Introduction

Malaria is caused by Plasmodium parasites and is one of the most important global infectious diseases. It accounts for over 249 million clinical cases and 600 000 deaths annually1. It primarily impacts tropical and subtropical regions, particularly sub-Saharan Africa. Young children and pregnant women are the most vulnerable populations in these regions. Malaria presents with a wide spectrum of clinical symptoms, ranging from a mild, flu-like illness to severe complications, including cerebral malaria (CM), severe malarial anemia, and malaria-associated acute respiratory distress syndrome (MA-ARDS)2,3. Over the past decade, the focus on malaria-associated acute kidney injury (MAKI) has steadily increased. Malaria is the primary cause of acute kidney injury (AKI) in sub-Saharan Africa, accounting for 50% of all AKI cases in the Democratic Republic of Congo4,5. MAKI is characterized by a sudden and rapid decline in renal function, resulting in an impaired ability of the kidney to effectively filter metabolic waste products, electrolytes, and excess fluids from the bloodstream6. MAKI has been associated with increased in-hospital mortality, post-discharge mortality, and neurologic deficits such as uremic encephalopathy5.

Traditionally, three main hypotheses have been proposed that could explain the development of MAKI: intravascular hemolysis, insufficient blood flow due to parasite sequestration and an exaggerated host response7. Recent studies have already demonstrated the detrimental role of heme in the development of MAKI8,9,10. Previously, we have shown that kidney pathology in the Plasmodium berghei NK65 (PbNK65) model is independent of parasite sequestration11. Nevertheless, the role of the host immune response in the development of MAKI remains unclear. Severe malaria triggers excessive proinflammatory cytokine production, such as Tumor Necrosis Factor-alpha (TNF-α), Interleukin-1β (IL-1β) and IL-6, in plasma of humans as well as in murine kidney tissues12,13. Additionally, a recent study found a correlation between circulating immune complexes and hemoglobinuria in children in Uganda5. Despite these insights, the full immunopathology of MAKI remains poorly understood.

Pro- and anti-inflammatory cytokines are essential to maintain the balance between immunity and pathology during malaria. This has been investigated and reviewed in detail in several mouse models of malaria3. Typically, in non-lethal models with a transient parasitemia peak such as P. chabaudi AS (PcAS) in C57BL/6 mice, proinflammatory cytokines such as IL-12 and Interferon-γ (IFN-γ) are required for antiparasitic immunity. In contrast, in models with lethal P. berghei ANKA or PbNK65 infections, these cytokines mediate lethal immunopathologies such as experimental CM and MA-ARDS14,15. Anti-inflammatory cytokines, such as IL-10, counterbalance these effects and downregulate both antiparasitic immunity and immunopathologies. IFN-γ is crucial for the T helper-type 1 (Th1) immune response and is produced by both the innate and adaptive immune system3. NK (T) cells and γδ T cells are the early producers of IFN-γ during a malaria infection, and subsequently the production shifts to activated CD4+ and CD8+ T cells as main producers of large amounts of IFN-γ3. While IFN-γ plays an important role in antimalarial immunity, its dysregulation can lead to complications14, such as CM and MA-ARDS15,16. Besides being an essential cytokine for the Th1 immune response during malaria, IFN-γ is crucial for the expression of chemokines and adhesion molecules and subsequent CD8+ T cell infiltrations in brain and lungs during experimental CM and MA-ARDS16,17,18. Furthermore, IFN-γ is essential for cross-presentation of malarial antigens by endothelial cells in these models15,19. In contrast, Th2 immune responses are typically less prominent in malaria20.

Clinical studies have described both protective and pathological associations regarding the role of IFN-γ in malaria14,21. Elevated levels of proinflammatory cytokines, including IFN-γ, have been linked to the severity of CM and anemia in Malawian children22. A recent meta-analysis has found a correlation between IFN-γ levels and malaria severity23. Furthermore, IFN-γ is implicated in the pathogenesis of kidney conditions like lupus nephritis24and renal ischemic injury25. By upregulating major histocompatibility complex (MHC) on renal cells, it enhances antigen presentation, potentially promoting autoimmunity26. Multiple studies have identified an increased renal expression of IFN-γ during malaria infection27,28. Despite these insights, the exact role of IFN-γ in conveying protection or exacerbating MAKI remains poorly understood.

In this study, we examined the effect of IFN-γ deficiency on kidney pathology in experimental models of severe and mild malaria using PbNK65 and PcAS, respectively. IFN-γ knockout (KO) mice infected with PbNK65 presented a delayed parasitemia development which did not reach similar levels as the wild type (WT) mice. Furthermore, these mice were protected from kidney pathology. In contrast, IFN-γ KO mice infected with PcAS presented with hyperparasitemia and exacerbated kidney pathology. These findings demonstrate that IFN-γ deficiency has a parasite strain-dependent effect on parasitemia and kidney pathology in murine malaria models.

Results

PbNK65-infected IFN-γ KO mice do not develop kidney pathology

Previously, we observed the development of MAKI in PbNK65-infected C57BL/6 mice, as evidenced by proteinuria, increased blood urea nitrogen (BUN) levels, renal inflammation and histopathological changes11. To determine the role of IFN-γ in the pathogenesis of experimental MAKI, both IFN-γ KO and WT mice were infected with PbNK65. We observed significantly decreased parasitemia levels in PbNK65-infected IFN-γ KO mice compared to WT mice from day 8 post infection (p.i.) (Fig. 1a). PbNK65-infected IFN-γ KO mice also showed a milder disease course compared to the infected WT mice, with significantly lower clinical scores (Fig. 1b) and no decrease in body weight (Fig. 1c). At 9 days p.i., mice were dissected and alveolar edema was determined. Upon infection, IFN-γ KO mice did not develop alveolar edema, while significantly increased alveolar edema was observed in the WT mice (Fig. 1d). Furthermore, we studied the effect of IFN-γ on the kidney pathology by determining proteinuria and BUN plasma levels (Fig. 1e, f). Upon infection, WT mice developed significantly increased proteinuria and plasma BUN levels. However, this was not observed in IFN-γ KO mice.

Fig. 1
figure 1

PbNK65-infected IFN-γ KO mice develop decreased parasitemia levels without kidney pathology. WT and IFN-γ KO C57BL/6 mice were infected with PbNK65. (a) Peripheral parasitemia was determined on blood smears. (b) The clinical score and (c) body weight was monitored daily from 6 days p.i. onwards. Infected mice were euthanized and dissected at 9 days p.i. and (d) protein content in the broncho alveolar lavage fluid (BALF) was determined as measure for alveolar edema. (e) Albumin/creatinine ratios were determined in urine samples. (f) BUN values were determined in blood collected via retro-orbital puncture. Asterisks above data points indicate significant differences compared to control mice, asterisks above a horizontal line show significant differences between infected groups, *p < 0.05, **p < 0.01, ***p < 0.001. Data of two experiments, Control IFN-γ WT: n = 6, Control IFN-γ KO: n = 8, PbNK65-infected IFN-γ WT mice: n = 12, PbNK65-infected IFN-γ KO mice: n = 14. In panels a-c, data are represented as means +/- SEM. Mann-Whitney U test with Holm-Bonferroni correction for multiple testing was performed (number of tests = 4 in panels c-f).

Full size image

To examine the contribution of IFN-γ signaling on renal histopathological changes during infection, we performed Periodic Acid Schiff (PAS) staining on kidney sections of control and PbNK65-infected WT and IFN-γ KO mice. Glomerulosclerosis was detected with collapse of some glomerular capillary tufts, in kidneys of PbNK65-infected WT mice at day 9 p.i. (Fig. 2a). Up to 9% of the glomeruli were affected in the PbNK65-infected IFN-γ WT mice, which was significantly higher compared to PbNK65-infected IFN-γ KO mice (Fig. 2b). Acute tubular injury, characterized by vacuolization of tubular epithelial cells and loss of brush border in proximal tubular epithelial cells, was present in infected WT mice, but not in IFN-γ KO mice (Fig. 2a). Intravascular accumulation of leukocytes was also observed in the kidneys of PbNK65-infected WT mice, but not in IFN-γ KO mice.

Fig. 2
figure 2

Less glomerular collapse in PbNK65-infected IFN-γ KO compared to WT mice. (a) Kidney sections were stained with PAS. Representative images are shown (original magnification x40, bar = 50 μm). Black frames indicate collapse of glomerular tufts. Black arrows indicate vacuolization of proximal tubular epithelial cells, red arrows indicate loss of brush border of proximal tubular epithelial cells. Blue arrows indicate vascular leukocyte accumulations. (b) Percentage of glomeruli with collapsed glomerular tufts counted on PAS-stained kidney sections. Asterisks above data points indicate significant differences compared to control mice, **p < 0.01. Data of two experiments, Control IFN-γ WT: n = 6, Control IFN-γ KO: n = 8, PbNK65-infected IFN-γ WT d9: n = 13, PbNK65-infected IFN-γ KO d9: n = 15. Mann-Whitney U test with Holm-Bonferroni correction for multiple testing (number of tests = 4) was performed in panel b.

Full size image

To determine whether kidney pathology in IFN-γ KO mice developed later during infection, mice were followed up until 16 days p.i. (Suppl. Fig S1). Parasitemia in IFN-γ KO mice slowly increased, but did not reach similar levels as in IFN-γ WT mice. At 16 days p.i. mild proteinuria was observed in IFN-γ KO mice in the absence of alveolar edema or body weight loss. A positive correlation between parasitemia and proteinuria was observed. PAS staining of kidney sections showed only minimal histopathological alterations at day 16 p.i., in IFN-γ KO mice, with no collapsing glomeruli. However, mild tubular injury such as interstitial edema was observed in some mice (Suppl. Fig S2). In conclusion, our findings highlight the pathogenic role of IFN-γ in the PbNK65 model, as its presence boosts parasitemia, and thereby exacerbates MAKI and MA-ARDS.

PcAS-infected IFN-γ KO mice develop hyperparasitemia and increased proteinuria

To examine the effect of IFN-γ on kidney pathology in the PcAS mouse model of mild malaria, we infected IFN-γ KO and WT mice with PcAS parasites. PcAS-infected IFN-γ KO mice developed significantly increased parasitemia levels compared to WT mice, with parasitemia levels up to 60% at day 9 and 10 p.i. (Fig. 3a). Furthermore, clinical scores were significantly higher in PcAS-infected IFN-γ KO mice compared to PcAS-infected WT mice (Fig. 3b). Body weight loss appeared more pronounced in IFN-γ KO mice, but this difference was only statistically different on day 9 p.i. (Fig. 3c). A subgroup of PcAS-infected IFN-γ KO mice developed proteinuria at peak parasitemia (day 9 or 10 p.i.) with significantly increased proteinuria at day 9 p.i. compared to control mice (Fig. 3d). Furthermore, these mice also produced dark urine at peak parasitemia.

Fig. 3
figure 3

PcAS-infected IFN-γ KO mice develop hyperparasitemia and increased kidney pathology. IFN-γ KO and WT C57BL/6 mice were infected with PcAS. (a) Peripheral parasitemia was determined on blood smears. (b) The clinical score and (c) body weight was monitored daily from day 6 p.i. onwards. Infected mice were euthanized and dissected at peak parasitemia at day 9 or 10 p.i. and (d) albumin/creatinine ratios were determined in urine samples. Asterisks above data points indicate significant differences compared to control mice, asterisks above a horizontal line show significant differences between infected groups. (e) Albumin/creatinine ratios were subdivided according to gender. Asterisks above data points indicate significant differences compared to IFN-γ WT infected mice of corresponding gender, asterisks above a horizontal line show significant differences between gender, *p < 0.05, **p < 0.01, ***p < 0.001. (f) Correlation analysis of parasitemia with albumin/creatinine ratios of urine of PcAS-infected IFN-γ WT and IFN-γ KO at day 7, 8, 9 and 10 p.i. Dark brown upside down triangles indicate PcAS-infected IFN-γ KO which produced dark urine. Spearman correlation test was performed, Spearman correlation coefficient r and p-values are shown. Data of two experiments, Control IFN-γ WT: n = 4, Control IFN-γ KO: n = 8, PcAS-infected IFN-γ WT mice: n = 13, PcAS-infected IFN-γ KO mice: n = 19. Mann-Whitney U test with Holm-Bonferroni correction for multiple testing (number of tests = 4) was performed in panels c-e.

Full size image

Interestingly, a clear gender difference was noted amongst the infected IFN-γ KO mice. Male IFN-γ KO mice developed higher parasitemia and increased proteinuria compared to female IFN-γ KO mice. In male mice, proteinuria and body weight loss were also significantly increased in the IFN-γ KO group compared to WT. This difference was not observed in the female mice (Fig. 3e and Suppl. Fig. S3). Increased parasitemia levels were positively correlated with proteinuria (Fig. 3f). Furthermore, no alveolar edema was observed, indicating the absence of MA-ARDS in both infected groups. (Suppl. Fig. S4).

PcAS-infected IFN-γ KO mice develop mild tubular injury and hyaline casts in kidneys

To study the renal histopathology in PcAS-infected IFN-γ KO and WT mice, PAS staining was performed on kidney sections (Fig. 4). Mild tubular injury, loss of the brush border, sloughing of epithelial cells and vacuolization were detected mainly in the cortex of kidneys of PcAS-infected IFN-γ KO mice. This was not observed in the PcAS-infected WT mice. No interstitial nephritis or interstitial edema was observed. The glomerular capillaries were open in both the PcAS-infected WT and IFN-γ KO mice, indicating the absence of glomerular collapse. In the kidney sections of most PcAS-infected IFN-γ KO mice, casts were detected in the lumen of the tubules (Fig. 5a). To confirm the presence of casts, hematoxylin-eosin (H&E) staining was performed on kidney sections from control mice and PcAS-infected mice, revealing hyaline casts primarily in the lumen of the distal tubules (Fig. 5b). Importantly, the number of hyaline casts in the PcAS-infected IFN-γ KO mice was significantly increased compared to the WT mice, further corroborating the increased renal injury (Fig. 5c). Interestingly, the largest number of hyaline casts were observed in infected male IFN-γ KO mice, further corroborating the gender difference in kidney pathology (Suppl. Fig S3.). Overall, these results indicate protective effects of IFN-γ in the PcAS model, as its presence inhibits parasitemia, and thereby diminishes kidney pathology.

Fig. 4
figure 4

PcAS-infection induces mild tubular injury in IFN-γ KO mice. At day 9 or day 10 p.i. (peak parasitemia), control, WT and IFN-γ KO C57BL/6 mice infected with PcAS were killed and dissected. PAS staining was performed (original magnification x40, bar = 50 μm). Black arrows indicate vacuolization of proximal tubular epithelial cells, red arrows indicate loss of brush border of proximal tubular epithelial cells. Blue arrows indicate vascular accumulation of leukocytes.

Full size image
Fig. 5
figure 5

Hyaline casts are formed in PcAS-infected IFN-γ KO mice. At day 9 or day 10 p.i. (peak parasitemia) IFN-γ KO C57BL/6 mice infected with PcAS were dissected. (a) PAS staining and (b) H&E staining were performed (original magnification x40, bar = 50 μm). Casts are indicated with a black arrow on both the PAS and H&E staining. (c) Casts on H&E stained sections were counted. The gender of the infected IFN-γ KO mice is indicated by dark red for males and bright red for females. Asterisks above data points indicate significant differences compared to control mice, asterisks above a horizontal line show significant differences between KO and WT infected groups, *p < 0.05, **p < 0.01, ***p < 0.001. Data of two experiments, Control IFN-γ WT: n = 4, Control IFN-γ KO: n = 8, PcAS-infected IFN-γ WT mice: n = 13, PcAS-infected IFN-γ KO mice: n = 19. Mann-Whitney U test with Holm-Bonferroni correction for multiple testing (number of tests = 4) was performed in panel c.

Full size image

Expression of kidney injury markers and cytokines increased during malaria infection

MAKI pathogenesis is associated with elevated kidney injury markers and increased renal cytokine expression11. Therefore, we measured the expression of kidney injury markers and cytokines in the kidneys of infected and non-infected mice by qRT-PCR. The mRNA expression of the following markers was determined: Neutrophil Gelatinase-Associated Lipocalin (NGAL, released by the renal tubular cells in response to injury), Heme Oxygenase 1 (HO-1, induced upon and protective against oxidative stress), C-X-C Motif Chemokine Ligand 10 (CXCL10, also designated as IP-10; pathogenic in other malaria-induced complications by attracting activated T cells) and TNF-α (key proinflammatory cytokine). Upon infection with PbNK65, the mRNA expression of NGAL was increased in both WT and IFN-γ KO mice at day 9 and 16 p.i., with significant differences between WT mice at 9 days p.i. and IFN-γ KO mice at 9 and 16 days p.i. (Fig. 6a). This is consistent with the minimal kidney injury in the PbNK65-infected IFN-γ KO mice. The mRNA of HO-1, which is typically induced by heme and protects against oxidative damage, was also increased in PbNK65-infected mice. Moreover, HO-1 was significantly more increased in WT than in IFN-γ KO mice at 9 days p.i., though no difference was observed anymore between WT mice at 9 days p.i. and IFN-γ KO mice at 16 days p.i. In PcAS-infected mice, both HO-1 and NGAL mRNA levels were increased compared to control mice, with only HO-1 being significantly higher in IFN-γ KO compared to WT (Fig. 6b). Both PbNK65 and PcAS infections led to increased renal CXCL10 mRNA expression, with significantly lower levels in IFN-γ KO mice compared to WT mice (Fig. 6c). TNF-α mRNA expression did not differ between WT and KO, but was significantly different from controls in KO-infected mice (Fig. 6d). These data confirm the presence of kidney pathology, and indicate that IFN-γ is important for the induction of CXCL10 in both models. Furthermore, the effect of IFN-γ on the expression of NGAL and HO-1 is dependent on the used parasite strain.

Fig. 6
figure 6

Kidney injury markers and cytokines increased during malaria infection. IFN-γ KO and WT C57BL/6 mice were infected with PbNK65 and PcAS. The kidneys were collected at day 9 p.i. and day 16 p.i. for PbNK65-infected mice and at day 9 or 10 p.i. (peak parasitemia) for PcAS-infected mice. RT-qPCR was performed on the kidneys. (a) mRNA expression of NGAL and (b) HO-1 were measured. (c) Inflammatory cytokines CXCL10 and (d) TNF-α were measured. Asterisks above data points indicate significant differences compared to control mice, asterisks above a horizontal line show significant differences between infected groups, *p < 0.05, **p < 0.01, ***p < 0.001. Mann-Whitney U test with Holm-Bonferroni correction for multiple testing (number of tests = 10 (no comparisons between parasite strains)) was performed in all panels.

Full size image

Materials and methods

Parasites and mice

Seven to eight weeks old WT and IFN-γ KO mice, 8 times backcrossed to C57BL/6 and obtained from Jackson Laboratories (strain #002287), were bred in the animal house of the Rega Institute for Medical Research, KU Leuven. All mice were housed in individually ventilated cages in a 12 h light and 12 h dark cycle in SPF animal facility. Drinking water was supplemented with 4-amino benzoic acid (0.422 mg/ml PABA, Sigma-Aldrich, Bornem, Belgium) and the mice had unlimited access to high energy food. IFN-γ WT and IFN-γ KO C57BL/6 mice were infected with P. berghei NK65 (PbNK65) or P. chabaudi AS (PcAS) parasites by intraperitoneal (IP) injection of 104iRBCs as described previously29.

Scoring of disease progression and parasitemia determination

Starting from day 6 p.i., parasitemia, body weight and clinical score were evaluated daily starting from day 6 p.i. Blood smears of tail blood were stained with 10% Giemsa (VWR, Heverlee, Belgium) and parasitemia was calculated by microscopic analysis. Urine collection was performed daily. The clinical score was calculated by evaluating different clinical parameters, including social activity (SA), limb grasping (LG), body tone (BT), trunk curl (TC), pilo-erection (PE), shivering (Sh), abnormal breathing (AB), dehydration (D), incontinence (I) and paralysis (P). A score of 0 (absent) or 1 (present) was given for TC, PE, Sh and AB and 0 (normal), 1 (intermediate) or 2 (most serious) for the other parameters. The total clinical score was calculated using the following formula: SA + LG + BT + TC + PE + 3 * (Sh + AB + D + I + P). The mice were euthanized when the body weight decrease was > 20% compared to day 0 p.i. or when clinical score reached 10 or more11,30.

Retro-orbital puncture and dissection

IFN-γ WT and IFN-γ KO mice infected with PbNK65, dissected at day 9 p.i., were anesthetized with 3% isoflurane (Iso-Vet, Dechra, Nortwhich, United Kingdom) before retro-orbital puncture was performed with a heparinized (LEO Pharma, Lier, Belgium) glass capillary tube (Hirschmann-Laborgeräte, Eberstadt, Germany). The collected blood was analyzed with the Epoc Blood Analysis System (Siemens, Munich, Germany) to determine BUN levels. After the blood collection, mice were euthanized by performing heart puncture under anesthesia with 3% isoflurane.

The PcAS-infected IFN-γ WT and IFN-γ KO mice were euthanized with Dolethal (Vétoquinol, Aartselaar, Belgium; 200 mg/mL, IP injection of 100 µL) followed by heart puncture at indicated time points. Broncho-alveolar lavage fluid (BALF) was collected by clamping off the left lung and catheterizing the bronchus of the mice. A total of 500 µl of PBS was infused through the catheter, then withdrawn after 30 s. This process was repeated, and the two samples were combined. The pooled BALF was then centrifuged at 314 g for 10 min at 4 °C, and the supernatant was collected. After a transcardial perfusion with 20 ml PBS, the left kidney was removed and laterally cut into two equal pieces and fixed in 4% formaldehyde for 48 h at 4 °C.

Kidney histology

After fixation for 48 h at 4 °C, kidney tissues were dehydrated by applying gradually increasing ethanol concentrations in the Excelsior MS tissue processor (Thermo Fisher Scientific, Waltham, USA). Next, the tissues were embedded in paraffin with the HistoStar Workstation (Thermo Fisher Scientific) and 5 μm thick tissue sections were made using a Microm HM 355 S microtome (Thermo Fisher Scientific). Tissue sections were stained with the Periodic Acid Schiff’s (PAS) staining kit (Carl Roth GmbH, Karlsruhe, Germany) and Hematoxylin and eosin (H&E) staining (Abcam, Ab245880). Histological assessment was performed with a Leica DM 2000 microscope and pictures were taken using the LAS V4.2 Software (Leica). The percentage of collapsed glomeruli and renal blood vessels with intravascular accumulation of one or more leukocytes was calculated on whole sections. To quantify the casts, the kidney sections were blinded and the number of casts were counted per kidney section.

Analysis of urine samples

Urine samples were collected in a 1.5 mL Eppendorf tube at the indicated time points in the morning. The albumin/creatinine ratio in the urine was determined to assess proteinuria and kidney function as described in Vandermosten et al.31.

Alveolar edema measurement

Alveolar edema formation was assessed by the quantification of the protein concentration in the supernatants of BALF using a Bradford assay (Bio-Rad).

Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)

RNeasy Mini Kit (Qiagen, Hilden, Germany) was used to extract RNA from the kidney after mechanical homogenization in RLT buffer. After extraction, RNA was quantified and cDNA was synthesized using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, USA). ABI Prism 7500 Sequence Detection System (Applied Biosystems) was used to perform qRT-PCR reaction on cDNA with specific primers (IDT, Leuven, Belgium, Table 1) in the TaqMan® Fast Universal PCR master mix (Applied Biosystems). The relative mRNA expression was determined as 2−ΔΔCT, normalized to the mean 2−CT value of the uninfected control mice and to the 2−CT value of the 18 S housekeeping gene.

Table 1 List of primers used for RT-qPCR.
Full size table

Statistical analysis

Statistical analysis was done using the GraphPad Prism software (GraphPad software, San Diego, USA, version 8.3.1). The non-parametric Mann–Whitney U test was used to determine the statistical significance between two groups. P-values smaller than 0.05 were considered statistically significant. P-values were defined as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. To correct for multiple testing, the Holm-Bonferroni method was applied and the number of comparisons are indicated in the figure legends. Unless otherwise specified, each dot represents an individual mouse. Horizontal lines represent group medians. Asterisks without horizontal lines represent significant differences compared to the control group, unless mentioned otherwise. Horizontal lines with asterisk on top indicate significant differences between infected groups, unless mentioned otherwise. Correlation analyses on non-normally distributed data were performed using the Spearman correlation analysis and linear regression was used for curve fitting.

Discussion

In this study, we investigated the role of IFN-γ, a crucial proinflammatory cytokine, in kidney pathology in experimental models of severe and mild malaria with PbNK65 and PcAS, respectively. Our results suggest that the effect of IFN-γ on the parasitemia indirectly affects kidney pathology (Fig. 7). These findings indicate that IFN-γ has a pathogenic effect in the PbNK65 model, since IFN-γ deficiency resulted in delayed parasitemia levels and milder kidney pathology. In contrast, IFN-γ has a protective effect in the PcAS model, as IFN-γ deficient mice developed higher parasitemia and more pronounced kidney pathology.

Fig. 7
figure 7

Graphical abstract. The effect of IFN-γ KO differs in the PcAS-infected versus PbNK65-infected mice. In PcAS-infected mice, IFN-γ deficiency result in hyperparasitemia, hemoglobin casts deposition, increased tubular damage and albuminuria. In contrast, IFN-γ deficiency in PbNK65-infected mice results in low parasitemia, reduced glomerular collapse, tubular damage and albuminuria.

Full size image

Consistent with our PcAS data, previous studies have shown that IFN-γ promotes parasite clearance in both liver and blood stages of malaria14,32,33. A meta-analysis in children confirmed the protective effect of IFN-γ against parasitemia and anemia34,35. In the PcAS model, IFN-γ is protective due to its proinflammatory roles in activating macrophages, promoting Th1 differentiation, and enhancing antigen presentation14. Moreover, A/J mice, which are highly susceptible to PcAS, expressed lower levels of IFN-γ compared to resistant C57BL/6 mice36.

In the PbNK65 model, we observed a stimulating effect of IFN-γ on parasitemia. This was surprising, since other studies with P. berghei models described no effect of IFN-γ deficiency on parasitemia17. Therefore, this phenotype might be specific for PbNK65. However, IFN-γ has been shown to mediate increased total parasite burden of P. berghei ANKA27, which may be explained by the induction of adhesion molecules that may mediate parasite sequestration16,17. Interestingly, Guthmiller et al. observed that IFN-γ signaling in B cells reduces the production of anti-plasmodial antibodies in a model with P. yoelii, suggesting that IFN-γ may also inhibit antimalarial immunity37.

Previously, the pathological role of IFN-γ has been observed using a monoclonal antibody against IFN-γ in a CM model38. Lung and brain pathology in the PbNK65 and PbANKA models are driven by CD8+ T cells, which has been demonstrated in multiple studies15,16,17,30. Notably, upon severe malaria infection, CD8+ T cells express high levels of IFN-γ15. The CD8+ T cell-driven pathology requires IFN-γ-induced cross-presentation of malarial antigens by endothelial cells. Furthermore, the IFN-γ-dependent induction of chemokines and adhesion molecules is also critical for the pathology15,16,39. In contrast, the PcAS model is less dependent on CD8+ T cell-mediated pathology40. Our study offers novel insight into the role of IFN-γ on parasitemia with an indirect immunopathological effect on MAKI. The immunopathogenesis in MAKI has been poorly investigated, although 50% of severe malaria patients present with MAKI5. Overall, we report improved versus exacerbated kidney pathology in the PbNK65 and PcAS model of malaria with deficiency of IFN-γ, respectively. Exacerbated kidney pathology was also marked by increased proteinuria, and an increase in mRNA expression of HO-1 in the PcAS model and NGAL in the PbNK65 model.

Our data suggest that parasitemia is strongly associated with kidney pathology in malaria. This may correspond with the known role of intravascular hemolysis in MAKI10,41,42. In particular, increased hyaline cast depositions were also observed in haptoglobin/hemopexin double KO mice and in mice with kidney-specific deletion of HO-1 or Ferritin-H upon P. chabaudi infection10,43. This suggests that the increased hemolysis associated with high parasitemia levels may overwhelm the haptoglobin / hemopexin / HO-1 / ferritin protective mechanisms, resulting in damage to the proximal epithelial tubules. The parasitemia in the PbNK65 model is lower than in the PcAS model. While we cannot exclude a role for a possibly higher load of sequestered PbNK65 parasites, the presence of glomerular collapse suggests a different pathogenesis, which is currently insufficiently understood. The absence of glomerular collapse in PbNK65-infected IFN-γ KO mice suggests either a direct role of IFN-γ, or a minimal parasitemia required to cause this glomerular pathology. Additionally, hemozoin is a known inducer of HO-1, CXCL10 and TNF-α in the liver and lungs44,45. Although the levels of hemozoin in the kidney is lower than in the liver and lungs46, hemozoin might also contribute to the glomerular and tubular damage.

Multiple studies in humans have investigated the cytokine pattern during malaria infections. Although, the majority report IFN-γ to have protective effects, many also report the opposite. A study performed in Sri Lanka described an inverse relationship between IFN-γ and parasitemia, suggesting an anti-parasitic effect34. On the other hand, IFN-γ profiles where higher in Malawian and Indian children with severe malaria compared to uncomplicated malaria22,47. In another study, adults also presented with increased levels of IFN-γ correlating with severe malaria48. Interestingly, a recent meta-analysis reported a positive association between IFN-γ levels and malaria severity. However, the degree of heterogeneity among included studies was high23. This suggests that the role IFN-γ plays during malaria remains bivalent and needs further investigation.

Overall, our study provides new insights in the complex role of IFN-γ in the development of kidney injury in relation to parasitemia levels. Our data demonstrate that the effect of IFN-γ on kidney pathology is dependent on the model, and is mostly related to its effects on parasitemia. Further investigations are required to delineate the precise contribution of local inflammation and immunopathology to the pathogenesis of MAKI.