8β antibody more efficiently engaged FcγRs in the presence of competing ICs than anti-CD4 and anti-CD8α antibodies, despite identical isotype and species of origin of the antibodies (rat IgG2b isotype) and comparable expression levels of target antigens. IgGs of all species and subclasses have a conserved N-linked glycosylation site N297 in the Fc portion, which is crucial for the binding to FcγRs. Fc glycans share a common biantennary heptasaccharide core with varying extensions modulating the affinity to FcγRs and rendering this region traditionally considered as invariant highly variable (19). We thus sought to determine whether the anti-CD8β antibody used in this study displayed a differential N-glycosylation pattern that could explain its superior activity during persistent LCMV infection. Compared with anti-CD90.2 and anti-CD4, which only contained three glycan structures (G0F, G1F, and G2F), anti-CD8β showed an increased heterogeneity in its N-linked glycans with five glycan structures (G0, G0F, G1F, G0FB, and G1FB) present (Fig. 5, A and B). Galactosylation levels (G1 + G2) varied substantially between the analyzed mAbs ranging from 42% down to 5% for anti-CD90.2 and anti-CD8β, respectively (Fig. 5C). In contrast to anti-CD90.2 and anti-CD4, which exclusively contained glycans with a core fucose, ~15% of anti-CD8β–derived glycans were afucosylated (Fig. 5D). Afucosylation has been shown to increase the affinity of IgG for mouse FcγRIV and human FcγRIII by at least one order of magnitude (4, 20). In addition, the anti-CD8β antibody contained a measurable fraction of bisecting glycans (~7%; Fig. 5E), which have been suggested to result in a modest increase in affinity for FcγRs (21). These data demonstrate that, compared with other antibodies used in this study, anti-CD8β was differentially glycosylated containing glycans commonly associated with increased FcγR affinity. These data thus suggest that the anti-CD8β antibody might be able to outcompete endogenous ICs and efficiently engage FcγRs during persistent LCMV infection due to increased affinity for FcγRs. However, we cannot exclude that other antibody-intrinsic features contribute to the observed efficacy.

Fig. 4 FcγR-dependent anti-CD8β efficiently depletes CD8+ T cells during persistent LCMV infection.

(A) Naïve and LCMV clone-13–infected mice (28 dpi) were injected twice intraperitoneally with PBS or 250 μg of anti-CD8β. Representative flow plots show the frequency of CD8α+ and CD4+ T cells among PBMCs 2 days after treatment initiation. (B and C) CD8α+ T cell depletion efficiency in peripheral blood and spleen. (D) Wt and FcRγ −/− mice were injected twice intraperitoneally with PBS or 250 μg of anti-CD8β, and the number of splenic CD8α+ T cells was assessed after 2 days. (E) Wt mice were injected with 200 μl of clodronate liposomes or PBS, followed by anti-CD8β treatment and analysis as described in (D). (F) Surface expression of CD8α, CD8β, and CD90.2 on CD8+ T cells and CD3+ T cells, respectively, in the spleen of LCMV clone-13–infected mice (28 dpi). Data (mean and SEM) from one representative experiment of at least two are shown (n = 3 to 5 mice per group). Unpaired two-sided Student’s t test was used for analyses shown in (D). Ordinary one-way ANOVA with post hoc Tukey’s test for multiple comparisons was used for analysis shown in (E) and (F). ns, not significant. ****P < 0.0001.

Afucosylated anti-CD4 mediates CD4+ T cell depletion during persistent LCMV infection

The glycosylation analysis of anti-CD8β suggested that afucosylation and bisection, two glycosylation patterns associated with increased FcγR affinity, could be beneficial for depleting antibodies during persistent LCMV infection. Of all tested Fc glycan variants, afucosylation has so far shown the greatest impact on the affinity of IgG for FcγRs, increasing the affinity for mouse FcγRIV and human FcγRIIIA 10- to 50-fold (4, 20). We thus sought to determine whether a glycoengineered antibody lacking the core fucose would be more efficiently depleting its target cells during persistent LCMV infection. We chose to generate an afucosylated version of the anti-CD4 antibody (clone GK1.5) because its parental fucosylated version failed to reduce the number of CD4+ T cells during persistent infection (Fig. 1A). The addition of the core fucose to the Fc glycan is mediated by the α-(1,6)-fucosyltransferase (FUT8), which can be inhibited by 2F-peracetyl-fucose (2F) (22). As previously described for Chinese hamster ovary cells, culturing of the GK1.5 hybridoma in the presence of increasing concentrations of 2F resulted in a dose-dependent reduction of surface fucosylation, as detected by Aurelia aurantia lectin (AAL) binding (Fig. 6A) (22). Twenty-five micromolar 2F resulted in a surface fucosylation level comparable with a FUT8-deficient GK1.5 hybridoma cell line we generated (GK1.5ΔFUT8). We confirmed the absence of the core fucose on antibodies purified from the supernatant of 2F-treated and FUT8-deficient GK1.5 hybridoma cultures by AAL blot analysis and mass spectrometry (Fig. 6, B and C). Besides the lack of fucosylation, the chemical inhibition of FUT8 by 2F or ablation of its expression (GK1.5ΔFUT8) did not change the overall glycan profile of anti-CD4, which was dominated by agalactosylated structures G0/G0F (~85%; Fig. 6C).

Fig. 5 Glycosylation analysis of monoclonal antibodies.

N-linked glycans released from purified anti-CD90.2, anti-CD4, and anti-CD8β by PNGase F were analyzed by mass spectrometry. (A) List of identified glycans and annotated mass spectra. a.u., arbitrary units. m/z, mass/charge ratio. (B) Summary of N-linked glycan composition (relative abundance). Relative abundance of (C) galactosylated, (D) fucosylated, and (E) bisecting glycans.

We next performed in vivo depletion assays using fluorescently labeled CD4+ T cells coated with antibody to directly compare the ability of fucosylated and afucosylated anti-CD4 antibodies to deplete their target cells within the same environment (Fig. 6D). CD4+ T cells coated with fucosylated and afucosylated anti-CD4 were efficiently depleted upon transfer into naïve mice, whereas only CD4+ T cells coated with afucosylated anti-CD4 were removed upon transfer into LCMV clone-13–infected mice (Fig. 6, D and E). We next sought to determine whether the administration of afucosylated anti-CD4 would result in efficient CD4+ T cell depletion in mice persistently infected with LCMV clone-13. We treated persistently infected animals on two consecutive days with either fucosylated or afucosylated anti-CD4 and assessed CD4+ T cell depletion in the peripheral blood 2 days after treatment initiation (Fig. 6F). Administration of afucosylated, but not fucosylated, anti-CD4 resulted in efficient depletion of CD4+ T cells in the peripheral blood (>97%; Fig. 6G). Moreover, although not as notable as in the peripheral blood, we observed a significant reduction in the number of splenic CD4+ T cells (~75%) within 2 days of treatment initiation with afucosylated anti-CD4 (Fig. 6H). Overall, these data demonstrate the superior activity of afucosylated anti-CD4 to deplete its target cells during persistent LCMV infection and suggest that it is possible to outcompete endogenous ICs for access to FcγRs if afucosylated antibodies are used.

Fig. 6 Afucosylated anti-CD4 (GK1.5) can deplete CD4+ T cells during persistent LCMV infection.

(A) Surface fucosylation of GK1.5 hybridoma cells cultured in the presence of 0 to 50 μM 2F analyzed by flow cytometry using biotinylated AAL and APC-labeled streptavidin. Streptavidin-only stained parental GK1.5 hybridoma is shown in gray. Graph shows percentage surface fucosylation of 2F-treated and FUT8-deficient (ΔFUT8) hybridoma compared with untreated parental GK1.5 hybridoma. (B and C) Anti-CD4 antibodies were purified from supernatant of untreated, 25 μM 2F-treated, and ΔFUT8 hybridoma. (B) Immunoblot (anti-rat IgG) and lectin blot (AAL) of IgG to assess fucosylation. (C) Composition (relative abundance) of N-linked IgG glycans was analyzed by mass spectrometry. MW, molecular weight. (D and E) In vivo depletion assay using CD4+ T cells coated with fucosylated and afucosylated anti-CD4 antibody. (D) Representative plots of transferred cell populations (CD3+ CD4+ CFSE+) before and 3 to 4 hours after transfer to naïve and LCMV clone-13–infected (28 dpi) recipients. Numbers indicate the percentage among transferred cells. (E) Depletion efficiency of CD4+ T cells coated with fucosylated and afucosylated anti-CD4 in naïve and LCMV clone-13–infected mice. (F to H) LCMV clone-13–infected mice (28 dpi) were injected twice intraperitoneally with PBS or 300 μg of fucosylated (fucos.) or afucosylated (afucos.) anti-CD4 and analyzed 2 days after treatment start. (F) Representative flow plots show frequency of CD4+ T cells among CD3+ T cells in peripheral blood. (G) CD4+ T cell depletion efficiency in peripheral blood. (H) Number of CD4+ T cells in the spleen. Data (mean and SEM) from one representative experiment of at least two are shown (n = 4 to 5 mice per group). Two-way ANOVA with post hoc Tukey’s test for multiple comparisons was used for analysis shown in (E). Unpaired two-sided Student’s t test was used for analyses shown in (G). Ordinary one-way ANOVA with post-hoc Tukey’s test for multiple comparisons was used for analysis shown in (H). ns, not significant. ***P < 0.001, ****P < 0.0001.

Afucosylated anti-CD8α mediates CD8+ T cell depletion during persistent LCMV infection

To further confirm our findings with the afucosylated anti-CD4 antibody, we generated an afucosylated version of the anti-CD8α antibody (clone 2.43) because its parental fucosylated version failed to efficiently deplete CD8+ T cells during persistent infection (Fig. 1A). In accordance with our observations using the GK1.5 hybridoma, culturing the 2.43 hybridoma in the presence of 25 μM of 2F resulted in the production of a completely afucosylated antibody as demonstrated by AAL blot analysis of purified anti-CD8α (fig. S3A). We first evaluated the depletion efficiency of afucosylated anti-CD8α in peripheral blood, spleen, liver, and the epithelium of the small intestine [intraepithelial lymphocyte (IEL)] of naïve mice. Whereas the afucosylated and the parental fucosylated anti-CD8α antibody caused efficient depletion of CD8+ T cells in both lymphoid and nonlymphoid tissues, afucosylated anti-CD8α exhibited a significantly higher depletion activity in all examined tissues (fig. S3, B to F). These data are in accordance with previous results and demonstrate the superior activity of afucosylated antibodies (4, 20). We next sought to determine whether afucosylated anti-CD8α would also result in efficient CD8+ T cell depletion in mice persistently infected with LCMV clone-13. Administration of afucosylated, but not fucosylated, anti-CD8α resulted in efficient depletion of CD8+ T cells from the peripheral blood and spleen with >90% depletion efficiency in peripheral blood and ~80% reduction in splenic CD8+ T cell numbers (Fig. 7, A to C). Of note, both afucosylated and fucosylated anti-CD8α resulted in a notable reduction of CD8+ T cells in the liver, although afucosylated anti-CD8α was more efficient compared with the parental fucosylated variant, demonstrating ~97 and ~70% depletion efficiency, respectively (Fig. 7, A and D). LCMV clone-13–infected mice exhibited a significant infiltration of CD8αβ+ T cells in the IEL with ~7-fold higher cell numbers compared with naïve controls (fig. S3G). Administration of neither afucosylated nor fucosylated anti-CD8α resulted in a significant depletion of CD8+ T cells in the IEL of infected animals despite the depleting antibody being bound as demonstrated by the blockade of the flow cytometric detection of CD8α (Fig. 7, A and F, and fig. S3H). These data suggest that, compared with naïve mice, CD8+ T cells in the IEL of persistently infected mice exhibit a higher degree of tissue residency, do not readily enter circulation, and are thus more resistant to antibody-mediated depletion because intravascular access has been previously shown to be a major determinant of susceptibility to antibody-mediated depletion (15). Overall, our data demonstrate that afucosylated antibodies exhibit superior depletion activity irrespective of the infection status. However, whereas in naïve mice both fucosylated and afucosylated antibodies efficiently deplete their target cells, during persistent LCMV clone-13 infection only afucosylated antibodies can efficiently deplete their target cells. Our data thus suggest that it is possible to outcompete endogenous ICs for access to FcγRs if antibodies with enhanced affinity for FcγRs such as afucosylated antibodies are used.

Fig. 7 Afucosylated anti-CD8α (2.43) can deplete CD8+ T cells during persistent LCMV infection.

LCMV clone-13–infected mice (28 dpi) were injected twice intraperitoneally with PBS or 300 μg of fucosylated or afucosylated anti-CD8α and analyzed 2 days after treatment start. (A) Representative flow plots show frequency of CD8β+ T cells among total lymphocytes in peripheral blood, spleen, liver, and the epithelium of the small intestine (IEL). CD8 T cell depletion efficiency in (B) blood, (C) spleen, (D) liver, and (E) IEL. Unpaired two-sided Student’s t test was used for analysis. ns, not significant. *P < 0.05, ***P < 0.001.

DISCUSSION

In this study, we present two distinct strategies to improve FcγR engagement of depleting antibodies in the presence of competing endogenous ICs. Efficient antibody-mediated depletion can be achieved by targeting highly abundant antigens such as CD90 or by modifying the Fc glycosylation of the depleting antibody to increase FcγR effector functions.

Impairment of FcγR functions by excessive IC formation has been reported during persistent LCMV infection and in the murine MLR/lpr lupus model (1, 2, 13, 14). The efficient depletion of target cells during persistent LCMV infection reported in this study supports a model in which endogenous ICs do not simply abrogate the engagement of FcγRs by depleting antibodies but rather compete with them for binding of available FcγRs. In this study, we observed that two antibodies specific for the highly expressed CD90 antigen efficiently depleted T cells in persistently infected mice; however, efficient depletion required increased amounts of surface-bound antibody compared with naïve mice. These data suggest that increased levels of bound antibodies on the cell surface can outcompete endogenous ICs for access to FcγRs and mediate efficient depletion. Our experiments with EL4 cell lines differentially expressing hCD20 further support this hypothesis. Two studies showed a positive correlation between CD20 expression levels and the response to rituximab treatment in B cell lymphoma patients (23, 24). Surface antigen expression levels might thus be an important factor to take into consideration when selecting a therapeutic antibody for a given patient because the expression level of CD20 and CD52, targeted by the therapeutic antibodies rituximab and alemtuzumab, respectively, can vary up to 20-fold between patients (25).

The Fc region of IgG and its interaction with FcγRs are crucial for the in vivo activity of several antitumor antibodies and neutralizing antibodies against influenza and HIV (5, 26, 27). Enhancing the affinity of the Fc region of IgG for activating FcγRs is thus an attractive approach to improve the therapeutic activity of antibodies and is intensively pursued with multiple antibodies in clinical trials (2831). Obinutuzumab, a glycoengineered anti-CD20 antibody, was recently approved for treatment of chronic lymphocytic leukemia. Increased FcγR binding of IgG can be achieved by introducing mutations into the backbone of the Fc region or by modifying the glycan moiety at the conserved N-glycosylation site N297 in the Fc region (20, 28). Human IgG1 lacking the core fucose shows about 50-fold higher affinity for the activating human FcγRIIIA, and a comparable albeit slightly smaller effect of afucosylation has been shown for the interaction of mouse IgG and murine FcγRIV (4, 20). However, previous studies demonstrating the superior in vivo activity of afucosylated antibodies were almost exclusively performed in animal models lacking excessive IC formation. The efficacy of an afucosylated antibody to deplete B cells in a murine hCD19 transgenic lupus model has been recently shown, although without directly comparing it with a fucosylated antibody and carefully examining the expression level of the hCD19 transgene (32). In this study, we show that in naïve mice, efficient depletion of CD4+ and CD8+ T cells can be achieved by administration of fucosylated antibodies, whereas during persistent LCMV infection, only afucosylated anti-CD4 and anti-CD8α antibodies efficiently deplete CD4+ and CD8+ T cells, respectively. Afucosylated IgG thus seems to outcompete endogenous ICs for access to activating FcγRs due to its higher affinity, resulting in FcγR-mediated effector functions in the presence of competing ICs. Thus, the use of Fc-engineered antibodies might allow efficient antibody-mediated depletion even if the target antigen is not highly expressed.

The alternative Fc glycosylation pattern and superior depletion activity of the anti-CD8β antibody during persistent LCMV infection led us to investigate whether afucosylated antibodies can efficiently deplete their targets during persistent LCMV infection. Although afucosylated antibodies directed against CD4 or CD8α showed superior depletion activity, these completely afucosylated antibodies were still inferior in their depletion activity (~75 to 80% in spleen) compared with the partially afucosylated anti-CD8β antibody (~90% in spleen) during persistent LCMV infection. These data suggest that additional antibody-intrinsic features, such as the location of the targeted epitope, could contribute to the superior activity of anti-CD8β. Of note, a recent study showed that the distance of an antibody-targeted epitope to the membrane is a crucial determinant for the efficiency with which an antibody can trigger FcγR-mediated effector functions, with ADCC and antibody-dependent cellular phagocytosis demonstrating divergent preferences (33).

Persistent LCMV infection and the MLR/lpr lupus model are widely used model systems to elucidate immunological mechanisms contributing to disease pathogenesis, and antibody-mediated depletion is a commonly used strategy to determine the role of specific cell populations. Of note, we initially observed impaired FcγR functions due to excessive IC formation during persistent LCMV infection by serendipity while attempting to deplete CD4+ T cells to study their role during persistent viral infection (1). Here, we demonstrate that afucosylated antibodies are superior in their in vivo activity in the presence of competing endogenous ICs and that afucosylated antibodies can be easily generated by both genetic and chemical approaches. Addition of the chemical FUT8 inhibitor 2F allows for the fast production of afucosylated antibodies without detectable cellular cytotoxicity or time-consuming genetic manipulations and selection. Achieving efficient antibody-mediated depletion by using afucosylated antibodies produced by the above-mentioned techniques will thus allow studying the role of various cell populations in model systems characterized by excessive IC formation.

In accordance with previous studies (4, 34), afucosylated antibodies exhibited increased depletion activity in naïve mice. However, the observed difference between afucosylated and fucosylated antibodies was more pronounced in persistently infected animals, in which fucosylated variants failed to induce significant depletion. Among activating murine FcγRs, only FcγRIV shows significantly increased affinity for afucosylated IgG (10-fold), suggesting that FcγRIV is crucially involved in the successful depletion of target cells mediated by afucosylated antibodies during persistent LCMV infection (4, 34). It is interesting that efficient antibody-mediated depletion during persistent infection required about 10-fold higher levels of opsonized antibody (quantity) or antibodies with a 10-fold higher affinity for FcγRIV (quality) and thus tempting to speculate that both approaches act independently but with the common goal to increase the chances of efficient FcγR engagement in the presence of competing ICs.

Afucosylation is the most widely used strategy to improve the activity of therapeutic antibodies depending on the engagement of activating FcγRs without inducing potential immunogenic mutations in the Fc region. Fc-engineered human IgG1 antibodies with amino acid substitutions in the Fc region aimed at enhancing affinity for activating human FcγRs have been developed and showed increased activity in a number of models including in vitro assays, nonhuman primate studies, and human FcγR transgenic mice (28, 31, 35). The increased activity of afucosylated antibodies during persistent viral infection is most likely due to increased affinity for FcγRIV, and thus Fc-engineered mAbs with increased affinity for activating FcγRs are expected to also show increased activity under these conditions. However, the experimental validation of Fc-engineered mAbs would require the use of human FcγR transgenic mice because modifications to the human IgG1 Fc region are aimed at increasing affinity for activating human FcγRs.

Prolonged treatment with extremely high doses of depleting antibody (10 mg weekly for 10 weeks) has been previously shown to efficiently deplete resistant B cells in the MLR/lpr lupus model (13). Although such an approach might theoretically prove successful in LCMV-infected animals due to the same underlying cause for resistance, that is, ICs, the relatively rapid kinetics of LCMV infection compared with the treatment cycle renders such an approach impractical for studying the role of certain subsets during the chronic phase of the infection. Furthermore, most depleting antibodies originated from rats and will thus induce potent “antidrug” antibody responses preventing efficient long-term treatments.

An additional approach to achieve efficient antibody-mediated depletion in the presence of excessive ICs, which might potentially synergize with the strategies of targeting highly abundant antigens and the use of Fc-engineered antibodies mentioned above, could be the simultaneous administration of a blocking anti-CD47 antibody. Blockade of CD47, a transmembrane protein overexpressed on many lymphomas and interacting with the inhibitory receptor signal-regulatory-protein (SIRP)α on macrophages, has been shown to synergize with rituximab and promote phagocytosis of NHL (3638). However, it is not known whether CD47 can reduce the critical threshold required for FcγR-mediated phagocytosis. Future studies should therefore address the question of whether blockade of the CD47-SIRPα pathway can promote FcγR-mediated phagocytosis in situations of suboptimal FcγR engagement such as persistent viral infections, autoimmune diseases, and cancer.

Our study using depleting antibodies for various lymphocyte subsets serves as a proof of concept demonstrating that, during persistent viral infection, FcγR-mediated effector functions can be efficiently triggered, although increased surface antigen expression or antibodies with increased affinity for activating FcγRs are required. However, one limitation of our study is that we assessed the FcγR-mediated depletion of T cell subsets with a limited number of monoclonal antibodies. Further experiments are thus required to confirm that depleting antibodies against other lymphocyte markers and antibodies directed against nonlymphoid target cells behave comparably. Furthermore, our experiments were predominantly focused on circulating lymphocytes in lymphoid and nonlymphoid tissues. The failure of even the afucosylated anti-CD8α antibody to deplete CD8 T cells in the intestinal epithelium of LCMV clone-13–infected animals suggests that resident cells in certain compartments of the body may be more resistant to FcγR-mediated depletion. Thus, additional experiments targeting various surface antigens and cellular targets are required to precisely assess the impact of tissue residency and differences between various compartments. These experiments should provide important insights, especially for sessile target populations such as solid tumors.

Future studies using LCMV coinfection models could provide novel insights into how persistent viral infections can modulate FcγR-mediated antibody effector functions involved in antimicrobial protection and provide a platform to evaluate Fc-engineered antibodies in a stringent experimental setting. Overall, our study presents two distinct and potentially synergistic strategies to achieve efficient antibody-mediated depletion during persistent LCMV infection, which is characterized by excessive IC formation resulting in compromised FcγR-mediated antibody effector functions. The quantity of the target antigen and the quality of the depleting antibody, that is, its affinity for activating FcγRs, might thus be both important factors to consider before treatment of autoimmune diseases, cancer, and persistent viral infections with antibodies relying on the engagement of FcγRs.

MATERIALS AND METHODS

Study design

The present study was designed to investigate strategies allowing the efficient depletion of target cells during persistent LCMV clone-13 infection, which is characterized by excessive IC formation and resistance to FcγR-mediated depletion. We performed depletion experiments by directly injecting depleting antibodies into mice or by adoptively transferring fluorescently labeled and antibody-coated cells, followed by fluorescence-activated cell sorting (FACS) analyses to assess depletion efficiency. The number of mice per experimental group ranged from three to eight, and experiments were repeated at least once, as indicated in the figure legends.

Mice and viruses

Female C57BL/6J mice were purchased from Jackson Laboratory. FcRγ−/− mice were a gift from S. R. Stowell, Emory University. All mice were housed at the Emory University School of Medicine animal facility under standard pathogen-free conditions. All experiments were performed in accordance with approved Institutional Animal Care and Use Committee protocols. For persistent LCMV infection, 6- to 8-week-old mice were infected intravenously with 2 × 106 plaque-forming units of the LCMV clone-13 strain. Depletion experiments were performed around 24 to 31 days postinfection (dpi).

Cell preparation

Lymphocytes from anticoagulated peripheral blood were isolated using Histopaque-1077 (Sigma-Aldrich). Splenocytes were isolated by mechanical disruption, followed by digestion with collagenase D (0.4 U/ml; Roche) for 30 min at 37°C with gentle shaking. EDTA was added to a final concentration of 5 mM, and spleens were homogenized by passing through a 70-μm cell strainer (BD Biosciences). Red blood cells were lysed using ACK Lysing buffer (Lonza). Cells were washed and resuspended in RPMI 1640 supplemented with 5% fetal bovine serum (FBS).

Flow cytometry and antibodies

Single-cell suspensions were stained for 30 min at 4°C in FACS staining buffer [2% FBS and 2 mM EDTA in phosphate-buffered saline (PBS)] with the following fluorochrome-conjugated antibodies (purchased from BioLegend): CD3 (145-2C11), CD4 (RM4-4), CD8α (53-6.7), CD8β (53-5.8), CD11b (M1/70), CD19 (6D5), CD44 (IM7), CD90.2 (30H12), and F4/80 (BM8). We used the anti-CD4 clone RM4-4 to ensure detection of CD4 expression on the cell surface in the presence of the depleting anti-CD4 antibody (GK1.5). Dead cells were excluded by staining with the LIVE/DEAD Fixable Near-IR Dead Cell Stain (Invitrogen). After staining, cells were fixed in 1% paraformaldehyde and acquired on a Canto II flow cytometer (BD Biosciences). Quantification of surface molecules was performed using Rainbow Calibration Particles (BioLegend). Data were analyzed using FlowJo (TreeStar).

Depletion of lymphocytes and phagocytes

Depletion of various lymphocyte populations and phagocytes was performed as described previously (1). Mice were injected intraperitoneally on two consecutive days with depleting antibody. Depletion efficiency in the peripheral blood and spleen was determined 2 days after treatment initiation. The following antibodies were purchased from BioXCell and used at the indicated doses: anti-CD4 (GK1.5; 300/500 μg), anti-CD8α (2.43; 500 μg), anti-CD90.2 (30H12; 250 μg), and anti-CD90 (T24/31; 250 μg). The anti-CD8β (H35-17.2; 250 μg) was produced in-house. The hybridoma H35-17.2 was a gift from A. E. Lukacher. The hybridoma was cultured in Hybridoma-SFM (Gibco) in CELLine bioreactors (INTEGRA Biosciences). Supernatants were purified using protein G sepharose, concentrated, and sterile-filtered. All antibodies were tested using a rat mAb isotyping test kit (Bio-Rad) and confirmed to be of the rat IgG2b isotype. Phagocytic active cells were depleted 24 hours before initiation of antibody treatment by intraperitoneal injection of 200 μl of clodronate liposomes (Chlophosome-A, FormuMax).

Generation of afucosylated antibodies

The GK1.5 and 2.43 hybridoma cell lines were obtained from American Type Culture Collection and cultured in Hybridoma-SFM. To produce afucosylated antibodies by chemical inhibition of FUT8, we cultured the cells in the presence of 25 μM 2F (EMD Millipore). To generate a FUT8-deficient GK1.5 hybridoma, we used the CRISPR-Cas9 system. Briefly, GK1.5 hybridoma cells were transfected using the Avalanche Sp2/0-Ag14 Cell Transfection Reagent (EZ Biosystems) with the plasmid pSp-Cas9(BB)-2A-GFP containing a FUT8 guide RNA specific for mouse and rat FUT8 (5′-CAGAATTGGCGCTATGCTAC-3′; GenScript). Two days after transfection, we single cell–sorted green fluorescent protein–positive (GFP+) cells into 96-well flat bottom plates using a FACSAria II, followed by clonal expansion in Hybridoma-SFM. We screened for the absence of FUT8 activity by analyzing surface fucosylation and by lectin blot analysis of purified antibodies.

Lectin blot and cytometry

Equal amounts of purified antibody were separated in reducing SDS–polyacrylamide gel electrophoresis, followed by transfer to polyvinylidene difluoride membranes. After blocking with tris-buffered saline with 0.1% Tween 20 + 3% bovine serum albumin (BSA), we incubated the membranes with either biotinylated AAL (200 ng/ml; Vector Laboratories) overnight at 4°C, followed by a 1-hour incubation with horseradish peroxidase (HRP)–conjugated streptavidin (20 ng/ml; Jackson ImmunoResearch), or as control overnight at 4°C with an HRP-conjugated goat anti-rat IgG (20 ng/ml; Jackson ImmunoResearch). Bound HRP was detected using the SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific).

The surface fucosylation of the GK1.5 hybridoma was analyzed by staining in Hanks’ balanced salt solution + 0.5% BSA with biotinylated AAL for 30 min at 4°C, followed by a 30-min incubation with allophycocyanin (APC)–labeled streptavidin (Thermo Fisher Scientific). Samples were then acquired on a Canto II flow cytometer, and mean fluorescence intensities were determined.

In vivo depletion assay

This assay was performed as described previously (1). We purified total T cells and CD4+ T cells from spleens of naïve C57BL/6J mice using the EasySep Mouse T Cell or EasySep Mouse CD4+ T Cell Isolation Kit (STEMCELL), respectively. Cells were labeled with different concentrations of CellTrace Violet and carboxyfluorescein diacetate succinimidyl ester (CFSE; both from Invitrogen), followed by coating with the indicated depleting antibodies. Labeled and coated cells were mixed to equal proportions and adoptively transferred into naïve or LCMV clone-13–infected mice by intravenous injection. Depletion was assessed 3 to 4 hours after transfer using flow cytometric analysis of the spleen. Half-maximal effective concentration (EC50) values for anti-CD90.2 experiments were calculated by fitting the data using a five-parameter logistic equation.

EL4 cell lines with differential expression of hCD20 were established by lentiviral transduction (pLVX-hCD20-IRES-ZsGreen1), followed by single-cell sorting of cells expressing low, medium, and high levels of hCD20 into 96-well flat bottom plates using a FACSAria II. After clonal expansion, three EL4 cell lines with differential hCD20 surface expression were selected, and the number of hCD20 molecules per cell was determined by using the BD Quantibrite PE system. To assess in vivo depletion, we labeled EL4 cell lines with different concentrations of CellTrace Violet and CellTrace FarRed (Invitrogen), mixed them to equal proportions, and intravenously transferred them into naïve or LCMV clone-13–infected mice, which received 250 μg of anti-hCD20 (2H7) intraperitoneally 4 hours earlier. Production and purification of the murine anti-hCD20 (2H7) have been described previously (1). Depletion was assessed 3 to 4 hours after transfer using flow cytometric analysis of the spleen.

Glycan analysis

N-linked glycans were released from purified antibodies under denaturing conditions with peptide-N-glycosidase (PNGase F), purified by solid phase extraction chromatography, and permethylated using NaOH–dimethyl sulfoxide slurry and methyl iodide, as described previously (39). Mass spectrometry data were acquired in positive mode using a 4800 Plus MALDI-TOF/TOF (Applied Biosystems) mass spectrometer, and data were analyzed using GlycoWorkBench software (40).

Statistical analysis

All statistical tests were performed using GraphPadPrism 6 (GraphPad). Values are expressed as means and SEM. As indicated in the figure legends, unpaired two-sided Student’s t test was used to compare two groups and ordinary one- or two-way analysis of variance (ANOVA) with post hoc Tukey’s test or Sidak’s test for multiple comparisons. *P < 0.05 was considered significant.

SUPPLEMENTARY MATERIALS

immunology.sciencemag.org/cgi/content/full/3/27/eaao3125/DC1

Fig. S1. Clodronate liposomes efficiently deplete phagocytic cells.

Fig. S2. CD90-specific antibodies can efficiently deplete T cells in naïve and persistently infected mice.

Fig. S3. Afucosylated anti-CD8α antibody exhibits superior depletion activity in naïve mice.

Table S1. Raw data sets.

REFERENCES AND NOTES

Acknowledgments: We thank members of the Ahmed laboratory for the insightful discussion and J. Gensheimer for the proofreading. We thank S. R. Stowell for providing us FcRγ−/− mice and A. E. Lukacher for providing the H35-17.2 hybridoma. Funding: This work was supported by NIH grant R01AI030048 to R.A. This study was also supported in part by the Emory Flow Cytometry Core, one of the Emory Integrated Core Facilities, and is subsidized by the Emory University School of Medicine. Additional support was provided by the National Center for Advancing Translational Sciences of the NIH under award number UL1TR000454. The content is solely the responsibility of the authors and does not necessarily reflect the official views of the NIH. Author contributions: A.W. conceived the study and designed, performed, and analyzed the experiments. A.O.K., R.M.V., J.-H.H., and X.X. helped with experiments and data analysis. B.P.C. performed the glycan analysis. R.A. supervised the study. A.W. and R.A. wrote the manuscript. All authors participated in editing the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The anti-CD8β hybridoma (H35-17.2) and the FUT8-deficient GK1.5 hybridoma cell line are available from A.W. All data needed to evaluate the conclusions are in the paper or in the Supplementary Materials.
View Abstract

Stay Connected to Science Immunology

Navigate This Article