In a mouse model of cerebral malaria, free heme molecules can induce inflammation and permeabilization of the blood brain barrier, leading to death. The enzyme heme oxygenase-1 or its product carbon monoxide can decrease free heme levels, offering a new therapeutic approach to this deadly complication.
When malaria parasites infect host erythrocytes, they feed on hemoglobin, releasing large quantities of free heme. By catalyzing free radical reactions, free heme is poisonous to both parasite and host. Malaria parasites express a polymerase that converts heme to nontoxic hemozoin. Chloroquine and some other antimalarial drugs interfere with this process, allowing heme to damage the parasite. In their response to free heme, host cells increase their expression of the anti-inflammatory enzyme heme oxygenase-1 (HO-1), which degrades heme to carbon monoxide, iron and biliverdin (Fig. 1). Now, Pamplona et al. show that HO-1 also shields a malaria-infected host from developing the potentially fatal complication cerebral malaria1, suggesting a new cerebral malaria therapy that is desperately needed.
Malaria causes 1–3 million deaths each year. Cerebral malaria is a major cause of death from Plasmodium falciparum, one of four malaria species that infect humans. Some individuals infected with P. falciparum develop this complication and become comatose. A quarter of these do not respond to antimalarial therapy and die without regaining consciousness. Although the precise pathogenesis of cerebral malaria is still controversial, it is increasingly believed that host protective mechanisms are critical in determining whether individuals infected with P. falciparum develop cerebral malaria2. The host responds to malaria infection with several strategies to target the parasite and protect its organs. These are regulated by the balance between pro- and anti-inflammatory cytokines, but a deregulated response can lead to the accumulation of monocytes and T lymphocytes in the small vessels of the brain. Together with parasitized erythrocytes, these immune cells can compromise the integrity of the blood-brain barrier, allowing cytokines and malarial antigens to enter the biochemical milieu in which neurons operate.
Pamplona et al.1 proposed that free heme generated during the malarial infection induces an inflammatory response that damages the host's brain endothelium, leading to cerebral malaria (Fig. 1). Under this hypothesis, expression of HO-1 by the host would neutralize heme toxicity and diminish this inflammatory response. The authors used a well characterized mouse model of cerebral malaria with infection by the ANKA strain of P. berghei2. They showed that HO-1 activity correlated with susceptibility to cerebral malaria after infection: a mouse strain with high HO-1 activity in the brain was protected, whereas a mouse strain with low HO-1 activity was vulnerable. Mice of a resistant strain could be made susceptible when HO-1 was deleted genetically or inhibited pharmacologically, and treatment of susceptible mice with an HO-1 inducer protected if given before or early during infection. These results show that HO-1 activity prevented the development of experimental cerebral malaria.
The carbon monoxide generated by HO-1 limits the production of free heme by preventing hemoglobin oxidation by reactive oxygen species (Fig. 1)1. Pamplona et al. show that, like HO-1 activity, carbon monoxide treatment prevented the development of experimental cerebral malaria1. These protective effects were titrated away when free heme was injected into carbon monoxide–treated mice.
In the presence of reactive oxygen species, free heme probably causes a mild permeabilization of the blood-brain barrier, a feature of early disease2 (Fig. 1). At the terminal stage of experimental cerebral malaria, CD8+ T lymphocytes induce the apoptosis of brain endothelial cells, disrupting the blood-brain barrier2,3. The authors showed that HO-1 deficiency does not make T lymphocyte–deficient SCID mice susceptible to cerebral malaria, indicating that the inflammatory immune response is necessary for the development of advanced experimental cerebral malaria1. Carbon monoxide treatment also had an anti-inflammatory effect, inhibiting the cerebral expression of proinflammatory cytokines and preventing the sequestration of leukocytes in the vasculature1. Together, these results indicate that free heme is a critical player in the pathogenesis of experimental cerebral malaria, and that its degradation can protect mice from this fatal complication by preventing the inflammatory immune response that would otherwise disrupt the blood brain barrier.
Several inducers of HO-1 have been characterized, including nitric oxide and the cytokine interleukin-10. These molecules can provide protection against cerebral malaria4,5 and, interestingly, experimental atherosclerosis6. HO-1 preserves vascular homeostasis with anti-inflammatory and antioxidant activities, including the protection of endothelial cells against heme toxicity7. Unfortunately, therapeutic HO-1 inducers are not yet available; compounds used experimentally are too toxic to be used to treat cerebral malaria in humans.
The use of metabolic products of HO-1 might be an alternative therapeutic approach6. The data of Pamplona et al. indicate that carbon monoxide may be an attractive candidate. Presently, tailored carbon monoxide–releasing molecules are being developed as potential therapeutics against inflammatory and cardiovascular diseases, though potential toxicity needs to be considered carefully. Pamplona et al.1 showed that after the protective carbon monoxide treatment, 23% of blood hemoglobin was bound by carbon monoxide, levels that would be toxic to humans. The nontoxic HO-1 metabolite biliverdin was not protective against experimental cerebral malaria1, although it was administered for only four days after infection, and plasma levels of biliverdin or its metabolite bilirubin were not measured to verify treatment efficacy. This might be worth revisiting, since biliverdin and bilirubin are effective antioxidants and could conceivably provide protection from heme toxicity6.
The study of Pamplona et al. raises several issues1. First, given the proposed central role of circulating heme, HO-1 induction would be expected in vascular endothelial cells7, yet expression in human cerebral malaria is seen in liver and lung macrophages rather than endothelial cells8. This may reflect a response to macrophage uptake of hemopexin-bound heme9. Hemopexin effectively binds extracellular heme and prevents its pro-oxidant and proinflammatory activities. Therefore, differences in hemopexin levels or heme-binding capacity could conceivably contribute to differences in susceptibility to cerebral malaria between mouse strains. Second, if carbon monoxide derived from HO-1 appreciably affected cerebral malaria pathogenesis, the level of carbon monoxide–bound hemoglobin in the blood would probably differ between infected mice from cerebral malaria–susceptible and resistant strains. This information is currently unknown. In a large study of infected children including many with cerebral malaria, however, malaria severity was found to be unrelated to the amount of carbon monoxide–bound hemoglobin in the blood10. Third, there is a polymorphism in the HO-1 gene promoter that is predicted to increase HO-1 gene transcription. Contrary to what would have been predicted from the Pamplona et al. study, this polymorphism is seen at a higher frequency in individuals with cerebral malaria, compared with individuals with uncomplicated malaria11. Fourth, high HO-1 activity in host tissue macrophages would be expected to increase iron levels in the blood. In human cerebral malaria, however, high blood iron levels are associated with longer time to recover consciousness after antimalarial therapy12. These reports are not readily reconciled with the data of Pamplona et al.1. Although the protective effect of HO-1 in experimental cerebral malaria is an intriguing finding with potential therapeutic implications, the role of free heme in the human disease requires further study.
Pamplona, A. et al. Nat. Med. 13, 703–710 (2007).
Hunt, N.H. & Grau, G.E. Trends Immunol. 24, 491–499 (2003).
Belnoue, E. et al. J. Immunol. 169, 6369–6375 (2002).
Gramaglia, I. et al. Nat. Med. 12, 1417–1422 (2006).
Kossodo, S. et al. Immunology 91, 536–540 (1997).
Stocker, R. & Perrella, M.A. Circulation 114, 2178–2189 (2006).
Abraham, N.G. et al. Proc. Natl. Acad. Sci. USA 92, 6798–6802 (1995).
Clark, I.A., Awburn, M.M., Harper, C.G., Liomba, N.G. & Molyneux, M.E. Malar. J. [online] 2, 41 (2003).
Hvidberg, V. et al. Blood 106, 2572–2579 (2005).
Cunnington, A.J., Kendrick, S.F., Wamola, B., Lowe, B. & Newton, C.R. Am. J. Trop. Med. Hyg. 71, 43–47 (2004).
Takeda, M. et al. Jpn. J. Infect. Dis. 58, 268–271 (2005).
Gordeuk, V.R. et al. Blood 85, 3297–3301 (1995).
Koury, M.J. & Ponka, P. Annu. Rev. Nutr. 24, 105–131 (2004).
Personal financial interest in form of consultancy fee, and the possibility of future research support. The authors declare that they are bound by confidentiality agreements that prevent them from further disclosing their financial interests in this work.
About this article
Cite this article
Hunt, N., Stocker, R. Heme moves to center stage in cerebral malaria. Nat Med 13, 667–669 (2007). https://doi.org/10.1038/nm0607-667
New insights into microvascular injury to inform enhanced diagnostics and therapeutics for severe malaria
Inhibition of hypoxia‐associated response and kynurenine production in response to hyperbaric oxygen as mechanisms involved in protection against experimental cerebral malaria
The FASEB Journal (2018)
Infection and Immunity (2015)
Heme-Mediated Induction of CXCL10 and Depletion of CD34+ Progenitor Cells Is Toll-Like Receptor 4 Dependent
PLOS ONE (2015)
PLoS ONE (2013)