Hemozoin-mediated inflammasome activation limits long-lived anti-malarial immunity

SUMMARY

During acute malaria, most individuals mount robust inflammatory responses that limit parasite burden. However, long-lived sterilizing anti-malarial memory responses are not efficiently induced, even following repeated Plasmodium exposures. Using multiple Plasmodium species, genetically modified parasites, and combinations of host genetic and pharmacologic approaches, we find that the deposition of the malarial pigment hemozoin directly limits the abundance and capacity of conventional type 1 dendritic cells to prime helper T cell responses. Hemozoin-induced dendritic cell dysfunction results in aberrant Plasmodium-specific CD4 T follicular helper cell differentiation, which constrains memory B cell and long-lived plasma cell formation. Mechanistically, we identify that dendritic cell-intrinsic NLRP3 inflammasome activation reduces conventional type 1 dendritic cell abundance, phagocytosis, and T cell priming functions in vivo. These data identify biological consequences of hemozoin deposition during malaria and highlight the capacity of the malarial pigment to program immune evasion during the earliest events following an initial Plasmodium exposure.

Graphical abstract

In brief

Using genetic, chimeric, and pharmacologic approaches, Pack et al. demonstrate that the parasite-derived crystal hemozoin erodes conventional CD8α⁺ type 1 dendritic cell (cDC1) abundance and function in an NLRP3 inflammasome-dependent manner, which results in reduced CD4⁺ T follicular helper cell differentiation and impaired anti-Plasmodium humoral immunity.

INTRODUCTION

Nearly half of the world’s population lives in areas endemic to the disease syndrome malaria, which is caused by infection with protozoan parasites of the genus Plasmodium. Children under the age of 5 account for two-thirds of all malaria-associated mortality (WHO, 2019). Repeated parasite infections generally fail to elicit sterilizing immunity, even following years of exposure among those living in regions of intense Plasmodium transmission (Tran et al., 2013). The emergence of parasite resistance to front-line antimalarial drugs (Das et al., 2018), coupled with the limited capacity of existing vaccines to elicit protection in endemic populations (Mensah et al., 2016; Rts, 2015), underscores the rationale for continued efforts to close knowledge gaps regarding the mechanisms, whether host or parasite associated, that restrict the development of durable, protective immune memory responses.

Following an initial Plasmodium infection and asymptomatic replication of parasites in the liver, schizont-induced rupture of parasite-infected red blood cells (pRBCs) liberates both host- and parasite-derived factors, including host RBC membranes, parasite antigens, and nucleic acids, that are recognized and often captured by host phagocytes. These stimuli rapidly induce the activation of macrophages and dendritic cells (DCs) that produce reactive oxygen and nitrogen species (Chua et al., 2013; Ranjan et al., 2016; Sobolewski et al., 2005; Sponaas et al., 2009), which are believed to limit infection intensity prior to the induction and orchestration of anti-Plasmodium cellular and humoral immunity. Phagocytes are also responsible, in part, for the production of high levels of proinflammatory cytokines such as interleukin (IL)-1β, IL-12, and interferon (IFN)-γ that both amplify the activation and recruitment of additional effector immune cells (Leisewitz et al., 2004; Perry et al., 2005; Wykes et al., 2007) and contribute to clinical malarial disease syndrome (deWalick et al., 2007). In addition to initiating and amplifying host immunity, DCs also contribute to the initiation of T follicular helper (T_FH) cell differentiation via MHC II-restricted antigen presentation and costimulation (Choi et al., 2011; Langenkamp et al., 2000). T_FH populations, in turn, are critical for orchestrating protective pathogen-specific humoral responses and deficiencies in anti-malarial humoral immunity have been linked to altered T_FH development and function (Hansen et al., 2017; Ryg-Cornejo et al., 2016; Zander et al., 2016; Zander et al., 2015). Whether DCs are distinctly programmed during Plasmodium blood-stage infection such that they inefficiently support the development of anti-Plasmodium T_FH responses and humoral immunity is not well defined.

During either in vivo blood-stage infection or following direct in vitro incubation with pRBCs, DCs exhibit an atypical maturation phenotype (Elliott et al., 2007; Götz et al., 2017; Urban et al., 1999), as well as reduced responsiveness to LPS stimulation and impaired capacity to initiate heterologous immune responses during an active Plasmodium infection (Lundie et al., 2010; Millington et al., 2006). Compared to DCs from uninfected subjects, myeloid DCs isolated from the peripheral blood of P. falciparum exposed individuals expressed lower levels of CD80, CD86, and HLA-DR following in vitro stimulation with blood-stage parasites (Turner et al., 2021). In vitro incubation of DCs with intact and dissociated pRBCs, along with RBC ghosts (membranes), identified that hemozoin, a crystalized parasite-derived hemoglobin degradation byproduct and NLRP3 inflammasome and CLEC12A agonist (Kalantari et al., 2014; Raulf et al., 2019; Shio et al., 2009), can alter DC function (Schwarzer et al., 1998; Skorokhod et al., 2004). However, whether modulating hemozoin levels in the context of natural Plasmodium blood-stage infection directly impacts DC-dependent development of anti-Plasmodium humoral responses has not been investigated.

In this report, we combined genetic, chimeric, and pharmacologic approaches to interrogate whether hemozoin exposure during experimental malaria impacts immune programming and the development of protective anti-Plasmodium humoral immune memory responses. Using both genetically modified Plasmodium berghei (Pb) parasites engineered to produce 10-fold lower levels of hemozoin (Lin et al., 2015) and wild-type P. yoelii infection of NLRP3-deficient mice, we show that hemozoin-induced, DC-intrinsic NLRP3-mediated inflammasome activation compromises T_FH differentiation, which in turn limits anti-Plasmodium memory B cell and the long-lived plasma cell responses and protective humoral immunity against malaria.

RESULTS

Durable anti-Plasmodium humoral immunity is compromised by hemozoin

To address whether hemozoin formation and deposition influences the generation of durable anti-malarial immunity, we established memory immune responses in wild-type C57BL/6 mice by infecting them with a genetically modified Plasmodium berghei ANKA variant lacking the aspartic proteases plasmepsin 4 and berghepain-2 (Δpm4Δbp2), genes that are necessary for efficient degradation of hemoglobin and crystallization of hemozoin (Lin et al., 2015; Spaccapelo et al., 2010). Mature schizonts of Δpm4Δbp2 parasites produce less hemozoin compared to wild-type (WT) P. berghei ANKA parasites with 50% reductions in hemozoin accumulation within the spleen as early as day 5 post-infection (p.i.) (Figure S1A). Because Δpm4Δbp2 parasites also exhibit reduced growth and replication in vivo compared to WT Pb ANKA parasites (Lin et al., 2015), other groups of mice were infected with a comparison mutant of growth-attenuated P. berghei lacking leucyl aminopeptidase (Δlap), a mutant that is not impaired in hemozoin production (Lin et al., 2015) but exhibits growth and clearance kinetics that are equivalent to Δpm4Δbp2 parasites (Figure S1B). These genetically modified parasites enabled us to directly address the in vivo role of hemozoin and avoid major caveats associated with injection of exogenous hemozoin, including the form (i.e., synthetic or natural), site of accumulation, and dose (Pham et al., 2021). Sixty days after the initial infection with either Δlap or Δpm4Δbp2 parasites, we challenged both groups of mice with WT (virulent) P. berghei ANKA parasites engineered to express the fusion protein GFP-Luciferase (Pb-ANKA-Luc) (Figure 1A). Compared to Δlap-immune mice, Δpm4Δbp2-immune mice exposed to lower levels of hemozoin during primary infection exhibited reduced parasitemia (Figure 1B) and total parasite biomass luminescence (Figure 1C) across all time points examined and whole-body bioluminescent imaging revealed substantial Pb-ANKA-Luc tissue accumulation in Δlap-immune mice that was absent in Δpm4Δbp2-immune mice (Figure 1C). Δpm4Δbp2-immune mice also exhibited 100% survival following WT Pb-ANKA-Luc challenge (Figure 1D).

Figure 1. Hemozoin limits humoral immune memory responses.

(A) C57BL/6 mice were infected with either Δpm4Δbp2 or Δlap parasites. After 60 days, immune mice were challenged with virulent WT Pb-ANKA-luc.

(B) The kinetics of peripheral parasite burden.

(C) Quantification (left) of parasite load as determined by whole-body bioluminescence following virulent Pb-ANKA-luc challenge. Images (right) on day 19 post-challenge.

(D) Survival of Δpm4Δbp2 and Δlap-immune mice following virulent Pb-ANKA-luc challenge.

(E) IgG was purified from the indicated immune mice and either purified IgG or the unbound flow-through was transferred via tail vein to naive WT mice 1 day prior to challenge with virulent Pb-ANKA.

(F) Parasite burden at day 4 post-challenge.

(G and H) Representative flow plots, frequency (G), and (H) number of mature plasma cells (B220^negCD19^negCD138^hiTACI^hi) in the bone marrow at day 40 p.i.

(I) Number of MSP1_1–19-specific plasma cells in the bone marrow at day 40 p.i.

(J and K) Representative flow plots (J) and quantification of (K) splenic IgD^negCD38⁺CD138^hi LLPC precursors expressing CXCR4 at day 14 p.i.

(L and M) Frequency of (L) bone-marrow IgD^negCD38⁺CD138^hiTACI^hi plasma cells and (M) MSP1_1–19-specific IgG secreting cells recovered from the bone marrow of immune mice at day 28 p.i.

Data (mean ± SEM) in (B)–(D) and (mean ± SD) in (I) and (M) are the summary data from two independent experiments. Data (mean ± SD) in (F) and (J)–(L) are representative of two independent experiments with n = 4. Data (mean ± SD) in (G) and (H) are representative of three experiments with n = 4 (see also Figure S1).

Because Δpm4Δbp2-immune mice exhibited enhanced control of WT Pb-ANKA as early as days 3–4 post-challenge, relative to Δlap-immune mice (Figure 1B), we further posited that the observed resistance to virulent Pb-ANKA-Luc challenge could be attributed to pre-existing, protective anti-Plasmodium antibodies. Consistent with this hypothesis, passive transfer of purified immunoglobulin G (IgG) from Δpm4Δbp2-immune donors to WT Pb-ANKA-infected recipient mice (Figure 1E) transiently resulted in a 75% reduction in parasitemia, compared to transfer of an equivalent amount of purified IgG from Δlap-immune donors (Figure 1F), although immune IgG transfer was not sufficient to protect recipients beyond the first week (Figure S1C). Strikingly, antibody titers directed against the 19 kDa fragment of Merozoite Surface Protein 1 (MSP1₁₉), an immunodominant and protective blood-stage antigen (John et al., 2004; O’Donnell et al., 2001; Terrientes et al., 1994), were not detectably different between Δlap and Δpm4Δbp2-immune mice (Figure S1D), and we found no significant alterations in the total IgG antibody avidity index or isotype class distributions (Figures S1E and S1F). Nevertheless, our passive transfer data are consistent with an enhanced long-lived antibody response in Δpm4Δbp2-immune mice, compared to Δlap-immune mice, which may be reflected by either quantitative or qualitative (specificity and functional) shifts that register below the sensitivity of these assays.

Secreted anti-Plasmodium antibodies that are present 60 days p.i. likely derive from long-lived plasma cells (LLPCs) that often reside in the bone marrow (Lightman et al., 2019). Given the enhanced protective capacity of Δpm4Δbp2-immune IgG, we quantified parasite-specific LLPC in the bone marrow during the memory phase. By day 40 p.i., we found approximately 2-fold greater numbers of CD138^hiTACI^hiCD19^loB220^lo mature resting LLPC (Pracht et al., 2017) (Figures 1G and 1H) and functional MSP1₁₉-specific IgG secreting cells (Figure 1I; Figure S1G) in the bone marrow of Δpm4Δbp2-immune mice, compared to Δlap-immune mice. Thus, we next evaluated the kinetics of parasite-specific LLPC formation. As early as day 14 p.i., the frequency and number of splenic CD38⁺IgD^negCD138^hi LLPC precursors expressing CXCR4, a chemokine receptor associated with bone marrow homing (Nie et al., 2004), was 30%–40% greater in Δpm4Δbp2-infected mice (Figures 1J and 1K). By day 28 p.i., the bone marrow of Δpm4Δbp2-immune mice presented with twice as many CD38⁺IgD^negCD138^hiTACI^hi LLPC (Figure 1L) and MSP1₁₉-specific IgG-secreting cells (Figure 1M), compared to Δlap-immune mice. Taken together, these data support that deposition of hemozoin during a primary blood-stage Plasmodium exposure impairs either the development or function of LLPC and durable anti-Plasmodium humoral responses.

Hemozoin compromises memory B cell responses

Given the deleterious effect of hemozoin on LLPC development and function following primary blood-stage infection, we additionally hypothesized that hemozoin may interfere with the development and function of memory B cell (MBC) populations. Consistent with this hypothesis, splenic CD19⁺CD38⁺IgD^negCD73⁺CD95^neg MBC and CD19⁺CD38⁺IgD^negCD73⁺CD80⁺PD-L2⁺ MBC, with CD80⁺PD-L2⁺ defining those most capable of rapid antibody production (Zuccarino-Catania et al., 2014), were elevated 2-fold by day 30 p.i. in Δpm4Δbp2-immune mice, compared to Δlap-immune mice (Figures 2A-2D). The chemokine receptor CCR6 is reported to enhance cell positioning within secondary lymphoid tissues for superior recall upon pathogen challenge (Elgueta et al., 2015; Suan et al., 2017) and by day 60 p.i., the frequency of splenic MBC cells expressing CCR6 were 2- to 3-fold greater in Δpm4Δbp2-immune mice, compared to Δlap-immune mice (Figures 2E-2G). To interrogate functional MBC responses, we evaluated MSP1₁₉-specific serum antibody titers in Δpm4Δbp2 and Δlap-immune mice both prior to and 5 days after challenge with WT Pb-ANKA. In this short interval, increases in MSP1₁₉-specific, isotype-switched IgG titers are most likely due to MBC activation and differentiation into plasma cells, rather than de novo humoral immune responses. Although Δpm4Δbp2-immune mice harbor 2-fold higher numbers of MBC (Figures 2A-2D), 5 days after challenge MSP1₁₉-specific IgG titers expanded by 10-fold, as compared to 3-fold in Δlap-immune mice (Figure 2H). Together, these data suggest that recall functions of MBC are impaired by the presence and deposition of hemozoin.

Figure 2. Hemozoin constrains the number and function of memory B cells.

C57BL/6 mice were infected with Δpm4Δbp2 and Δlap parasites.

(A–D) Representative flow plots (A) and quantification (B) of splenic CD19⁺CD38⁺IgD^negCD73⁺CD95^neg MBC and the frequency (C) and number (D) of splenic CD19⁺CD38⁺IgD^negCD73⁺CD80⁺PD-L2⁺ MBC by day 30 p.i.

(E) Representative histograms describing frequency (F) and number (G) of CCR6⁺CD19⁺CD38⁺IgD^neg MBC.

(H) Δpm4Δbp2 and Δlap-immune mice were challenged with virulent Pb-ANKA. Sera were collected before (pre) and 5 days after (post) challenge and MSP1_1–19-specific IgG endpoint titers were quantified. Data (mean ± SD) in (A)–(D) and are representative of three independent experiments with n = 4–5. Data (mean ± SD) in (E)–(H) are the summary data of two independent experiments.

CD4 T cell help and germinal center B cell responses are impaired by hemozoin

LLPC and MBC populations derive from germinal center (GC) B cell responses (Mesin et al., 2016; Weisel et al., 2016) and hemozoin deposition was associated with reduced LLPC and MBC development as early as days 14 and 30 p.i. in Δpm4Δbp2-infected mice, time points that encompass the peak of GC B cell responses following experimental Plasmodium (Krishnamurty et al., 2016; Vijay et al., 2020). Indeed, by day 14 p.i. the frequency and number of CD95⁺GL-7⁺ GC B cells were elevated by ~30% in Δpm4Δbp2-infected mice, compared to Δlap-infected mice (Figures 3A and 3B). Because GC B cell responses depend on the presence and activity of T_FH cells (Kräutler et al., 2017), we next postulated that hemozoin impairs either the induction, stepwise differentiation, or function of T_FH populations. To test this hypothesis, we transferred Plasmodium-specific CD4 T cell receptor transgenic cells (PbTII [Fernandez-Ruiz et al., 2017]) to WT mice, infected recipients 1 day later with either Δpm4Δbp2 or Δlap parasites and then quantified both polyclonal, endogenous and PbTII CD4 T cells exhibiting characteristics of either mature CXCR5^hiPD-1^hi GC-T_FH cells that localize to follicular GC reactions or their extrafollicular CD150^loCXCR5^int T_FH precursors (Choi et al., 2011; Crotty, 2014). By day 14 p.i., we found 30%–50% increases in the frequencies and numbers of both CD150^loCXCR5^intT_FH precursors (Figures 3C and 3D) and mature CXCR5^hiPD-1^hi GC-T_FH subsets (Figures 3E and 3F) among PbTII cells in Δpm4Δbp2-infected mice, compared to PbTII cells in Δlap-infected mice. We also observed 2-fold greater numbers of CXCR5^hiPD-1^hi GC-T_FH cells among polyclonal endogenous CD4 T cells recovered from Δpm4Δbp2-infected mice (Figures 3G and 3H), supporting the view that this effect extends to the endogenous CD4 T cell receptor repertoire. Bcl-6 is a transcriptional repressor that is essential for T_FH differentiation and function (Hatzi et al., 2015) and T-bet⁺ T_H1-like T_FH and Foxp3⁺ T regulatory cells are reported to constrain anti-Plasmodium humoral immunity (Kurup et al., 2017; Obeng-Adjei et al., 2015). However, on 14 p.i., we identified only modestly reduced levels of T-bet and Foxp3, and equivalent expression levels of Bcl-6, in mature GC-T_FH cells recovered from Δpm4Δbp2-infected mice, compared to mature GC-T_FH cells recovered from Δlap-infected animals (Figure 3I; Figure S2A). Despite these modest cell-intrinsic transcriptional differences, we additionally tested whether mature GC-T_FH cells recovered from Δpm4Δbp2-infected mice differed in their capacity to promote GC B cell differentiation. We sort-purified mature CXCR5^hiPD-1^hiGC-T_FH from each group and co-cultured cells with an equivalent number with naive B cells for 5 days (Figure 3J). We observed significant increases in the proportions of B cells exhibiting a CD95⁺GL-7⁺ GC phenotype when cells were cultured with mature GC-T_FH recovered from Δpm4Δbp2-infected mice, compared to mature GC-T_FH cells recovered from Δlap-infected mice (Figures 3K and 3L). Moreover, T_FH cells sort-purified on day 7 p.i. from the spleens of Δlap-infected mice were also impaired in their capacity to drive GC responses following transfer to Tcrα^−/− recipients, compared to T_FH recovered from Δpm4Δbp2-infected mice (Figures 3M-3P; Figures S2B-S2D). Together, these data support that the presence of elevated levels of hemozoin impairs the differentiation and cell-intrinsic function of T_FH cells with subsequent reductions in GC B cell responses.

Figure 3. Hemozoin impairs the germinal center response.

Parasite-specific CD4 T cells (PbTII) were transferred to C57BL/6 mice 1 day prior to infection with either Δpm4Δbp2 or Δlap parasites. Splenic T_FH and GC B cell responses were evaluated.

(A–F) Representative plots (A) and frequency and number (B) of GC B cells (B220⁺CD19⁺CD95⁺GL-7⁺) on d14 p.i. T_FH precursors (CD150^loCXCR5^int) (C and D) and mature GC-T_FH (CXCR5^hiPD-1^hi) expansion (E and F) among donor-derived PbTII cells.

(G and H) Frequency (G) and summary quantification (H) of mature GC-T_FH (PD-1^hiCXCR5^hi) among polyclonal endogenous CD4 T cells.

(I) Relative expression (gMFI) of Foxp3, T-bet, and Bcl-6 in polyclonal GC-T_FH cells on day 14 p.i.

(J) Polyclonal GC-T_FH cells were sort purified from spleens day 14 p.i. and co-cultured for 6 days with purified naive B cells.

(K and L) Representative flow plots (K) and summary (L) of GC B cell (B220⁺CD19⁺CD95⁺GL-7⁺) frequency following co-culture.

(M) Polyclonal T_FH cells were sort purified from spleens of Δpm4Δbp2 and Δlap-infected mice day 7 p.i. and transferred to Tcrα^−/− mice. Tcrα^−/− mice were infected with Δlap parasites and GC responses were assessed on 14 p.i.

(N–P) Representative flow plots (N) and summary of GC B cell (B220⁺CD19⁺CD95⁺GL-7⁺) frequency (O) and number (P).

Data (mean ± SEM) in (A)–(H) are representative of four experiments with n = 4–5. Data (mean ± SD) in (J)–(L) are the summary data from two independent experiments. Data (mean ± SD) in (M)–(P) are representative of two independent experiments with n = 5 (see also Figure S2).

Conventional DC functions are compromised by hemozoin

CD11c⁺B220^neg conventional dendritic cells (cDCs), particularly the CD11b^negCD8α⁺ conventional type 1 DC (cDC1) subset, are critical for initiating protective T cell responses against blood-stage Plasmodium infection (Fernandez-Ruiz et al., 2017; Voisine et al., 2010). Total CD11c⁺B220^neg cDC numbers, phagocytic functions, and cytokine expression are reported to progressively decline during the course of experimental and clinical malaria, which has been described as immunoparalysis (Loughland et al., 2016; Lundie et al., 2010; Pinzon-Charry et al., 2013; Woodberry et al., 2012). Moreover, purified hemozoin can down-modulate DC reactivity following direct co-incubation ex vivo (Millington et al., 2006; Skorokhod et al., 2004). However, whether the deposition and/or accumulation of hemozoin specifically impairs either total cDC or cDC1 functions during blood-stage Plasmodium infection in vivo has not been directly tested. To investigate this, we examined cDC subset abundance, costimulatory receptor expression, and phagocytic function in Δpm4Δbp2 and Δlap-infected mice (representative gating, Figure S3A). As early as day 4 p.i., we found elevated numbers of CD11c⁺B220^neg cDCs in Δpm4Δbp2-infected mice, compared to their Δlap-infected counterparts (Figures 4A and 4B). These increases were primarily associated with elevated frequencies and numbers of CD11b^negCD8α⁺ cDC1 in Δpm4Δbp2-infected mice (Figures 4C-4E) expressing elevated levels of MHCII (Figures 4F and 4G). Although expression of the costimulatory molecule CD80 was unaffected by hemozoin abundance (Figures 4H and 4I), CD86 expression was significantly increased on CD11b^negCD8α⁺ cDC1 recovered from Δpm4Δbp2-infected mice (Figures 4J and 4K). Relative to cells from Δlap-infected mice, similar frequencies of cDC1 from Δpm4Δbp2-infected mice exhibited the capacity to phagocytose fluorescent carboxylate beads and OVA protein, but the quantities of each phagocytosed reagent (gMFI) were significantly elevated (Figures 4L-4O). Enhanced cDC1 function was also observed in antigen processing, as CD11b^negCD8α⁺ cDC1 DCs recovered from Δpm4Δbp2-infected mice exhibited more efficient ovalbumin (OVA) protein cleavage and endosomal accumulation, compared to cDC1 recovered from Δlap-infected mice (Figures 4P and 4Q). Compromises in cDC antigen accumulation and processing and MHCII and costimulatory receptor expression were further magnified by the second week of primary infection in Δlap-infected mice (Figures S3B-S3I). Of note, CD11b⁺CD8α^neg cDC2 functions were also impacted by the abundance of hemozoin, although these impacts were more modest (Figures S3J-S3L). Finally, and despite the caveats associated with dose, schedule, and formulation, we also found that direct injection of synthetic hemozoin into Δpm4Δbp2-infected mice reversed the enhanced cDC1 abundance (Figures S4A-S4C), MHC II expression (Figures S4D and S4E), and antigen processing (Figures S4F and S4G). These data phenocopy parameters observed in Δlap-infected mice and serve as an additional complementary approach to show that hemozoin impairs cDC1 functions during Plasmodium infection.

Figure 4. Hemozoin limits dendritic cell numbers and functions.

C57BL/6 mice were infected with either Δpm4Δbp2 or Δlap parasites, spleens were harvested on day 4 p.i., and DCs were liberated by digestion as described in STAR Methods.

(A–E) Frequency and number of CD11c⁺MHCII⁺B220^neg conventional DCs (A and B) and CD11b^negCD8α⁺ cDC1 (C–E).

(F and G) Representative flow plots (F) and summary of MHC II expression (gMFI) (G) on CD11b^negCD8α⁺ cDC1.

(H–L) Representative flow plots and summary data of CD80 (H and I) and CD86 (J and K) expression on CD11b^negCD8α⁺ cDCs.

(L and M) Representative flow plots (L) and quantification of bead⁺CD11b^negCD8α⁺ cDCs (left) and fluorescence intensity of phagocytosed microbeads (gMFI) (right) (M).

(N and O) Representative flow plots of OVA-488 uptake (N), frequency of OVA⁺CD8α⁺ cDC (O, left), and fluorescence intensity of phagocytosed OVA (gMFI) (O, right).

(P and Q) Flow plots and summary data of DQ-OVA protein processing assay. Data (mean ± SD) in all experiments are representative of two independent experiments with n = 5 (see also Figure S3).

Hemozoin limits early CD4 T cell expansion and differentiation

Given the changes we observed in cDC number, phenotype, and function, we further tested whether hemozoin impaired the capacity of cDC to prime Plasmodium-specific CD4 T cells. PbTII activation and division, as evidenced by CTV dilution, was first evaluated on day 4 p.i., which corresponds to a time point matching our kinetic analyses of impaired cDC function (Figure 4) and the reported development of cDC immunoparalysis (Lundie et al., 2010). Consistent with our cDC phenotypic and functional analyses, reduced levels of hemozoin in Δpm4Δbp2-infected mice increased the capacity of antigen-presenting cells (APCs) to support parasite-specific T cell proliferation (Figure 5A). There were nearly 20% increases in the proportion of PbTII that proliferated and nearly 50% increases in the proportion of divided PbTII cells that differentiated into CD150^loCXCR5^intT_FH precursor cells (Figures 5B and 5C). APC:CD4 T cell interactions are also crucial for helper cell migration into the GC, and this transition is regulated in part by either the strength of TCR stimulation (Benson et al., 2015; Goenka et al., 2011; Moran et al., 2011; Zemmour et al., 2018) or dwell time with APC (González et al., 2005; Govern et al., 2010). To investigate these aspects of cDC function, we transferred Nur77-GFP PbTII reporter cells to both Δpm4Δbp2 and Δlap-infected mice and found that reduced levels of hemozoin in Δpm4Δbp2-infected mice significantly enhanced TCR stimulation strength (Figures 5D and 5E), which likely contributes to the expanded number and proportion of T_FH cells in Δpm4Δbp2-infected mice (Figure 3 and Figures 5B and 5C). In contrast to day 14 p.i. CXCR5^hiPD-1^hi mature GC-T_FH (Figure 3I), we found elevated Bcl-6 expression among CD150^loCXCR5^int T_FH precursor cells on days 7 and 10 p.i. (Figure 5F) in mice exposed to reduced levels of hemozoin. Importantly, the skewing of CD4 T cells toward a GC T_FH fate in Δpm4Δbp2-infected mice was cDC intrinsic, as ex vivo co-culture of GP_66–77 peptide-pulsed cDCs isolated from Δpm4Δbp2-infected mice resulted in both enhanced proliferation (Figures 5G and 5H) and trending upregulation of PD-1 and CXCR5 by GP_66–77-specific TCR transgenic SMARTA CD4 T cells (Figures 5I and 5J). Notably, and similar to the cDC datasets, injection of synthetic hemozoin into Δpm4Δbp2-infected mice reversed the enhanced parasite-specific T_FH differentiation phenotypes (Figures S4H and S4I). Taken together, these data support that cDC are numerically and functionally compromised by hemozoin and this alters their capacity to optimally induce CD4 T cell responses and T_FH differentiation during blood-stage Plasmodium infection.

Figure 5. Optimal T_FH expansion is impaired by hemozoin.

CTV-labeled PbTII cells were transferred into mice 1 day prior to infection with either Δpm4Δbp2 or Δlap-infected parasites and harvested 4 days later.

(A) Representative histograms and quantification of CTV dilution among PbTII cells.

(B and C) Flow plots (B) and summary (C) of CD150 and CXCR5 expression by CTV diluted PbTII.

(D) gMFI of Nur77 in CD150^loCXCR5^intT_FH and CD150^hiCXCR5^loT_H1 PbTII CD4⁺ T cells on day 7 p.i. with light blue corresponding to increased gMFI. Numbers on flow plots are gMFI.

(E) Summary of Nur77 gMFI among CD150^loCXCR5^intT_FH

(F) Normalized Bcl-6 expression by PbTII T_FH collected on days 7 and 10 p.i. with Δpm4Δbp2 or Δlap. CTV-labeled SMARTA cells were cocultured for 4 days with GP-66 pulsed cDCs isolated from either Δpm4Δbp2 or Δlap-infected mice.

(G and H) Histogram of CTV dilution (G) and summary (H) data collected from SMARTA cells on d4 of in vitro culture.

(I and J) Flow plots (I) and summary (J) of PD-1 and CXCR5 expression by CTV diluted SMARTA cells.

Data (mean ± SD) are representative of two independent experiments with n = 4–5 (see also Figure S4).

NLRP3 activation prevents the development of optimal T_FH and humoral responses

To interrogate the mechanisms by which hemozoin erodes cDC numbers and function, with subsequent impacts in T_FH cell development and humoral immunity, we focused on the NLRP3 inflammasome. In vitro data support that hemozoin engages and activates the NLRP3 inflammasome (Kalantari et al., 2014; Shio et al., 2009; Strangward et al., 2018), and we found that serum IL-1β levels were on average 75% lower in mice infected with Δpm4Δbp2 parasites, compared to mice infected with Δlap parasites (Figure S5A). Thus, we next tested whether the enhanced T_FH cell differentiation phenotypes we observe in Δpm4Δbp2-infected mice were formally linked to NLRP3 inflammasome activation. To do this, we first transferred WT PbTII CD4 T cells to both C57BL/6 (WT) and NLRP3 deficient (KO) mice and subsequently infected the recipients with either Δlap or Δpm4Δbp2 parasites. Consistent with our previous observations (Figure 3), we found reduced frequencies of both PbTII CD150^loCXCR5^int T_FH precursors and CXCR5^hiPD-1^hi mature GC-T_FH populations in Δlap-infected WT mice, compared to Δpm4Δbp2-infected WT mice (Figure 6A, left panels). Strikingly, PbTII T_FH and GC-T_FH development in both Δlap-infected NLRP3 KO mice and WT mice treated from days 0 to 6 p.i. with the NLRP3 inflammasome inhibitor MCC950 phenocopied PbTII T_FH differentiation in Δpm4Δbp2-infected WT mice (Figure 6A, right panels). NLRP3 activation also impaired the capacity of GC-T_FH to promote GC B cell responses, as mature GC-T_FH cells sort-purified from Δlap-infected NLRP3 KO mice induced 2-fold greater numbers of GC B cell in ex vivo co-culture assays, compared to GC-T_FH cells recovered from Δlap-infected WT mice (Figure 6B). In addition to the deleterious impact of NLPR3 activation on T_FH differentiation and function, the frequency and number of splenic GC B cells (Figures S5B and S5C) and CD95^negCD73⁺CD80⁺PD-L2⁺ MBC (Figures 6C and 6D) were significantly decreased by either the presence of hemozoin or NLRP3 inflammasome activity.

Figure 6. Hemozoin-mediated impairments in humoral immunity are governed by NLRP3.

PbTII cells were transferred to C57BL/6 and NLRP3^−/− mice 1 day prior to infection with either Δpm4Δbp2 or Δlap parasites. Splenic GC responses were evaluated on day 14 p.i.

(A) Representative flow plots of CD150^loCXCR5^int T_FH and CXCR5^hiPD-1^hi GC-T_FH populations.

(B) CXCR5^hiPD-1^hi GC-T_FH cells were sort purified on day 14 p.i., co-cultured with naive B cells, and the frequency of CD95⁺GL-7₊ GC B cells was quantified after 5 days.

(C and D) Representative flow plots (C) and frequency (D) of CD80⁺PD-L2⁺CD73⁺CD38⁺CD95^neg MBC on day 28 p.i.

(E) C57BL/6 and NLRP3^−/− mice were immunized with Δpm4Δbp2 and Δlap parasites 60 days prior to challenge with a virulent Pb-ANKA strain.

(F) Frequency of immune mice that survived virulent Pb-ANKA challenge.

(G) MSP1_1–19-specific IgG endpoint titers 1 day prior to (pre) and 5 days after (post) virulent Pb-ANKA challenge.

Data (mean ± SD) in (A)–(D) are representative of three experiments with n = 3–4. Data (mean ± SD) in (E)–(G) are representative of two independent experiments with n = 3–4 per experimental group (see also Figure S5).

To directly address whether hemozoin-mediated NLRP3 activation compromises long-term protective immunity in vivo, groups of both WT and NLRP3 KO mice were immunized with either Δpm4Δbp2 or Δlap parasites and then challenged with WT Pb-ANKA after 60 days (Figure 6E). Both WT and NLRP3 KO mice immunized with Δpm4Δbp2 exhibited 100% survival and reduced parasite burden (Figures 6F, S5D, and S5E) through day 35 p.i. However, more than 75% of Δlap-immune NLRP3 KO mice survived, compared to only 30% of Δlap-immune WT mice (Figure 6F). We also found functional MBC recall antibody responses were substantially improved in Δpm4Δbp2-immune WT mice and Δlap-immune NLRP3 KO mice, as evidenced by 38- and 169-fold boosts in parasite-specific IgG titers 5 days after Pb-ANKA challenge, respectively (Figure 6G), in contrast to the modest 7-fold boost in Δlap-immune WT mice. Collectively, these experiments show that the infection of NLRP3-deficient mice with Δlap parasites effectively phenocopies elevated GC-T_FH and MBC numbers and functions observed following infection of WT mice with Δpm4Δbp2 parasites, further supporting that hemozoin-mediated NLRP3 activation impairs protective anti-Plasmodium humoral immunity.

Cell-intrinsic NLRP3 activation impairs cDC number and function

Our genetic and pharmacological approaches support that hemozoin-mediated NLRP3 inflammasome activation constrains T_FH dependent humoral immunity. To determine whether these phenotypes are mechanistically linked to cDC-intrinsic NLRP3 activity and confirm that our findings extend beyond study of the mutant Pb ANKA parasites, we first evaluated cDC numbers and functions and T_FH development in WT and NLRP3 deficient mice infected with Plasmodium yoelii parasites (Figure 7A). We observed that the total numbers of CD11b^negCD8α⁺ cDC1 (Figures 7B and 7C), and their relative MHC II expression (Figure 7D) were significantly increased in the absence of NLRP3 activation. Moreover, the capacity of cDC1 from P. yoelii-infected mice to both phagocytose multiple carboxylate beads (gMFI) (Figure 7E) and functionally process and accumulate cleaved OVA protein (Figures 7F and 7G) was markedly enhanced in the absence of hemozoin-mediated NLRP3 inflammasome activation, which closely resembles our findings with the mutant Pb ANKA parasites (Figure 4). Both PbTII and polyclonal CD4 T cell responses were also skewed toward a more robust T_FH phenotype in P. yoelii-infected NLRP3-deficient mice, compared to P. yoelii-infected WT mice (Figures 7H and 7I). Importantly, P. yoelii-immune NLRP3-deficient mice were substantially more resistant to challenge with WT Pb ANKA (Figures S6A-S6C), phenotypes that additionally associated with enhanced MBC recall function (Figure S6D). Together, these data provide additional support that NLRP3 activation compromises cDC function and results in suboptimal T_FH responses and protective humoral immunity.

Figure 7. Hemozoin impairs dendritic functions via cell-intrinsic NLRP3 activity.

(A) Experimental design for samples depicted in (B)–(I) from C57BL/6 (WT) and NLRP3^−/− (KO) mice infected with P. yoelii 17XNL parasites.

(B and C) Summary CD11b^negCD8α⁺ cDC1 frequency (B) and number (C) in WT and NLRP3 KO mice.

(D) Histogram and summary of MHC II expression in CD8α⁺ cDC1.

(E) Histograms of bead uptake, summary graph of the frequency of bead positive CD8α⁺ cDC1 (center), and gMFI of phagocytosed beads (right).

(F and G) Flow plots (F) and summary (G) of DQ-OVA antigen processing assay. The frequency of CD11b^negCD8α⁺ cDC1 in each quadrant is depicted.

(H and I) Flow plots (H) and number (I) of CD150^loCXCR5^int T_FH among parasite-specific (left) and polyclonal (right) CD4 T cells on day 6 p.i.

(J) Experimental schematic for C57BL/6:NLRP3^−/− (50:50) bone-marrow chimera generation.

(K) Ratio of CD11b⁺CD8α^neg cDC2 and CD11b^negCD8α⁺ cDC1 recovered from chimeras on d11 p.i.

(L and M) Histograms and quantification of MHC II expression by total cDC and CD11b^negCD8α⁺ cDC1.

(N) Flow plots of bead uptake (left), summarized frequency of fluorescent bead⁺ CD11b^negCD8α⁺ cDC1 (center), and gMFI of fluorescent beads are quantified (right).

(O) DQ-OVA antigen-processing assay.

(P) The frequency of CD11b^negCD8α⁺ cDC1 in each quadrant is quantified.

(Q) C57BL/6 (WT) mice were administered either MCC950 or PBS i.p. during the first 6 days of infection, and bone marrow was collected on day 28 for quantification of MSP₁₉ LLPC via ELISPOT.

(R and S) Representative ELISPOT photos (R) and summary data (S).

Data (mean ± SD) in (A)–(I) and (Q)–(S) are representative of two independent experiments with n = 5. Data (mean ± SD) in (K)–(P) are the summary data collected from two independent experiments (see also Figure S6).

Finally, to compare WT and NLRP3 KO cDC1 development and function in the same Plasmodium-infected host, we generated and infected WT:NLRP3 KO (50:50) mixed bone-marrow chimeras with Δlap parasites (Figure 7J). Importantly, this approach controls for potential differences in parasite load, antigen, and inflammation in Plasmodium-infected WT and KO mice. Although only modest shifts in cDC1:cDC2 ratios were observed among NLRP3 KO cells (Figure 7K), we observed that NLRP3 KO cells exhibited significantly elevated MHC II expression among both total cDC and CD11b^negCD8α⁺ cDC1 populations (Figures 7L and 7M), compared to WT cDC and cDC1 populations. These data provide direct in vivo evidence that hemozoin-mediated cDC-intrinsic NLRP3 activation, rather than extrinsic and environmental cues that include inflammatory cytokines, such as IL-1β, regulate the expression of MHC II. Consistent with these findings, NLRP3 KO CD11b^negCD8α⁺ cDC1 exhibited significantly enhanced phagocytotic (Figure 7N) and protein processing functions (Figures 7O and 7P), compared to WT cDC1 recovered from the same Plasmodium infection environment. Last, to complement these genetic studies and confirm that early NLRP3 activation leads to impaired humoral immunity, we treated Δlap-infected WT mice with MCC950 during the initial stages of DC activation and CD4 T cell priming (days 0 to 6 p.i.) and measured the development of MSP1₁₉-specific plasma cells. Pharmacologic inhibition of NLRP3 inflammasome activity resulted in 2-fold greater numbers of parasite-specific plasma cells by day 28 p.i. (Figures 7Q-7S). Altogether, our data provide direct in vivo evidence that hemozoin-mediated inflammasome activation during an initial blood-stage Plasmodium exposure limits cDC1 numbers and functions, which in turn impairs the development of protective anti-Plasmodium cellular and humoral immunity.

DISCUSSION

The adaptive immune response elicited by Plasmodium infection is robust, but the resilience of humoral memory populations is suboptimal. Waning immune-mediated resistance to malarial disease in individuals who emigrate from endemic areas and the failure of vaccines in endemic populations highlight that new mechanistic understanding of why humoral responses following Plasmodium exposure are suboptimal remains a priority. Our study revealed that these defects may be programmed early after an initial Plasmodium exposure via hemozoin-mediated, NLRP3 inflammasome-dependent impacts on the cDC compartment with significant reductions in phagocytosis, antigen processing, and MHC II expression occurring as early as day 4 p.i. We further linked these cDC re-programming events to impairments in the stepwise differentiation of T_FH, which resulted in reduced GC B cell responses and lower numbers of MBC and LLPC capable of producing protective, anti-parasite antibodies. Further investigation supported that humoral memory is suppressed by DC-intrinsic inflammasome activation. Importantly, hemozoin-induced NLRP3 activation limited humoral immunity irrespective of the strain of parasites (P. yoelii or P. berghei ANKA) used to initiate cDC reprogramming, suggesting that sustained hemozoin-induced compromise of CD8α⁺ cDC numbers and function may be a conserved pathophysiologic feature of blood-stage Plasmodium infection.

Despite the genetic and life-cycle complexity of Plasmodium parasites, hemozoin is currently the only described malaria-associated NLRP3 inflammasome agonist. The capacity for inflammasome activation to alter cDC gene expression and phenotype has been evaluated ex vivo, but studies designed to evaluate the contribution of DC-intrinsic inflammasome activation during vaccination and infection are limited. A recent report showed that inflammasome activation during Salmonella infection is detrimental to T cell priming and that reduced levels of inflammasome machinery limit splenic DCs from undergoing inflammasome-mediated pyroptosis (McDaniel et al., 2020). It was further reported that cDC activation by inflammasomes rarely occurs early during infection because of IRF4 and IRF8-mediated suppression of inflammasome genes, but this observation is not consistent with those from the Plasmodium field where robust expression of inflammasome genes and caspase 1 activation are observed following both rodent and human infection (Ataide et al., 2014). Indeed, the magnitude of inflammasome activation that occurs during the blood stage of Plasmodium infection is supported by high levels of serum IL-1β, which contribute to malarial symptoms that include fever, chills, and malaise. Deleterious functions for hemozoin during malaria are also suggested by strong correlations between hemozoin accumulation and severe clinical manifestations and inflammation in the central nervous system and placenta, where increased accumulation of monocytes and platelets can exacerbate disease (Hochman et al., 2016; Milner et al., 2013; Moore et al., 2004; Nguyen et al., 1995; Sarr et al., 2006). We found that reducing cDC exposure to hemozoin limited splenic IL-1β levels by day 10 p.i. and significantly increased DC maturation kinetics and function. Our in vivo studies showed that hemozoin-induced, NLRP3-dependent inflammasome activation impaired cDC phagocytosis, antigen processing, and expression of MHC II. Reported discrepancies between experimental studies of ex vivo and in vivo contributions of inflammasome activation and resultant programming of DC fate and function may be linked to the severity of malaria or virulence of specific species of Plasmodium parasites, as well as the tissue-specific deposition and abundance of hemozoin. Our complementary genetic, biochemical, and pharmacologic studies support that NLRP3 activation exerts a similarly deleterious impact on humoral immunity following either P. yoelii or P. berghei ANKA primary infections.

The functional defects we observed in cDC following hemozoin-mediated NLRP3 inflammasome activation were further linked to negative impacts on the Plasmodium-specific T cell response. T cell priming, the emergence of CD150^loCXCR5^intT_FH precursors, and the eventual differentiation of mature CXCR5^hiPD-1^hiGC-T_FH cells were all reduced by hemozoin-mediated NLRP3 activation in cDC. Notably, alum is a well-documented NLRP3 agonist and adjuvant reported to promote the initiation of T cell and humoral responses. Alum’s capacity to dissociate from vaccine components upon injection differs significantly from that of hemozoin, which is retained within the DCs. This difference may serve as one explanation for disparities in reported outcomes following examination of NLRP3-mediated DC function. Our high-resolution analyses revealed that, in the absence of NLRP3 activation, parasite-specific T cells exhibited evidence of enhanced TCR stimulation, which is reported to support increased CXCR5 expression and reinforce T_FH differentiation. Whether the presence of hemozoin and altered DC programming impairs a limited number of GC-T_FH functions or reduces the expression of multiple co-stimulatory molecules and cytokines is not currently known. Elucidation of the precise mechanisms by which hemozoin limits the capacity of cDC to promote the development of optimal GC-T_FH function remains of interest and will also be relevant to our understanding of how Plasmodium infection influences responses to either heterologous infection or vaccination.

Our results support the hypothesis that hemozoin and the associated programming of DC number and function contribute to suboptimal humoral responses following Plasmodium infection. During a blood-stage Plasmodium infection with reduced levels of hemozoin, enhancements in humoral immunity begin early in the spleen, where we observed a coordinated shift in CXCR4⁺CD138⁺ LLPC precursors and accumulation of mature LLPC in the bone marrow. Our data support a model wherein the abundance of LLPC or reactivation of memory B cells mediate protection from lethal parasite challenge, as opposed to either qualitative shifts in either the isotype or affinity of parasite-specific antibodies in mice with reduced hemozoin exposure. Limitations of our study include the sensitivity of assays used to probe the quantity, quality, and functions of secreted antibody responses, particularly in our studies showing enhanced survival yet equivalent immune memory baseline titers in NLRP3 KO mice and WT mice infected with Δpm4Δbp2 parasites. Future molecular and biophysical studies designed to interrogate critical qualitative attributes and effector functions of the secreted antibody responses are warranted. Another notable aspect of our studies that we did not assess is whether hemozoin accumulation within the bone marrow impacts or erodes the maintenance of either vaccine-induced or pathogen-specific LLPC. The bone marrow serves as an infection reservoir for P. vivax parasites, and the presence of hemozoin in the bone marrow pushes the cytokine environment toward a more proinflammatory state (Lee et al., 2017). Understanding whether either inflammation or cytokine and chemokine perturbations contribute to reductions in the frequency of mature LLPC following infection with Plasmodium may shed light onto additional factors that restrain the quality and/or durability of anti-Plasmodium humoral responses.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Noah Butler (noah-butler@uiowa.edu).

Materials availability

• This study did not generate unique reagents.

Data and code availability

• All data reported in this paper will be shared by the lead contact upon request.

• This study did not generate/analyze datasets or code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

C57BL/6 wild-type (CD45.2 and CD45.1), C57BL/6-Tg(Nr4a1-EGFP/cre)820Khog/J (Nur77^GFP Stock No. 016617) and Nlrp3^tm1Bhk (NLRP3 KO stock No. 021302) were either purchased from The Jackson Laboratory or acquired from colonies at the University of Iowa. For the generation of WT: NLRP3^−/− chimeras, CD45.1 wild-type recipient mice were irradiated with 475 rads twice, separated by 4 h. Bone marrow cells from CD45.1 (C57BL/6) and CD45.2 (NLRP3 KO) were mixed 50:50 and 1 × 10⁷ were injected i.v. Mice were maintained on a Uniprim diet (Envigo) for 2 weeks. Chimerism was assessed at 8 weeks post-irradiation. Age and sex-matched mice between 6 and 13 weeks were used for all experiments. Ethics oversight for animal use was provided by The University of Iowa IACUC.

Parasite strains and infections

Two wild-type P. berghei ANKA parasite lines were used: the Pb-ANKA line, obtained from MR4 (American Type Culture Collection) and the reporter, reference P. berghei ANKA line 676m1cl1 (Pb-ANKA-Luc), which constitutively express GFP and luciferase that was obtained through BEI Resources (MRA-868) and was contributed by Chris J. Janse and Andrew P. Waters. The Plasmodium berghei ANKA mutants Δpm4Δbp2 and Δlap have been described in Lin et al. (2015) and were obtained from Dr. Chris Janse (Leiden University Medical Center, LUMC, the Netherlands). P. yoelii (clone 17XNL) were obtained from the Malaria Research and Reference Reagent Resource Center (MR4, American Type Culture Collection). All parasite lines were maintained by passage in NIH Swiss Webster mice. Primary infections were initiated by an intravenous (IV) transfer of 1 *× 10⁶ parasitized red blood cells (RBCs) derived from an infected donor mouse. Challenge of mice with virulent Pb-ANKA or Pb-ANKA-luc were initiated by intravenous (IV) injection of 5 × 10⁶ donor-derived parasitized RBCs.

Quantification of parasite biomass

Percentage of infected RBC was monitored in tail blood samples by flow cytometric analysis of the frequency of Hoechst⁺ Diethydiumbromide⁺Ter119⁺CD45^neg cells (Malleret et al., 2011). Parasite loads, quantified as total flux (p/s/cm²/sr), following IP injection of 250 μg of D-luciferin (Promega) were determined using an IVIS Lumina S5 and Living Image software (Perkin Elmer).

Treatments

Synthetic hemozoin (sHZ) (InvivoGen; 750 μg in PBS) was injected every second day IV (tail vein). MCC950 (Sigma, 250 μg in PBS) was administered via intraperitoneal (IP) injection every 2 days beginning at the time of infection.

METHOD DETAILS

Immune cell phenotyping

For analysis of splenic cells, single cell suspensions were generated by forcing mouse spleens through a 70 μm mesh. For experiments enumerating cDC subsets, spleens were incubated with collagenase IV (100 U/mL) and DNase I (10 μg/mL) for 30 mins at 37°C prior to preparation of single cell suspensions. RBC were lysed prior to Fc receptor blockade (CD16/32-clone 2.4G2) in FACS buffer (PBS+ 0.09% sodium azide + 2% FCS) for 15 min at 4°C prior to surface staining. For identifying T_FH like cells, cells were stained with purified rat anti-mouse CXCR5 (BD Biosciences) in T_FH staining buffer (PBS+ 0.09% sodium azide + 2% FCS+ 0.05% BSA+ 2% mouse serum) for 1h at 4 °C. Cells were washed then stained with biotin-conjugated AffiniPure Goat anti-rat IgG (H+L) (Jackson Immunoresearch) in T_FH buffer for 30 min. Cells were then washed and surfaced stained in an antibody cocktail containing (Anti-CD4, -PD-1, -CD150, -CD11a, -CD44, streptavidin) for at least 30 min at 4°C. Anti-Bcl-6, -T-bet and -Foxp3 (BioLegend) transcription factor staining was performed using the Foxp3 Staining Kit per manufacturer’s protocol (BD Biosciences). For B cell staining, cells were surface stained in FACS buffer for at least 30 min in the following cocktail solution containing anti-CD19, -B220, -CD267, -CD138, -CD38 (BioLegend), -CD95 (BD PharMingen), -IgD and -CXCR4 (BioLegend) to identify plasma cells, while a cocktail of anti-CD19, -B220, -CD38, -CD80, -CD273, -CD73, -IgD, -CCR6 (BioLegend) was used to identify memory B cells. cDCs were identified using a surface stain mix containing Anti-CD8, -F4/80, -CD80, -CD86, -CD19, -B220, -CD11c (BioLegend) and anti-CD11b and MHC II (Thermofisher) following Fc-block for 10 mins at 4°C.

Standard and antibody affinity ELISA

For standard ELISA, Nunc Immunosorb Plates were coated with 1 μg/mL of recombinant MSP19 overnight at 4°C. Plates were blocked for 2h at RT with PBS+2.5%BSA+ 2%FCS prior to addition of serially diluted sera and incubation overnight at 4°C. Wells were washed with PBS + 0.05% Tween20 prior to incubation with HRP conjugated anti-mouse IgG for 2h at RT. For affinity ELISA, after plates were coated overnight as described above, plates were washed with PBS-Tween prior to blocking with PBS+1% BSA+0.05% Tween for 1h at 37 °C. Diluted serum was incubated for 2h at RT in buffer (2.5% BSA+5% FCS in PBS). Wells were washed extensively with PBS+Tween20 prior to addition of diluted ammonium thiocyanate and incubation for 15 minutes. Wells were washed prior to incubation with HRP conjugated anti-mouse IgG for 1h at 37°C. All plates were reacted with SureBlue TMB Peroxidase and TMB Stop Solution (Sera Care) prior to absorbance measurement at an OD of 450 nm using a plate reader (Biotek). The concentration of serum IL-1β was assessed using an IL-1β LEGENDplex immunoassay (Biolegend) according to the manufacturer’s protocol.

ELISPOT

White plates (Nunc MaxiSorp) were coated with 1 μg/mL of recombinant MSP-19 diluted in 0.1M Na₂HPO₄ binding solution and incubated overnight at 4°C. Plates were washed with PBS and blocked for 2h at RT with 10% FCS in RPMI 1640. Bone marrow cells were diluted serially in 10% FCS in RPMI 1640 and incubated for 18 h at 37 °C with 5% CO₂. Plates were thoroughly washed then incubated with HRP-conjugated anti-mouse IgG in PBS+5% FCS overnight at 4 °C. Spots were then developed with 3-amino-9-ethylcarbazole.

IgG purification

The method for IgG purification from immune serum using Protein G Sepharose4 Fast Flow columns was adapted from manufacturer’s protocol and Akter et al. (Akter et al., 2019). Briefly, following column equilibration with binding buffer (20 mM sodium phosphate), diluted serum (1:4) was ran over column and flow-through was collected prior to complex elution. IgG was eluted from the column using 0.1M glycine+0.01% sodium azide and the pH was neutralized using 1M Tris-HCl. Amicon Ultra-4 Membrane Units were used for buffer exchange to PBS.

CTV labeling of T cells

CTV labeled CD4+ T cells were prepared according to manufacturer’s protocol (Invitrogen) with minor protocol modification. Splenocytes were washed and suspended in serum free media supplemented 1:1000 with Cell Trace Violet. Cells were incubated at 37°C for 20 min prior to dilution with RPMI 1640+10% FCS and incubation for 5 mins. Cells were pelleted and incubated again for 5 mins with RPMI 1640+10% FCS at 37°C prior to surface staining.

Ex vivo cultures

GC-T_FH (CXCR5^hiPD-1^hi) were sort purified from Δpm4Δbp2- and Δlap-infected mice on day 14 p.i. and suspended in IMDM+10% FCS. Naive B cells, enriched from naive mice using a B cell Isolation Kit (Miltenyi Biotec), were mixed with GC-T_FH at a 1:1 ratio and co-incubated for 5 days at 37°C with 5% CO₂. Cell cultures were supplemented with anti-IgM (5 μg/mL) (Jackson Immunoresearch) and anti-CD3 (5 μg/mL) (Biolegend). For cDC-T cell cultures, magnetically enriched cDCs were pulsed with 1 nM of GP-66 for 1h, washed, and incubated at a 1:10 ratio with CTV-labeled SMARTA cells for 4 days.

Phagocytosis assays

To examine the phagocytic capacity of cDC, splenocyte suspensions were incubated with Fluoresbrite carboxylate microsphere YG (0.5 μm-Polysciences) at a ratio of 1:100 microspheres/cell or in 25 μg/mL solution OVA-488 (Molecular Probes) for 1h at 37°C. To assess antigen processing by cDC, splenocytes were incubated in a 25 μg/mL solution of DQ-OVA (Molecular Probes) for 1h at 37°C. Cells were then stained with Fc-block for 10 mins at 4°C prior to surface staining.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data centers in summary data graphs represent the mean. Data dispersion and tests of statistical significance are detailed in the figure legends. Graphs were generated and statistical analyses were performed using Prism Graphpad 9 software (GraphPad Software). For analyses comparing only two groups, Mann-Whitney non-parametric tests were used to determine p value. For comparison of 3 or more groups, Dunn’s test for nonparametric multiple comparisons were employed and adjusted P values are reported. Details can be found in respective figure legends.

KEY RESOURCES TABLE

Reagent or resource	Source	Identifier
Antibodies
Purified anti-mouse CD3	Biolegend	Cat#100202; RRID: AB_312659
Anti-mouse CD16/32 (Fc-Block)	BioXCell	Cat#BE0307, RRID: AB_2736987
Biotin SP-conjugated Affinipure Goat-anti-IgG (H+L) F(ab)	Jackson Immunoresearch	Cat#112-065-167, RRID: AB_2338179
Purified rat anti-mouse CXCR5	BD Biosciences	Cat#551961, RRID: AB_394302
Anti-mouse CXCR4-BV605 (clone: L276F12)	Biolegend	Cat#146519, RRID: AB_2687359
Anti-mouse CD150-PeCy7 (clone:TC15-12G12.2)	Biolegend	Cat#115914, RRID: AB_439797
Anti-mouse CD267-PE (clone 8F10)	Biolegend	Cat#133403, RRID: AB_2203542
Anti-mouse IgD-PeDazzle 594 (clone: 11-26c.2a)	Biolegend	Cat#405742, RRID: AB_2571985
Anti-mouse CD86-BV605 (clone:GL1)	Biolegend	Cat#105037, RRID: AB_11204429
Anti-mouse CD80-BV510 (clone:16-10A1)	Biolegend	Cat#104741, RRID: AB_2810337
Anti-mouse CD80-PerCPCy5.5 (clone:16-10A1)	Biolegend	Cat#104722, RRID: AB_2291392
Anti-mouse CD11c-PE Cy7 (clone:N418)	Biolegend	Cat#117318, RRID: AB_493568
Anti-mouse CD38-Pac Blue (clone:90)	Biolegend	Cat#102720, RRID: AB_10613468
Anti-mouse CD95-FITC (clone: Jo2)	BD Biosciences	Cat#554257, RRID: AB_395329
Anti-mouse MHCII-AF700 (clone:M5/114.15.2)	Biolegend	Cat#107622, RRID: AB_493727
Anti-mouse CCR6-AF647 (clone:29-2L17)	Biolegend	Cat#129808, RRID: AB_1227497
Anti-mouse CD138 Pac Blue (clone 281-2)	Biolegend	Cat#142523, RRID: AB_2565621
Anti-mouse B220-PerCP-Cy5.5 (clone:RA3-6B2)	Biolegend	Cat#103236, RRID: AB_893354
Anti-mouse B220-PE (clone:RA3-6B2)	Invitrogen	Cat#12-0452-83, RRID: AB_465672
Anti-mouse CD11b-APC (clone: M17/70)	Biolegend	Cat#101212, RRID: AB_312795
Anti-mouse CD4-PerCP-Cy5.5 (clone:GK1.5)	Biolegend	Cat#100434, RRID: AB_893324
Anti-mouse CD8-PerCP-Cy5.5 (clone: 53-6.7)	Biolegend	Cat#100734, RRID: AB_2075238
Anti-mouse CD19-AF700 (clone: 6D5)	Biolegend	Cat#115528, RRID: AB_493735
Anti-mouse CD45.2-PE (clone: 104)	Biolegend	Cat#109808, RRID: AB_313445
Anti-mouse CD273-PE (clone: TY25)	Biolegend	Cat#107206, RRID: AB_2162011
Anti-mouse F4/80-PE (clone: BMB)	Biolegend	Cat#123110, RRID: AB_893486
Anti-mouse Foxp3-PeCy7 (clone: 3G3)	Tonbo	Cat#60-5773-U100, RRID: AB_2621869
Anti-mouse Bcl-6-AF647 (clone IG191E/A8)	Biolegend	Cat#648305, RRID: AB_2565298
Anti-mouse Tbet-BV605 (clone: 4B10)	Biolegend	Cat#644817, RRID: AB_11219388
Anti-mouse CD95-PeCy7 (clone: Jo2)	BD PharMingen	Cat#557653, RRID: AB_396768
Anti-mouse CD73-FITC (clone: TY/11.8)	Biolegend	Cat#127220, RRID: AB_2716076
Anti-mouse PD-1-BV510 (clone: RPM1-30)	Biolegend	Cat#135241, RRID: AB_2715761
Anti-mouse PD-1-FITC (clone: RPM1-30)	Thermo Fisher Scientific	Cat#11998181, RRID: AB_465466
Anti-mouse CD45.1-PacBlue (clone: A20)	Biolegend	Cat#110722, RRID: AB_492866
Anti-mouse CD90.1-PeDazzle (clone: OX-7)	Biolegend	Cat#202524, RRID: AB_1595635
Anti-mouse CD44-AF700 (clone:IM7)	Biolegend	Cat#103026, RRID: AB_493713
Anti-mouse CD44-FITC (clone:IM7)	Biolegend	Cat#103006, RRID: AB_312957
Anti-mouse CD11a-FITC (clone:M17/4)	Biolegend	Cat#101106, RRID: AB_312779
Anti-mouse GL7-FITC (Clone:GL7)	Biolegend	Cat#144603, RRID: AB_2561696
Chemicals, peptides, and recombinant proteins
Hemozoin	InvivoGen	Cat#tlrl-hz
MCC950 (NLRP3 Inhibitor)	Sigma	Cat#256373-96-3
D-luciferin	Promega	Cat#P1043
Collagenase IV	GIBCO	Cat#17104019
DNase I	Roche	Cat#10104159001
Cell Trace Violet	Invitrogen	Cat#C34557
streptavidin-BV421	BD Biosciences	Cat#563259
streptavidin-APC	BD Biosciences	Cat#405207
Fluoresbrite carboxylate microsphere YG 0.5 μm	Polysciences	Cat#15700
OVA-488	Molecular Probes	Cat#O34781
DQ-OVA	Molecular Probes	Cat#D12053
B Cell Isolation Kit	Miltenyi Biotec	Cat#130-090-862
GP66-77 peptide	BioSynthesis Inc.	Cat#Custom
Recombinant MSP119 (yPyMSP119XL)	BEI Resources	Cat#MRA-48
Critical commercial assays
SureBlue TMB Peroxidase	Sera Care	Cat#5120-0077
TMB Stop Solution	Sera Care	Cat#5150-0021
LEGENDplex immunoassay	Biolegend	Cat#740150
Experimental models: Organisms/strains
C57BL/6J	Jackson Laboratories	Cat#RRID:IMSR_JAX:000664
C57BL/6-Tg(Nr4a1-EGFP/cre)820Khog/J	Jackson Laboratories	Cat#RRID:IMSR_JAX:016617
B6.129S6-Nlrp3tm1Bhk/J	Jackson Laboratories	Cat#RRID:IMSR_JAX:021302
B6.SJL-Ptprca Pepcb/BoyJ	Jackson Laboratories	Cat#RRID:IMSR_JAX:002014
B6.Cg-Ptprca Pepcb Tg(TcrLCMV)1Aox/PpmJ	Jackson Laboratories	Cat#RRID:IMSR_JAX:030450
P. berghei ANKA	BEI-MR4	Cat#MRA-871
P. berghei ANKA line 676m1cl1	BEI-MR4	Cat#MRA-868
P. yoelii (clone 17XNL)	BEI-MR4	Cat#MRA-593
P. berghei ANKA Δpm4Δbp2 and Δlap	A gift from Dr. Chris Janse (Leiden University)	https://rupress.org/jem/article/212/6/893/41818/Replication-of-Plasmodium-in-reticulocytes-can
Software and algorithms
Flowjo v10.5.0	Treestar	https://www.flowjo.com/solutions/flowjo/downloads
Prism v9	Graphpad	https://www.graphpad.com
Excel v16.16.20 9	Microsoft	https://www.microsoft.com/en-us/microsoft-365/excel
Living Image	Perkin Elmer	https://www.perkinelmer.com/product/spectrum-200-living-image-v4series-1-128113
Other
Biotek plate reader	Biotek	N/A
BD LSR II	BD	N/A
BD FACSVerse	BD	N/A
IVIS Lumina S5	Perkin Elmer	N/A

Highlights.

• Hemozoin engages NLRP3 to reduce CD8α⁺ dendritic cell (DC) number and function

• Impaired DC responses compromise the B cell helper functions of CD4⁺ T cells

• The accumulation of hemozoin reduces memory B cell and plasma cell responses

• NLRP3 deficiency boosts humoral immune memory-mediated protection