Reversal of Type 2 Diabetes in Mice by Products of Malaria Parasites: I. Effect of Inactivated Parasites K.M. Elased, J.B. de Souza, and J.H.L. Playfair C57BL/ KsJ-db/ db and C57BL/ KsJ-ob/ ob mice are good models for studies on human obesity and type 2 diabetes. We have previously shown that infection with blood-stage malaria or injection of extracts from malaria-parasitized red blood cells induces hypoglycemia in normal mice and normalizes hyperglycemia in mice made moderately diabetic by streptozotocin. In the present study, w e show that a single intravenous (IV) injection of Formalin-fixed Plasmodium yoelii YM (FFYM) preparation decreases blood glucose in db/ db mice from an initial value of 19 mmol/ L to a normal value of 7 mmol/ L (P F .0001) for at least 24 hours and reduces food intake. Plasma insulin concentrations in db/ db mice were not altered. FFYM was also active in normal and ob/ ob mice, an effect associated with an increase in plasma insulin. Although the rate of weight gain in lean ob/  and lean db/  was not altered by this treatment, there was a significant reduction in weight gain in db/ db and ob/ ob mice (P F .001). We suggest that malaria-derived molecules, when fully characterized, may provide structural information for the development of new agents for the management of type 2 diabetes. Copyright  2000 by W.B. Saunders Company H YPOGLYCEMIA is a common complication of Plasmo- dium falciparum malaria, affecting primarily children and pregnant women. 1 Its mechanism of induction in malaria remains incompletely understood. Hypoglycemia can be repro- duced in murine models of blood-stage malaria, and using these, we have shown that malaria infection or the injection of parasite extracts can induce hypoglycemia and hyperinsulin- emia in normal mice, 2,3 while a parasite extract is also able to enhance insulin-mediated glucose uptake into adipocytes in vitro. 4 Furthermore, malaria-induced hypoglycemia can be prevented by agents that inhibit insulin secretion, namely diazoxide, somatostatin, and adrenaline. 5 In a model of type 1 diabetes (streptozotocin-induced), malaria infection reduced blood glucose to normal levels, provided some residual islet function was present. 6 However, in mice made severely diabetic with streptozotocin, malaria para- site extracts did not affect hyperglycemia significantly. 6 It has also been reported that a European diabetic patient with non–insulin-dependent diabetes mellitus (NIDDM) developed hypoglycemia during an infection with P falciparum in Kenya (30% parasitemia). The patient received prophylactic doses of glucose for 3 days and the hypoglycemia resolved after eradication of the parasitemia. 7 We therefore thought it of interest to determine whether malaria parasite–derived molecules also affect diabetic param- eters in obese diabetic hyperglycemic hyperinsulinemic C57BL/ Ks-db/db and C57BL/6J-ob/ob mice, widely regarded as good models for human NIDDM. 8-11 MATERIALS AND METHODS Mice Obese diabetic C57BL/Ks-db/db, C57BL/6J-ob/ob, and lipopolysac- charide-unresponsive (C3H/HeJ) mice were obtained from Harlan Olac (Bicester, UK). We used female mice aged 8 to 12 weeks, when both blood glucose and insulin levels are markedly elevated. 11 Controls were lean age-matched, normoglycemic, heterozygous littermates (db/+ and ob/+). The normal (C57BL/6  BALB/c) F1 mice were bred in our animal colony. Mice were allowed to acclimatize for at least 7 days before the study started. All animals had free access to water and were fed ad libitum with normal laboratory chow. Parasites The lethal YM line of P yoelii strain 17 (from Dr A. Holder, National Institute for Medical Research, London, UK) was maintained in our mice by blood passage of parasitized red blood cells. Mice were bled 5 to 7 days after intravenous (IV) infection with 10 4 parasites. Parasitized red blood cells (90% parasitemia) were lysed by incuba- tion in 0.01% saponin for 5 minutes at 37°C. Freed parasites were washed 3 times with phosphate-buffered saline (PBS), fixed overnight in 0.06% Formalin (Formalin-fixed P yoelii YM [FFYM]), washed again with PBS, counted by phase-contrast microscopy, resuspended at 5  10 8 /mL in PBS containing streptomycin (100 μg/mL) and penicillin (100 IU/mL), and stored at 4°C. FFYM was either injected IV or IP (10 7 to 10 8 parasites in 0.2 mL) or fed orally (1 to 2  10 8 parasites) in a 0.2-mL vol via a gastric tube. Control preparations containing the same number of normal red blood cell ghosts were made by treating blood from uninfected mice with saponin and Formalin in exactly the same way. Blood Glucose Glucose concentrations were determined using Glucostix and an Ames Glucometer (Miles, Stoke Poges, UK) on a drop of tail blood collected between 10 AM and midday, or at intervals thereafter as indicated. Results are expressed as the mean  SEM. Immunoreactive Insulin Blood was collected into heparinized tubes from the trunk after decapitation. Plasma was separated by centrifugation and frozen at -20°C. Immunoreactive insulin concentrations were determined in 50-μL vol of plasma by a double-antibody radioimmunoassay (kit supplied by ICN Biomedical, Irvine, CA) using a crystalline rat insulin standard (Novo, Bagsvaerd, Denmark). The results are expressed as the geometric mean with 95% confidence limits. From the Rademacher Group, London; and Department of Immunol- ogy, UCL Medical School, London, UK. Submitted September 9, 1999; accepted January 28, 2000. Address reprint requests to K.M. Elased, PhD, Rademacher Group, 6th Floor, Arthur Stanley House, 40-50 Tottenham St, London W1P 9PG, UK. Copyright  2000 by W.B. Saunders Company 0026-0495/00/4907-0022$10.00/0 doi:10.1053/mt.2000.6756 Metabolism, Vol 49, No 7 (July), 2000: pp 937-941 937