- Email: [email protected]
Does malarial tolerance, through nitric oxide, explain the low incidence of autoimmune disease in tropical Africa?
THE LANCET
Hypothesis
Does malarial tolerance, through nitric oxide, explain the low incidence of autoimmune disease in tropical Africa?
I A Clark, F M Al-Yaman, W B Cowden, K A Rockett Autoimmune disease is generally rare in tropical rural populations. Plasma concentrations of nitrite plus nitrate (reactive nitrogen intermediates), reflecting high nitric-oxide production somewhere in the body, can be high in patients who have cerebral malaria, but even higher in symptom-free parasitised individuals, who are termed malaria-tolerant. We propose that the nitric oxide causing high serum levels of reactive nitrogen intermediates in malaria-tolerant individuals is generated in macrophages during the establishment and maintenance of malarial tolerance, and makes autoimmune disease rare in many tropical rural populations by minimising proliferation of autoreactive T cells. Conversely, innately low levels of nitric-oxide generation in these populations, selected by malarial disease in tropical areas, could rationalise their high frequency of autoimmune disease and hypertension when living in western societies. In 1968 Greenwood1 drew attention to the infrequent occurrence of autoimmune disease in tropical Africa but not in Africans living in North America, and proposed that this difference was caused by multiple parasitic infections since childhood altering the immunological status of the tropical African population. We offer an explanation for this phenomenon based on the immunosuppressive effects of the pleotropic secondary messenger, nitric oxide, being chronically increased when nitric oxide is upregulated during the establishment and maintenance of malarial tolerance.
Nitric oxide and cerebral malaria As part of the debate on our proposal2 that nitric oxide, generated soon after schizogony in blood-vessel walls near sequestered parasites, might mediate the reversible coma of cerebral malaria, laboratories have measured serum nitrite plus nitrate (reactive nitrogen intermediates [RNI]) as a crude indicator of overall nitric-oxide generation in patients with cerebral malaria. Results have varied, depending in part on the degree of malarial tolerance in the population.3 For instance, RNI levels correlated positively with the presence of coma in groups in Vietnam and Brazil that can be safely assumed (from their ages) not to be part of a malaria-tolerant population.3 Moreover, in a study of 92 children with cerebral malaria in the Madang Province, a hyperendemic area in Papua New Guinea, we found plasma RNI, although low, to correlate positively with depth and duration of coma, as well as fatal outcome.4 On the basis of the objections by Anstey and co-workers,5 we took serum creatinine, a marker for glomerular filtration rate into account, but this did not alter our conclusions.6 The presence of severe disease in these children implies that they had not yet achieved, or had temporarily lost, malarial tolerance. Lancet 1996; 348: 1492–94 Division of Biochemistry and Molecular Biology, School of Life Sciences, Australian National University, Canberra, ACT 0200, Australia (I A Clarke DSc , F M Al-Yaman PhD); John Curtin School of Medial Research, Australian National University, Canberra (W B Cowden PhD); and Department of Paediatrics, John Radcliffe Hospital, Oxford, UK (K A Rockett PhD) Correspondence to: Dr I A Clark
1492
Nitric oxide and malaria tolerance A different picture is emerging in plasma from malariatolerant individuals (parasitaemia but no illness). Their plasma RNI is higher than that from cerebral malaria cases. We made this observation in the highly endemic Madang Province, comparing 81 cerebral malaria children7 with 88 new individuals (41 children, mean age 4, with mild malaria and 37 children, mean age 5, with asymptomatic parasitaemia). At least half of the children were matched for village, and the rest came from nearby locations. The concentrations of tumour necrosis factor were 61·8, 65·7, and 25·8 pg/mL for children with cerebral, mild, and asymptomatic infections, respectively (p=0·0026). The corresponding median RNI values were 19·9, 20·6, and 40 µmol/L (p=0·00014). A recent study from coastal Tanzania5 showed that inducible nitric-oxide synthase (iNOS), measured in peripheral blood mononuclear cells by western blot, correlated with the serum RNI increases across several groups. This finding included those with coma, who had the lowest RNI levels and no detectable iNOS in their peripheral blood mononuclear cells. Compared with other human cells or the equivalent cells from mice, human peripheral blood mononuclear cells are usually poor at being stimulated to make nitric oxide.7 The Tanzanian children severely ill with malaria6 and normal adults from the USA8 both conform to this pattern, as, from their RNI levels (median 12·6 µmol/L, range 9·5–34·9), do healthy adults (presumed from their ages no longer to be tolerant, but immune, to malaria) in coastal Papua New Guinea (n=12, age range 18–40). We therefore think it is most significant that Anstey et al found considerable iNOS activity in peripheral blood mononuclear cells from their healthy local children, both the nine with and the 25 without parasitaemias on the day they were examined. These children were at the age at which malaria tolerance is normal in such populations. Similarly, a study in Gabon9 included a group of 28 patients in whom plasma RNI was measured a month after admission and treatment for severe falciparum malaria, when levels were found to be higher than they had been when the children were severely ill. From the studies of the kinetics of change in pyrogenic threshold collected by Kitchen10 these children can be expected to have begun to acquire malarial tolerance by this time.
Vol 348 • November 30, 1996
THE LANCET
These observations raise the question of the origin and significance of increased plasma RNI in malaria-tolerant children, compared with those who are not tolerant but severely ill. The results from Anstey et al5 suggest that the RNI in tolerant children indicates increased generation of nitric oxide in cells of mononuclear phagocyte lineage. In lymphoid organs these cells and the nitric oxide they produce would be close enough to T lymphocytes to influence their function. By contrast, our model for cerebral malaria envisages the RNI in the plasma of cerebral malaria patients arising from nitric oxide generated in cerebral vascular endothelium, near sites of sequestration, and influencing nearby neurons.2 As discussed below, we propose that the nitric oxide causing high serum RNI in malaria-tolerant individuals, and apparently generated in mononuclear phagocytes during the establishment and maintenance of malarial tolerance, makes autoimmune disease rare in many tropical rural populations by minimising proliferation of adjacent autoreactive T cells. Even a first malarial infection routinely generates tolerance to the harmful effects of both malarial parasites10 and bacterial endotoxin.11 Antibodies directed against the products of malarial schizogony do not cross-react with endotoxin,12 so a humoral explanation for this dual tolerance is most unlikely. As reviewed by ZieglerHeitbrock,13 endotoxin tolerance is now seen as an active process, with the transcription factor NF-kappaB still being mobilised, but the complex consisting predominantly of p50p50 homodimers, instead of the usual p50p65 heterodimers, causing the profile of gene expression to differ from that seen on first exposure to endotoxin. Gene products downregulated during endotoxin tolerance include tumour necrosis factors, and those upregulated include the inducible form of nitricoxide synthase.14,15 This idea is consistent with tumour necrosis being lowest and RNI levels highest in the asymptomatic parasitaemia groups in both our Papua New Guinean and the Tanzanian studies,5 and argues that endotoxin tolerance and malarial tolerance are fundamentally the same biological process. Moreover, it is clear that this increased nitric oxide has an active role in endotoxin tolerance, mediating the reduction in production of tumour necrosis factor.14,16
Malarial immunosuppression The immunosuppression that accompanies malaria infections is well-documented and of practical importance, with malaria-infected children having more severe gastrointestinal and respiratory infections than normal children, and impaired efficacy of vaccination.17 This immunosuppression has been associated with a poor mitogenic response of lymphocytes,18 when lymphocytes are in close proximity to macrophages releasing nitric oxide.19 Recent in-vitro and in-vivo evidence in mouse malaria argues that the poor proliferative response of lymphocytes in this model is mediated through nitric oxide,20 most likely by nitric oxide’s ability to inhibit ribonucleotide reductase. The new data on increased plasma RNI in malaria-tolerant children in Papua New Guinea and Tanzania5 makes us confident to broaden this mechanism to longer-term human malarial immunosuppression, as well as during acute infections. We note that Williamson and Greenwood17 found that the immune response to meningococcal vaccine was still
Vol 348 • November 30, 1996
impaired a month after a malarial attack, a time when, in the Gabon study,9 plasma RNI was still high in children who had recovered from severe malaria. Thus much of malarial immunosuppression can be viewed as a sideeffect of the nitric oxide generated during the establishment and maintenance of malarial tolerance.
Nitric oxide and autoimmune disease Similar to other immunological diseases, multiple sclerosis is rare in the tropical world. Experimental allergic encephalomyelitis, the usual laboratory model for multiple sclerosis, is aggravated when the ability of the macrophages in affected rats to generate nitric oxide is inhibited by arginine analogues.21 Since the development of experimental allergic encephalomyelitis depends on T-cell proliferation, and is suppressed when this proliferation is prevented,22 this result implies that the lesions were exacerbated in these experiments21 because the nitric oxide that would have limited the proliferation of autoreactive T cells was absent. Our data in this model, in which an arginine analogue permitted T-cell proliferation commensurate with its ability to enhance lesion severity, are consistent with this reasoning. Conversely, when macrophage nitric oxide is chronically enhanced, such as during the establishment and maintenance of malarial tolerance, such autoreactive T cells would be strictly controlled, greatly reducing the frequency of autoimmune diseases. Greenwood’s original suggestion1 was that many different parasitic infections, not just malaria, were responsible for the rarity of autoimmune disease in Nigeria. Our proposal is consistent with this broader picture, in that the suppression of T-cell responses in Trypanosoma brucei23 and staphylococcal24 infections have also been shown to be mediated by nitric oxide. A related unexplained occurrence is the higher prevalence of systemic lupus erythematosus in people of African descent living in the UK than in whites.25 A corollary to our hypothesis is to predict that despite the high RNI levels during malarial tolerance, Africans possess an innately lower ability to generate nitric oxide from the inducible enzyme, developed over generations of selection for survival in the face of malaria disease, for which there is increasing evidence implicating nitric oxide as a mediator. Thus Africans may innately generate less nitric oxide than whites under the same conditions. By lessening the normal brake, provided by this mediator, on proliferation of autoreactive and other T cells, this could explain the higher prevalence of systemic lupus erythematosus and other autoimmune diseases, such as sarcoidosis, that Africans experience in the UK and other countries where malaria does not occur. This would also be consistent with the higher pyrogenic thresholds reported in non-immune Afro-American volunteers infected with Plasmodium falciparum than in whites under the same conditions,26 since inhibiting nitric oxide suppresses lipopolysaccharide-induced fever.27 In addition, were this proposed innately lower production of nitric oxide to include that produced by endothelial cells, it could contribute to the greater tendency of AfroAmericans to develop salt-sensitive hypertension, a condition associated experimentally with low nitric-oxide production, than do whites in the same circumstances.28 These concepts could be explored in various ways, including establishing whether the reduced T-cell 1493
THE LANCET
proliferative ability and endotoxin tolerance seen in patients after a malarial attack are associated with increased production of nitric oxide. A further prediction is a tendency for T cells from Africans born in North America or the UK to proliferate more readily, when under the influence of autologous mononuclear phagocytes, than T cells from suitable white controls. Our research receives financial support from the Australian National Health and Medical Research Council and the Ben Brown Anti-Malaria Fund.
References 1
Greenwood BM. Autoimmune disease and parasitic infections in Nigerians. Lancet 1968; ii: 380–82. 2 Clark IA, Rockett KA, Cowden WB. Possible central role of nitric oxide in conditions clinically similar to cerebral malaria. Lancet 1992; 340: 894–96. 3 Nussler AK, Eling W, Kremsner PG. Patients with Plasmodium falciparum malaria and Plasmodium vivax malaria show increased nitrite and nitrate plasma levels. J Infect Dis 1994; 169: 1418–19. 4 Al Yaman F, Genton B, Mokela D, et al. Association between serum levels of reactive nitrogen intermediates and coma in children with cerebral malaria in Papua New Guinea. Trans R Soc Trop Med Hyg 1996; 90: 270–73. 5 Anstey NM, Weinberg JB, Hassanali M, et al. Nitric oxide in Tanzanian children with malaria: inverse relationship between malaria severity and nitric oxide production/nitric oxide synthase type 2 expression. J Exp Med 1996; 184: 557–67. 6 Al-Yaman F, Awburn M, Clark IA. Serum creatinine levels and reactive nitrogen intermediates in children with cerebral malaria in Papua New Guinea. Trans R Soc Trop Med Hyg (in press). 7 Murray HW, Teitelbaum RF. L-arginine dependent reactive nitrogen intermediates and the antimicrobial effect of activated human mononuclear phagocytes. J Infect Dis 1992; 165: 513–17. 8 Weinberg JB, Misukonis MA, Shami PJ, et al. Human mononuclear phagocyte inducible nitric oxide synthase (iNOS): analysis of iNOS mRNA, iNOS protein, biopterin, and nitric oxide production by blood monocytes and peritoneal macrophages. Blood 1995; 86: 1184–95. 9 Kremsner PG, Winkler S, Wildling E, et al. High plasma levels of nitrogen oxides are associated with severe disease and correlate with rapid parasitological and clinical cure in plasmodium falciparum malaria. Trans R Soc Trop Med Hyg 1996; 90: 44–47. 10 Kitchen SF. Falciparum malaria. In: Boyd MF, ed. Malariology. Philadelphia: WB Saunders, 1949: 966–94. 11 Rubenstein M, Mulholland JH, Jeffery GM, et al. Malaria-induced endotoxin tolerance. Proc Soc Exp Biol Med 1965; 118: 238–87.
1494
12 Bate CAW, Taverne J, Dave A, et al. Malaria exoantigens induce T-independent antibody that blocks their ability to induce TNF. Immunology 1990; 70: 315–20. 13 Ziegler-Heitbrock HWL. Molecular mechanism in tolerance to lipopolysaccharide. J Inflamm 1995; 45: 13–26. 14 Fahmi H, Charon D, Mondange M, et al. Endotoxin-induced desensitization of mouse macrophages is mediated in part by nitric oxide production. Infect Immun 1995; 63: 1863–69. 15 Zingarelli B, Halushka PV, Caputi AP, et al. Increased nitric oxide synthesis during the development of endotoxin tolerance. Shock 1995; 3: 102–08. 16 Rojas A, Padron J, Caveda L, et al. Role of nitric oxide pathway in the protection against lethal endotoxemia afforded by low doses of lipopolysaccharide. Biochem Biophys Res Commun 1993; 191: 441–46. 17 Williamson WA, Greenwood BM. Impairment of the immune response to vaccination after acute malaria. Lancet 1978; i: 1328–29. 18 Taylor DW, Siddiqui WA. Effect of falciparum malaria infection on the in vitro mitogen responses of spleen and peripheral blood lymphocytes from owl monkeys. Am J Trop Med Hyg 1978; 27: 738–42. 19 Mills CD. Molecular basis of “suppressor” macrophages: arginine metabolism via the nitric oxide synthetase pathway. J Immunol 1991; 146: 2719–23. 20 Rockett KA, Awburn MM, Rockett EJ, et al. Possible role of nitric oxide in malarial immunosuppression. Parasite Immunol 1994; 16: 243–49. 21 Ruuls SR, Vanderlinden S, Sontrop K, et al. Aggravation of experimental allergic encephalomyelitis (EAE) by administration of nitric oxide (NO) synthase inhibitors. Clin Exp Immunol 1996; 103: 467–74. 22 von Muller CS, Spitler LE, LeCocq J. Experimental allergic encephalitis: study of cellular immunity during disease suppression. Eur J Immunol 1978; 8: 771–76. 23 Sternberg J, McGuigan F. Nitric oxide mediates suppression of T-cell responses in murine Trypanosoma brucei infection. Eur J Immunol 1992; 22: 2741–44. 24 Isobe KI, Nakashima I. Feedback suppression of staphylococcal enterotoxin-stimulated lymphocyte-T proliferation by macrophages through inductive nitric oxide synthesis. Infect Immun 1992; 60: 4832–37. 25 Johnson AE, Gordon C, Palmer RG, et al. The prevalence and incidence of systemic lupus erythematosus in Birmingham, England. Arthritis Rheum 1995; 38: 551–58. 26 Powell RD, McNamara JD, Reickman KH. Clinical acquisition of immunity to falciparum malaria. Proc Helminth Soc Wash 1972; 39: 51–66. 27 Scammell TE, Elmquist JK, Saper CB. Inhibition of nitric oxide synthase produces hypothermia and depresses lipopolysaccharide fever. Am J Physiol 1996; 40: R333–38. 28 Rutledge DR. Race and hypertension: what is clinically relevant? Drugs 1994; 47: 914–32.
Vol 348 • November 30, 1996
Hypothesis
Does malarial tolerance, through nitric oxide, explain the low incidence of autoimmune disease in tropical Africa?
I A Clark, F M Al-Yaman, W B Cowden, K A Rockett Autoimmune disease is generally rare in tropical rural populations. Plasma concentrations of nitrite plus nitrate (reactive nitrogen intermediates), reflecting high nitric-oxide production somewhere in the body, can be high in patients who have cerebral malaria, but even higher in symptom-free parasitised individuals, who are termed malaria-tolerant. We propose that the nitric oxide causing high serum levels of reactive nitrogen intermediates in malaria-tolerant individuals is generated in macrophages during the establishment and maintenance of malarial tolerance, and makes autoimmune disease rare in many tropical rural populations by minimising proliferation of autoreactive T cells. Conversely, innately low levels of nitric-oxide generation in these populations, selected by malarial disease in tropical areas, could rationalise their high frequency of autoimmune disease and hypertension when living in western societies. In 1968 Greenwood1 drew attention to the infrequent occurrence of autoimmune disease in tropical Africa but not in Africans living in North America, and proposed that this difference was caused by multiple parasitic infections since childhood altering the immunological status of the tropical African population. We offer an explanation for this phenomenon based on the immunosuppressive effects of the pleotropic secondary messenger, nitric oxide, being chronically increased when nitric oxide is upregulated during the establishment and maintenance of malarial tolerance.
Nitric oxide and cerebral malaria As part of the debate on our proposal2 that nitric oxide, generated soon after schizogony in blood-vessel walls near sequestered parasites, might mediate the reversible coma of cerebral malaria, laboratories have measured serum nitrite plus nitrate (reactive nitrogen intermediates [RNI]) as a crude indicator of overall nitric-oxide generation in patients with cerebral malaria. Results have varied, depending in part on the degree of malarial tolerance in the population.3 For instance, RNI levels correlated positively with the presence of coma in groups in Vietnam and Brazil that can be safely assumed (from their ages) not to be part of a malaria-tolerant population.3 Moreover, in a study of 92 children with cerebral malaria in the Madang Province, a hyperendemic area in Papua New Guinea, we found plasma RNI, although low, to correlate positively with depth and duration of coma, as well as fatal outcome.4 On the basis of the objections by Anstey and co-workers,5 we took serum creatinine, a marker for glomerular filtration rate into account, but this did not alter our conclusions.6 The presence of severe disease in these children implies that they had not yet achieved, or had temporarily lost, malarial tolerance. Lancet 1996; 348: 1492–94 Division of Biochemistry and Molecular Biology, School of Life Sciences, Australian National University, Canberra, ACT 0200, Australia (I A Clarke DSc , F M Al-Yaman PhD); John Curtin School of Medial Research, Australian National University, Canberra (W B Cowden PhD); and Department of Paediatrics, John Radcliffe Hospital, Oxford, UK (K A Rockett PhD) Correspondence to: Dr I A Clark
1492
Nitric oxide and malaria tolerance A different picture is emerging in plasma from malariatolerant individuals (parasitaemia but no illness). Their plasma RNI is higher than that from cerebral malaria cases. We made this observation in the highly endemic Madang Province, comparing 81 cerebral malaria children7 with 88 new individuals (41 children, mean age 4, with mild malaria and 37 children, mean age 5, with asymptomatic parasitaemia). At least half of the children were matched for village, and the rest came from nearby locations. The concentrations of tumour necrosis factor were 61·8, 65·7, and 25·8 pg/mL for children with cerebral, mild, and asymptomatic infections, respectively (p=0·0026). The corresponding median RNI values were 19·9, 20·6, and 40 µmol/L (p=0·00014). A recent study from coastal Tanzania5 showed that inducible nitric-oxide synthase (iNOS), measured in peripheral blood mononuclear cells by western blot, correlated with the serum RNI increases across several groups. This finding included those with coma, who had the lowest RNI levels and no detectable iNOS in their peripheral blood mononuclear cells. Compared with other human cells or the equivalent cells from mice, human peripheral blood mononuclear cells are usually poor at being stimulated to make nitric oxide.7 The Tanzanian children severely ill with malaria6 and normal adults from the USA8 both conform to this pattern, as, from their RNI levels (median 12·6 µmol/L, range 9·5–34·9), do healthy adults (presumed from their ages no longer to be tolerant, but immune, to malaria) in coastal Papua New Guinea (n=12, age range 18–40). We therefore think it is most significant that Anstey et al found considerable iNOS activity in peripheral blood mononuclear cells from their healthy local children, both the nine with and the 25 without parasitaemias on the day they were examined. These children were at the age at which malaria tolerance is normal in such populations. Similarly, a study in Gabon9 included a group of 28 patients in whom plasma RNI was measured a month after admission and treatment for severe falciparum malaria, when levels were found to be higher than they had been when the children were severely ill. From the studies of the kinetics of change in pyrogenic threshold collected by Kitchen10 these children can be expected to have begun to acquire malarial tolerance by this time.
Vol 348 • November 30, 1996
THE LANCET
These observations raise the question of the origin and significance of increased plasma RNI in malaria-tolerant children, compared with those who are not tolerant but severely ill. The results from Anstey et al5 suggest that the RNI in tolerant children indicates increased generation of nitric oxide in cells of mononuclear phagocyte lineage. In lymphoid organs these cells and the nitric oxide they produce would be close enough to T lymphocytes to influence their function. By contrast, our model for cerebral malaria envisages the RNI in the plasma of cerebral malaria patients arising from nitric oxide generated in cerebral vascular endothelium, near sites of sequestration, and influencing nearby neurons.2 As discussed below, we propose that the nitric oxide causing high serum RNI in malaria-tolerant individuals, and apparently generated in mononuclear phagocytes during the establishment and maintenance of malarial tolerance, makes autoimmune disease rare in many tropical rural populations by minimising proliferation of adjacent autoreactive T cells. Even a first malarial infection routinely generates tolerance to the harmful effects of both malarial parasites10 and bacterial endotoxin.11 Antibodies directed against the products of malarial schizogony do not cross-react with endotoxin,12 so a humoral explanation for this dual tolerance is most unlikely. As reviewed by ZieglerHeitbrock,13 endotoxin tolerance is now seen as an active process, with the transcription factor NF-kappaB still being mobilised, but the complex consisting predominantly of p50p50 homodimers, instead of the usual p50p65 heterodimers, causing the profile of gene expression to differ from that seen on first exposure to endotoxin. Gene products downregulated during endotoxin tolerance include tumour necrosis factors, and those upregulated include the inducible form of nitricoxide synthase.14,15 This idea is consistent with tumour necrosis being lowest and RNI levels highest in the asymptomatic parasitaemia groups in both our Papua New Guinean and the Tanzanian studies,5 and argues that endotoxin tolerance and malarial tolerance are fundamentally the same biological process. Moreover, it is clear that this increased nitric oxide has an active role in endotoxin tolerance, mediating the reduction in production of tumour necrosis factor.14,16
Malarial immunosuppression The immunosuppression that accompanies malaria infections is well-documented and of practical importance, with malaria-infected children having more severe gastrointestinal and respiratory infections than normal children, and impaired efficacy of vaccination.17 This immunosuppression has been associated with a poor mitogenic response of lymphocytes,18 when lymphocytes are in close proximity to macrophages releasing nitric oxide.19 Recent in-vitro and in-vivo evidence in mouse malaria argues that the poor proliferative response of lymphocytes in this model is mediated through nitric oxide,20 most likely by nitric oxide’s ability to inhibit ribonucleotide reductase. The new data on increased plasma RNI in malaria-tolerant children in Papua New Guinea and Tanzania5 makes us confident to broaden this mechanism to longer-term human malarial immunosuppression, as well as during acute infections. We note that Williamson and Greenwood17 found that the immune response to meningococcal vaccine was still
Vol 348 • November 30, 1996
impaired a month after a malarial attack, a time when, in the Gabon study,9 plasma RNI was still high in children who had recovered from severe malaria. Thus much of malarial immunosuppression can be viewed as a sideeffect of the nitric oxide generated during the establishment and maintenance of malarial tolerance.
Nitric oxide and autoimmune disease Similar to other immunological diseases, multiple sclerosis is rare in the tropical world. Experimental allergic encephalomyelitis, the usual laboratory model for multiple sclerosis, is aggravated when the ability of the macrophages in affected rats to generate nitric oxide is inhibited by arginine analogues.21 Since the development of experimental allergic encephalomyelitis depends on T-cell proliferation, and is suppressed when this proliferation is prevented,22 this result implies that the lesions were exacerbated in these experiments21 because the nitric oxide that would have limited the proliferation of autoreactive T cells was absent. Our data in this model, in which an arginine analogue permitted T-cell proliferation commensurate with its ability to enhance lesion severity, are consistent with this reasoning. Conversely, when macrophage nitric oxide is chronically enhanced, such as during the establishment and maintenance of malarial tolerance, such autoreactive T cells would be strictly controlled, greatly reducing the frequency of autoimmune diseases. Greenwood’s original suggestion1 was that many different parasitic infections, not just malaria, were responsible for the rarity of autoimmune disease in Nigeria. Our proposal is consistent with this broader picture, in that the suppression of T-cell responses in Trypanosoma brucei23 and staphylococcal24 infections have also been shown to be mediated by nitric oxide. A related unexplained occurrence is the higher prevalence of systemic lupus erythematosus in people of African descent living in the UK than in whites.25 A corollary to our hypothesis is to predict that despite the high RNI levels during malarial tolerance, Africans possess an innately lower ability to generate nitric oxide from the inducible enzyme, developed over generations of selection for survival in the face of malaria disease, for which there is increasing evidence implicating nitric oxide as a mediator. Thus Africans may innately generate less nitric oxide than whites under the same conditions. By lessening the normal brake, provided by this mediator, on proliferation of autoreactive and other T cells, this could explain the higher prevalence of systemic lupus erythematosus and other autoimmune diseases, such as sarcoidosis, that Africans experience in the UK and other countries where malaria does not occur. This would also be consistent with the higher pyrogenic thresholds reported in non-immune Afro-American volunteers infected with Plasmodium falciparum than in whites under the same conditions,26 since inhibiting nitric oxide suppresses lipopolysaccharide-induced fever.27 In addition, were this proposed innately lower production of nitric oxide to include that produced by endothelial cells, it could contribute to the greater tendency of AfroAmericans to develop salt-sensitive hypertension, a condition associated experimentally with low nitric-oxide production, than do whites in the same circumstances.28 These concepts could be explored in various ways, including establishing whether the reduced T-cell 1493
THE LANCET
proliferative ability and endotoxin tolerance seen in patients after a malarial attack are associated with increased production of nitric oxide. A further prediction is a tendency for T cells from Africans born in North America or the UK to proliferate more readily, when under the influence of autologous mononuclear phagocytes, than T cells from suitable white controls. Our research receives financial support from the Australian National Health and Medical Research Council and the Ben Brown Anti-Malaria Fund.
References 1
Greenwood BM. Autoimmune disease and parasitic infections in Nigerians. Lancet 1968; ii: 380–82. 2 Clark IA, Rockett KA, Cowden WB. Possible central role of nitric oxide in conditions clinically similar to cerebral malaria. Lancet 1992; 340: 894–96. 3 Nussler AK, Eling W, Kremsner PG. Patients with Plasmodium falciparum malaria and Plasmodium vivax malaria show increased nitrite and nitrate plasma levels. J Infect Dis 1994; 169: 1418–19. 4 Al Yaman F, Genton B, Mokela D, et al. Association between serum levels of reactive nitrogen intermediates and coma in children with cerebral malaria in Papua New Guinea. Trans R Soc Trop Med Hyg 1996; 90: 270–73. 5 Anstey NM, Weinberg JB, Hassanali M, et al. Nitric oxide in Tanzanian children with malaria: inverse relationship between malaria severity and nitric oxide production/nitric oxide synthase type 2 expression. J Exp Med 1996; 184: 557–67. 6 Al-Yaman F, Awburn M, Clark IA. Serum creatinine levels and reactive nitrogen intermediates in children with cerebral malaria in Papua New Guinea. Trans R Soc Trop Med Hyg (in press). 7 Murray HW, Teitelbaum RF. L-arginine dependent reactive nitrogen intermediates and the antimicrobial effect of activated human mononuclear phagocytes. J Infect Dis 1992; 165: 513–17. 8 Weinberg JB, Misukonis MA, Shami PJ, et al. Human mononuclear phagocyte inducible nitric oxide synthase (iNOS): analysis of iNOS mRNA, iNOS protein, biopterin, and nitric oxide production by blood monocytes and peritoneal macrophages. Blood 1995; 86: 1184–95. 9 Kremsner PG, Winkler S, Wildling E, et al. High plasma levels of nitrogen oxides are associated with severe disease and correlate with rapid parasitological and clinical cure in plasmodium falciparum malaria. Trans R Soc Trop Med Hyg 1996; 90: 44–47. 10 Kitchen SF. Falciparum malaria. In: Boyd MF, ed. Malariology. Philadelphia: WB Saunders, 1949: 966–94. 11 Rubenstein M, Mulholland JH, Jeffery GM, et al. Malaria-induced endotoxin tolerance. Proc Soc Exp Biol Med 1965; 118: 238–87.
1494
12 Bate CAW, Taverne J, Dave A, et al. Malaria exoantigens induce T-independent antibody that blocks their ability to induce TNF. Immunology 1990; 70: 315–20. 13 Ziegler-Heitbrock HWL. Molecular mechanism in tolerance to lipopolysaccharide. J Inflamm 1995; 45: 13–26. 14 Fahmi H, Charon D, Mondange M, et al. Endotoxin-induced desensitization of mouse macrophages is mediated in part by nitric oxide production. Infect Immun 1995; 63: 1863–69. 15 Zingarelli B, Halushka PV, Caputi AP, et al. Increased nitric oxide synthesis during the development of endotoxin tolerance. Shock 1995; 3: 102–08. 16 Rojas A, Padron J, Caveda L, et al. Role of nitric oxide pathway in the protection against lethal endotoxemia afforded by low doses of lipopolysaccharide. Biochem Biophys Res Commun 1993; 191: 441–46. 17 Williamson WA, Greenwood BM. Impairment of the immune response to vaccination after acute malaria. Lancet 1978; i: 1328–29. 18 Taylor DW, Siddiqui WA. Effect of falciparum malaria infection on the in vitro mitogen responses of spleen and peripheral blood lymphocytes from owl monkeys. Am J Trop Med Hyg 1978; 27: 738–42. 19 Mills CD. Molecular basis of “suppressor” macrophages: arginine metabolism via the nitric oxide synthetase pathway. J Immunol 1991; 146: 2719–23. 20 Rockett KA, Awburn MM, Rockett EJ, et al. Possible role of nitric oxide in malarial immunosuppression. Parasite Immunol 1994; 16: 243–49. 21 Ruuls SR, Vanderlinden S, Sontrop K, et al. Aggravation of experimental allergic encephalomyelitis (EAE) by administration of nitric oxide (NO) synthase inhibitors. Clin Exp Immunol 1996; 103: 467–74. 22 von Muller CS, Spitler LE, LeCocq J. Experimental allergic encephalitis: study of cellular immunity during disease suppression. Eur J Immunol 1978; 8: 771–76. 23 Sternberg J, McGuigan F. Nitric oxide mediates suppression of T-cell responses in murine Trypanosoma brucei infection. Eur J Immunol 1992; 22: 2741–44. 24 Isobe KI, Nakashima I. Feedback suppression of staphylococcal enterotoxin-stimulated lymphocyte-T proliferation by macrophages through inductive nitric oxide synthesis. Infect Immun 1992; 60: 4832–37. 25 Johnson AE, Gordon C, Palmer RG, et al. The prevalence and incidence of systemic lupus erythematosus in Birmingham, England. Arthritis Rheum 1995; 38: 551–58. 26 Powell RD, McNamara JD, Reickman KH. Clinical acquisition of immunity to falciparum malaria. Proc Helminth Soc Wash 1972; 39: 51–66. 27 Scammell TE, Elmquist JK, Saper CB. Inhibition of nitric oxide synthase produces hypothermia and depresses lipopolysaccharide fever. Am J Physiol 1996; 40: R333–38. 28 Rutledge DR. Race and hypertension: what is clinically relevant? Drugs 1994; 47: 914–32.
Vol 348 • November 30, 1996