PAD4 controls chemoattractant production and neutrophil trafficking in malaria

Peptidylarginine deiminase 4 (PAD4) is a key regulator of inflammation but its function in infections remains incompletely understood. We investigate PAD4 in the context of malaria and demonstrate a role in regulation of immune cell trafficking and chemokine production. PAD4 regulates liver immunopathology by promoting neutrophil trafficking in a Plasmodium chabaudi mouse malaria model. In human macrophages, PAD4 regulates expression of CXCL chemokines in response to stimulation with TLR ligands and P. falciparum. Using patient samples, we show that CXCL1 may be a biomarker for severe malaria. PAD4 inhibition promotes disease tolerance and may represent a therapeutic avenue in malaria.

Most PAD4 studies to date have focused on neutrophils, abundant myeloid cells with both protective and detrimental roles in inflammation. Neutrophils respond to infectious or inflammatory stimuli by phagocytosis, degranulation, or extrusion of chromatin in the form of neutrophil extracellular traps (NETs). 9 NETs trap and prevent dissemination of microbes but can also propagate inflammation via their immunostimulatory properties. 9 In neutrophils, PAD4-mediated citrullination of histones regulates chromatin structure 1 and has been implicated in both enhancement of NET immunogenicity 10 and direct control of NET release, 6,11 although the latter has been contested. 12,13 The function of PAD4 in other myeloid cells is less clear. Citrullination is not restricted to histones but can also modulate activity of various transcription factors 14,15 and apoptosis regulators such as p53. 16 In macrophages, citrullination was reported to regulate the assembly of the NLRP3 inflammasome. 17 To better characterize the function of PAD4 in systemic inflammation, we examined the role of the enzyme in malaria, a disease caused by replication of Plasmodium parasites in red blood cells (RBCs).
Malaria annually affects over 200 million people and leads to approximately 400,000 deaths. 18 The majority of individuals infected with P. falciparum, the most lethal parasite species, suffer from uncomplicated malaria. A portion of patients, primarily children in sub-Saharan Africa, develop severe symptoms such as metabolic acidosis, severe anemia, or cerebral malaria. Organ failure in severe malaria is initiated by cytoadhesion of infected RBCs (iRBCs) in the microvasculature, causing endothelial activation and recruitment of inflammatory cells, which leads to vascular leak or occlusion. 18 Neutrophils and CD8+ T cells both contribute to malaria immunopathology. NET release in malaria promotes iRBC cytoadhesion, 19,20 while neutrophil and CD8+ T cell infiltration of organs causes tissue damage. 19,[21][22][23][24] Understanding chemoattractant regulation is therefore essential for understanding malaria immunopathology.

Human samples
Collection of blood samples was approved by the Comité d'Ethique Régional Indépendent de Lambaréné (Gabon) and NHS REC 18/EE/0265. All participants provided written informed consent.
Patient characteristics are described in Ref. 19 .

Histology
Livers were harvested, stained, and scored as previously described. 19 Neutrophils were detected using an in-house antibody directed against neutrophil calgranulin A. 19 CXCL1 was immunostained with Protein-tech 12335-1-AP and macrophages with anti-F4/80 antibody Proteintech 28463-1-AP primary antibodies, followed by anti-rabbit IgG secondary antibody (ThermoFisher A-11012). CD8+ T cells were stained with monoclonal antibody SP16 (Thermo Fisher MA5-16345) and antirabbit AP-Red conjugated secondary. Image acquisition was performed on Leica DMI 6000D at 20× magnification, with LAS X software. The total CXCL1 area was calculated in Fiji by manually thresholding the image, followed by normalization to total DAPI area.

Quantification of sequestration
Mice were infected with a luciferase-expressing strain of P. chabaudi. 24 Mice were culled at nighttime, which is the time of maximum schizont sequestration and perfused systemically by injection of 10 ml of PBS into the heart. Organs were removed and processed for luminescence as previously described. 19

Flow cytometry
Quantification of circulating neutrophils was performed as previously described. 19 Neutrophils were counted as CD45+, CD115−, and

In vitro macrophages and neutrophils
Mouse bone marrow-derived macrophages (BMDMs) were prepared with 10% final volume L929 cell-conditioned medium. Mouse neutrophils were isolated from bone marrow as previously described, 25 using EasySep™ Mouse Neutrophil Enrichment Kit (Stem Cell). Human macrophages were prepared as previously described, 19

Biomarker quantifications
Aspartate aminotransferase (AST) was quantified in mouse plasma by the veterinarian service laboratory at SYNLAB.vet GmbH (Berlin, Germany). NETs were quantified as before, 19 by combining the Cell Death Detection ELISA Kit (Roche) and anti-MPO antibody (1 μg/ml final concentration, HM1051BT; Hycult). Human and mouse CXCL1 and CXCL2 were quantified using DuoSet ELISA (R&D Systems).

Quantitative PCR
RNA was extracted with RNeasy kit (Qiagen). cDNA was pre- Real time PCR was carried out with SYBR green (Thermo) on QuantStudio 3 machine. Relative transcript abundance was calculated using the 2-ΔΔCT method.

RESULTS/DISCUSSION
We infected wildtype and PAD4−/− mice with asexual stage P.
chabaudi, which predominantly sequesters in the liver, inducing inflammation and hepatocyte necrosis. 19,24 Parasites proliferated at similar rates in wildtype and knockout animals, indicating that PAD4 does not control the host antiparasitic response (Figure 1(A)). On the other hand, liver damage was strikingly reduced in PAD4−/− animals, as measured by both circulating AST (Figure 1(B)) and by histopathologic analysis (Figure 1(C)).
Liver damage in the P. chabaudi model results from accumulation of both parasites and inflammatory cells. 19,24 We tested which of these is affected in the PAD4 knockout. Quantification of P. chabaudi cytoadherence using a luciferase expressing parasite revealed no difference in sequestration rates or patterns (Figure 1(D)). On the other showing that NET formation is not altered. This is consistent with a previous report of unchanged in vitro NET formation by PAD4−/− neutrophils in response to heme/TNF, 19 although conflicting findings have also been reported for macrophage inhibitory factor-induced NETosis in malaria. 11 Neutrophil migration is controlled by gradients of chemokines such as CXCL-1, -2, -5, and -8. [28][29][30][31] We quantified CXCL1 by immunofluorescence in liver sections and found reduced levels in PAD4−/− mice compared with controls (Figures 3(A) and Figure S3(A)). Importantly, macrophage numbers did not differ significantly between WT and PAD4−/− mice ( Figure S3(B)). CXCL1 concentration in plasma was also reduced in PAD4−/− mice (Figure 3 Figure S4(B)).
To test if PAD4 controls chemokine production in human cells, we treated monocyte-derived macrophages (n = 5-6 donors) with GSK484, a PAD4 pharmacologic inhibitor. As with mouse macrophages, PAD4 inhibition significantly reduced CXCL1 (Fig-ure 4(A)) and CXCL2 (Figure 4(B)) secretion in TLR7-stimulated human macrophages (Figure 4(A)), consistent with a reported role for PAD4 in TLR7 responses in lupus. 4,27 We obtained similar results with human monocytes ( Figure S4(D)). Finally, we isolated P. falciparum trophozoite and schizont iRBCs from in vitro culture and used them to stimulate human macrophages in the presence of GSK484. PAD4 inhibition led to reduced CXCL1 secretion in response to malaria parasites, demonstrating a key role for PAD4 in regulation of this chemoattractant ( Figure 4(B)). Similar results were obtained by antibody-mediated depletion of neutrophils 22 and by neutralization of the neutrophil growth factor GCSF. 19 Neutrophils are also pathogenic in the P. berghei lung injury model. 21 In humans, NETs are specifically detected in the retinal neurovasculature of cerebral malaria patients. 19 Neutrophil infiltration is also observed in placentas of women suffering from pregnancy associated malaria. 34 Our study has several limitations. Experiments were performed in the P. chabaudi AS infection model, which only induces mild pathology, despite demonstrating true iRBC cytoadhesion. 24 It would be of value to test the role of PAD4 in more virulent models, such as P. berghei, where neutrophils also have a detrimental effect on pathology. 21 Similarly, we only analyzed neutrophil trafficking up to day 13 postinfection, making it unclear to what extent neutrophil infiltration is impaired versus delayed.
Our study suggests that targeting PAD4 may be a viable therapeutic opportunity in malaria, where adjunctive therapies to treat inflammatory pathology are badly needed. Investigation of PAD4 in other rodent malaria models, in which brain and lung inflammation are prominent, would help clarify the therapeutic potential of PAD4 inhibition. Further studies are warranted to fully explore the fundamental and complex role of PAD4 in myeloid cell biology.

ACKNOWLEDGMENTS
We thank the patients, physicians, and nurses at the Albert-Schweitzer Hospital. contributed equally to this work.

DISCLOSURE
The authors declare no conflict of interests.