INFECTION AND IMMUNITY, Jan. 1983, p. 1-6 0019-9567/83/010001-06$02.00/0 Vol. 39, No. 1 Copyright © 1983, American Society for Microbiology Evidence for Reactive Oxygen Intermediates Causing Hemolysis and Parasite Death in Malaria I. A. CLARK* AND N. H. HUNT John Curtin School of Medical Research, Australian National University, Canberra, Australian Capital Territory 2601, Australia Received 6 July 1982/Accepted 5 October 1982 A rapid reduction in parasitemia associated with damage to intraerythrocytic parasites was observed in Plasmodium vinckei-infected mice after they had received a single intravenous injection of alloxan. This was not prevented by prior injection of glucose, but was prevented by desferrioxamine or diethyldithiocarba- mate. Prior injection of propanol partially blocked the phenomenon. A transient hemolysis was observed in malaria-infected mice, but not in controls, after injection of alloxan. This was also blocked by desferrioxamine, but not by glucose. Both the fall in parasitemia and hemolysis occurred, but less dramatical- ly, when phenylhydrazine or hydrogen peroxide was injected into parasitized mice. Again, the hemolysis was blocked by desferroxamine. These observations are consistent with the parasite death and hemolysis being mediated by reactive oxygen species, possibly hydroxyl radicals, and have implications for our understanding of hemolysis, endothelial damage, and parasite suppression in acute malaria. Our evidence that malaria parasites are susceptible to free oxygen radicals supports the view that high intraerythrocytic oxidative stress may contribute to the high frequencies in malarial areas of genes for certain erythro- cyte-related traits and suggests that some antimalarial drugs may suppress parasites partly through oxidative damage. Acute malaria is a complex disease, with nature of alloxan-induced hemolysis of tocoph- much pathology not easily attributable to the erol-deficient erythrocytes and damage to islet presence of protozoa inside circulating erythro- P-cells, it seems likely that reactive oxygen cytes. We have proposed that release of soluble species mediate these changes. mediators from mononuclear phagocytes may explain many of these changes and also the MATERIALS AND METHODS concurrent suppression of the parasites (2). This Mice and parasites. CBA/CaH mice of either sex and model includes release of arachidonate metabo- between 6 and 10 weeks old were used in these lites (3, 6; manuscript in preparation), as do experiments. Neither sex nor age had an apparent current explanations for damage to vascular effect on the outcome. P. vinckei subsp. vinckei (strain endothelial cells in sepsis (44), hypertension V52, from F. E. G. Cox) was stored at -96°C and (26), and trauma (49). Because these authors before use had been passaged several times in CBA/ argue that oxygen-derived free radicals released CaH mice. All infections were initiated by the intra- peritoneal injection of 106 parasitized erythrocytes and during the formation of arachidonate metabo- monitored by examining Giemsa-stained thin smears. lites might mediate this endothelial cell damage, Reagents. Alloxan monohydrate (Sigma Chemical we investigated whether this family of reactive Co.) was dissolved in normal saline at 12.5 mg/ml, oxygen species could also ca'use pathology and stored in the dark at 4°C, and injected without further parasite inhibition associated with acute malar- dilution. This remained stable for at least 6 weeks. ia. We were encouraged by evidence that eryth- Freeze-dried desferrioxamine (Desferal, Ciba Pharma- rocyte membranes (28) and endothelial cells ceuticals) was used within 1 week of reconstitution (25), both of which are damaged in malaria, are with distilled water to 100 mg/ml, having been stored very susceptible to damage by free radicals, and at 4°C. Diethyldithiocarbamate (DDC) (Sigma Chemi- cal Co.), n-propanol, hydrogen peroxide, glucose by reports that other protozoan parasites are (BDH Chemicals Ltd.), and phenylhydrazine hydro- susceptible to toxic oxygen species (33, 36). We chloride (Fluka AG) were freshly dissolved in saline have found that when alloxan is injected into before each experiment. Desferrioxamine was injected Plasmodium vinckei-infected mice there is a at stock concentration, and other reagents were used transient hemolysis, and the parasites are rapid- as follows: DDC as 200 ,ul of a 75-mg/ml solution, n- ly damaged. In keeping with arguments on the propanol as 500 Ill of a 10% (vol/vol) solution, hydro- 1 2 CLARK AND HUNT INFECT. IMMUN. ently healthy. This effect was total, without seeming to harm the mice, with doses between 40 and 800 mg/kg (data not shown). Alloxan-induced chemiluminescence in a cell- free system appears to be a hydroxyl radical- mediated phenomenon, and the avid metal che- lator DDC is as inhibitory as desferrioxamine in this system (20). At all doses we tested (0.3 to o 20 1.2 g/kg), pretreatment with DDC totally inhibit- OCl ed the action of alloxan against malaria parasites (Fig. 2, 0.6 g/kg shown). Figure 2 also shows that pretreatment with propanol (1.6 g/kg) partially 01 inhibited the killing of malarial parasites by alloxan. Propanol prevents alloxan diabetes, probably by scavenging hydroxyl radicals (23). 60 120 180 360 Neither desferroxamine, DDC, nor propanol alone had any apparent effect on parasites over a MINUTES AFTER ALLOXAN 3-h period. FIG. 1. Effect of intravenous injection of alloxan A transient hemolysis, evidenced by hemoglo- (50 mg/kg) on P. vinckei infections. Symbols: (0) binuria and lowered hematocrits, accompanied alloxan; (0) saline. Standard deviations are indicated the damage to parasites. Hemolysis was not by vertical bars; n = 35 mice at each point. seen in normal mice given the same dose of gen peroxide as 200 ,ul of a 1% (vol/vol) solution, alloxan. However, parasite damage was not glucose as 200 Ill of a 250-mg/ml solution, and phenyl- dependent on hemolysis since many degenerat- hydrazine as 100 or 200 ,ul of a 75-mg/ml solution. The ing intraerythrocytic parasites were present (30 stock hydrogen peroxide solution was 30% (wt/vol). to 40 per field). A comparison of parasitemia and Hematocrits were measured after a 5-min spin in a degree of hemolysis, together with the presence Clay Adams Readacrit centrifuge. of degenerating parasites within intact erythro- cytes, indicated that hemolysis was not confined RESULTS to parasitized erythrocytes. Marked hemoglo- Injection of alloxan into malaria-infected mice. binuria 30 min after alloxan injection indicated On day 6 after infection with P. vinckei, when parasitemias were between 30 and 60%, mice received an intravenous injection of alloxan. Figure 1 shows the outcome of seven experi- ments with an alloxan dose of 50 mg/ml. Doses ranging from 25 to 100 mg/kg gave very similar 40 results, and two or three doses of 50 mg/kg each several days apart, when monitored for a further month, seemed to have cured the infections 30 completely (data not shown). The effects were w clearly evident within 30 min of injection, when U) blood smears contained many degenerating in- 20 traerythrocytic parasites. These effects were independent of the diabe- 0. togenic action of alloxan since injection of glu- cose immediately beforehand, which prevents 10 alloxan from causing diabetes (30), did not inhib- it its action against parasites (Fig. 2). Even 1 2 3 without glucose injections, the fluctuations in HOURS AFTER ALLOXAN blood glucose levels were minor when parasite damage occurred. In contrast, pretreatment with FIG. 2. Effect of pretreatment with glucose, pro- desferroxamine, a specific iron chelator which panol, desferrioxamine, or DDC on antimalarial action inhibits formation of hydroxyl radicals from of alloxan (50 mg/kg, intravenous). Symbols: (U) D- glucose (2 g/kg) given intravenously immediately be- superoxide and hydrogen peroxide by the iron- fore alloxan; (A) n-propanol (1.6 g/kg), (A) desferriox- catalyzed Haber-Weiss reaction (21), complete- amine (200 mg/kg), or (O) DDC (0.6 g/kg) given ly prevented the antimalarial activity of alloxan. intraperitoneally 30 min before alloxan; (0) alloxan When 200 mg of desferroxamine per kg was alone; (0) saline. Standard deviations are indicated by injected intraperitoneally 30 min before alloxan, vertical bars; n = 5 at each point. Representative the parasites (Fig. 2) remained intact and appar- outcome of two to four experiments is shown. VOL. 39, 1983 REACTIVE OXYGEN INTERMEDIATES IN MALARIA 3 60, parasite densities, on day 3 after injections be- fi, -Ii gan. When we attempted to see whether desfer- H rioxamine would inhibit phenylhydrazine or hy- 50 L I drogen peroxide-induced parasite suppression, we noted that during the time required to moni- 40t.. + tor these slower-acting agents, desferrioxamine ic- alone inhibited the parasites, perhaps by limiting 0 0 their nutritional iron. This is consistent with a I-30'L recent report about desferrioxamine added to a2 cultured Plasmodium falciparum (C. Raventos, * 20 S. Pollack, and R. L. Nagel, Clin. Res. 30:377A, 1982). 10 0 A DISCUSSION Here we present evidence that alloxan will kill P. vinckei in vivo (Fig. 1) and, in the doses used, causes hemolysis in malaria-infected mice but UNPARASITIZED P. VINCKEI - INFE(CTED not in controls (Fig. 3). Although most of the FIG. 3. Hematocrit (2 h after alloxan injection) of studies of alloxan concern its capacity to induce unparasitized mice and others with 41 ± 7% parasit- diabetes, alloxan also causes hemolysis in to- emia of P. vinckei. Symbols: (0) controls; (0) intrave- copherol-deficient rats (42) and produces an nous alloxan (50 mg/kg) alone; (*) intraperitoneal edema characterized desferrioxamine (200 mg/kg) alone; (A) desferriox- experimental pulmonary amine given 30 min before alloxan; (O) intravenous by damage to vascular endothelial cells (8). glucose (2 g/kg) given immediately before alloxan. Subsequent studies of this hemolysis incriminate Standard deviations are indicated by vertical bars; n = toxic oxygen species (12, 43). The hemolysis we 5 for each group.