CD4 T cell-dependent and -independent roles for IFN-gamma in blood-stage malaria¹

Abstract

Production of interferon-γ by CD4 T-cells is widely theorized to control Plasmodium parasite burden during blood-stage malaria. Surprisingly, the specific and crucial mechanisms through which this highly pleiotropic cytokine acts to confer protection against malarial disease remain largely untested in vivo. Here we used a CD4 T-cell-restricted Cre-Lox IFN-γ excision mouse model to test whether and how CD4 T-cell-derived IFN-γ controls blood-stage malaria. While complete absence of IFN-γ compromises control of acute and the chronic, recrudescent blood-stage infection with P. c. chabaudi, we identified a specific, albeit modest role for CD4 T-cell-derived IFN-γ in limiting parasite burden only during the chronic stages of P. c. chabaudi malaria. CD4 T cell IFN-γ promoted IgG antibody class-switching to the IgG2c isotype during P. c. chabaudi malaria in C57Bl/6 mice. Unexpectedly, our data do not support gross defects in phagocytic activity in IFN-γ-deficient hosts infected with blood-stage malaria. Together, our data confirm CD4 T cell-dependent roles for IFN−γ but suggest CD4 T cell-independent roles for IFN−γ in immune responses to blood-stage malaria.

Introduction

A highly effective malaria vaccine would greatly reduce the formidable toll malaria currently imposes on global health (1). However, the ability to strategically manipulate key determinants of immunity to the Plasmodium infections that cause malaria in the form of an efficacous vaccine remains frustratingly elusive.

The inflammatory cytokine interferon-γ (IFN-γ) has been widely implicated in control of both the asymptomatic liver-stage and the morbidity-driving blood-stage of malaria (2). Herein, we focus upon the blood-stage of malaria, during which Plasmodium parasites cause disease by replicating within and then destroying red blood cells (RBCs). Accordingly, control of parasite burden during blood-stage malaria requires identification and removal of Plasmodium-infected RBCs, and/or extracellular parasites transiting between host RBCs. A crucial role for IFN-γ in suppressing blood-stage parasite replication has been clearly demonstrated in murine models, where elevated parasitemia was seen in IFN-γ-deficient Ifng^−/− mice (3, 4), receptor-deficient Ifngr^−/− mice (5, 6), or upon antibody-mediated IFN-γ neutralization (7).

CD4 T-cells are also required for clearing parasitemia during blood-stage malaria in mice (8, 9). As IFN-γ production is the hallmark of Th1-type CD4 T-cells (10), Th1 cells emerged as obvious candidates for producing protective IFN-γ during blood-stage malaria. Early work posited a biphasic model where Th1-type CD4 T-cells suppress early parasitemia, and later emerging Th2-type CD4 T-cells reduce chronic parasitemia (11). Later work, however, revealed a more complex and plastic CD4 T-cell response, and that follicular T helper (Tfh) cells rather than Th2 cells likely control parasitemia during chronic blood-stage malaria (11–14).

Despite the clear complexity of the CD4 T-cell response to blood-stage malaria, a crucial role for CD4 T-cell derived IFN-γ in controlling blood-stage parasite burden has remained a largely untested assumption. Indirect evidence supporting this model includes studies in humans that identified IFN-γ-producing CD4 T-cells as correlating with effective anti-malarial immunity (15, 16). In mice, CD4 T-cells clearly produce IFN-γ during blood-stage malaria, with plasma IFN-γ content markedly reduced upon depletion of CD4 T-cells (17). However, numerous leukocytes are competent IFN-γ producers during blood-stage malaria, including NK cells, gamma-delta T-cells, and CD8 T-cells (18, 19). To date, the most direct test of the theorized role for CD4 T-cell derived IFN-γ was a singular study where reconstitution of CD4 T-cell-depleted mice with an exogenously raised IFN-γ-producing Th1-type CD4 T-cell clone facilitated parasite clearance (20). Whether the endogenous CD4 T-cell compartment mediates control of parasite burden in an IFN-γ-dependent manner remains untested.

Here we sought to directly test if CD4 T-cell derived IFN-γ is crucial for controlling parasite burden during blood-stage murine malaria. To this end, we generated transgenic CD4CreER^T2 Ifng^flox/KO (for short, CD4CreIfng) mice where a tamoxifen-responsive CreER^T2 protein is driven by CD4 promoter elements (21) with a Cre-excisable Ifng^flox/KO locus. This results in a mouse where tamoxifen inducible excision of Ifng is restricted to mature CD4 T-cells. Unlike prior studies that non-specifically abolished all IFN-γ signaling, our system minimizes off target effects by only curtailing IFN-γ production in CD4 T-cells during a limited temporal window.

We selected the Plasmodium chabaudi chabaudi CB isolate (PccCB) model of murine malaria to interrogate the role of CD4 T-cell derived IFN-γ in controlling blood-stage malaria, as Pcc malaria is the only rodent system that recapitulates both the acute and chronic blood-stage infections that are seen in humans. In the PccCB model, acute malaria comprises an initial wave of high parasitemia that lasts ~1 week, which is followed by a a chronic infection where low-grade recrudescent parasitemia persists for months (22). Additional advantages of the P. c. chabaudi model also include a robust circadian synchrony of parasitized RBC lysis and sequestration of parasitized RBCs into endothelial vessels – both features of human malaria theorized to be relevant to pathology (22–24).

In this study, we show that ablating CD4 T-cell IFN-γ production had no discernable impact on controlling acute PccCB parasitemia and only moderately exacerbated parasite burden during chronic PccCB infection. Excision of the Ifng^flox/KO locus resulted in a marked reduction of CD4 T-cells competent to produce IFN-γ that persisted for weeks and translated into significantly reduced serum IFN-γ levels at the peak of CD4 T-cell activation (d7pi). Compromised control of chronic parasitemia in the CD4 T-cell IFN-γ-ablated mice was not associated with reduced numbers or diminished phagocytic capacity of myeloid cell subsets. Instead, loss of CD4 T-cell derived IFN-γ paired with reduced class switching of parasite-specific antibodies to the IgG2c isotype, suggesting that reduced control of chronic parasitemia could be driven by functional impairments in the quality of the antibody response. In concert, our results highlight the importance of IFN-γ-independent CD4 T-cell effector mechanisms, and alternative sources of IFN-γ in controlling parasite burden during primary episodes of blood-stage malaria in mice.

Materials and Methods

Mice and tamoxifen treatment

Female C57BL6 (B6) mice and Ifng^−/− mice on the B6 background (25) were sourced from Jackson laboratories. Ifng^−/− mice were maintained in-house at the University of Iowa. Ifng^flox/wt mice were produced at the University of Iowa Genome Editing Facility, using a CRISPR knock-in strategy (26) (guides: TAAACTCAACAAAGCTGACG, CATGCTGGGTAAATGCACTG) to introduce LoxP flanking sites to B6SJLF1/J × B6 (Jackson Labs) embryos. After backcross of H-2 ^b/b founders to B6, Ifng^flox/wt mice were intercrossed to produce Ifng^flox/flox progeny. The CD4 T-cell-specific inducible IFN-γ ablation mice were generated by crossing Ifng^flox/flox mice to B6.CD4CreER^T2 mice (Jackson Labs) (21), producing litters of CD4CreER^T2+/− Ifng^flox/flox mice, which were then crossed with Ifng^−/− mice to yield the final CD4CreER^T2 Ifng^flox/KO mice, hereafter called CD4CreIfng. MHC haplotype of the final CD4CreIfn mice was determined to be H-2^b/b. The breeding scheme produces litters of CD4CreER^T2 -positive and -negative mice. In all experiments with CD4CreIfng mice, Cre⁺ and Cre⁻ littermates were identified and compared to each other, and genotypes confirmed by PCR. Tamoxifen (Sigma-Aldrich) treatments were performed by intraperitoneal injection of tamoxifen suspended in corn oil and 10% ethanol. 3 doses of 4mg were given the week prior to infection (day −7,−5,−3 relative to infection). TAM-induced excision was confirmed by intracellular cytokine stain of either peripheral blood or spleen samples after PMA/ionomycin stimulation for all experiments.

Infections

All infections were performed by intravenous (IV) injection at the retro-orbital sinus. For P. c. chabaudi CB (PccCB) or P. c. chabaudi AS (PccAS) experiments, mice were infected with 1 × 10⁵ PccCB-infected RBCs freshly harvested from BALB/c donors that were previously infected with cryopreserved stabilites. PccCB and PccAS were obtained from Patrick Duffy (NIAID). For P. yoelii 17XNL (Py17XNL) experiments, mice were injected IV with 1×10⁶ cryopreserved infected RBCs. Py 17XNL was obtained from BEI and maintained as cryopreserved stabilites produced from a C57BL/6 mouse infected from sporozoites (27). ActA- Listeria monocytogenes strain DP-L1942 was the gift of Dan Portnoy (UC Berkeley) (28) and maintained as previously described (25). 1×10⁷ colony forming units were used for IV challenge (25).

Parasitemia and anemia

Parasitemia and anemia were assessed by flow cytometry on glutaraldehyde-fixed peripheral blood as previously described (29), with infected RBCs defined as Hoechst⁺/ CD45⁻/Ter-119⁺cells.

Intracellular cytokine stain

Splenocytes were harvested and stimulated with 5 μg/mL PMA (phorbol 12-myristate 13-acetate), 5 μg/mL ionomycin and 5 μg/mL brefeldin A, followed by staining using the FoxP3 transcription factor kit (Tonbo) for permeabilization.

IFN-gamma ELISA

Serum IFN-γ concentrations were determined in duplicate technical replicates using the IFN-γ Femto High Sensitivity Mouse Uncoated ELISA kit (ThermoFisher).

Parasite end point titer ELISA

PccCB-specific endpoint antibody titers were determined via ELISA by probing sera samples against Nunc Maxi-sorp plates coated with 5 μg/mL crude parasite antigen prepared as a lysate from infected peripheral blood as previously described (30), HRP-conjugated IgG and IgG2c antibodies (Novus) were used for detection. Samples were processed in technical duplicate and sigmodal 4PL curve fit used to calculate end-point titers relative to naive sera.

Phagocytosis and myeloid population assay

Organs were collected 20 min post IV injection of 9×10⁹ Fluoresbrite BB Caryboxylate Microspheres (0.5 micrometer diameter) (Polysciences) via IV injection. Splenocytes were prepared by homogenization over 70 micron filters. Livers were homogenized via GentleMACS (Milyteni), then 30 min 37C digestion in Liver Digest Buffer (Gibco), filtration through 100 micron filters, and removal of debris with 35% Percoll gradient. RBCs were removed with Vitalyse (CMDG).

Flow Cytometry

Samples were run on an LSR Fortessa (BD). Staining panels are outlined in Tables 1–3.

Table 1.

Parasitemia and anemia

Marker	Clone	Fluorochrome	Manufacturer
CD45	30-F11	APC	Tonbo
Ter-119	Ter-119	APC/Cy7	Biolegend
Parasite DNA		Hoechst stain	Thermofisher

Table 3.

Myeloid panel

Marker	Clone	Fluorochrome	Manufacturer
Live-Dead	n/a	Near IR 780	Invitrogen
CD3e	145–2C11	FITC	BD
CD19	1D3	FITC	eBiosciences
Ly-6g	1A8	AlexaFLuor700	Biolegend
Ly-6c	HK1.4	PE/Dazzle 594	Biolegend
CD11b	M1/70	Brilliant Violet 785	Biolegend
CD11c	N418	APC	Tonbo
MHCII	114.15.2	PE/Cy7	ThermoFisher
F4/80	BM8.1	PerCP/Cy5.5	Tonbo

Statistics

All statistical analysis was performed in Graphpad Prism, using paired T-tests to compare two groups. ANOVA was used to compare multiple groups, with normally distributed data analyzed by Welch ANOVA with Brown-Forsythe test, and non-normally distributed data analyzed by ANOVA with Kruskal-Wallis post-test. Statistically significant (< 0.05) P-values are indicated.

Results

Ablation of IFN-γ production by CD4 T-cells

To specifically investigate the role of CD4 T-cell derived IFN-γ in controlling blood-stage malaria, we created a transgenic CD4CreIfng mouse model (21) where IFN-γ production can be inducibly ablated in mature CD4 T-cells. In this system, C57BL/6 CD4CreER^T2 Ifng^flox/KO mice (hereafter, CD4CreIfng) harbor a single floxed Ifng allele that can be efficiently excised via tamoxifen-inducible activity of CreERT2 expressed from the CD4 promoter.

To assess the extent to which CD4 IFN-γ production is reduced in our CD4CreIfng mice, we measured IFN-γ production in splenocytes harvested at d24pi with PccCB. Because MHC-II epitopes recognized by substantial populations of CD4 T-cells have not been identified for blood-stage murine malaria, we focused our analysis on dually CD11a^hiCD44^hi antigen-experienced cells (27) as a proxy for CD4 T-cells responding to PccCB (Fig. 1A), and an alternative CD11a^hiCD8a^int expression pattern for the identification of antigen-experienced CD8 T-cells (32). We observed a robust 14-fold decrease in the proportion of antigen-experienced CD4 T-cells producing IFN-γ in Cre⁺ vs Cre⁻ littermates, with no significant reductions in TNF-alpha production by CD4 T-cells or IFN-γ production by CD8 T-cells (Fig. 1B–D). Although cre activity in these mice is overwhelmingly restricted to mature CD4 T cells (21) we cannot formally exclude whether deletion of IFNγ in cells such as CD4 expressing monocytes may also contribute to the observed parasitemia alterations.

FIGURE 1.

Cytokine production in d24pi splenocytes from PccCB infected mice stimulated with PMA/ionomycin. (A) Gating strategy and representative cytokine staining of splenocytes from tamoxifen-treated ifngfl/fl CD4ERT2cre- and cre+ mice. Proportions of CD11a^hi/CD44^hi CD4 T-cells producing IFN-γ (B) and TNF-α (C), and CD11a^hi/CD8^lo CD8 T-cells producing IFN-γ (D), with P-values from Mann-Whitney tests indicated. (E-F) Expansion kinetics for peripheral blood antigen-experienced CD4 T-cells shown as group means +/− SEM (E). IFN-γ was measured by serum ELISA in wild-type C57BL/6 mice, vs a d1pi Listeria monocytogenes control (F), and PccCB d7pi and d14pi in CD4CreIfng mice (G) with P-values from Kruskal-Wallis post-test and one-way ANOVA indicated. Experiments in A-F show data from 2 experiments; G shows data from 4 experiments. Each symbol represents one mouse.

To test whether disrupting CD4 IFN-γ production led to a gross alteration of the CD4 response to PccCB, we tracked the kinetics of CD11a^hiCD44^hi antigen-experienced CD4 T-cells in the peripheral blood. No discernable differences were detected comparing Cre⁺ and Cre⁻ mice (Fig. 1E), suggesting the overall CD4 T-cell response remains intact even when IFN-γ production is ablated.

We next asked whether the achieved reduction CD4 T-cell IFN-γ production is sufficiently potent to alter systemic IFN-γ levels. In wild-type C57BL/6 mice, we observe a peak in serum IFN-γ content at PccCB d7pi (Fig. 1F), coinciding with maximal expansion of CD11a^hiCD44^hi CD4 T-cells in the peripheral blood (Fig. 1E). Serum IFN-γ content during PccCB malaria was orders of magnitude lower than that observed early during acute infection with Listeria monocytogenes (Fig. 1F). Nevertheless, we still detected a statistically-significant 5-fold decrease in serum IFN-γ content at PccCB d7pi in Cre⁺ vs Cre⁻ CD4CreIfng mice (Fig. 1G). Together, these data demonstrate the fidelity of our model for inducible deletion of IFN-γ from CD4 T cells.

CD4 T-cell derived IFN-γ and control of parasite burden during chronic P. c. chabaudi CB malaria

To specifically investigate the role of CD4 T-cell derived IFN-γ in controlling blood-stage malaria, we challenged our CD4CreIfng mice and Cre-negative littermate controls with blood-stage PccCB parasites to test the widely-hypothesized crucial role of CD4 T-cell-derived IFN-γ in controlling parasite burden during blood stage malaria. Mice were pre-treated with tamoxifen, and parasitemia, anemia and survival monitored during blood-stage PccCB malaria (Fig. 2A). During the first acute parasitemia wave, we observed no discernable impact of Cre-mediated CD4 IFN-γ ablation on parasitemia (Fig. 2B). During the chronic phase of PccCB malaria (d9pi+), we observed a consistent but modest increase in parasite burden in Cre⁺ mice, compared to Cre⁻ littermates (Fig. 2B). Anemia was not impacted by Cre-mediated CD4 IFN-γ ablation during either acute or chronic malaria (Fig. 2C).

FIGURE 2.

(A) CD4CreIfng mice were pre-treated with tamoxifen and then challenged with blood-stage malaria (Parasitemia (BFG), anemia (C) and survival (H) post challenge with 1×10⁵ PccCB -infected red blood cells (RBC) of tamoxifen-treated Cre⁺ vs. Cre⁻ CD4CreIfng mice (B,C), or age-matched (non-tamoxifen-treated) C57BL/6j and Ifng^−/−, females (D,E). Plot shading indicates acute (yellow) and chronic (no shading) infection stages. Mean parasitemia measurements calculated for groups (error bars, SEM) are shown as kinetic curves (B-E) (P values, Holm-Sidak’s test; error bars, SEM) and maximum parasitemia observed in single mice during acute (d0–d9pi, F) or chronic (d11–17pi, G) stages of infection (P values, Brown-Forsythe test with Welch ANOVA). Survival curves (H) (P values, Gehan-Breslow-Wilcoxon test). Data from 3 experiments with C57BL/6 (N=18) and Ifng^−/− (N=14) mice and 4 experiments with CD4CreIfng mice (N=18) are shown. Parasitemia (I) and anemia (J) post challenge with 1×10⁶ P. yoelli 17XNL-infected red blood cells (RBC) of tamoxifen-treated Cre⁺ vs. Cre⁻ CD4CreIfng mice. P. yoelii data is from 1 experiment, in which all mice survived and no statistically significant differences between Cre⁺ (N=7) and Cre⁻ (N=4) littermates were observed.

The relatively minor impact of ablating IFN-γ production in CD4 T-cells on parasitemia was unexpected. Thus, we next queried the impact of complete disruption of IFN-γ production on PccCB malaria by comparing parasitemia and anemia in wild-type C57BL6/J and Ifng^−/− mice. Consistent with prior reports (4), we observed a dramatic increase in acute and chronic parasitemias in the Ifng^−/− mice (Fig. 2D). Anemia was also exacerbated in the Ifng^−/− mice during the chronic phase of infection (Fig. 2E), and survival subtly but significantly decreased (Fig. 2H). Non-tamoxifen-treated C57BL/6 mice consistently developed lower acute parasitemia than those observed in experiments with tamoxifen treated Cre⁻ mice (Fig. 1 B vs D), likely due to previously-reported ability of tamoxifen to exacerbate parasitemia during blood-stage malaria¹.

Because the recrudescent parasitemia episodes of PccCB malaria occur asynchronously, the magnitude of these episodes can be obscured by tracking mean parasitemia in groups of mice. To account for this limitation, we compared the maximum observed parasitemia in individual mice across the entire time course of the infection (Fig. 2F,G). During the acute infection, Ifng^−/− mice again displayed significantly exacerbated maximum parasitemia values relative to C57BL6, whereas Cre⁺ vs Cre⁻ CD4CreIfng mice did not significantly differ from one another (Fig. 2F). However, during the chronic phase of infection, Cre⁺ CD4CreIfng mice developed elevated parasite burdens (vs Cre⁻ littermate controls) of similar severity to the elevated parasitemia of Ifng^−/− mice (Fig. 2G).

Combined, these results suggest that IFN-γ is indeed a crucial player in suppressing parasite burden during blood-stage PccCB malaria. However, IFN-γ production by CD4 T-cells only contributes to managing parasite burden during the chronic phase PccCB malaria and appears functionally irrelevant to limiting acute parasitemia. The IFN-γ crucial for controlling acute parasitemia that leads to dramatically elevated acute parasitemia in Ifng^−/− mice must be produced by other cell populations, of which several candidates have emerged in other work (18).

To probe whether the limited ability of CD4 T-cell-derived IFN-γ to control acute parasitemia is a quirk of the PccCB model, we compared parasitemia and anemia kinetics in the CD4CreIfng mice in another acute resolving model of murine blood-stage malaria, P. yoelii 17XNL (31). In this system, we could not discern any exacerbation of parasitemia in Cre⁺ CD4 IFN-γ-ablated mice, and instead observed a non-significant trend towards improved control of parasitemia and anemia in Cre⁺ mice, vs Cre⁻ littermate controls (Fig. 2I,J).

Assessment of phagocytosis

Effective control of parasite burden during blood-stage malaria requires identification and removal of parasitized RBCs. The spleen is well recognized as a prime site for the removal of parasite RBCs due to its inherent blood filtration function 4, 32–34), while the liver is infrequently assessed as a player during blood stage malaria but also well positioned as a site for parasitized RBC clearance as blood and phagocytes are closely juxtaposed in the liver sinusoids. (33, 34). IFN-γ is classically considered an activator of macrophage phagocytosis, but can also inhibit phagocytosis in some scenarios (35), such as non-opsonic phagocytosis (36, 37). During P. c. chabaudi AS malaria, IFN-γ was found to enhance the effector function of splenic macrophages (4). Accordingly, we hypothesized that a failure for IFN-γ to activate phagocytic mechanisms may explain the poor control of parasite burden observed in Ifng^−/− mice and the CD4 IFN-γ-ablated mice.

We tested whether the strikingly poor control of chronic parasitemia observed in Ifng^−/− mice (Fig. 2D) could be explained by deficient IFN-γ signaling leading to alternations in the presence of phagocytic myeloid cells in the spleen or liver, or phagocytic capacity of such cells. At d14pi with PccCB, fluorescent beads were introduced I.V. into Ifng^−/− mice and C57BL6/j mice, and splenic and liver myeloid populations were assessed for abundance and bead capture (Fig. 3A). We observed no significant differences in the abundance of Ly6c⁺CD11b⁺ or Ly6c⁻CD11b⁺ monocytes or macrophages, neutrophils, or dendritic cells in the spleen or liver (Fig. 3BC), although trending decreases in Ly6c⁺CD11b⁺ monocytes and neutrophils were seen in the spleens of Ifng^−/− mice. Instead, we observed increased proportions of bead-positive CD11b+ monocytes and dendritic cells in both the spleen and liver in Ifng^−/− mice vs wild-type C57BL/6J controls (Fig. 3DE), and increased absolute quantities of bead-positive myeloid cells per organ (Fig. 3FG). Enhanced phagocytosis in Ifng^−/− mice does not support a model where compromised phagocytosis leads to the poor control of parasitemia we observed in Ifng^−/− mice. Of note, phagocytosis of parasitized RBCs may occur in a mechanistically distinct manner that is not reflected by our bead-based experiment, in particular in an antigen-specific process such as opsonization-dependent phagocytosis and further experiments would be necessary to address this possibility. Accordingly, we focused our remaining efforts on investigating disruptions in humoral immunity in the absence of IFN-γ, given that altered parasite control in Cre⁺ CD4CreIfng mice only manifested after the first week of infection.

FIGURE 3.

(A) Experimental schematic and gating strategy for assessing myeloid populations and phagocytosis. (B-C) Myeloid populations shown as number of cells per organ in the spleens (B) and livers (C) at d14pi. (D-G) Bead capture shown as percent (DE) or absolute number of bead positive cells per organ (FG, spleen; DF, liver) at d14pi. P-values from Brown-Forsythe post-test and Welch Anova indicated. Data is from 2 experiments, each with 4 mice/group.

Disruption of antibody class switching

Abundant evidence implicates IFN-γ in the tuning of antibody class-switching towards IgG2a (35, 38, 39). Hence, we measured PccCB-specific IgG2c (the functional equivalent of IgG2a expressed by C57BL/6 mice (40)) antibody titers at d24pi. As expected, IgG2c-specific titers were reduced in sera from Ifng^−/−, vs C57BL/6 mice (Fig. 4A). IgG2c titers were similarly reduced in sera from Cre⁺ vs. Cre⁻ CD4CreIfng mice (Fig. 4A). The similarly reduced IgG2c titers in Cre⁺ Cre⁻ CD4CreIfng and Ifng^−/− mice suggest that CD4 T-cell derived IFN-γ is indeed crucial in driving class switching to the IgG2c isotype and further supports the robust excision of Ifng in our mice. Importantly, we did not observe significant differences in PccCB-specific titers of total IgG isotypes, comparing C57BL/6 mice to Ifng^−/−, or Cre⁺ and Cre⁻ CD4CreIfng mice (Fig. 4B). The lack of titer reduction for all IgG supports a model where class-switching is specifically reduced, rather than an overall inhibition of parasite-specific humoral responses.

FIGURE 4.

(A-B) Pcc-CB specific IgG2c (A) and total IgG (B) endpoint titers in d24pi sera. P-values from one-way ANOVA and Kruskal-Wallis post-test are shown. (C-G) Splenic follicular T helper (Tfh) cells (CXCR5^hi PD^hi; C, gating for Tfh, following pre-gating for single lymphocytes by FSC/SSC and CD11a-hi/Cd44-hi CD4+ cells) at PccCB d24pi, with proportion (DF) or absolute numbers (EG) indicated (DE-Tfh; FG- GC Tfh) indicated. Data from 2 independent experiments are shown. (H-J) Sera from PccCB d24pi or naïve mice was transferred into Ifng^−/− mice at peak acute P. c. chabaudi AS parasitemia (H). Parasitemia (normalized to time of sera transfer) shown as maximum observed values in individual mice (I) or mean values +/− SEM for groups (J). Hash tags (#) indicate mortality. P-values indicate Kruskal Wallis post-test and one-way ANOVA. Data from 3 experiments are shown.

We next asked whether the ablation of CD4 IFN-γ production could disrupt humoral responses by disrupting efficient formation of CD4+ follicular T helper (Tfh) cells important for development of Plasmodium-specific immunity. No differences in splenic Tfh or GC Tfh (defined by CXCR5 and PD1 expression, Fig. 4C) abundance, in absolute or proportional terms, were observed at PccCB d24pi (Fig. 4D–G).

We finally asked if reduced IgG2c-switching in Cre⁺ CD4CreIfng mice had functional consequences for the control of parasite burden, using Ifng^−/− mice infected with a P. c. chabaudi strain, AS, that requires humoral immunity for control (20), in order to highlight CD4-derived IFN-γ-dependent changes in antibody-mediated control. After the peak of acute parasitemia, P. c. chabaudi AS-infected Ifng^−/− mice received sera produced by either naïve mice, or tamoxifen-treated CD4CreIfng mice at PccCB d24pi (Fig. 4H). As expected, control of parasite burden was least effective in mice receiving sera from naïve donors (Fig. 4I,J). Compared to naïve sera controls, the maximum observed parasitemia following transfer was significantly reduced only for mice receiving sera from Cre⁻ CD4CreIfng mice (Fig. 4I,J). A non-significant trend towards enhanced parasitemia in Cre⁺ vs Cre⁻ CD4CreIfng mice was also observed (Fig. 4I,J). Combined, these results suggest that antibody responses to P. c. chabaudi infection that develop in the absence of CD4 T-cell-derived IFN-γ are very subtly inferior at conferring control of parasite burden.

Discussion

In total, our results reveal a striking role for IFN-γ in suppressing parasite burden during PccCB malaria. Consistent with previous studies (4), total disruption of the Ifng locus dramatically exacerbated parasitemia during acute and chronic PccCB malaria (Fig. 1EF).

To build upon this finding, we used an inducible CD4-specific Cre/Lox excision system to define the contribution of CD4 T-cell IFN-γ production to IFN-γ-dependent parasitemia suppression. Although the peak of CD4 T-cell activation occurred contemporaneously to the first peak of acute parasitemia at ~d7pi, we could not discern any role for CD4 T-cell-derived IFN-γ in controlling this first parasitemia wave (Fig. 1B). Inefficient ablation of CD4 T-cell IFN-γ production is unlikely to explain the lack of an acute parasitemia phenotype, as ablation was sufficiently potent to 1) be robustly detected at d24pi in splenic CD4 T-cells, nearly a full month post tamoxifen-induced Ifng excision (Fig. 1B), 2) significantly reduce serum IFN-γ content at the d7pi peak of CD4 T-cell activation and serum IFN-γ accumulation (Fig. 1G), and 3) significantly alter antibody class-switching (Fig. 4A).

Accordingly, we conclude that other sources of IFN-γ are crucial to controlling parasite burden during the acute phase of PccCB malaria. Future studies should investigate whether IFN-γ-producers identified in other studies, including NK, NKT, gamma-delta T-cells, and CD8 T-cells (18) are crucial sources of this protective cytokine, either as sole or cooperative actors.

Our CD4CreIfng mice did reveal a moderate yet significant role for CD4 T-cell-derived IFN-γ in controlling recrudescent parasitemia during the chronic PccCB malaria (Fig. 2B,G). More efficient suppression of parasite burden observed in CD4 T-cell-IFN-γ-intact mice may stem from a superior humoral response via more efficient antibody class-switching. Although this defect in isotype class switching manifested as enhanced parasitemia during the chronic phase of PccCB malaria, the crucial window of activity for CD4 T-cell-derived IFN-γ likely lies earlier, at ~d7pi when serum IFN-γ content and the expansion of activated CD4 T-cells both peak (Fig. 1E,F,G), and significant reductions in serum IFN-γ were discernable between Cre⁺ vs Cre⁻ mice in the CD4CreIfng model (Fig. 1G). Notably, during blood-stage malaria, we never observed accumulation of IFN-γ in the serum at levels similar to that observed early in other infections, such as Listeria monocytogenes (Fig. 1F). This observation may suggest that IFN-γ signaling acts on primarily a local, rather than system level during blood-stage malaria, or simply that serum IFN-γ is rapidly consumed during blood-stage malaria.

An intriguing question for further follow-up is whether IFN-γ production by memory-phenotype CD4 T-cells would more rapidly or potently control parasite burden during subsequent blood-stage malaria episodes. Several studies in mice show that IFN-γ producing memory CD4 T-cells can confer potent immunity (41, 42). As humans residing in malaria-endemic regions experience many episodes of malaria throughout their life and accordingly develop expansive and complex memory lymphocyte compartments, it is likely memory CD4 T-cells are the true analogs of the protective IFN-γ producing memory cells observed in human studies. Further use of the mouse model to refine the mechanism by which IFN-γ, produced both by CD4 T-cells and other cell populations, should help to guide efforts to design efficacious vaccines targeting the blood-stage of malaria.

Table 2.

Intracellular cytokine stain and follicular T helper stain

Marker	Clone	Fluorochrome	Manufacturer
CD4	RM4–5	APC/eFluor 780	eBiosciences
CD8	53–6.7	PE/Dazzle-594	Biolegend
CD11a	M17/4	Brilliant Violet 510	BD Biosciences
CD44	IM7	RedFluor 710	Tonbo
PD1	RMP10–30	PE/Cy7	Biolegend
CXCR4	L128D7	Brilliant Violet 785	Biolegend
IFN-gamma	XMG1.2	PE	Tonbo
TNF-alpha	MP6-XT22	PerCP/ eFluor710	eBiosciences

Key Points.

IFN−γ contributes to control of acute and chronic blood-stage Pcc malaria

CD4 T cell IFN-γ contributes to control of chronic Pcc blood-stage malaria

IFN-γ has CD4 T cell independent roles in control of blood-stage Pcc malaria