Differential induction of malaria liver pathology in mice infected with Plasmodium chabaudi AS or Plasmodium berghei NK65

Background Cerebral malaria and severe anaemia are the most common deadly complications of malaria, and are often associated, both in paediatric and adult patients, with hepatopathy, whose pathogenesis is not well characterized, and sometimes also with acute respiratory distress syndrome (ARDS). Here, two species of murine malaria, the lethal Plasmodium berghei strain NK65 and self-healing Plasmodium chabaudi strain AS which differ in their ability to cause hepatopathy and/or ARDS were used to investigate the lipid alterations, oxidative damage and host immune response during the infection in relation to parasite load and accumulation of parasite products, such as haemozoin. Methods Plasma and livers of C57BL/6J mice injected with PbNK65 or PcAS infected erythrocytes were collected at different times and tested for parasitaemia, content of haemozoin and expression of tumour necrosis factor (TNF). Hepatic enzymes, antioxidant defenses and lipids content and composition were also evaluated. Results In the livers of P. berghei NK65 infected mice both parasites and haemozoin accumulated to a greater extent than in livers of P. chabaudi AS infected mice although in the latter hepatomegaly was more prominent. Hepatic enzymes and TNF were increased in both models. Moreover, in P. berghei NK65 infected mice, increased lipid peroxidation, accumulation of triglycerides, impairment of anti-oxidant enzymes and higher collagen deposition were detected. On the contrary, in P. chabaudi AS infected mice the antioxidant enzymes and the lipid content and composition were normal or even lower than uninfected controls. Conclusions This study demonstrates that in C57BL/6J mice, depending on the parasite species, malaria-induced liver pathology results in different manifestations, which may contribute to the different outcomes. In P. berghei NK65 infected mice, which concomitantly develop lethal acute respiratory distress syndrome, the liver tissue is characterized by an excess oxidative stress response and reduced antioxidant defenses while in P. chabaudi AS infected mice hepatopathy does not lead to lipid alterations or reduction of antioxidant enzymes, but rather to inflammation and cytokine burst, as shown earlier, that may favour parasite killing and clearance of the infection. These results may help understanding the different clinical profiles described in human malaria hepatopathy. Electronic supplementary material The online version of this article (10.1186/s12936-017-2159-3) contains supplementary material, which is available to authorized users.

the virulence of the parasites have to be considered [2,3]. Hepatic involvement in cases of malarial infection is a well-recognized entity. Malaria hepatocellular dysfunction is characterized by an increase in the serum levels of bilirubin and aminotransferases, exceeding three times the upper limit of normal values. Malarial hepatopathy is associated with a higher incidence of cerebral malaria, shock, acute respiratory distress syndrome (ARDS) and acute kidney injury [4]. It occurs both in adult and paediatric patients, although it is more common in adults [3,4]. One suspected cause of malarial hepatopathy is related to the activation of liver macrophages that phagocytize haemozoin or parasitized red blood cells [5]. Focal hepatocyte necrosis, cholestasis and granulomatous lesions have been documented in all malarial nodules [4]. In malaria patients with jaundice, histopathological changes in the form of damaged hepatocytes, congestion of liver cells, haemozoin deposition, inflammatory infiltrates, and cholestasis have been demonstrated [6,7], as well as hyperplastic Kupffer cells [8].
The molecular basis of different manifestations of malaria using two species of murine Plasmodium, namely Plasmodium berghei strain NK65, and Plasmodium chabaudi strain AS have been recently investigated. Plasmodium berghei NK65-infection results in lethal ARDS [9], whereas P. chabaudi AS-infection in C57Bl/6J is selfhealing after an initial peak of parasitaemia and recrudescence, and does not induce lung pathology [10,11]. Plasmodium berghei NK65 and P. chabaudi AS parasites produce different amounts of haemozoin (Hz), and it has been demonstrated that Hz is pathogenic in the lungs by inducing pulmonary inflammation [12,13]. Moreover, in mice infected by P. berghei NK65, the development of ARDS is associated with biochemical modifications and lipid alterations of lung tissue and disruption of the molecular organization and lipid composition of alveolar surfactant [12].
Excess Hz deposition was shown in the livers of P. berghei NK65-infected mice compared to P. chabaudi AS mice, while hepatomegaly due to inflammation and cell recruitment was higher in P. chabaudi AS mice. Just as in the lungs, liver dysfunction with elevated serum ALT and AST levels and production of both pro-and anti-inflammatory cytokines were associated with the amount of haemozoin present in the livers of P. chabaudi AS-infected mice. It was intriguing that inflammation, measured as cell recruitment and cytokine production, was much less pronounced in P. berghei NK65 livers that had higher amounts of haemozoin than P. chabaudi AS livers. Since haemozoin, due to its haem iron content, is strongly pro-oxidant and immunosuppressive [14,15] the present work aimed at investigating the extent of oxidative damage, the lipid alterations and the antioxidant defences present in the liver tissue of C57BL/6J mice infected with P. berghei NK65, compared to mice infected with P. chabaudi AS.

Animals
The infection of C57BL/6J mice were performed by intraperitoneal injection of 10 4 P. berghei NK65 or P. chabaudi AS infected red blood cells, as described [9,12]. Three different experiments were performed; 6 or 8 mice/group (control and infected mice) were used for each time point and used at day 8 or 10 after infection. The time intervals and the dose of parasites were chosen based on previous experience. All the data were in line with and confirmed our previous study [16]. Mice were euthanized by intraperitoneal injection of a lethal dose of sodium pentobarbital (200 µl at 60 mg/ml Nembutal) at day 8 or 10 post infection and blood was removed by heart puncture and centrifuged to separate plasma and RBC.

Preparation of biological specimens
Mice were perfused with 0.15 M NaCl containing 0.2 mM butylhydroxytoluene (BHT) as an antioxidant, and livers were removed, weighted and homogenized in Precellys tubes in six volumes of a solution containing 20 mM Tricine, 250 mM sucrose, 5 mM EDTA, 0.2 mM BHT (pH 7.4) and a protease inhibitor cocktail (Sigma, Milan, Italy).

Haemozoin determination
The content of Hz in the perfused livers was analysed by haem-enhanced chemoluminescence determination according to Deroost et al. [13,17]. The amount of Hz (pmol/mg tissue) was multiplied by the total weight of the liver and expressed as pmol or nmol haematin/organ.

Lipid analyses
The extraction of lipids in the liver was performed according to Folch [18]. Dried lipid extracts were dissolved in chloroform and saved at − 80 °C until further analysis. The content of phospholipid (PL) phosphorus was determined by the Bartlett procedure [19]. Cholesterol (Cho), triglycerides (TG) and cholesterol esters (ChoE) were quantified by densitometric analysis (Camag Reprostar 3) after separation by HPTLC in hexane/ diethyl ether/acetic acid (90:10:1 by vol.). A solution of 10% CuSO 4 in 8% H 3 PO 4 was used to visualize the spots. The fatty acid composition of the lipid fractions was analysed by gas liquid chromatography according to Corsetto et al. [20]. The levels of thiobarbituric acid reactive substances were used to determinate the lipid peroxidation, as previously reported [21] and expressed as pmoles of malondialdehyde (MDA).

Antioxidant enzymes and other analyses
Antioxidant enzymes and total glutathione (GSH + GSSG) were determined in the supernatant of liver homogenates obtained after a centrifugation at 12000×g for 10 min. Protein concentration was determined by the Bradford assay (Bio-Rad, Hercules, CA) [22]. Catalase (CAT), glutathione reductase (GR) and superoxide dismutase (SOD) activity was determined according to Aebi et al. [23], Pinto et al. [24] and by a assay kit (Cayman Michigan, USA), respectively. Total GSH was determined spectrophotometrically by titration with 5-5′-dithiobis-2 nitro benzoic acid (DTNB) according to Beutler [25]. The collagen content of livers, was determined by measuring hydroxyproline in the tissue homogenates [26]. Liver damage was determined by measuring plasma levels of serum alanine transaminase (ALT) and aspartate transaminase (AST) enzymes, according to manufacturer's protocol (Teco Diagnostics, California, USA).

Quantitative reverse transcription-polymerase chain reaction
The liver homogenates were used to extract and quantify total RNA and to synthetize cDNA. A quantitative PCR was performed on 25 and 12.5 ng cDNA with primer and probe sets from Applied Biosystems according to Deroost et al. [13]. Data were normalized to 18S ribosomal RNA levels.

Statistical analysis
The data are reported as mean ± standard deviation (SD). ANOVA test followed by Tukey's multiple comparison test with GraphPad Prism 6.0 (Graphpad Software Inc., La Jolla, CA, USA) were performed to compare the groups and replicates. The differences were considered significant when p < 0.05. To eliminate possible artefacts (over-or under-estimation) due to hepatomegaly and hepatic oedema, all the biochemical parameters have been expressed as total content/liver.

Malaria infection in mice
C57BL/6J mice were infected intraperitoneally with 10 4 infected RBC (either P. berghei NK65 or P. chabaudi AS) and the development of parasitaemia was monitored every day whereas the liver pathology was checked on day 8 and 10 after the start of infection. These time intervals and the parasite dose were chosen on the basis of the indications from previous experiments [9]. Parasitaemia was lower in P. chabaudi AS compared to P. berghei NK65 infected mice at both time points (Fig. 1a), whereas hepatomegaly was evident in P. chabaudi AS, but not in P. berghei NK65-infected mice, corroborating previous data [16] (Fig. 1b). Similarly, elevated levels of AST and ALT were detected in the serum of infected mice at both time points (Additional file 1, Panel A and B). Of notice is the observation that in P. berghei NK65 infected mice, AST levels were particularly increased over ALT levels, leading to an AST/ALT ratio of 3.9 and 5.9 at day 8 and 10, respectively. Differently, in P. chabaudi AS infected mice the AST/ALT ratios was 1.5 and 1.2 at the same time intervals (Additional file 1, Panel C). These findings are consistent with the levels of parasitaemia in P. berghei NK65 mice since elevated AST levels may also reflect haemolysis in addition to liver damage [27]. Previous reports in a larger set of experiments confirm these observations. [16].
Differently from P. chabaudi AS-infected mice, livers from P. berghei NK65-infected mice also showed a darkbrown coloration in agreement with the significantly higher hepatic Hz deposition (Fig. 2a). In parallel with these modifications, also the proinflammatory status of the liver changed during the infections with the different Plasmodium species. The mRNA levels of TNF increased during the progression of the disease in the livers of both P. berghei NK65 and P. chabaudi AS infected mice (Fig. 2b), but with a different kinetic and values. Collagen deposition, measured as hydroxyproline content, was significantly higher with both parasite species at day 8 post infection with the highest levels in P. berghei NK65 livers. At day 10, the values returned to normal in both groups of mice, as those of control (Fig. 2c). Lipid peroxidation, expressed as total pmoles MDA/liver, was found to be significantly and time-dependently increased in mice infected with both Plasmodium species showing higher levels in P. berghei NK65 infected mice at either day post infection (Fig. 2d).

Oxidative stress response in the liver
The above findings, namely the increased levels of Hz and increased levels of MDA suggestive of oxidative stress in P. berghei NK65-livers, prompted us to investigate the status of the antioxidant defence systems in the infected livers as a determinant of tissue susceptibility to oxidative damage. The pattern of response in the two groups of infected mice was again different. As expected from the higher levels of lipoperoxidation, the liver extracts of P. berghei NK65 infected mice were characterized by a lower CAT activity and lower levels of total glutathione (GSH + GSSG) at both days post infection (Fig. 3a, b). SOD activity was significantly enhanced at day 10 post infection, whereas GR was not different from uninfected mice (Fig. 3c, d). On the contrary, P. chabaudi AS-infected mice showed CAT and SOD activity and GSH levels similar to control at both days post infection, whereas GR activity was higher at both time points (Fig. 3d).

Species-specific lipid changes in infected livers
Liver dysfunction is generally associated with lipid alterations. The lipid profile of infected livers was significantly Fig. 2 Hepatic haemozoin levels, oxidative damage and fibrosis in liver tissue of P. berghei NK65 or P. chabaudi AS infected mice. Hz content in liver tissue (pmol Hz/mg liver tissue) of P. berghei NK65 or P. chabaudi AS infected C57Bl/6J mice at day 8 and 10 post infection (a). Hepatic mRNA expression levels of TNF-α in non-infected and infected mice at day 8 or day 10 post infection determined by quantitative RT-PCR and normalized to 18S ribosomal RNA levels (b). OH-proline content (μg/mg protein) in non-infected and infected mice at day 8 or day 10 post infection (c); MDA content (pmol/liver) in non-infected and infected mice at day 8 or day 10 post infection (d). n = 6 mice/for each time point and group. *p < 0.05; **p < 0.01; ***p < 0.0001 vs Control; °°p < 0.05; °°°p < 0.01 P. berghei NK65 vs P. chabaudi AS modified compared to control mice and different between the two species. TG levels were significantly higher in the P. berghei NK65-compared to P. chabaudi AS-infected livers (Fig. 4a). In addition, higher levels of ChoE, and lower levels of free Cho were seen with PbNK65 at both day 8 and 10 post infection, compared to control mice, whereas P. chabaudi AS-infected mice had higher levels of free cholesterol (Fig. 4b, c). Compared to control mice, PL were more strongly increased in P. chabaudi ASinfected livers than in P. berghei NK65 at both days post infection (Fig. 4d), whereas no differences were found in the composition of the PLs due to infection (Fig. 5).
Modification in the fatty acid distribution between P. berghei NK65 vs P. chabaudi AS livers was also observed. Palmitic acid was significantly increased in P. chabaudi AS livers compared to control, whereas this was decreased in P. berghei NK65 together with palmitoleic acid. Among the polyunsaturated fatty acids, significantly higher amounts of linoleic acid (C18:2 n − 6) and lower levels of eicosatrienoic (C20:3 n − 6) and arachidonic (C20:4 n − 6) acids were seen in P. berghei NK65 compared to control (Table 1). These differences were less pronounced in P. chabaudi AS livers and limited to the changes of arachidonic acid (C20:4 n − 6) and docosapentaenoic acid (C22:5 n − 3). As a consequence of these changes, also the C18:2/C20:4 ratio was significantly increased in P. berghei NK65 compared to control or P. chabaudi AS.

Discussion
The pathogenesis of hepatic dysfunction in malaria is complex and not completely understood. Histopathological findings show congested and swollen hepatocytes, Kupffer cell hyperplasia, deposition of brown malarial pigment, mononuclear cell infiltration, micro-occlusion by parasitized RBC and, less frequently, steatosis and spotty and submassive necrosis [2,7,13,16,28]. Infection, inflammation and oxidative stress have been shown to induce multiple alterations in hepatic lipid and lipoprotein metabolism [29]. The present study demonstrates that blood stage malaria infection by P. chabaudi AS or P. berghei NK65 exerts different effects on the liver of C57BL/6J mice. Hepatomegaly was confirmed in P. chabaudi AS mice, as previously shown [16] as well as liver dysfunction suggested by the serum increase of the hepatic enzymes AST and ALT, especially at day 10 post infection. The high AST levels lead to an AST/ALT ratio significantly higher in P. berghei NK65 compared to P. chabaudi AS. The AST/ALT index is used to discriminate the type of liver damage observed [27,30] and a progressive increase in the AST/ALT ratio correlated with a decrease in liver functions, like in cirrhosis and fibrosis Fig. 3 Enzymes activity in the liver of C57Bl/6J mice infected by P. berghei NK65 or P. chabaudi AS. Catalase activity (a), total GSH (b), SOD activity (c) and glutathione reductase activity (d) were determined in the livers of non-infected and infected mice at day 8 and day 10 post infection. n = 8 mice/group. Data are from one representative experiment out of three performed in the same conditions; *p < 0.05; **p < 0.01; ***p < 0.0001 vs control [31]. Since plasma clearance of AST is modulated by the activity of sinusoidal liver cells, during progressive fibrosis and cirrhosis, the functions of these cells are progressively impaired resulting in a relative increase in AST levels [32]. In the case of malaria, in P. berghei NK65infected mice, an alternative hypothesis can be proposed. Elevated AST levels in absence of elevated ALT, in fact, may indicate haemolysis, since RBC contain AST. This has already been reported in sickle cell disease where the AST/ALT ratio has been used as haemolytic index [33]. In the models used for this work, the AST/ALT ratio may well indicate initial fibrosis in both species, as confirmed by the OH-proline increase at day 8 post infection, and increased haemolysis in P. berghei NK65-infected mice that develop higher parasitaemia than P. chabaudi AS mice. The latter leads to higher levels of Hz in the liver of P. berghei NK65-than in P. chabaudi AS-infected mice [16]. It is well known that Hz plays a crucial role in generating oxidative damage [12,14,15]. The higher content of Hz, free (pro-oxidant) haem derived from intravascular haemolysis and high parasitaemia are likely to contribute to excessive oxidative stress response in P. berghei NK65 livers. TNF may also participate. A marginal role seems to be played by the inflammatory infiltrate, which is much less abundant in P. berghei NK65-infected livers than in P. chabaudi AS mice, in which increased liver Fig. 4 Alteration in the lipid profiles of the livers from uninfected, P. berghei NK65 or P. chabaudi AS infected mice. Mice infected with P. berghei NK65 or P. chabaudi AS were dissected at day 8 and 10 post infection. The total content of liver triglycerides (TG) and phospholipids (PL) were determined as described in "Methods" (a and c). The content of ChoE and free Cho is shown in b and d, respectively. n = 8 mice for each condition. Data are from one representative experiment out of three performed in the same conditions *p < 0.05; **p < 0.01; ***p < 0.0001 vs Control; °p < 0.05; °°°p < 0.001 P. berghei NK65 vs P. chabaudi AS Fig. 5 Phospholipids composition in liver from uninfected, P. berghei NK65 or P. chabaudi AS infected mice. The distribution of liver PL (µg PL/mg protein) was determined in uninfected and in P. berghei NK65or P. chabaudi AS-infected mice at day 8 or 10 post infection. Data are from one representative experiment out of three performed in the same conditions. LPC lysophosphatidylcholine, SM sphingomyelin, PC phosphatidylcholine, PS phosphatidylserine, PI phosphatidylinositol, PE phosphatidylethanolamine, CD cardiolipine weight, histology and the correlation between hepatic inflammation, enzymes and haemozoin were demonstrated [16].
The higher production of ROS in P. berghei NK65infected mice was confirmed by the elevated concentration of MDA, which is an index for the loss of structure and cell membrane integrity, and the concomitant depletion of the antioxidant defence system, such as CAT and total GSH. GSH constitutes the first line of defence against free radicals and is a critical determinant of tissue susceptibility to oxidative damage. The increase of SOD, shown also in other models of liver damage [34], reveals the necessity of minimizing oxidative stress, which largely derives from the production of superoxide anion radicals. These changes seems to be associated with P. berghei NK65 infection, since they were not present in P. chabaudi AS-infected mice that showed only an increase of GR at both days post infection.
An important change observed in P. berghei NK65 vs P. chabaudi AS mice was the augmentation of the liver content of TG and cholesterol esters that indicates a potent stimulation of hepatic lipogenesis by P. berghei parasites [35]. This is in agreement with literature data showing that livers from P. berghei-infected mice contain lipid droplets and myelin-like figures [36]. A significant, although limited increase of TG was also present in P. chabaudi AS infected mice in agreement to the small neutral lipid inclusions observed by Seixas et al. [37]. Elevated TG are a characteristic of liver pathologies of different aetiology, clinically defined as non-alcoholic fatty liver disease in which TG accumulate and plasma albumin and protein decrease.
On the contrary, higher amounts of PL and free Cho were present in P. chabaudi AS-infected mice compared to contr or P. berghei NK65-infected mice. This is possibly related to reticuloendothelial hyperplasia and/or cholestasis, similar to those reported in the same or other models of malaria infection [7,38]. The fatty acid distribution in the liver tissue of P. chabaudi AS or P. berghei NK65 infected mice at different days after the infection confirms the impairment (possibly induced by ROS) of the elongation/desaturation pathway from linoleic to arachidonic acid and may explain the higher linoleic/arachidonic acid ratio in the liver of P. berghei NK65-infected mice [39]. These observations also contribute to explain the changes in the lipid content and composition that was reported in the lungs of P. berghei NK65 infected mice [12].
All together, these findings are in agreement with other models of P. berghei infection and support the hypothesis that ROS may play a critical role in the liver dysfunction caused by P. berghei NK65 parasites [35,40]. Consistent with this hypothesis are also the signs of fibrosis, revealed by the elevated hepatic levels of hydroxyproline at day 8 post infection in P. berghei NK65-infected mice. Indeed, it is well known that oxidative stress is one of the major stimuli of fibrosis [41]. Excess of ROS may induce severe liver damage followed by a phase of repair [42] during which TGF-β1 on one hand limits the proliferative response of hepatocytes, and on the other hand, increases the production of collagen and other extracellular matrix proteins. Submicroscopical signs of fibrosis were evident also in P. chabaudi AS-infected mice, although at a lower extent. Interestingly, at day 10 post infection, the amount of hydroxyproline decreased with both parasite species, becoming similar to control mice. The hypothesis that inflammation may lead to the activation of matrix metalloproteinases (MMPs) is under investigation. MMPs are proteolytic enzymes able to degrade different proteins of the extracellular matrix (e.g. collagen, laminin, fibronectin) and modulate cytokine and chemokine activity in cases of severe inflammation. Previous data indicated that, in the liver of C57Bl/6J mice infected with P. berghei ANKA, the activity of MMPs is significantly increased [43]. Furthermore, several in vitro studies have shown a TNF-dependent induction of MMPs at both mRNA and protein levels in monocytes fed with natural Hz [16,44].

Conclusion
In conclusion, this study demonstrates that, in C57BL/6J mice, P. berghei NK65 and P. chabaudi AS parasite species induce liver pathology in a different manner. In particular, in P. berghei NK65 infected mice, which succumb of malaria ARDS, it seems that an excess oxidative response predominates, as demonstrated by the alteration of the liver anti-oxidant enzymes, the increased MDA levels and the accumulation of TG. These changes are related to the high parasitaemia, the liver Hz deposition and the modification in protein content and lipid profile. These alterations are less pronounced with P. chabaudi AS, which develops a self-limiting infection, with lower parasitaemia and Hz deposition, and thus less oxidative stress. The cell infiltrate and the cytokine burst occurred at the peak of the infection may suggest an inflammatory liver response, which precedes parasite clearance [16].