¹The Walter and Eliza Hall Institute of Medical Research, Melbourne, VIC, Australia

²Menzies School of Health Research, Darwin, NT, Australia

³Nuffield Department of Clinical Laboratory Science, University of Oxford, UK and National Blood Service, UK

⁴Division of Infectious Diseases, University of Utah, Salt Lake City, UT, USA

⁵VA & Duke University Medical Centers, Departments of Medicine and Immunology, Durham, NC, USA

⁶Kenya Medical Research Institute, Kilifi, Kenya

⁷Muhimbili Medical Centre, Dar es Salaam, Tanzania

Correspondence to: G Morahan, The Walter and Eliza Hall Institute of Medical Research, PO Royal Melbourne Hospital, Victoria 3050, Australia. E-mail: morahan@wehi.edu.au

^†This work was financially supported by the National Health and Medical Research Council of Australia and by a Special Program Grant from the NHMRC and the Juvenile Diabetes Research Foundation; National Institutes of Health RO1 AI41764, NHMRC Scholarship (CSB) and the Wellcome Trust, UK. DJR is a Howard Hughes International Research Scholar. No author has a commercial or other association that might pose a conflict of interest.

Interleukin-12 (IL-12) is an important regulatory cytokine in infection and immunity. Administration of IL-12 may reduce complications of severe malaria in rodents. Polymorphisms in IL12B, the gene encoding the IL-12 p40 subunit, influence the secretion of IL-12 and susceptibility to Type 1 diabetes. We therefore investigated whether IL12B polymorphisms may affect the outcome of severe malaria. Homozygosity for a polymorphism in the IL12B promoter was associated with increased mortality in Tanzanian children having cerebral malaria but not in Kenyan children with severe malaria. Furthermore, homozygotes for the IL12B promotor polymorphism had decreased production of nitric oxide, which is in part regulated by IL-12 activity. These studies suggest that IL12B polymorphisms, via regulation of IL-12 production, may influence the outcome of malaria infection in at least one African population.

Genes and Immunity (2002) 3, 414-418. doi:10.1038/sj.gene.6363909

interleukin-12; immune regulation; Th1-Th2; severe malaria

Introduction

Childhood malaria is a major cause of global morbidity and mortality.¹ Syndromes of severe and fatal malaria include cerebral malaria, metabolic acidosis, anemia and hypoglycemia.¹ One method of dissecting the complex pathophysiological processes involved in severe malaria is to identify genetic traits that confer protection from severe and fatal malaria and use these associations to investigate the role of particular mechanisms in acute infection. In this context, one particularly interesting gene to investigate is IL12B, which encodes the p40 subunit of interleukin-12 (IL-12), a heterodimer in which p35 is the second subunit (see ref. ² for review). Studies have suggested that IL-12 has a protective role in malaria,^3,4 possibly through its ability to promote interferon- production and development of Th1 cells. Thus, administration of IL-12 may prevent malaria-induced anemia³ and IL-12 levels are decreased in severe malaria in African children.⁵ However, the role of functional genetic polymorphisms of the genes encoding IL-12 in determining the outcome of clinical infection has not been described.

We have defined polymorphisms in and around IL12B,⁶ and have shown that a polymorphism in the IL12B 3' untranslated region (3'UTR) is associated with susceptibility to development of the autoimmune disease, Type 1 diabetes (T1D) in British and Australian families.⁷ The polymorphism associated with T1D susceptibility ('A' at position 16974 of the genomic sequence with Genbank code AY008847; referred to hereafter as the IL12B 3'UTR allele 1) was associated with higher levels of basal IL-12 expression in vitro.⁷ Therefore, individuals with this genotype may have an increased likelihood of mounting immune responses with a bias toward the Th1 phenotype. In contrast, diseases characterized by lower IL-12 production may show an association with the alternative IL12B 3'UTR allele. That is, the allele that confers susceptibility to T1D may confer resistance to other diseases.⁶ Recently, we have described another polymorphism, which resides upstream of the IL12B gene in a region which may have promoter activity. We therefore examined the influence of these two polymorphisms of the IL12B gene in the outcome of severe malaria in two independent case-control studies of severe malaria in children from East Africa.^8,9

Given our previous findings in T1D, we hypothesized that IL12B 3'UTR-2 homozygosity, associated with low levels of IL-12 secretion, would increase the risk of death from cerebral malaria. We were unable to measure IL-12 levels in these cohorts; so to investigate the functional significance of the IL12B polymorphisms, we examined whether different genotypes correlated with production of nitric oxide (NO), a downstream mediator of IL-12 activity.¹⁰

Results

Association between IL12B polymorphisms and outcome from cerebral malaria in Tanzanian subjects

Genotyping data were generated for the 178 Tanzanian subjects for whom DNA samples were available (Table 1). Each was typed for the IL12B 3'UTR polymorphism⁶ and for a complex insertion-deletion polymorphism 3 kb upstream of the transcriptional start site (see the Methods section and Morahan et al, submitted for publication), referred to here as the IL12B promoter polymorphism, IL12Bpro. The longer allele IL12Bpro is designated as IL12Bpro-1; IL12Bpro-2 is 4 bp shorter. There were no significant differences in IL12B genotypes between the two control groups (ie children exposed to malaria but not currently affected, and children who were experiencing a malaria attack without more severe complications; see the Methods section). However, the IL12Bpro polymorphism was associated with a significantly different likelihood of death from cerebral malaria (CM) in a 3´2 ² analysis (P=0.024). Individuals homozygous for IL12Bpro-1 were over-represented amongst the children who died from CM. As there was no difference in the likelihood of death from CM between IL12Bpro heterozygotes and IL12Bpro-2 homozygotes, these subjects were considered together as a single group. The unadjusted odds ratio (OR) for death from CM in IL12Bpro-1 homozygotes was 5.04 (P=0.013 (Table 1)). This finding remained significant after controlling for potentially confounding variables such as IL12B 3'UTR genotype, age and parasitemia (adjusted OR 5.09; 95% CI: 1.03-25.3; P=0.046).

Considering the IL12B 3'UTR polymorphism, the OR for death from CM in 3'UTR-2 homozygotes was 3.2 (95% CI: 1.0-10.7; P=0.055) after controlling for age and parasitemia. When the promoter genotype was added to the multivariate model, the OR for 3'UTR-2 homozygosity was reduced to 1.81 (95% CI: 0.45-7.21; P=0.40), suggesting that the 3'UTR effect may be explained by linkage disequilibrium with IL12Bpro-1. If so, it would be important to consider haplotypes based on the two IL12B polymorphisms typed.

There are four potential haplotypes. Based on the above analyses, it was hypothesized that homozygosity for the IL12Bpro-1/3'UTR-2 haplotype would confer greater susceptibility to death from CM. Seven of the nine individuals (ie 88%) with CM having this genotype died, compared to 15 deaths from 72 (21%) CM cases having all other genotypes (Table 2). Controlling for age and parasitemia, the OR for death from CM in subjects homozygous for the IL12Bpro-1/3'UTR-2 haplotype was 18.1 (95% CI: 2.87-115; P=0.002).

Association between IL12B polymorphisms and NO production

NO_x excretion, an indicator of NO production by NOS, has previously been shown to be significantly inversely associated with disease severity in these children, with levels lowest in children with CM (both fatal and non-fatal).⁸ iNOS production is stimulated in part by the Th1 cytokine interferon-. IL12B genotype could have an effect on NO production via increased IL12B gene expression, Th1 cell development and interferon- production. Therefore, fasting urinary excretion of the NO metabolites, nitrate+nitrite (NO_x), was measured in all Tanzanian subjects. In a univariate analysis, the geometric mean urinary NO_x/creatinine ratio was 0.11 (95% CI: 0.07-0.15) in IL12pro-1 homozygotes compared to 0.16 (95% CI: 0.14-0.19) in all other subjects (P=0.027). IL12BPro-1 homozygosity was independently predictive of lower NO_x excretion after controlling for the previously described disease association⁸ (P=0.012; Figure 1A). The significant predictive effect of IL12B genotype was maintained after controlling for additional potentially confounding variables (age, parasitemia, plasma creatinine and IL12B 3'UTR genotype; P=0.042). Mean urinary NO_x excretion was lowest in IL12B 3'UTR-2 homozygotes in each disease category (Figure 1B), but this was not statistically significant, either in a univariate model (P=0.066) or after controlling for disease severity (P=0.077).

Association of IL12B polymorphisms with outcome from severe malaria in Kenyan subjects

In the Kenyan children, no significant association was found between mortality from severe malaria and either the IL12Bpro-1 or IL12B 3'UTR-2 genotypes (data not shown). Similarly, there was no association between these IL12B genotypes and outcome in the subgroups with cerebral and non-cerebral severe malaria (data not shown). However, there was a weak association of the IL12Bpro-1/3'UTR-2 haplotype and another manifestation of severe disease, hypoglycemia: of 130 children who developed hypoglycemia, 16 (12%) were homozygous for this haplotype, while amongst those who did not become hypoglycemic, 31 of 477 (6%) had this genotype (P=0.03).

Discussion

The association we have found between the IL12Bpro genotype and cerebral malaria mortality in Tanzanian children supports a role for differential expression of IL-12 affecting outcome in cerebral malaria. How may this effect be mediated? IL-12 production has been associated with protective immune responses to both exoerythrocytic and erythrocytic stages of P. falciparum.^4,11 Administration of IL-12 ameliorates disease in a number of rodent models.^3,4,12 Levels of IL-12 are also inversely related to malaria disease severity in Gabonese children.⁵ IL-12 influences the development of interferon- secreting T cells, which are associated with a 'Th1' cellular immune response and the recruitment and activation of mononuclear cells² including induction of inducible nitric oxide synthase (iNOS). Mononuclear cell iNOS expression and systemic NO production have previously been associated with protection from cerebral malaria in Tanzania⁸ and with protection from severe malaria in other populations.^13,14

We have previously demonstrated that children in the Tanzanian cohort with CM produced less NO than healthy control children or children with uncomplicated malaria.⁸ In these same children, IL12Bpro-1 homozygosity was functionally associated with significantly lower NO production. The magnitude of the effect of IL12Bpro genotype on NO production in children with uncomplicated or cerebral malaria may reflect the influence of parasite factors¹⁵ and other cytokines such as IL-10 that negatively regulate NO production, as indices of NO production were markedly lower in these disease groups.⁸

Our results suggest that differential expression of genetically variable IL-12 production may be one of the mechanisms underlying the differential NO production and outcome seen in these Tanzanian children. Although IL12B genotypes did not affect increased risk of having CM, the findings are consistent with the hypothesis that once CM occurs, homozygosity of the IL12Bpro-1 allele increases risk of fatal outcome through impaired production of mediators such as NO.

Although there was a strong association between IL12B genotype and fatal outcome in the Tanzanian children, no such association was found in the Kenyan cohort. We cannot exclude an effect due to chance but we think this unlikely because of the stronger association that was observed with a particular IL12B haplotype: if the original observation was due to chance there was no reason for IL12Bpro-1 alleles that showed association with CM susceptibility to also show allelic association with IL12B 3'UTR allele 2. Furthermore, the functional correlation of protective genotype and NO production, which has been associated with disease protection,⁸ supports a role for IL12B genotypes in CM outcome. Similar differences in genetic associations with severe malaria susceptibility/outcome have been shown previously among different African populations and polymorphisms for ICAM-1, TNF- and haptoglobin 1-1, and with HLA types (reviewed in ref. ¹⁶). Although geographically close, the two populations studied here are ethnically distinct. It is possible that the IL12Bpro-1 polymorphism is in linkage disequilibrium with one or more polymorphisms that also influence gene expression and downstream NO production. It was not possible to determine the functional relationships between IL12B promoter polymorphisms and NO production in the Kenyan case-control study. Other functional mutations in or around the IL12B gene may also exist in the Kenyan population, and further parallel functional and genetic studies are required in both populations. It is also possible that IL12B alleles interact differentially with other genetic elements, especially the genes encoding HLA. The effect of IL12B in T1D was observed in HLA identical but not mismatched sib pairs,⁷ suggesting that a recessive HLA-linked gene might interact with IL12B. HLA types are known to differ between Kenya and Tanzania¹⁷ so this could conceivably account for some of the difference.

Differences in malaria epidemiology have also been proposed to account for temporal and geographic variation in protective effect of host polymorphisms¹⁶ yet the different observations of IL12Bpro polymorphism occurred here in coastal areas with similar malaria epidemiology. Both sites have rainfall-related seasonal peaks, and a relatively high proportion of CM, although with higher transmission in Kilifi than in urban Dar es Salaam. It is possible that unknown parasite-determined virulence factors may also differ between the two study areas.

The different explanations of the associations of the IL12B polymorphisms in the outcome of malaria in two distinct African populations are not mutually exclusive, and a number of effects may be operating. How should we proceed with the investigation of the role of IL12B genotypes in malaria? The strong association of the IL12Bpro-1 polymorphisms in the outcome of malaria infection in Tanzania and the functional association of this polymorphism with urinary NO_x excretion and thus, by inference, IL-12 and NO production are consistent with the hypothesis that decreased IL-12 production may be harmful in severe malaria. Lack of an association of the same polymorphisms in Kenya can be explained in a variety of ways, as we have indicated. Perhaps the most plausible explanation in the light of all the data, including the observations of the role of IL-12 in animal malaria and other studies of acute malaria, is that the critical IL12B polymorphism(s) which regulate susceptibility to severe malaria are in linkage disequilibrium with different marker alleles in different populations. Such a crucial susceptibility allele should be found on the IL12B haplotype we have defined in the Tanzanian CM subjects. It will be important to determine if the alleles studied here, or other IL12B polymorphisms, influence the outcome of malaria infection in other populations, as well as the relationship of IL12B genotypes to IL-12 secretion and NO production.

Complex diseases with polygenic resistant factors will not easily be understood by single case-control studies. We therefore believe it is more helpful to present all the available data to allow the significance of the positive finding in a single study to be honestly evaluated and, crucially, to allow the planning of future studies in the light of as much relevant data as possible. In particular, our studies emphasize the importance of combining genetic and functional studies wherever possible and to examine the putative associations in multiple studies and/or distinct populations whenever and wherever possible. Only with such an integrated approach can we hope to progress from descriptive epidemiology to a reliable appreciation of pathophysiology.

Materials and methods

Subjects

DNA was genotyped from two previously described cohorts of subjects from Tanzania⁸ and Kenya.⁹ The Tanzanian cohort included 191 children admitted to Muhimbili Medical Centre, Dar es Salaam, with and without CM. Of these, 86 had cerebral malaria (CM) with unrousable coma, 55 had uncomplicated clinical malaria (UM) and 50 were asymptomatic malaria-exposed heal-thy control (HC) children (with or without subclinical parasitemia at the time of sample collection).⁸ The Kenyan case-control study⁹ comprised 693 cases of severe malaria and 693 ethnically matched community controls from consecutive primagravidae attending the antenatal clinic in Kilifi. Of those with severe malaria, 413 had CM, and 280 had severe anemia, respiratory distress and/or hypoglycemia as previously defined.⁹ Blood samples (and urine samples in Tanzania) were obtained with informed consent from parents of children and from adult controls. The Ethics Committee of each institution approved these studies.

Genotyping

DNA from Tanzanian children was amplified from blood filter spots using whole genome amplification (Hobbs et al., submitted for publication).¹⁸ DNA samples were genotyped as described previously for the single nucleotide polymorphism at the IL12B 3'UTR.⁶ The IL12Bpro polymorphism was found by extending our previous search for IL12B polymorphisms.⁶ This variant is an insertion/deletion polymorphism (Morahan et al, submitted for publication). Alleles were genotyped as follows. DNA samples were amplified using the following primers (5'3': TCAGACACATTAACCTTGCA and TAATGTGGTCATTGGCAGGT, one of which was radiolabelled with ³²P-ATP) in PCR reactions with the following conditions: 50 ng DNA, 25 ng each primer, 200 nM dNTP, 2 mM MgCl₂, 0.3 U Taq polymerase (Gibco); DNA was denatured at 95°C for 3 min; then PCR was performed for 35 cycles of 95°C 20 s, 55°C 20 s, 72°C 30 s, with a final 2 min extension. Alleles were detected after separation on polyacrylamide gels. Allele 1 was designated as the 4 bp larger (ie insertion) allele; the smaller product was designated as allele 2.

No production

Levels of fasting urinary excretion of the NO metabolites, nitrate+nitrite (NO_x), had been previously measured on a low-nitrate diet in each Tanzanian disease group.⁸ Because of variability in urine concentration among subjects, NO_x levels were expressed as a ratio of [NO_x] to [creatinine].

Statistical analyses

Predicted outcomes from CM were modelled using logistic regression, controlling for factors expected to influence outcome including age and level of parasitemia. Linear regression was used to model the relationship between IL12B genotypes and NO production (log-transformed urine NO_x/creatinine ratio), controlling for factors expected to influence urine NO_x excretion including age group (0-3.5, 3.5-7, >7 year), parasitemia and renal function (plasma creatinine as percentage of normal for age). P values of <0.05 were considered to indicate statistical significance. Analyses were performed using Intercooled Stata 6.0 (Stata Corporation, TX).

Acknowledgements

We thank Zhiqiang Wang for statistical advice, and Mushtaq Hassanali and Denis Manyenga for assistance with data collection. This paper is published with the approval of the Director of KEMRI.

1 Marsh K, Forster D, Waruiru C et al. Indicators of life-threatening malaria in African children. N Engl J Med 1995; 332: 1399-1404. Article MEDLINE

2 Gately MK, Renzetti LM, Magram J, Stern AS, Gubler U, Presky DH. The interleukin-12/interleukin-12-receptor system: role in normal and pathologic immune responses. Annu Rev Immunol 1998; 16: 495-521. Article MEDLINE

3 Mohan K, Sam H, Stevenson MM. Therapy with a combination of low doses of interleukin 12 and chloroquine completely cures blood-stage malaria, prevents severe anemia, and induces immunity to reinfection. Infect Immun 1999; 67: 513-519.

4 Stevenson MM, Su Z, Sam H, Mohan K. Modulation of host responses to blood-stage malaria by interleukin-12: from therapy to adjuvant activity. Microbes Infect 2001; 3: 49-59.

5 Perkins DJ, Weinberg JB, Kremsner PG. Reduced interleukin-12 and transforming growth factor-beta 1 in severe childhood malaria: relationship of cytokine balance with disease severity. J Infect Dis 2000; 182: 988-992.

6 Huang D, Cancilla M, Morahan G. Complete primary structure, chromosomal localisation, and definition of polymorphisms of the gene encoding the human interleukin-12 p40 subunit. Genes Immun 2000; 1: 515-520. Article MEDLINE

7 Morahan G, Huang D, Ymer S et al. Linkage disequilibrium of a Type 1 diabetes susceptibility locus with a regulatory IL12B allele. Nat Genet 2001; 27: 218-221. Article MEDLINE

8 Anstey NM, Weinberg JB, Hassanali MY 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: 557-567.

9 Pain A, Urban BC, Kai O et al. A non-sense mutation in Cd36 gene is associated with protection from severe malaria. Lancet 2001; 357: 1502-1503. Article MEDLINE

10 Levesque MC, Hobbs MR, Anstey NM et al. A review of polymorphisms in the human gene for inducible nitric oxide synthase (NOS2) in patients with malaria. Sepsis 2001; 4: 217-231.

11 Sedegah M, Finkelman F, Hoffman SL. Interleukin 12 induction of interferon gamma-dependent protection against malaria. Proc Natl Acad Sci USA 1994; 91: 10700-10702.

12 Su Z, Stevenson MM. IL-12 is required for antibody-mediated protective immunity against blood-stage Plasmodium chabaudi AS malaria infection in mice. J Immunol 2002; 168: 1348-1355.

13 Perkins DJ, Kremsner PG, Schmid D, Misukonis MA, Kelly MA, Weinberg JB. Blood mononuclear cell nitric oxide production and plasma cytokine levels in healthy gabonese children with prior mild or severe malaria. Infect Immun 1999; 67: 4977-4981. MEDLINE

14 Chiwakata CB, Hemmer CJ, Dietrich M. High levels of inducible nitric oxide synthase mRNA are associated with increased monocyte counts in blood and have a beneficial role in Plasmodium falciparum malaria. Infect Immum 2000; 68: 394-399.

15 Taramelli D, Basilico N, Pagani E et al. The heme moiety of malaria pigment (beta-hematin) mediates the inhibition of nitric oxide and tumor necrosis factor-alpha production by lipopolysaccharide-stimulated macrophages. Exp Parasitol 1995; 81: 501-511.

16 Craig A, Hastings I, Pain A, Roberts DJ. Genetics and malaria¾more questions than answers. Trends Parasitol 2001; 17: 55-56.

17 Hall P, Levin A, Entwistle C, Knight S, Wasunna A, Brubaker G. B15 heterogeneity in East African blacks. Tissue Antigens 1980; 16: 326-332.

18 Hobbs M, Pole A, Anstey N. Whole genome amplification from field collected blood spots on filter paper. submitted for publication.

Figure 1 Association between IL12B polymorphisms and NO production (as reflected in urinary NO_x excretion, expressed as the natural logarithm of urinary NO_x:creatinine ratio), in Tanzanian children in each disease category:healthy malaria-exposed controls (HC), uncomplicated malaria (UM) and cerebral malaria (CM). The top, bottom, and middle lines of each box correspond to the 75th percentile, 25th percentile, and median, respectively; the mean for each group is also indicated (). Whiskers extend to the top 90th percentile and to the 10th percentile. (A) Individuals homozygous for IL12B promoter allele 1 (IL12Bpro-1; unshaded bars) compared to all other genotypes (shaded bars). (B) Individuals homozygous for 3'UTR allele 2 (unshaded bars) vs all other genotypes (shaded bars).

Table 1 Frequency of IL12B polymorphisms and genotypes within each disease category in Tanzanian children and odds ratio for death from CM

Table 2 Effect of IL12B haplotype on CM survival