Liver changes in severe Plasmodium falciparum malaria: histopathology, apoptosis and nuclear factor kappa B expression

Specimens were classified into two groups according to the level of total bilirubin (TB): TB ≥ 51.3 μmol/L (3 mg/dl) (12 cases) and TB < 51.3 μmol/L (10 cases), based on laboratory data. Liver specimens with normal histology obtained from fatal accident cases (10 cases), served as the control group. The specimens were obtained from autopsied cases at the Department of Tropical Pathology, Faculty of Tropical Medicine, Mahidol University, Bangkok, Thailand. Control cases came from the same region. Patients with hepatitis B co-infection, and patients with glucose-6-phosphate dehydrogenase (G6PD) deficiency, were excluded from the study. The use of left-over liver specimens and the study protocol were reviewed and approved by the Ethics Committee of the Faculty of Tropical Medicine, Mahidol University (MUTM 2011-025-01 and MUTM 2011-025-02).

Left-over liver tissues in paraffin-embedded blocks were re-embedded with new paraffin medium using standard histological techniques. Liver tissues were sectioned at 4 μm thickness and placed onto glass slides for histopathological examination and onto positively charged slides for immunohistochemistry study against cleaved caspase-3 and NF-κB p65.

Overall liver histopathology was classified based on the severity of six histological criteria, namely fatty change, hyperplastic Kupffer cells, portal tract inflammation, bile duct proliferation, sinusoidal congestion and haemozoin deposition. Quantification was performed under high (400x) magnification. The severity level of each histopathological change was graded on a scale from 0 to 3, according to semi-quantitative assessment (Table 1). The highest possible total score was 18 (6 histological criteria x 3 as highest scale). Score 0 meant no histopathological change and score 18 referred to most severe histopathological change. Microscopic evaluation was done by two observers (PV and CP). A third person was asked to score the histopathological changes if more than 2 grading variations were observed.

The 4 μm paraffin sections were deparaffinized and rehydrated by sequential immersion in a graded series of alcohol, then transferred into water for 5 mins. To inhibit endogenous peroxidase activity, the sections were incubated with 3% H₂O₂ for 5 mins. The sections were then heated in a microwave oven (in 0.1 M sodium citrate buffer, pH 6.0 for cleaved caspase-3 and 0.1 M Tris–HCl buffer, pH 9.0 for NF-κB p65) for 20 mins for epitope retrieval. After washing with phosphate buffered saline (PBS), pH 7.4, sections were incubated for 30 mins with normal serum as blocking solution to reduce the non-specific background, then cooled to room temperature while still immersed in buffer. The following protocol was realized using avidin-biotin alkaline phosphatase complex (VECTASTAIN® ABC-AP kit (Rabbit IgG) # AK-5001) for cleaved caspase 3 and avidin-biotin peroxidase complex (VECTASTAIN® ABC kit (Mouse IgG) # PK-4002) for NF-κB p65 (Vector Laboratories, Inc., USA) according to the manufacturer’s protocol. The sections were incubated with primary antibody; rabbit polyclonal anti-cleaved caspase-3 (Asp175) antibody (1:200 dilution) (Cell Signaling Technology, USA) and mouse monoclonal anti-NF-κB p65 (1:50 dilution) (Santa Cruz Biotechnology Inc., Santa Cruz, CA) and incubated overnight at 4˚C in a humidity chamber. The following day, sections were washed three times with PBS, and incubated with anti-mouse/rabbit biotinylated secondary antibody (Vector Laboratories, Inc., USA) for 30 mins at room temperature and reacted with avidin-biotin complex (ABC) conjugated with alkaline phosphatase (AP)/horseradish peroxidase (HRP) (Vector Laboratories, Inc., USA) for 30 mins. After washing, enzyme activity was visualized by Vector® Red substrate kit (Vector Laboratories, Inc., USA) for cleaved caspase-3, resulting in the formation of a red colour at the antigen sites and by 3,3′- diaminobenzidine (DAB) (Vector Laboratories, Inc., USA) for NF-κB p65, presenting as a brown colour. Subsequently, sections were counterstained with haematoxylin for 1 min, and mounted with a coverslip.

The presence of immunopositive cells for cleaved caspase-3 and NF-κB p65 was recorded as percentages. In addition, immunostaining intensity was scored from 0–3 (0-negative staining, 1- mild, 2- moderate, and 3- strong immunostaining). Total score was calculated by multiplying percentage immunopositive cells and intensity, a method used by Punsawad et al. 2013[13].

Ten representative views of H&E-stained liver sections were randomly photographed at 400x magnification using an Olympus Bx41 light microscope (Olympus, Tokyo, Japan) connected to an Olympus DP20 digital camera (Olympus, Tokyo, Japan). Kupffer cell length was measured with the UTHSCSA Image Tool program (developed at the University of Texas Health Science Center at San Antonio, TX; freely available from the Internet).

Data were expressed as mean ± standard error of the mean (SEM). The normality of distribution was determined by the Kolmogorov-Smirnov test. Differences between groups were analyzed by Mann Whitney U- test. In addition, the correlations of each variable within groups and pertinent clinical data were calculated using Spearman’s rank correlation (r _s ). Statistical analysis was performed using SPSS version 17.0 software (SPSS, IL, USA). A p value < 0.05 was considered significantly different.

Twenty-two liver specimens were collected from P. falciparum malaria cases, consisting of 12 cases with hyperbilirubinaemia (total bilirubin (TB) ≥ 51.3 μmol/L or 3 mg/dl) and 10 cases without hyperbilirubinaemia (TB < 51.3 μmol/L). Ten cases with normal liver histopathology served as controls. Table 2 summarizes the mean age, sex, days of fever pre-admission, parasitaemia and important liver function tests of the malaria patients. Significant differences between two groups were noted in the levels of albumin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase, total and direct bilirubin (p value < 0.05). Common associated complications were cerebral malaria (70.00% for non-hyperbilirubinaemia and 33.33% for hyperbilirubinaemia groups) and acute kidney injury (33.33% for hyperbilirubinaemia group).

Figure 1 displays histopathological changes of the liver in severe P. falciparum malaria. Morphologically, hepatocytes and endothelial cells (ECs) in the liver were generally unaffected in severe P. falciparum malaria. Fatty change and bile duct proliferation were not features of P. falciparum infection. Hyperplastic Kupffer cells, portal tract inflammation, sinusoidal congestion and haemozoin pigment deposition were important pathological hallmarks related to higher TB levels (Table 3). The total histopathological grading score showed the overall changes and was found to be highest among the group of malaria patients with hyperbilirubinaemia (12/18). Figure 1 (A-C) illustrates a normal portal tract, consisting of hepatic vein, hepatic artery, and bile duct, with surrounding hepatocytes. Very few inflammatory cells are noted within the portal tract. Minimal fatty change is sometimes observed (Figure 1C). Liver tissues from malaria patient without hyperbilirubinaemia show enlarged sinusoidal area, haemozoin pigment within the hyperplastic Kupffer cells, and inflammatory cells within the portal tract (Figure 1, D-F). In hyperbilirubinaemia group (Figure 1, G-I), liver tissues show dense inflammatory cells infiltration in the portal tract. At higher magnification, hyperplastic Kupffer cells are visible and contain haemozoin pigment. Sinusoidal areas are often congested. Central vein contains numerous PRBCs.

Fatty change, hyperplastic Kupffer cells and portal tract inflammation were quantified in each group (Table 4). There was no significant difference between fatty changes in the livers of the P. falciparum malaria groups and the normal controls (all p > 0.05). However, hyperplastic Kupffer cells and portal tract inflammation were significantly higher in the malaria groups compared with the normal controls and highest in the livers of the P. falciparum patients with hyperbilirubinaemia. The number of hyperplastic Kupffer cells were increased in the malaria group with hyperbilirubinaemia (52.21 ± 2.32/high power field (HPF)), compared with the normal liver group (9.65 ± 0.67/HPF) (p < 0.001) and the malaria group without hyperbilirubinaemia (32.05 ± 3.34/HPF) (p < 0.001). Generally, Kupffer cell contains packed haemozoin pigment within the cytoplasm. Positive correlations were evident between TB and the number of Kupffer cells (r _s  = 0.551, p = 0.018) and between TB and % lymphocytes in the portal tract (r _s  = 0.743. p = 0.020). In terms of Kupffer-cell length (Figure 2), the size of Kupffer cells as measured and analysed by Image Tool software showed significant numbers of reactive Kupffer cells in the malaria group with hyperbilirubinaemia (more than three times), compared with the normal liver group (p < 0.001) and the non-hyperbilirubinaemia group (p < 0.001).

Apoptosis was evaluated using monoclonal antibody against cleaved caspase-3, the final apoptotic pathway. The occurrence of apoptosis in the hepatocytes was negligible even in the severe P. falciparum malaria with hyperbilirubinaemia group. Figure 3 A-C illustrates the occurrence of apoptosis in the liver tissues of severe P. falciparum malaria cases, compared with the normal controls. Percentage apoptosis for Kupffer cells and lymphocytes within the portal tracts is shown in Figure 4. The data show that both Kupffer cell and lymphocyte apoptosis were significantly increased in group with severe P. falciparum malaria with hyperbilirubinaemia compared with the non-hyperbilirubinaemia groups (p = 0.030 and p = 0.009, respectively) and the normal controls (all p < 0.001) (Figure4A and B). The Kupffer cells and lymphocytes in the non-hyperbilirubinaemia group also expressed significantly higher levels of apoptosis than the normal controls (p = 0.005 and p = 0.073, respectively). Since immunostaining intensity is an important factor in evaluating the degree of protein expression, a total score incorporating percentage apoptosis and immunostaining intensity was also used to compare apoptosis in the liver. The total scores for Kupffer cell and lymphocyte apoptosis are shown in Figure 4. A similar trend to the percentage data was observed.

NF-κB p65 expression and apoptosis in the livers of severe P. falciparum malaria were investigated. NF-κB p65 expression in Kupffer cells and portal lymphocytes is shown in Figure 5. In both Kupffer cells and lymphocytes, NF-κB p65 expression was significantly increased in the group with severe P. falciparum malaria with hyperbilirubinaemia, compared with non-hyperbilirubinaemia group (p = 0.030 and p = 0.009, respectively) and normal controls (all p < 0.001). The malaria group without hyperbilirubinaemia showed significantly higher levels of NF-κB p65 expression in the Kupffer cells and lymphocytes than the normal controls (p = 0.005 and p = 0.007, respectively).

A significant positive correlation was found between cleaved caspase-3 and NF-κB p65 immunopositive cells in Kupffer cells (r _s  = 0.713, p < 0.001), and lymphocytes in the portal tracts (r _s  = 0.741, p < 0.001) (Figure 6). Both the expression of cleaved caspase-3 and NF-κB p65 were positively correlated with TB (Kupffer cells: r _s  = 0.707, p = 0.001 and r _s  = 0.853, p < 0.001, respectively) (lymphocytes: r _s  = 0.490, p = 0.039 and r _s  = 0.636, p = 0.008, respectively). Hyperplastic Kupffer cells were positively correlated with cleaved caspase-3 (r _s  = 0.617, p = 0.001) and NF-κB p65 (r _s  = 0.568, p = 0.001) expression. The degree of portal tract inflammation was also positively correlated with cleaved caspase-3 (r _s  = 0.612, p = 0.001) and NF-κB p65 (r _s  = 0.519, p = 0.002) expression. On the other hand, no significant association was seen between both NF-κB p65 and cleaved caspase-3 expression of Kupffer cells and portal lymphocytes, and clinical data (age, sex, days of fever pre-admission, parasitaemia and important liver function tests.

The present study shows a rise in liver transaminases and alkaline phosphatase in the malaria group with hyperbilirubinaemia (TB ≥ 51.3 μmol/L). Clinical jaundice in P. falciparum can be caused by several factors, i.e. intravascular haemolysis from parasitized red blood cells (PRBCs), G6PD deficiency-related haemolysis or anti-malarial drugs, disseminated intravascular coagulation (DIC) or co-existing septicaemia-induced hepatitis[14, 15]. The histopathology of the liver in severe malaria has been previously studied. However, the present study demonstrated certain morphological variations in the liver from other reports, such as an abundant chronic inflammatory cell response and an absence of liver cell necrosis. Liver changes in severe malaria often include hyperplastic Kupffer cells[4, 5, 16‐18], fatty change[16, 17], portal tract inflammation[17], cholestasis[16, 17], liver cell necrosis[4, 16, 18], sequestration of PRBCs and the deposition of haemozoin pigment[4, 5, 16, 18]. The present study documented hyperplastic Kupffer cells with scattered haemozoin deposition and portal inflammation as the most common histological changes in the livers of severe P. falciparum malaria cases. The enlarged Kupffer cells were confirmed quantitatively using Image Tool software. The immune response in the liver to PRBCs primarily involves the activation of Kupffer cells. The recruitment and activation of Kupffer cells and macrophages in the spleen and bone marrow are important for the clearance of malaria parasites[19]. Portal tract inflammation consists mainly of lymphocytes and a few plasma cells (Table 4), in contrast with mild inflammation (portal and lobular lymphocytic infiltrates) reported earlier[5]. An acute inflammatory process involving neutrophils is not seen. The minimal fatty change noted here was similar to a previous finding[5]. Liver cell necrosis in P. falciparum was not a striking finding in this study. It has also been reported as a rare event in some other studies[5, 18]. However, the incidence of hepatic necrosis may be as high as 41%[4] and severe cases of centrizonal necrosis have been documented[18]. This change has been reported to be secondary to suppression of bilirubin excretion by PRBCs or metabolic acidosis rather than hepatitis per se[14]. Sequestration of PRBCs is a common finding and depends on malaria parasite load.

Among various cells evaluated in the liver tissue, Kupffer cells and inflammatory cells show significant apoptotic changes in severe P. falciparum malaria. Hepatocytes, bile ducts and ECs failed to show significant apoptotic changes. P. falciparum has been shown to induce apoptosis in human cells, such as in lymphocytes[13, 20‐22], neurons[13], glial cells[13], brain ECs[13] and lung ECs[23]. This may be responsible for clinical manifestations and progression to severe disease. In animal models, malaria-induced apoptosis was evident in astrocytes[24], lymphocytes[25], liver and spleen[8, 25]. During the liver stage, however, hepatocytes are usually spared from the apoptotic process, allowing merozoites to be released into the circulation, suggesting that malaria sporozoites can block pro-apoptotic pathways[26]. The present study focuses on liver changes in the erythrocytic stage, which is far beyond the liver stage of parasite development. Nevertheless, hepatocytes remain morphologically unaffected and protected from apoptosis. Hepatocytes are defended by a barrier of Kupffer cells, endothelial cells, stellate cells, space of Disse (perisinusoidal space) and Kupffer cells. Moreover, PRBCs localized within the sinusoidal area are in close contact with Kupffer cells, the first-line immune defense in the liver. Recognized as foreign bodies, PRBCs and haemozoin are primarily engulfed by the Kupffer cells. In contrast, apoptosis in the hepatocytes has been reported in animal models, linked to activation of the mitochondrial pathway, release of reactive oxygen species[7, 8] and induction by glycosylphophatidylinositol (GPI), a major membrane-associated protein of P. falciparum[27].

Apoptosis in the Kupffer cells was evidenced by strong caspase-3 expression. The loaded haemozoin within the cytoplasm of Kupffer cells can be toxic to these immune cells. In humans, during the erythrocytic stage, malaria parasites degrade haemoglobin to produce haem and haemozoin which are harmful. Haemozoin can be deposited in the liver and primarily phagocytized by Kupffer cells, where it can induce oxidative stress[7], a possible mechanism for the induction of apoptosis in Kupffer cells. During the liver stage, Kupffer cell apoptosis has also been detected in a murine model after incubation with Plasmodium yoelii sporozoites[9]. Apoptosis in malaria is reportedly mediated by the Fas-ligand in lymphocytes[20] and in murine astrocytes[24].

NF-κB has been shown to regulate various cellular processes such as inflammation, immunity, cell proliferation and apoptosis[28]. A previous study documented NF-κB activation and pro-inflammatory response in human brain ECs exposed to PRBCs[29]. A recent study of human brain tissues has demonstrated that NF-κB is one of the signaling molecules that modulates apoptosis in brain ECs and intravascular leukocytes in fatal cerebral malaria[13]. The present study documented NF-κB mediating apoptosis in Kupffer cells and lymphocytes within the portal tract in severe P. falciparum infection.