MiRNA-146a Polymorphism Was Not Associated with Malaria in Southern India

Understanding the immune components that predispose individuals to susceptibility to and severe manifestation of malaria is a key component of controlling and eliminating this complex disease. Micro-RNAs (miRNAs) are small, noncoding RNA molecules in a hairpin structure, and they are involved in gene regulation on a posttranslational level.¹ There is increasing interest in miRNAs because of their regulatory role at various steps in the innate and adaptive immune network, which could open up novel clinical applications. Micro-RNA-146a is an important regulator in pattern recognition receptor signaling, such as the Toll-like receptor machinery.^1,2 Micro-RNA-146a suppresses NF-kB activation by a negative regulation loop, thereby altering several inflammatory cytokines.^3,4 It also participates in hematopoietic cell regulation,⁵ which might be relevant in malaria, considering its strong interference with blood cell homeostasis.

A common single-nucleotide polymorphism (SNP) in miRNA-146a, rs2910164 G>C, has been associated with modified risks for various types of cancer,⁶ autoimmune disorders,⁷ and infectious diseases.^8–10 The presence of the polymorphism results in reduced amounts of mature miRNA-146a in vitro.¹¹ Also, the SNP gives rise to a miRNA-146a molecule that is thought to target another set of genes than wild-type miRNA-146a, and, therefore, SNP heterozygosity might have a very different functional effect compared with either of the homozygous genotypes.¹²

Previously, we found that this SNP greatly increased the odds of malaria in pregnancy, particularly in primigravidae.⁸ Malaria in pregnancy has a distinct pathophysiology and immunity, which differs from those of nonpregnant hosts.¹³ Also, miRNAs have been shown to play a specific role at the maternal–fetal interface,^14,15 that is, the predilection site of malaria in pregnancy. Here, we examined whether miRNA-146a rs2910164 G>C also affects susceptibility to and/or manifestation of malaria in nonpregnant hosts including those with Plasmodium vivax infection.

The present study was conducted from June to December 2015 in Mangaluru, Karnataka, coastal southwestern India. Mangaluru is a harbor city of some 500,000 inhabitants exhibiting the peculiarity of urban malaria in the otherwise less affected state of Karnataka. Wenlock Hospital is the largest governmental health facility in Mangaluru particularly serving the socioeconomically deprived portion of the population of the Mangaluru conglomeration. Patients presenting at the outpatient department (OPD) of the Wenlock Hospital with malaria-like symptoms were referred to the malaria diagnostic unit for diagnosis using Giemsa-stained thick blood films. Patients with confirmed malaria diagnosis were consecutively recruited during the operating hours (08:00–16:00) of the OPD. Patients presenting beyond that time were not regarded. Detailed recruitment procedures, clinical and laboratory examinations, and patient characteristics have been published elsewhere.¹⁶ Afebrile, unmatched community controls were randomly recruited across Mangaluru City Corporation in parallel, and they provided a finger-prick blood sample.¹⁷ Informed written consent was obtained from all adult participants and from parents or legal guardians of minors. The study protocol was reviewed and approved by the Institutional Ethics Committee of Kasturba Medical College, Mangaluru, Manipal University, and permission to conduct the study was granted by the Directorate of Health and Family Welfare Services, Government of Karnataka. In brief, a medical history was obtained, and a clinical examination was performed for all patients. Parameters assessed or calculated included weight, height, body mass index, axillary temperature, fever (> 37.5°C), hemoglobin, white blood cell and thrombocyte counts, geometric mean parasite density, parasite species (polymerase chain reaction [PCR]-confirmed), and concentrations of creatinine and bilirubin.¹⁶ Severe malaria was defined based on the current WHO definition with some modifications,¹⁶ and severe thrombocytopenia as < 50,000/mcL.¹⁸ DNA was extracted (Qiamp blood mini kit, Qiagen, Hilden, Germany), and miRNA-146a rs2910164 was genotyped by melting curve analysis using commercially available primers and probes (TIB Molbiol, Berlin, Germany). Analysis was limited to patients and controls originating from Karnataka avoiding potential geographic impact on the genetic association study. Data analysis was performed using R version 3.4.3. Genotype frequencies in patients and controls were compared by a two-tailed Fisher’s exact test, and odds ratios, 95% CIs and P-values were produced. Clinical parameters of patients were compared between genotypes using a two-tailed Fisher’s exact test, Welch’s t-test, Wilcoxon signed-rank test, or Kruskal–Wallis test as applicable. Adjusted odds ratios (aORs) for malaria, clinical parameters, and self-reported symptoms were derived from a logistic regression model including age and gender. In addition, all analyses were stratified by Plasmodium spp., that is, P. vivax mono-infection, Plasmodium falciparum mono-infection, and mixed spp. infection. A P-value < 0.05 was considered statistically significant.

The mean age (SD) of the 449 malaria patients was 33.7 (13.0) years, and it was 35.0 (15.2) years in 666 community controls. Among the patients, only 11.8% (53/449) were female, but 47.9% (319/666) among the controls (P < 0.001). Plasmodium vivax malaria was present in 67.5% (303/449) of patients, P. falciparum malaria in 10.5% (47/449), and 22.0% (99/449) had a mixed P. vivax–P. falciparum infection. Overall, 58.9% (657/1115) of the study participants were carriers of the miRNA-146a SNP (46.8% heterozygous and 12.1% homozygous). Of note, the SNP distribution did not differ between patients and controls, irrespective of stratification by parasite species (Tables 1 and 2). Geometric mean parasite density (95% CI) in individuals with a wild-type genotype, and heterozygous and homozygous polymorphism was 3,422 (2,734–4,283), 4,057 (3,310–4,971), and 3,256 (2,092–5,068), respectively (P = 0.4). Likewise, absolute thrombocyte counts were similar in malaria patients with three genotypes (median [range]: 110,000 [6,000–273,000], 103,000 [6,000–658,000], and 113,000 [10,000, 299,000]; P = 0.6), but severe thrombocytopenia was least common in wild-type patients (9.6%, 17/177) and more common in heterozygous (19.1%, 36/188) and homozygous patients (21.2%, 11/52, P = 0.01). In multivariate analysis, severe thrombocytopenia was increased in heterozygous (aOR [95% CI], 2.3 [1.2–4.3]; P = 0.01) and homozygous patients (aOR [95% CI], 2.4 [1.0–5.5]; P = 0.04). After stratification by parasite species, this risk increase became nonsignificant. Severe malaria was present in 3.4% (15/443) of all patients. It tended to occur more frequently in heterozygous (3.5%, 7/202) and homozygous patients (7.3%, 4/55) than those with wild-type alleles (2.2%, 4/186; P = 0.2). No further differences with respect to clinical parameters and self-reported symptoms were observed (data not shown).

We have previously shown that carriage of the miRNA-146a rs2910164 SNP is present in 67% of a study population of pregnant women in Ghana. It is associated with 1.6-fold increased odds of malaria and with almost 6-fold increased odds in homozygous primigravidae and primiparae.⁸ In the present Indian study population, the prevalence was 59%, but the SNP was not associated with malaria. Several reasons may explain these discrepant findings. First, all pregnant African women in the previous study had P. falciparum infection, whereas P. vivax predominated in the Indian patients who largely were adult males. The observed absence of an effect of the SNP may consequently stem from the scarcity of P. falciparum among the Indian patients. Considering the sample size, the present study was powered to detect increased odds of falciparum malaria, at a level of almost double the value previously observed among pregnant African women.⁸ Moreover, acquired immune mechanisms in the Indian adult patients may have blurred an effect of an innate immune mechanism provided by miRNA-146a variation. Malaria in pregnancy is caused by specific placenta-adhering strains of P. falciparum, which induce strain-specific immune mechanisms. These are absent or low in first pregnancies.¹³ Potential concealment of a SNP effect is consequently less likely, particularly in first-time pregnant women who can be considered relatively immune-naive. In addition, the malaria-related effect of miRNA-146a might be pronounced in malaria in pregnancy: placental tissue exhibits a specific miRNA expression pattern, and placental miRNAs take part in gene regulation and inflammation control at the fetal–maternal interface.^14,15 Micro-RNA-146a has been hypothesized to be involved in the reaction on cell stress in the placenta.¹⁹ In pregnancy, P. falciparum adheres to the placental syncytiotrophoblast, thereby evading splenic elimination.¹³ This epithelium lining the intravillous space is known to produce extracellular vesicles containing miRNAs.¹⁵ This peculiarity may partially explain the discrepant finding of the miRNA-146a SNP affecting malaria risk in pregnant women but not in predominantly adult Indian males. Furthermore, considering both, genetic diversity between Africans and Indians and the complex interaction of miRNA-146a in immune regulation,^1,3,4 it is possible that the SNP yields different effects per se in these two populations. Last, human miRNA transcripts have been shown to impair Plasmodium function via gene regulation²⁰ even though the role of the miRNA-146a SNP in this context is unknown. The scarcity of data on the actual function of this polymorphism also impedes the interpretation of the weak associations of the polymorphism with thrombocytopenia (nonsignificant following Bonferroni correction for 16 comparisons) and even more so with respect to severe malaria. Conflicting findings on thrombocyte counts in miRNA-146a knockdown and knockout mice have been reported.⁵ Assuming that miRNA-146a negatively regulates inflammatory processes, a SNP-dependent alteration might result in more intense pro-inflammatory response to infection. This in turn may increase the probability of thrombocytopenia and severe malaria. However, this interpretation should be considered with care because we did not assess inflammatory mediators in the present study. Definitely, this as well as the absence of an effect on malaria risk as seen here needs to be confirmed in larger study samples and different populations, and functional studies are needed to decipher the underlying mechanisms. Still, the present work emphasizes the role of various factors, which might influence the results of genetic association studies with respect to malaria.