Exploring pancreatic pathology in Plasmodium falciparum malaria patients

Hypoglycaemia is an important complication of Plasmodium falciparum malaria infection, which can be lethal if not treated. A decrease in blood sugar (BS) level has been correlated with disease severity, parasitaemia and the use of certain antimalarial drugs. This study explored the relationship between pancreatic pathology, including the expressions of insulin and glucagon in the islets of Langerhans, and the BS levels in P. falciparum malaria patients. Pancreatic tissues from malaria patients were divided into three groups, namely those with BS < 40 mg/dl, BS = 40–120 mg/dl, and BS > 120 mg/dl. In P. falciparum malaria, pancreatic tissues showed numerous parasitised red blood cells (PRBCs) in the capillaries, oedema, acinar necrosis and the presence of inflammatory cells. The islet size and the expression of insulin were significantly increased in P. falciparum malaria patients with hypoglycaemia. In addition, insulin expression was positively correlated with islet size and negatively correlated with BS levels. This pioneer study documents an increase in insulin expression and an increase in islet size in hypoglycaemic patients with P. falciparum malaria. This could contribute to the pathogenesis of hypoglycaemia and provides evidence for the potential need to effectively manage the hypoglycaemia seen in malaria infection.

pathology in P. falciparum malaria and to investigate the expressions of insulin and glucagon in the islets of Langerhans in parallel.
Histopathological changes of pancreatic tissues in P. falciparum malaria patients. Pancreatic tissues from P. falciparum malaria patients usually showed oedema, acinar necrosis and inflammatory reactions, in addition to the presence of parasitised red blood cells (PRBCs). However, there was no difference in the overall total score of histopathological changes between malaria patients and control group (P = 0.392). The histopathology of pancreatic tissues from malaria patients, compared to normal pancreatic tissues is illustrated in Fig. 1.
Normal histology of the pancreas shows intact islets of Langerhans and acinar cells forming exocrine glands (Fig. 1A). The number of pancreatic islets in malaria patients is similar to the control group (1-2 cells/mm 2 ). Numerous PRBCs were frequently seen in the capillaries of pancreas from severe malaria patients (Fig. 1B).
Scientific REPORTS | (2018) 8:10456 | DOI:10.1038/s41598-018-28797-w Oedema was characterised by widening of the pancreatic interlobular and interglandular spaces (Fig. 1C). There was a significant difference between pancreatic oedema in malaria patients and the control group (P = 0.002), especially in malaria patients with BS < 40 mg/dl (P = 0.004) and those with BS = 40-120 mg/dl (P = 0.004). Necrosis of acinar cells is considered in pale staining cells with the loss of nuclei and visible cellular distortion, causing deformation in pancreatic parenchyma (Fig. 1D). Acinar necrosis was evident in all groups of malaria patients compared to the control group (P < 0.001). The presence of inflammatory cells, particularly lymphocytes, neutrophils and occasionally eosinophils in the interlobular septum and interglandular areas were significantly higher in all malaria patients compared to the control group (P = 0.011) (Fig. 1E). Fibrosis was not observed in the malarial pancreas. Areas of haemorrhage in pancreatic parenchyma and fat necrosis in peripancreatic tissue were occasionally seen in malaria patients compared to normal pancreas (P = 0.196 and P = 0.067, respectively). The quantification of total histopathological changes in the pancreatic tissues of malaria patients compared with the control group are shown in Table 1.
Evaluation of pancreatic islet size. The comparative size of islets of Langerhans is depicted in Fig. 2A.

Expressions of insulin and glucagon.
Insulin-and glucagon-expressing cells were detected as brown color in the cytoplasm of the cells in the islets of Langerhans. Generally, insulin expression was demonstrated in the center of the pancreatic islets, whereas glucagon expression was located at the periphery of the islets (control, Fig. 3A,B). In this study, insulin expression was prominent in the group of P. falciparum malaria patients with hypoglycaemia (Fig. 3C), and those with BS = 40-120 mg/dl (Fig. 3E), compared to the control group (P = 0.014 and P = 0.011, respectively). A trend towards a reduction in the expression of insulin was revealed in malaria with hyperglycaemia (Fig. 3G). For glucagon expression, weak positive staining was observed in malaria patients with  Table 1. Quantification of pancreatic histopathology in P. falciparum malaria patients compared to the control group. a Compared between malaria patients. b Significant difference between malaria patients and control group.  Fig. 4. There was a significant negative correlation between BS levels and the total insulin expression score (r s = −0.728; P = 0.002) (Fig. 5A). In addition, insulin expression was directly correlated with islet size (r s = 0.696; P = 0.004) (Fig. 5B).

Discussion
Histopathological changes documented in pancreatic tissues of P. falciparum malaria patients include oedema in the interlobular and interglandular spaces, acinar necrosis and inflammatory reactions. Although these changes were frequently found in P. falciparum malaria patients with hypoglycaemia (oedema and necrosis), similar histopathological changes were reported in other infectious diseases including dengue infection 11 , leptospirosis 10 and babesia 9 . Limited studies have been reported on the pathology of pancreatic tissues in severe P. falciparum malaria. Previous rare autopsy findings showed an enlarged pancreas with minimal fluid in the peritoneal cavity, related clinically to pancreatitis 12 , and necrotic pancreatic tissue 13 . Our findings are similar to reports in animals infected by other plasmodium species. In a study of P. berghei in mice, pancreatic tissues revealed acute inflammatory reactions of acini, islets cells and interlobular duct 14 . Another study of Macaca mulatta monkeys infected with P. knowlesi showed acinar necrosis and the disappearance of pancreatic islets 15 . In addition, hypoglycaemia   was observed in animal models infected with P. knowlesi 16 . The inflammatory cells such as lymphocytes and neutrophils are crucial in the control of malarial parasite replication, which contribute to the subsequent elimination and resolution of the infection through the process of phagocytosis and/or production of inflammatory mediators 17 . Fibrosis, areas of haemorrhages and fat necrosis in the pancreatic tissues were not features of P. falciparum malaria patients. The presence of PRBCs within the capillaries can cause an increase in blood viscosity. In addition, malaria parasites in the red blood cells (RBCs) can cause decrease in RBC membrane deformity. These circulatory disturbances can lead to tissue hypoxia, release of pancreatic acinar enzymes, impaired capillary and venous drainage, which in turn can lead to haemorrhagic pancreatic necrosis 18 . Fibrosis is characterized by the accumulation of extracellular matrix and an increase in fibroblastic reaction. Though fibrosis is not significantly present in P. falciparum malaria patients, possibly due to the acute nature of malaria infection, a previous study has reported that high BS levels can stimulate the proliferation of pancreatic stellate cells and subsequent increased in fibroblast production 19 .
Whether the increase in islet size is caused by malaria infection per se, or is secondary to antimalarial drugs, needs to be further investigated. Nevertheless, the enlarged pancreatic islets can contribute to hyperinsulinemia and subsequent hypoglycaemia seen in P. falciparum malaria patients.
The present study used immunohistochemical methods to demonstrate the expressions of insulin and glucagon in terms of distribution and intensity in the pancreatic tissues of P. falciparum malaria patients. Similar to previous reports, insulin immunopositive cells were found in greater density in the central part of the islets of Langerhans, compared to islet cells expressing glucagon, which are frequently detected around the periphery of pancreatic islets 20 .
The increased expression of insulin in P. falciparum malaria patients with hypoglycaemia can be secondary to the malaria parasite itself, the use of antimalarial drugs, alteration in the gluconeogenesis pathway, the effects of cytokines during malaria infection and renal injury 6 . Malaria parasites in the circulation need glucose from the host in order to produce energy, mainly via aerobic glycolysis, for their survival 4 . A previous report showed that high parasitaemia is associated with high glucose requirements by malaria parasites, which can lead to clinical hypoglycaemia 21 . Antimalarial drugs, such as quinine and quinidine have an effect on glucose metabolism in malaria patients, causing a decrease in BS level 5 . Most of our malaria patients were treated with quinine. Quinine has been reported to be a more potent stimulant of insulin secretion than quinidine 5 . These drugs act similarly to glucose on potassium permeability of the beta cells membrane, causing calcium influx and subsequent release of insulin to reduce BS levels 5 . We accept that quinine/quinidine have an effect in lowering insulin level and is still used in endemic areas and in places where artesunate is not available. Our data suggests the need for precautious use of quinine/quinidine in these circumstances, and recommend closed monitoring of the BS level of malaria patients. Other antimalarial drugs, such as chloroquine, amodiaquine, mefloquine and halofantrine have no direct effect on insulin secretion 22 . Interestingly, the gluconeogenesis pathway is abnormal in severe P. falciparum malaria patients due to insufficient galactose substrate, resulting to low BS levels 7 . The effect of cytokines released during P. falciparum malaria infection, such as tumour necrosis factor (TNF), interleukin (IL)-1 and 6 can inhibit the activity of phosphoenolpyruvate carboxykinase in the gluconeogenesis pathway and also mediate a reduction in hepatic glycogen content, hence the lower BS levels 23,24 . In addition, IL-1 and 6 have been reported to stimulate islet cell hyperplasia 25 . It can be hypothesised that IL-1 and 6 released during malaria infection can cause islet cell hyperplasia resulting to enlargement of pancreatic islets and subsequent increase in insulin expression. Furthermore, the kidney is an important organ involved in glucose regulation. Occurrence of acute kidney injury in severe P. falciparum malaria can affect the net glucose release and reabsorption, resulting in hypoglycaemia 26 .
The maintenance of a normal BS level is a critically important process in the metabolism in living organisms. Other parasitic infections have been reported to cause hypoglycaemia, e.g. trypanosomiasis in both human and animals 27 and Chagas's disease in a murine model 28 . Although these studies have shown that hypoglycaemia is not a major cause of death, in patients with severe P. falciparum malaria it is an important clinical manifestation that may well contribute significantly to high mortality rates. This pioneer study documents an increase in insulin expression and an increase in islet size in P. falciparum malaria patients with hypoglycaemia. The study will be useful for clinicians interested in understanding and managing the process of hypoglycaemia in severe P. falciparum malaria. The paper sets out recognizable histopathologic and immunologic changes associated with malarial pancreatic tissues.

Materials and Methods
Pancreatic tissues. Pancreatic tissues from P. falciparum malaria patients and control pancreas were obtained from the Department of Tropical pathology, Faculty of Tropical Medicine, Mahidol University, Bangkok, Thailand. For malarial cases, pancreatic tissues were divided into three groups based on the BS levels before death, namely BS < 40 mg/dl, BS between 40-120 mg/dl, and BS > 120 mg/dl. Control pancreatic tissues were obtained from accidental death cases, which showed normal islets of Langerhans, secretory acini, blood vessels, ducts and interstitial area. The study protocol was approved by the Ethics Committee of the Faculty of Tropical Medicine, Mahidol University, Thailand (MUTM 2015-041-01 and MUTM 2015-041-02). Methods were performed in accordance with the relevant guidelines and regulations of the above committee. Written informed consent from closest relatives was obtained for all malaria patients. The Ethics Committee of the Faculty of Tropical Medicine, Mahidol University, Thailand approved the use of control pancreatic tissues obtained from discarded autopsy specimen.
Histopathological preparation and evaluation. Left-over pancreatic tissues in paraffin blocks were re-embedded and processed using the standard histological protocol. Pancreatic tissues were sectioned at 4 μm thickness for histopathology and immunohistochemistry studies. Histopathological sections of pancreatic tissues were stained with Mayer's haematoxylin and eosin (H&E). To study the average size of islets of Langerhans, islets were measured and evaluated by randomisation of 40 pancreatic islets under ×100 magnification using the "ImageJ" software program, developed by the National Institutes of Health, Bethesda, Maryland, USA. In addition, the pancreatic tissues were interpreted based on seven histological criteria, namely presence of parasitised red blood cells (PRBCs), oedema, haemorrhage, inflammatory infiltration, acinar necrosis, fat necrosis and fibrosis, under ×200 magnification. The changes were graded on a scale of 0-3. All criteria were assessed in the lobule, interlobular and intergladular areas, except for fat necrosis where the occurrence was at the peripancreatic tissue level. The grading criteria were based on previously described studies 29,30 , with minor modifications ( Table 2).

Immunohistochemistry study of insulin and glucagon. The expressions of insulin and glucagon
were detected by immunohistochemical staining, using an avidin-biotin peroxidase complex method (Vector Laboratories, Inc., USA). Primary antibodies used were rabbit polyclonal anti-insulin and anti-glucagon (Cell Signaling Technology, Inc., USA). Pancreatic tissues at 4 µm thickness were placed on an adhesive poly-L-lysine slide and heated in a hot oven at 56 °C for 30 minutes. Tissue sections were de-paraffinised and rehydrated in xylene and a decreasing series of graded alcohol, respectively. Then, the pancreatic tissue sections were incubated in 3% hydrogen peroxide diluted in distilled water for 30 minutes at 37 °C in order to block the endogenous peroxidase activity in the tissue. To retrieve the antigen, tissue sections were incubated with 0.1 M sodium citrate buffer at pH 6.0 in a 800 W microwave for 20 minutes. When the slides cooled down to room temperature (RT), the sections were washed in PBS buffer at pH 7.4. To reduce the non-specific background, the sections were incubated with diluted normal goat serum for 30 minutes. Then the pancreatic sections were incubated with a primary antibody (rabbit polyclonal anti-insulin or anti-glucagon (1:200 dilution) in a moisture chamber overnight at 4 °C. The following day, sections of pancreas were washed with PBS buffer at pH 7.4 for three time then incubated with diluted biotinylated goat anti-rabbit Ig G for 30 minutes at RT and treated with avidin-biotin peroxidase complex (ABC) conjugated with horseradish peroxidase (HRP) for 30 minutes. After all sections were washed, the tissue sections were reacted with 0.05% 3,3′ diaminobenzidine (DAB) and 0.01% H 2 O 2 for 3 minutes, resulting in a brown coloration, which indicated the target antigen. Subsequently, pancreatic sections were counterstained with Mayer's haematoxylin and mounted with a coverslip for microscopic evaluations.
Evaluation of immunohistochemical staining. The expressions of insulin and glucagon were evaluated in the cytoplasm of the islet cells in the pancreatic islets. Twenty islets from each pancreatic section were randomly observed at ×400 magnification. The percentages of insulin-and glucagon-expressing cells were determined by counting the number of positive cells over the total number of cells in each islet, multiplied by 100. The intensity of staining with both antibodies was subjectively scored as follows: 0 = no staining, 1 = weakly positive, 2 = moderately positive, and 3 = strongly positive. The total score (TS) was calculated by multiplying the percentage of positive cells (%) and staining intensity (I), according to previous study 31 . Statistical analysis. Data were recorded into a computer database and analysed with SPSS software version 18.0. All quantitative data were represented as mean ± standard error of mean (SEM). The test for normality of distribution was calculated using Kolmogorov-Smirnov test. Comparisons of the difference in the size of pancreatic islets between groups were analysed by one way ANOVA. Correlations between variables of histological and immunohistochemical analyses and clinical data were analysed by Spearman's rank correlation (r s ). Mean differences in the total scores of histopathological and immunohistochemical studies between experimental groups were analysed by Mann-Whitney U test. The P-value < 0.05 was considered to be statistically significant.   Table 2. Histopathological changes and grading criteria for pancreatic tissue evaluation.