Chronic HIV Infection Impairs Nonopsonic Phagocytosis of Malaria Parasites

Abstract

Malaria-specific immune responses are altered in HIV/malaria-coinfected individuals and are associated with higher parasite burdens and more severe clinical disease. Monocyte/macrophage phagocytosis is a major mechanism of malaria parasite clearance. We hypothesized that phagocytosis of malaria-parasitized erythrocytes is impaired in coinfected individuals and could contribute to the increased parasite burdens observed. We show that nonopsonic phagocytosis of Plasmodium falciparum parasitized erythrocytes is impaired in monocytes isolated from HIV-infected individuals. The observed defects in phagocytic capacity were rescued after 6 months of antiretroviral therapy, demonstrating the importance of HIV treatment and immune reconstitution in the context of coinfection.

INTRODUCTION

Malaria and HIV coinfection is a global health priority. Mounting evidence indicates that HIV and malaria negatively interact. Malaria increases HIV replication; conversely, HIV increases the risk of clinical and severe malaria.¹ One manifestation of increased disease severity in coinfected individuals is a higher malaria parasite burden. Impaired antiparasitic antibody and phagocytic responses, and altered malaria-specific cytokine responses may contribute to the higher parasitemia levels.

Monocytic cells are key effectors of cell-mediated immunity. During malaria infection, they produce inflammatory cytokines to stimulate innate and adaptive immune responses, and they phagocytose parasitized erythrocytes (PEs) to decrease parasite burden. Opsonin-dependent clearance of PEs occurs through Fcγ receptor (FcγR),² whereas opsonin-independent phagocytosis of PEs occurs primarily through the pattern-recognition receptor CD36.³

During HIV infection, monocyte/macrophage functions become impaired.⁴ Macrophages from HIV-infected (HIV+) individuals and in vitro HIV-infected monocyte-derived macrophages are less efficient for both FcγR-mediated and complement-mediated phagocytosis, although conflicting data exist. Evidence of impaired nonopsonic phagocytosis has also been reported for some organisms, for example, Aspergillus spp and Pneumocystis spp, but not others, for example, Salmonella typhimurium.⁴ Phagocytic impairment may be reversible with combination antiretroviral therapy (cART).⁵

Most studies on the impact of HIV on malaria parasite clearance have focused on opsonic phagocytosis. HIV infection was associated with lower opsonizing antibodies for pregnancy-associated parasite variant antigens.^2,6 Additionally, in vitro infection of human monocyte-derived macrophages with HIV resulted in reduced phagocytic uptake of opsonized PEs.⁷ These lines of evidence suggest that HIV impairs both the acquisition of opsonizing antibodies against PEs and the phagocytic ability of monocyte/macrophages for opsonized PEs. Since opsonic phagocytic responses are compromised in the presence of HIV, nonopsonic clearance of PEs might be expected to play an increased role in HIV+ individuals. However, the impact of HIV infection on nonopsonic phagocytosis of PEs is unknown.

Our aim was to elucidate the impact of HIV infection on nonopsonic phagocytosis of PEs and assess the effect of cART initiation on this process. We hypothesized that in treatment-naive HIV+ individuals, nonopsonic phagocytosis of PEs would be compromised but would recover with cART.

MATERIALS AND METHODS

Study Population and Ethics Statement

University Health Network and University of Toronto Ethics Review Boards approved this study. Study participants gave written informed consent. HIV+ participants were chronic progressors (HIV infected for >1 year, CD4⁺ T-cell count decline of >50 cells·mm⁻³·y⁻¹) as previously described.⁸ Blood was collected before initiation of cART (M0), at 3 months (M3), and at 6 months (M6) post-cART. CD4⁺ T-cell counts, viral load, and drug regimen are shown in Table S1 (see Supplemental Digital Content, https://links.lww.com/QAI/A594). HIV-uninfected (HIV−) donors were recruited in the same demographic area and with similar age and sex profile.⁸ When possible, each HIV+ donor was matched to the same HIV− individual for all 3 time points.

Plasmodium falciparum Culture

The laboratory line ITG was cultured in vitro as described.⁹ Cultures were routinely treated with mycoplasma removal agent (ICN-VWR, Mississauga, ON, Canada) and tested negative for mycoplasma by polymerase chain reaction.

Phagocytosis Assays

Peripheral blood mononuclear cells (PBMCs) were purified from venous blood using Ficoll gradients. Monocytes were purified by adherence to glass coverslips in 24-well plates (1.5 × 10⁶ PBMCs/well) and cultured for 3 days in RPMI-1640 with 10% FBS, MEM nonessential amino acids, sodium pyruvate, β-mercaptoethanol, and gentamicin (Invitrogen, Life Technologies, Burlington, ON, Canada) before use in phagocytosis assays. Each HIV+ participant was matched with an HIV− donor for each phagocytosis experiment.

For nonopsonic phagocytosis, synchronized schizont-stage PEs in 500 μL of RPMI-10% FBS with L-glutamine were added to monocytes at a target:effector ratio of 20:1. For opsonic phagocytosis assays, human RBCs were opsonized using a rabbit antihuman RBC IgG antibody (MP Biomedicals/Cappell, Solon, OH; total protein 4.7 μg/mL, hemagglutinin titer 1:6400) for 1 hour at room temperature. Opsonized RBCs were added to monocytes at a target:effector ratio of 20:1. Plates were rotated gently for 2 hours (nonopsonic assay) or 15 minutes (opsonic assay) at 37°C, 5% CO₂. Coverslips were washed in ice-cold distilled water for 30 seconds to lyse and remove bound but not internalized PEs/RBCs, and fixed and stained using Diff-Quick. Phagocytosis was quantified microscopically, counting 500 monocytes per coverslip. Phagocytic index, defined as the total number of internalized PEs/RBCs per total monocytes counted expressed as a percentage, was calculated. Phagocytic index was used to assess correlation between phagocytic uptake and CD36 surface expression levels. To assess differences between HIV+ and uninfected controls, the phagocytic index for each HIV+ donor was normalized to their respective control.

Multicolor Flow Cytometry

To assess monocyte CD36 expression, freshly isolated PBMCs were stained with fluorophore-conjugated monoclonal antibodies to CD14, CD3, and CD36 (BioLegend, San Diego, CA). Analysis was done on a portion of samples due to sample limitations. Cells were analyzed within 24 hours using FACSCanto or LSRII system (BD Biosciences, Mississauga, ON, Canada). At least 10,000 CD14⁺ cells were collected. Data were analyzed using FlowJo (Tree Star, Inc., Ashland, OR).

Statistical Analysis

Phagocytosis data for HIV+ donors were normalized to their respective HIV− controls, and statistical comparisons were performed using the Wilcoxon signed-rank test. Table S2 (see Supplemental Digital Content, https://links.lww.com/QAI/A594) shows the normalized nonopsonic and opsonic phagocytic indices for each HIV+ donor at each time point and indicates all missing data. CD36 mean fluorescence indexes (MFIs) were compared using the Wilcoxon matched pairs test. Correlation between phagocytic index and CD36 MFI was assessed using the Spearman r test. Analyses were performed using GraphPad Prism (GraphPad, La Jolla, CA).

RESULTS

Monocytes From Chronic HIV+ Individuals Are Impaired in Their Ability to Phagocytose Nonopsonized PEs and Opsonized Erythrocytes, but These Functions Are Restored With cART

To assess the impact of HIV infection on nonopsonic phagocytosis of PEs, we performed phagocytosis assays using monocytes isolated from HIV+ and uninfected donors. To eliminate potential morphological and functional changes associated with using frozen cells, we performed all our assays on freshly isolated cells. Monocytes were isolated from each HIV+ donor before cART initiation (M0), and at 3 months (M3) and 6 months (M6) after cART initiation.

Nonopsonic phagocytosis of PEs was impaired in HIV+ treatment-naive donors compared with uninfected controls (Fig. 1A), showing approximately a 20% decrease in phagocytosis {median normalized phagocytic index: 0.79 [interquartile range (IQR), 0.58–0.92]; P = 0.0054}. This defect was still evident after 3 months of cART [median normalized phagocytic index: 0.73 (IQR, 0.63–1.09); P = 0.0105]. However, after 6 months of cART, nonopsonic PE phagocytosis was similar between HIV+ and uninfected controls [median normalized phagocytic index: 0.91 (IQR, 0.69–1.11); P = 0.25].

FIGURE 1: Nonopsonic phagocytosis of P. falciparum parasitized erythrocytes is impaired in HIV+ donors and recovers after 6 months of cART. Monocytes (purified by glass adhesion from freshly isolated PBMCs) were exposed to either nonopsonized P. falciparum trophozoite stage parasitized erythrocytes (PEs) (A) or antibody-coated human erythrocytes (B) at a ratio of 20 PEs/RBC per monocyte. Phagocytic index, defined as the total number of internalized PEs/RBCs per total number of monocytes counted expressed as a percentage, was calculated. Each HIV-infected (HIV+) individual was normalized to his/her HIV-negative (HIV−) control on the day of sample collection. Testing occurred pre-cART (M0), 3 months post-cART (M3), and 6 months post-cART (M6). A, Monocytes from chronic HIV+ individuals (black squares with median shown) are less able to phagocytose PEs (n = 23 per group, P = 0.0054) than those from HIV− individuals (normalized to 1—shown as a gray line). After 3 months of cART, phagocytic ability remains impaired (n = 19 per group, P = 0.0105). Six months post-cART, phagocytic ability is restored (n = 20 per group, P > 0.05). B, Monocytes from chronic HIV+ individuals (black squares with median shown) are less able to phagocytose opsonized RBCs (n = 9 per group, P = 0.0156) than those from HIV− individuals (normalized to 1—shown as gray line). After 3 months of cART, phagocytic ability was restored (n = 15 per group, P > 0.05). After 6 months of cART, phagocytic ability remained similar to HIV− controls (n = 17 per group, P > 0.05). All comparison by Wilcoxon signed-rank test assessing difference from controls normalized to a value of 1.

We also assessed differences in phagocytosis of opsonized RBC between HIV+ donors and uninfected controls (Fig. 1B). In agreement with previous reports,^5,7 opsonic phagocytosis was impaired in monocytes from treatment-naive HIV+ donors compared with uninfected controls [median normalized phagocytic index: 0.62 (IQR, 0.54–0.99); P = 0.0156], but recovered to control levels after 3 months of cART [median normalized phagocytic index: 0.94 (IQR, 0.66–1.55); P = 0.84].

No correlation was observed between viral load and either nonopsonic or opsonic phagocytosis. Nonopsonic PE phagocytosis did not correlate with CD4⁺ T-cell count at any of the times tested. However, opsonic phagocytosis directly correlated with CD4⁺ T-cell count in treatment-naive HIV+ donors (r² = 0.65; P = 0.008), suggesting that immune status may impact, at least partly, the degree of opsonic phagocytic defect observed. This correlation was lost after cART initiation, when opsonic phagocytic levels recovered to control levels. We were unable to detect any correlation between phagocytosis and type of cART therapy; however, our sample size was not powered for such analysis.

We also examined the change in nonopsonic phagocytic index in HIV+ donors over time. There was a significant increase in phagocytic index from M0 to M6 in the HIV+ group (P = 0.018) by Wilcoxon signed-rank test, see Table S2 Supplemental Digital Content, https://links.lww.com/QAI/A594. Three different types of responses were observed: 39% of participants showed no improvement in phagocytic index over time; for 33% of participants, phagocytic index improved by M3; and for 28%, phagocytic index improved by M6 (see Table S2, Supplemental Digital Content, https://links.lww.com/QAI/A594). We did not observe any correlation between CD4⁺ T cell, viral load, or drug regimen and improvement in phagocytic index, although our sample size was underpowered for such analyses.

Monocyte CD36 Levels Do Not Differ Between HIV+ and HIV− Donors, but the Correlation Between CD36 Levels and Nonopsonic PE Phagocytosis Is Lost in HIV+ Donors

CD36 is the major phagocytic receptor for nonopsonized PEs in vitro.³ We assessed monocyte CD36 levels by flow cytometry. CD36 MFI did not differ significantly between HIV+ and uninfected donors at any of the times tested (M0, M3, or M6) [HIV− vs. HIV+ median (IQR); for M0: 17.6 (10.0–43.7) vs. 15.9 (12.7–32.2), P = 0.5; for M3: 17.9 (10.7–24.2) vs. 17.1 (13.1–29.6), P = 0.6; and for M6: 23.1 (13.4–36.2) vs. 15.4 (9.9–25.5), P = 0.4]. In uninfected controls (Fig. 2A), CD36 levels were significantly correlated with phagocytic uptake of nonopsonized PEs (r² = 0.47; P = 0.0071). This correlation was not observed in HIV+ donors at M0 and M3, when a defect in nonopsonic phagocytosis of PEs existed (r² = 0.053; P = NS) (Fig. 2B). However, a significant correlation between CD36 levels and phagocytic uptake of nonopsonized PEs was restored in HIV+ donors after 6 months of cART (r² = 0.59; P = 0.016) (Fig. 2C), coinciding with the recovery of nonopsonic phagocytosis of PEs to control levels. No correlation was observed between CD36 levels and opsonic phagocytosis for either HIV+ or uninfected controls (Figs. 2D–F).

FIGURE 2: A correlation between CD36 surface levels and nonopsonic phagocytosis of P. falciparum parasitized erythrocytes is lost in HIV+ donors pre-cART and 3 months post-cART, but is regained after 6 months of cART. Monocyte CD36 surface levels were assessed by flow cytometry. CD36 MFI was plotted against nonopsonic parasitized erythrocyte phagocytic index (A–C) or opsonized erythrocyte phagocytic index (D–F). Data for HIV-negative donors (showing all data over the 3 sampling time points) are shown in A and D. Data for HIV-infected donors are shown in B and E [for pre-cART (M0) and 3 months post-cART (M3) combined], and in C and F (for 6 months post-cART). Correlation was assessed by Spearman r test.

DISCUSSION

HIV/malaria-coinfected individuals are more likely to have higher parasite burdens and develop severe malaria complications.¹ In malaria infection, the innate immune system plays an important role in controlling parasitemia and disease severity through innate clearance mechanisms and induction of proinflammatory cytokine responses. We have previously shown that malaria-induced IFNγ and TNF responses from innate immune cells are impaired in HIV+ donors pre- and post-cART.⁸ Here, we demonstrate that phagocytic uptake of nonopsonized PEs is also impaired in treatment-naive HIV+ individuals, highlighting the broad impact of HIV infection on innate immune responses to malaria. However, unlike the sustained defect in innate cytokine responses,⁸ the defect in nonopsonic PE phagocytosis was rescued after 6 months of cART.

Innate phagocytic defects have been previously reported in the context of HIV infection. Innate clearance of Pneumocystis jiroveci was impaired in alveolar macrophages from HIV+ donors compared with controls, a defect that correlated with downregulation of mannose receptor (reviewed in ^{Ref. 4}). We are the first to report an innate phagocytic defect for Plasmodium falciparum PEs. The major phagocytic receptor for nonopsonized PEs in vitro is CD36,³ a scavenger and pattern-recognition receptor. CD36 levels did not differ between HIV+ donors and uninfected controls, suggesting that the phagocytic defect was not related to surface levels of CD36. However, our data suggest a potential dysregulation of CD36-mediated uptake of PEs in untreated HIV infection. Although we observed a significant correlation between CD36 levels and nonopsonic PE phagocytosis in uninfected controls, this correlation was absent in HIV+ donors at the points where a phagocytic defect was detected (M0 and M3). The correlation was recovered at M6, when nonopsonic phagocytic levels were similar between HIV+ donors and controls. These data may imply a dysregulation in CD36-mediated signaling or CD36-binding partner interactions that impact on phagocytic ability, a hypothesis that merits further investigation. CD36 exists in complexes that can include integrins, tetraspanins, and FcγR.¹⁰ The presence of FcγR in these complexes allows CD36 to interact with Syk and promote actin remodeling and internalization of oxidized low-density lipoprotein. Whether CD36 engages this pathway in the uptake of PEs has not been determined. Impaired FcγR-mediated Syk phosphorylation has been observed in HIV infection and has been postulated to contribute to FcγR-mediated phagocytic impairment.¹¹

CD36 co-operates with TLR2 to mediate phagocytosis of PEs, and CD36-mediated PE clearance can be enhanced with exogenous TLR2 activation.⁹ Since malaria is a strong TLR2 stimulator, the impaired phagocytosis observed in HIV+ donors may result from impaired TLR2 signaling. However, HIV infection has been associated with increased TLR2 expression and no indication of TLR2 dysfunction.^12,13

To study the impact of HIV on phagocytosis, most in vitro research has focused on macrophages. However, monocytes and macrophages differ in their susceptibility to HIV infection; monocytes are highly refractory, whereas macrophages are fully permissive.¹¹ Since only a small proportion of blood monocytes (<1%) are HIV infected,¹¹ the impaired phagocytosis we and others have observed, may be an indirect consequence of infection, reflecting the dysregulation of cytokine/chemokine production by monocytes and/or other cells. Defective phagocytosis by neutrophils (not targets for HIV infection) from HIV+ individuals has also been reported.¹⁴

We have shown that nonopsonic phagocytosis of PEs is impaired in the context of HIV infection and that this defect is reversed with cART. This underlines the importance of providing cART to HIV+ individuals especially in light of recent studies implicating astrocyte functional impairment in blood–brain barrier dysfunction,¹⁵ a central component of the pathogenesis of cerebral malaria.

ACKNOWLEDGMENTS

The authors thank all the participants who took part in the study, as well as the staff at the Maple Leaf Medical Clinic, in particular Roberta Halpenny and Tigist Kidane. The authors are grateful to staff and patients in the apheresis unit at Princess Margaret Hospital (Toronto, Canada) for providing human serum for parasite culture medium. The authors also thank Dr Erdman and Dr Hawkes for their help with data acquisition.