Abstract
ABSTRACT Our understanding of how the host immune response kills Plasmodium, the causative agent of malaria, is limited and controversial. One widely held belief is that reactive oxygen species are crucial for controlling parasite replication. One of the hallmarks of blood-stage malaria is the cyclic rupture of erythrocytes by the parasite, which releases free hemoglobin into the circulation. We propose that this free hemoglobin, as well as the hemoglobin within the erythrocyte and surrounding the parasite, effectively shields Plasmodium from reactive oxygen species well in excess of those achievable in vivo.
Malaria, a devastating disease that kills over 2 million people each year, is caused by infection with parasites of the genus Plasmodium.1,2 Plasmodium is transmitted by the bite of an infected mosquito vector, and this initiates first the liver and then the blood stage of the infection. The liver stage is asymptomatic and is beyond the scope of this article. The blood stage comprises parasites (merozoites) invading erythrocytes, developing within the erythrocyte, and producing new progeny every 48 or 72 hours, depending on the species. The blood stage of the infection causes clinical complications, such as fever and chills, as well as life-threatening multiorgan failure (brain, lungs, and kidney). Plasmodium falciparum is the most virulent of the species of Plasmodium that infect humans, and this species adheres to activated venular endothelium to sequester itself from the lymphoid and filtration organs.3Despite considerable research effort, an effective vaccine against Plasmodium infection remains elusive perhaps because our understanding of the mechanism(s) underlying parasite killing by the immune system is poorly defined and controversial. The observation by Jensen and colleagues of crisis forms of the parasite (parasites exhibiting deteriorating morphology) immediately prior to declines in parasitemia4suggests that a soluble factor is responsible for parasite killing, rather than erythrophagocytosis. One potential effector mechanism that is widely believed to be critical for this parasite killing is the immune production of reactive oxygen species (ROS), including nitric oxide (NO•), superoxide, and peroxynitrite. These ROS are produced mainly by activated phagocytic cells, including macrophages and neutrophils, although T helper 1 cells are also reported to produce NO• during experimental malaria.5
NO•, a highly toxic free radical, is produced during the enzymatic conversion of l-arginine to l-citrulline by members of the nitric oxide synthase (NOS) family.6To date, three members of the NOS family have been identified, specifically endothelial (eNOS or NOS3), neuronal (nNOS or NOS1), and inducible (iNOS or NOS2).7-9All three isoforms have similar NO• production rates of about 1 μM/min/mg protein.10The iNOS isoform is responsible for the high-level production of NO• because it is highly expressed after activation of phagocytes, comprising up to 1% of total protein.11
Activated phagocytes also produce superoxide by the enzymatic reduction of molecular oxygen by reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (pHox family). On activation, the NADPH oxidase is assembled from its membrane and cytosolic pHox subunits (p91, p22, p67, p47, p40) and superoxide is formed inside a phagosome or outside the cell.12When both NO• and superoxide are produced in close proximity, for example, by activated phagocytes, this results in the formation of peroxynitrite by the rapid reaction between NO• and superoxide.10
NO• is well established as a major cytotoxic molecule in infectious diseases.13Because NO• and ROS are important in the control of other parasitic diseases, it is generally accepted that they play a critical role in killing the malarial parasite. In fact, this hypothesis is supported in all 15 recent reviews addressing the role of NO• or ROS in malaria. However, evidence from animal models challenges this notion: mice deficient in the iNOS gene or p91phox gene exhibit similar Plasmodiumberghei parasitemia as controls, indicating that iNOS-derived NO•, pHox-derived superoxide, and peroxynitrite do not serve to control parasitemia.14-18Further, the role of hemoglobin (Hb) as a ROS quencher has not been taken into account, leading us to propose that NO• and ROS do not play a significant role in malarial parasite killing.
ROLE OF HB
Hb is released when Plasmodium ruptures the erythrocyte, and our preliminary experimental data suggest that the plasma concentration of free Hb on day 6 of P. berghei infection (when the animal becomes moribund) can reach 100 μM. The in vivo effects owing to free natural molecular Hb in blood are well defined in the research community interested in the development of molecular Hb-based oxygen carrying blood substitutes.19,20Free Hb is a powerful in vivo scavenger of NO•, leading to vasoconstriction and impaired microvascular blood perfusion, which, in turn, are major determinants of tissue and organism survival.21Free Hb is almost 1,000-fold more efficient at scavenging NO• than Hb packaged in red blood cells (RBCs),22so this free Hb may be protective by significantly quenching ROS before they can diffuse into the erythrocyte. Besides free Hb, Plasmodium resides in a parasitophorus vacuole within the erythrocyte and is surrounded by about 0.5 fmol Hb. The combination of free molecular Hb in blood plasma and Hb in the erythrocyte likely protects the parasite because of complex ROS scavenging reactions.
The chemical reactions of NO• with Hb depend on the form and state of the Hb molecule. When Hb passes through the lung, it changes from the taut or T state without oxygen (deoxyhemoblogin) to the relaxed or R state with oxygen bound to heme moiety (oxyhemoglobin).23,24In the venous microcirculation, in which sequestered parasites such as P. falciparum and murine-infecting P. berghei and Plasmodium yoelii reside, about 60% of Hb within erythrocytes is oxyhemoglobin and 40% is deoxyhemoglobin. In the arterial microcirculation, most of the Hb is oxyHb (> 99%). NO• reacts with the oxygen in oxyHb, generating methemoglobin [Hb(Fe3+)] and nitrate (Figure 1, reaction 1). NO• also reacts with deoxyhemoglobin, forming Hb(Fe2+)NO (see Figure 1, reaction 2), which, in turn, reacts with oxygen to form methemoglobin and nitrate (see Figure 1, reaction 3). Further, oxyhemoglobin is capable of quenching superoxide to form molecular oxygen, hydrogen peroxide, and methemoglobin (see Figure 1, reaction 4).25Oxyhemoglobin can also rapidly react with peroxynitrite to ultimately form nitrate and methemoglobin (see Figure 1, reaction 5).26Thus, Hb effectively renders exogenous free radicals incapable of directly damaging the parasite.
Hb not only acts as a scavenger of radical species, it can also undergo redox transitions to higher oxidation states, making it a potent oxidant.27First, oxyhemoglobin can slowly auto-oxidize, forming methemoglobin and superoxide, and the latter can form hydrogen peroxide via dismutation. However, this process within the erythrocyte is balanced within the erythrocyte by the presence of superoxide dismutase, catalase, and methemoglobin reductase.28,29Second, both oxyhemoglobin and methemoglobin can react with peroxides to form the highly oxidant ferryl species, which damages many biomolecules.30Again, the presence of catalase within the erythrocyte is likely to mitigate this. Third, Hb may be able to act as a Fenton reagent, catalyzing the formation of the hydroxyl radical from hydrogen peroxide,31,32but it is likely that the heme must be free (ie, released from the Hb molecule) for this to occur.33,34Finally, recently Cosby and colleagues and Nagababu and colleagues proposed that deoxyhemoglobin within the erythrocyte can reduce nitrite, resulting in NO• production.35,36However, this reaction is limited by the influx of nitrite into the erythrocyte and has a low yield of NO•, which may bind to the excess deoxyhemoglobin, according to Figure 1, reaction 2.
Although the intraerythrocytic Hb may generate oxidants and radicals, the extent of exposure of the Plasmodium parasite to oxidant stress is mitigated by the presence of superoxide dismutase, catalase, and methemoglobin reductase within the erythrocyte. However, Clark and colleagues argue that the intraerythrocytic environment is not favorable to the parasite and that the requirement of 3% oxygen tension for parasite culture in vitro is also indicative of the parasite being susceptible to oxidant stress.37Indeed, in vivo the Plasmodium parasite sequesters itself on the venous endothelium, where oxygen tension is low. However, hypoxic conditions actually result in greater radical production by Hb auto-oxidation.38This observation, combined with the nitrite reductase role of deoxyhemoglobin proposed by Gladwin and Rifkind, suggests that oxidative stress may be increased when the parasite sequesters. A more likely explanation for parasite sequestration in venules is that cell adhesion molecules (CAMs) are increased under proinflammatory conditions in venules but not arterioles. To combat ROS exposure, Plasmodium has adapted by acquiring host superoxide dismutase39and possibly synthesizing its own.40We conclude that within the erythrocyte, Hb is more likely to act as a ROS scavenger than a source of oxidant stress, thus serving to protect the parasite.
MAXIMAL IN VIVO LEVELS OF NO• AND ROS DURING MALARIA
We envision two maximal scenarios for exposure of P. falciparum parasitized erythrocytes to NO•: (1) a parasitized erythrocyte passing through the splenic shunt between arterioles and venules in the red pulp and between activated red pulp macrophages41and (2) a sequestered parasite adherent in the venous microcirculation to an activated endothelial cell with a macrophage adherent close by. For the case of the splenic red pulp, the erythrocyte passes within 20 to 30 μm of two layers of macrophages.41Assuming a monolayer of cells surrounding the shunt, the NO• production of the macrophages should be about 2 μM.42However, the splenic red pulp is filled with erythrocytes in close proximity to the macrophages, resulting in tremendous NO• scavenging capacity, which likely would prevent the majority of the NO• from reaching the parasitized RBC. Thus, the effective dose of NO• experienced by the parasite is likely lower than the 1 μM.
The likely worst-case scenario for the parasite occurs when the erythrocyte is sequestered and there is an activated macrophage in very close proximity (Figure 2). The first source of NO• is the activated venous endothelial cell, which can produce approximately 0.17 fmol/h of NO• in venules exposed to shear stress of 1.8 dyne/cm2.43In arterioles, the blood flow results in a 1 to 10% cell-free region near the luminal surface of the endothelium, where NO• is scavenged only by the reaction with oxygen, resulting in a NO• half-life in the range of several minutes to hours, depending on the NO• and oxygen concentrations.44In venules, on the other hand, no axial migration of erythrocytes is seen,45so NO• is quenched by Hb throughout the vessel. Although models estimate that the arterial endothelium can maintain approximately 100 nM NO• at its luminal surface primarily owing to the cell-free region,46a fraction of this concentration is likely at the luminal surface of venules because of the presence of erythrocytes and free Hb. In C57BL/6 mice, we have detected a marked increase in plasma free Hb on day 6 of P. berghei infection compared with healthy animals. Therefore, the bulk of the NO• will come from the area of direct contact of the erythrocyte with the single endothelial cell. Based on the endothelial cell to erythrocyte surface area ratio of ˜50, as well as the fact that half of the 0.17 fmol/h NO• produced will diffuse away from the lumen, we estimate that 1.7 amol/h NO• will be bioavailable within the RBC (0.17/50/2 fmol/h). If the P. falciparum-infected erythrocyte is sequestered for 24 hours, this yields 40.8 amol of NO•, which is well below the 2 fmol of heme within an RBC. Thus, it appears that the endothelium alone cannot produce sufficient levels of NO• to overcome the quenching by the Hb inside the erythrocyte.
Any scenario involving a macrophage is highly dependent on the relative positioning of the macrophage and erythrocyte with respect to the flow (see Figure 2). The macrophage must be extremely close and upstream to have an effect, but if the macrophage touches the parasitized erythrocyte, it will likely phagocytize the erythrocyte, and the parasite will not be able to replicate because its progeny merozoites cannot reach new erythrocytes. An activated macrophage sustains up to 10 fmol/h of NO• production and 1 fmol/h of superoxide production for several hours.42This yields an NO• concentration of about 1 μM of NO• near the surface of the macrophage,42but the resultant flow-driven NO• “plume” is rapidly scavenged by the surrounding erythrocytes and free Hb. Thus, it is unlikely that the sequestered erythrocyte will experience any marked elevation of NO• levels, and it will remain sheltered from NO• by erythrocytic Hb.
Activated macrophages also produce superoxide, but at a 10-fold lower level than NO•.47Owing to the much higher NO• levels and the fact that both radicals come from the same source, all of the superoxide rapidly reacts with the NO•, yielding ˜1.5 nM peroxynitrite at the macrophage surface. In addition, Hb in other erythrocytes and plasma Hb quench superoxide, making the likelihood of superoxide reaching the parasite almost nil.
The peroxynitrite resultant from the reaction of superoxide and NO• produced by the macrophage is a highly reactive radical and at physiologic pH has an extremely short half-life.47This results in a very short diffusion distance and near-zero concentration (< 0.2 nM) at 10 μm from the macrophage. Thus, a sequestered erythrocyte is unlikely to be exposed to significant concentrations of peroxynitrite (< 1 nM), and peroxynitrite is quenched by oxyhemoglobin, to form methemoglobin, nitrite, and oxygen. Based on the low production and its scavenging by Hb, we propose that peroxynitrite does not play a significant role in parasite killing in vivo.
CURRENT EVIDENCE IN SUPPORT OF NO• - AND ROS-MEDIATED PLASMODIUM KILLING
NO• is reportedly produced at high levels during the course of P.falciparum malaria in humans.48Further, the blood from patients with uncomplicated malaria contains high levels of iNOS messenger ribonucleic acid (mRNA), with the increased expression primarily in monocytes, compared with those with severe malaria.49One concern with this study is that the two patient groups also received different drug treatments, and antimalarial agents have been shown to affect monocyte function.50Several studies report that polymorphisms in the promotor region of the NOS2 gene (CCTTT repeat, G-945C or Lambarene mutation, and C-1173T) result in higher baseline NOS activity in the peripheral blood mononuclear cells and are associated with protection from malaria.51-54On the one hand, these investigators proposed that NO• killing of the parasite explains this genetic association with protection from malaria. On the other hand, the results have been inconsistent,55-58and we speculate that these findings may be explained by the possible anti-inflammatory, homeostatic, and vasodilatory roles of NO• rather than direct parasite killing. NO• is a potent anti-inflammatory molecule that plays a role in P. falciparum sequestration, so increased NO• may down-regulate endothelial activation and CAM expression.59Further, NO• quenching by Hb during sickle cell crisis results in vasoconstriction, which can be attenuated by NO• gas or sodium nitroprusside treatment.60The reduced NO• bioavailability during sickle cell anemia may also be responsible for the observed increase in soluble vascular cell adhesion molecule 1 expression.60Similarly, NO• quenching by Hb is likely to occur during malaria, and increased NO• production may ameliorate microcirculatory complications. Finally, it is possible that the chronic exposure to increased levels of free Hb and NO• quenching during sickle cell anemia serves as an adaptive mechanism, contributing to the resistance to malaria conferred by sickle cell anemia.
Several in vitro coculture studies are also cited as evidence for NO• and ROS killing the malarial parasite. One in vitro study reported that interferon-γ-treated macrophages kill P. falciparum during coculture, and this parasite killing is markedly reduced by NG-monomethyl-l-arginine, a NOS inhibitor.61However, another in vitro study with cultured P.falciparum reported that a saturated solution of NO• does not kill the parasite.62
Similarly, superoxide reportedly kills P. falciparum when produced in vitro by monocyte-derived macrophages.63,64Peroxynitrite also reportedly kills P. falciparum in vitro.65
However, these studies do not take into account the role of free Hb in the in vivo situation. Ockenhouse and colleagues used indirect methods (3H-hypoxanthine incorporation when the parasite replicates its deoxyribonucleic acid [DNA] at the trophozoite stage) to assess parasitemia.64Because coculture of P. falciparum with activated monocyte-derived macrophages shortens the replication time from 48 hours to 24 hours, it is likely that lower 3H-hypoxanthine incorporation reflects poor labeling, despite the presence of viable parasites, because the parasites are past the stage of DNA incorporation when the radiolabel is added.66Fritsche and colleagues detected parasite killing only when they added to the parasite-macrophage cocultures, which are already producing ROS, additional supraphysiologic concentrations (1 mM) of SIN-1 (3-morpholinosydnonimine, a NO• and superoxide donor that reacts to produce peroxynitrite) for 24 hours, resulting in > 1 mM cumulative peroxynitrite production.65
EVIDENCE AGAINST THE ROLE OF NO• AND ROS KILLING OF THE MALARIAL PARASITE
Direct evidence using animal models of malaria argues against the role of NO• in parasite killing. Although NO• is produced at high levels during experimental malaria, mice deficient in the iNOS gene exhibit similar P. berghei and Plasmodiumchabaudi parasitemia as controls, indicating that iNOS-derived NO• and peroxynitrite are not required in parasite killing.14-16,67Treatment with the iNOS inhibitor aminoguanidine does not alter the time course of P. chabaudi or P. berghei parasitemia.14,16,67,68However, a compensatory mechanism by one or both of the remaining NOS genes in these mice may account for the lack of an effect. Injection of killed Propionibacteriumacnes prior to P. chabaudi infection results in a marked (> 100-fold) increase in NO• production that is sustained for more than a week, but this high level of NO• has no detectable effect on the time course of parasitemia in mice when compared with similarly treated iNOS knockout mice lacking the elevated NO• production.16
Similar to the lack of detectable in vivo killing of malarial parasites by NO•, the results from animal models also question the in vivo role of superoxide and peroxynitrite in parasite killing. P-mice (a strain of mice that have a spontaneous mutation in the pHox gene, resulting in reduced superoxide production) and p91pHox knockout mice have similar P. chabaudi and P. berghei parasitemias as controls,17,18,67indicating that superoxide and peroxynitrite are not required for parasite killing.
In fact, the results obtained from both the NOS- and pHox-deficient animals can be explained by the ROS quenching capability of Hb. Because Hb is present at such a high concentration within the blood, it can easily quench even the elevated ROS production that occurs during malaria,69-73resulting in the effective amount of ROS reaching the parasite being near zero. Thus, modulating ROS production via gene knockout or inhibitor treatment cannot further lower the already low effective dose of ROS to which the parasite is exposed.
Finally, we assessed the ability of ROS to kill the P. berghei parasites in a defined ex vivo situation with careful analysis of the Hb oxidation state.74Parasitized RBCs were treated ex vivo with 1.5 μM NO•, 15 μM NO•, 150 μM NO•, 1.2 mM NO•, 1.5 mM NO•, or 1 mM SIN-1 for 10 minutes, and the ROS-treated inoculum (1 million pRBC) was injected into groups of mice. Supraphysiologic NO• treatments up to and including 150 μM do not reduce parasite viability in the inocula. The 150 μM NO• treatment was bioactive inside the erythrocyte because it converts all of the oxyhemoglobin to methemoglobin. The fact that the parasite remains viable, even if treated with NO• (up to 150 μM) or peroxynitrite (up to 220 μM cumulative production), far in excess of the Hb concentration indicates that (1) the theoretical in vivo levels of ROS are insufficient to kill and (2) Plasmodium likely has an intrinsic protective mechanism to deal with exogenous (immune) ROS.
SUMMARY
Despite its complex redox chemistry, Hb, which is a potent NO• and ROS scavenger and is present in abundance around the parasite, is likely to protect the malarial parasite. The high levels of free Hb in the plasma during the course of disease are likely to minimize the ability of NO• and ROS to enter the erythrocyte. Intraerythrocytic Hb is also likely to be protective because oxidant stress relating to Hb redox chemistry is enzymatically controlled within the red cell. Finally, the malarial parasite has developed mechanisms to cope with the oxidant nature of the free heme it produces, such as packaging it in the hemozoin granule and confiscating erythrocyte superoxide dismutase.
Although NO• does not kill the parasite, it may protect against the development of disease. NO• plays important roles in maintaining the homeostasis of (1) the immune system, (2) the endothelium, and (3) the coagulation system, whereas pathologic activation of these systems is required for malarial pathogenesis.3,75,76Thus, NO•'s function may be as an antioxidant and protective molecule,77as well as a potent anti-inflammatory molecule that minimizes P. falciparum sequestration by down-regulating endothelial activation and CAM expression.59
ACKNOWLEDGMENT
Figure 2 was generated by Karen Landon.