Variants of MIRNA146A rs2910164 and MIRNA499 rs3746444 are associated with the development of cutaneous leishmaniasis caused by Leishmania guyanensis and with plasma chemokine IL-8

Leishmania are intracellular protozoan parasites that cause a wide spectrum of clinical manifestations in genetically susceptible individuals with an insufficient or balanced Th1 immune response to eliminate the parasite. MiRNAs play important regulatory role in numerous biological processes including essential cellular functions. miR146-a acts as an inhibitor of interleukin 1 receptor associated kinase 1 (IRAK1) and tumour necrosis factor (TNF) receptor associated factor 6 (TRAF6) present in the toll-like receptors pathway while miR499a modulates TGF-β and TNF signalling pathways. Here, we investigated whether MIRNA146A rs2910164 and MIRNA499 rs3746444 variants are associated with the development of L. guyanensis (Lg)-cutaneous leishmaniasis (CL). The variants MIR146A rs2910164 and MIR499A rs3746444 were assessed in 850 patients with Lg-CL and 891 healthy controls by polymerase chain reaction and restriction fragment length polymorphism (PCR-RFLP). Plasma cytokines were measured using the BioPlex assay. Carriers of rs2910164 CC genotype have 30% higher odds of developing CL (ORadjage/sex = 1.3 [95%CI 0.9–1.8]; Padjage/sex 0.14) compared to individuals with the genotype GG (ORadjage/sex = 0.77 [95%CI 0.56–1.0]; Padjage/sex 0.14) if exposed to Lg-infection. Heterozygous GC individuals also showed lower odds of developing CL (ORadjage/sex = 0.77 [95%CI 0.5–1.1]; Padjage/sex 0.09). Homozygosity for the allele C is suggestive of an association with the development of Lg-CL among exposed individuals to Lg-infection. However, the odds of developing CL associated with the CC genotype was evident only in male individuals (ORadjage = 1.3 [95% CI = 0.9–2.0]; Padjage = 0.06). Individuals homozygous for the G allele tend to have higher plasma IL-8 and CCL5. Similarly, for the MIR499A rs3746444, an association with the G allele was only observed among male individuals (OR = 1.4 [1.0–1.9]; P = 0.009). In a dominant model, individuals with the G allele (GG-GA) when compared to the AA genotype reveals that carriers of the G allele have 40% elevated odds of developing Lg-CL (ORadjage = 1.4 [1.1–1.9]). Individuals with the GG genotype have higher odds of developing Lg-CL (ORadjage/sex = 2.0 [95%CI 0.83–5.0]; Padjage = 0.01. Individuals homozygous for the G allele have higher plasma IL-8. Genetic combinations of both variants revealed that male individuals exposed to Lg bearing three or four susceptible alleles have higher odds of developing Lg-CL (OR = 2.3 [95% CI 1.0–4.7]; p = 0.017). Both MIR146A rs2910164 and MIR499A rs3746444 are associated with the development of Lg-CL and this association is prevalent in male individuals.

Introduction Leishmaniasis, a vector-borne disease caused by protozoan parasites from the genus Leishmania, is endemic in the tropical and subtropical regions, including more than 98 countries. Nearly, one billion of people are at risk of infection [1]. Leishmania species cause a spectrum of clinical forms of the disease: visceral (VL), cutaneous (CL), diffuse cutaneous, disseminated cutaneous and mucocutaneous leishmaniasis (ML) [2].
CL is considered the most common form of Leishmania-infection. Approximately 0.7 million to 1.2 million human beings are affected by this disease [1,3]. In Brazil, the main species involved in the etiology of CL are L. braziliensis (Lb), L. guyanensis (Lg), L. lainsoni, L. amazonensis, L. shawi, L. naiffi and L. lindbergi [4]. In the Amazonas state of Brazil, Lg is responsible for 95% of CL cases [5].
MicroRNAs (miRNAs) are small, single-stranded, untranslated endogenous RNAs, composed of 18-26 nucleotides that regulate gene expression. miRNAs bind to their target messenger RNA (mRNA) through base complementarity mechanism resulting in the regulation or degradation of protein translation [8,9]. More than 30% of human protein-coding genes are under post-transcriptional control of miRNAs [10]. miRNAs can regulate many physiological processes such as cell cycle, metabolism, and apoptosis. miRNAs play a crucial role in haematopoiesis, immune cells development and differentiation, inflammation, and immunomodulation [11,12].
Single nucleotide polymorphisms (SNPs) in miRNA precursors may affect the miRNA biogenesis, causing a reduction of mature miRNA expression levels [22]. SNPs in mature miRNA may affect miRNA target specificity and leads to altered cellular protein levels [23,24]. These SNPs or variants in miRNA may by their actions alter the course of various diseases.
MIR499A and MIR146A genes are located on 20q11.22 and 5q33.3 chromosomes, respectively. The variant rs3746444 located in the seed region of mature miR499-a leads to the disruption of miRNA-mRNA interactions and creation of new gene targets [20]. The variant rs2910164 present in the seed sequence of miR146-a precursor results in low production of mature miR146-a and consequently to a decrease inhibition of TRAF6 and IRAK1, leading to a higher production of pro-inflammatory cytokines upon TLRs activation [25]. miRNA146-a and miR499-a have been associated with susceptibility to multiple types of cancer, psoriasis and rheumatoid arthritis [26][27][28][29].
Taken into account the potential role of miR146-a and miR499-a in modulating the T helper cell immune response, we investigated whether the variants MIR146A rs2910164 and MIR499A rs3746444 are associated with susceptibility or protection to the development of Lg-CL in the Amazonas. The influence of the MIR146A and MIR499A genotypes on plasma cytokine levels were also investigated.

Area of study and population
The study was conducted at the Fundação de Medicina Tropical Doutor Heitor Vieira Dourado (FMT-HVD), the referral regional center for treatment of leishmaniasis. The study population and the endemic area of recruitment of the participant of the study are described elsewhere [30]. Briefly, all the participants are from the perirural areas of Manaus, the capital city of the Amazonas State where L. guyanensis is endemic. Patients with active CL were followed at the FMT-HVD. The healthy controls with no history of CL and devoid of any scar suggestive of CL are from the same endemic area of the patients with CL.

Ethical statement
This study was conducted according to the Declaration of Helsinki and approved by the Research Ethics Committee of the FMT-HVD granted under the file number CAAE:09995212.0.0000.0005 on 31 May 2013. All the participants or their responsible party for individuals less than 18 years of age provided written informed consent for the collection of samples and subsequent analysis.

Sample size calculation
Sample size calculation is described elsewhere [31]. Briefly, using the Genetic Power calculator developed at Harvard University (http://pngu.mgh.harvard.edu/~purcell/gpe), we assumed a minor allele frequency of 5%, disease prevalence of 5%, complete linkage disequilibrium 1 between marker and case-control discrete trait, case-control ratio 1, and 5% type 1 error rates with an odds ratio of 1.5 and 2.0 for heterozygosity and homozygosity, respectively. For 80% power, the genetic allelic model provided a sample size of 789 individuals for cases and 789 for controls.

Leishmania spp identification from biopsy specimens and DNA extraction from whole blood for SNP typing
All the patients with CL provided a biopsy specimen from the cutaneous lesion for the isolation of parasite DNA. For the identification of the Leishmania spp., the discrimination of the Leishmania Viannia subgenus specific PCR was in accordance with established protocols [32,33]. Identification of Leishmania spp. was performed by nucleotide sequencing of a fragment of HSP 70 and miniexon genes [34,35]. Venous blood was drawn from all participants and collected into EDTA-containing Vacutainers (Becton Dickinson, Brazil) to DNA extraction and cytokine assay. Genomic DNA was extracted by the salting out method [36].

SNP genotyping
The SNPs MIR146A rs2910164 and MIR499A rs3746444 were performed by PCR-RFLP with the restriction enzymes HpyCH4 III and HpyCH4 IV (New England Biolabs, Ipswich, MA, United States), respectively. The respective pairs of primers for the amplification the region flanking the SNPs and the fragments generated for alleles discrimination by the restriction enzymes as well as the cycling protocols for PCR are shown in S1 Table. The pair of primers for MIR499A was designed from the reference sequence NC 000020.11 from NCBI. The underline G nucleotide from the forward primer substitutes A from the reference sequence to eliminate a site of restriction for the HpyCH4IV. Primers for the MIR146A were designed from the reference sequence NC 000005.10 and the underline A nucleotide substitutes C from the reference sequence to create restriction site when the G allele is present.
The PCR mix contains: 0.2 μM of each primer (Thermofisher, MA USA), 40 nM of dNTP (Thermofisher, MA USA), 1.0 mM of MgCl2 (Thermofisher, MA USA), 1 U of Taq DNA polymerase (Thermofisher, MA USA), 1X of 10X Taq polymerase buffer containing 500 mmol/L KCl and 100 mol/L of Tris-HCL (pH 8.3) and 50 ng of DNA in a final volume of 25 μL. A volume of 10 μL of the PCR products was digested with 1 unit of the respective restriction enzyme and buffer in a final volume of 20 uL and size separated in a 3% agarose (Ultrapure Agarose, Thermofisher, MA USA) gel electrophoresis.

Cytokine assay
Cytokine assay of IL-1β, TNF-α, IL-2, IL-6, IL-8, IL-17, IFN-γ, CCL2 and CCL5 in the plasma were measured using the multiplex cytokine commercial kit Human Cytokine Grp I Panel 27-Plex (Bio-Rad, USA) according to the instructions of the manufacturer in the Bio-plex 200 Protein Array System (Luminex Corporation, USA).

Statistical analysis
The genotype and allele frequencies were calculated by direct gene counting. For calculation of Hardy-Weinberg equilibrium (HWE), the website http://ihg.gsf.de/cgibin/hw/hwa1 was used, that also compared cases with the control groups by logistic regressions analysis to determine associations to susceptibility or resistance for the different genotypes and alleles by χ2 test along with OR and 95% confidence interval. The correlation of the different genotypes of MIR146A rs2910164 and MIR499A rs3746444 to the concentration of circulating plasma cytokines was performed by the R software version 4.0.0 of SNPassoc package for quantitative traits analysis and ggplot2 package for visualizing. P values for the correlations of cytokines by genotypes were corrected for false discovery rate (FDR) of Benjamini-Hochberg.

Study population
The study population is the same as described previously [30]. A total of 850 patients with Lg-CL and 891 healthy controls (HC) were included in the study. The HC are from the same endemic area of the patients with Lg-CL. Among the patients with Lg-CL, 639 (75%) patients were male (mean age 34.4 ± SD 13.7 years) and 211 (25%) were females (37.5± SD 15.7 years). In the controls group, 608 (68%) were male (42± SD 17.5 years) and 283 (32%) were female (40± SD 17.4 years). Overall, the mean age among the patients with Lg-CL and controls is 35.17± SD 14.25 and 41.4± SD 17.5 years, respectively. The HC is older than the cases (P<0.0001). Men were older among the HC group compared to group of male patients with Lg-CL (P<0.0001) while there was no age difference among females (P<0.077). All the participants of the study were devoid of HIV and the patients had fewer or equal to six lesions and treatment naïve at the time of enrolment. Pregnant women were excluded from the study. Of note, there are more females in the HC compared to the group of patients with CL (P<0.0013).
Association MIR146A rs2910164 and MIR499A rs3746444 with susceptibility to Cutaneous Leishmaniasis MIR146A rs2910164 was assessed in 826 patients with Lg-CL and 886 controls. Genotype and allele frequencies for the two variants are demonstrated in Table 1. rs2910164 was in Hardy-Weinberg equilibrium (HWE) among the patients with Lg-CL and HC. The MIR146A rs2910164 CC genotype was prevalent in patients with Lg-CL (12%) compared with HC (9%). Carriers of rs2910164 CC genotype have 30% higher odds of developing CL (ORadj age/sex = 1.3 [95%CI 0.9-1.8]; Padj age/sex = 0.14) compared to individuals with the genotype GG (ORadj age/ sex = 0.77 [95%CI 0.56-1.0]; Padj age/sex = 0.14) if exposed to Lg-infection. Heterozygous GC individuals also have lower odds of developing CL compared with homozygous carriers of the C allele (ORadj age/sex = 0.77 [95%CI 0.55-1.1]; Padj age/sex = 0.09). In a recessive model, when homozygous individuals for the C allele are compared with individuals carrying a G allele (CC versus GC-GG), carrier of G allele have 23% lower odds of developing Lg-CL (ORadj age/sex = 0.77 [95%CI 0.56-1.0]; Padj age/sex = 0.098). Homozygous for the C allele have 30% higher odds of developing CL (ORadj age/sex = 1.3 [95%CI 0.9-1.8]; Padj age/sex = 0.098). The C allele is suggestive of an association with the development of Lg-CL. As male individuals were prevalent in both patients with Lg-CL and HC, we stratified according to sex as shown in Table 2. The odds of developing CL associated with the CC genotype was evident only in male individuals (OR adjage = 1.3 [95% CI = 0.9-2.0 P adjage = 0.06]) compared with GG genotype. Individuals with GG genotype had lower odds of developing Lg-CL (R adjage = 0.77 [95%CI 0.5-1.0]; p = 0.06). The genotype CC is suggestive of an association with the development of CL compared to GG genotype among Lg-infected individuals.
MIR499A rs3746444 was genotyped in 818 patients with Lg-CL and 851 HC. The variant was in HWE in both groups of patients with Lg-CL and HC. The distribution of genotypes among patients with Lg-CL and HC was different as shown in Table 1, revealing a common odds ratio of 1.3 for the G allele compared with the A allele (p = 0.04). The rs3746444 AA genotype was more prevalent among the HC (83%) compared to 79% in the patients with Lg-CL group. Comparison of the genotype AA to GG revealed that individuals with the GG genotype have 100% higher odds of developing Lg-CL, with a 95% CI ranging from a decreased odds of 17% to an elevated odd of 400% (ORadj age/sex = 2.0 [95%CI 0.83-5.0]; Padj age/sex = 0.15). Similarly, the G allele confers 27% elevated odds of developing Lg-CL suggesting that the G allele contribute to susceptibility to the development of Lg-CL (OR = 1.27 (95% CI 1.0-1.6]; p = 0.035). In a dominant model, individuals with the G allele (GG-GA) when compared to the AA genotype reveals that carriers of the G allele have 30% elevated odds of developing Lg-CL (ORadj age/sex = 1.3 [95% CI 1.0-1.6]; Padj age/sex = 0.067).
We stratified into male and female individuals to look for the strength of the association. Like the MIR146A, the association was more evident among the male individuals as shown in Table 3. Male individuals with GG genotype compared with AA genotype revealed an elevated odds of 50% with a 95% confidence interval ranging from a decreased odds of 40% to an increased odds of 310% to the development of Lg-CL (OR adj/age = 1.5 [0.6-4.1]; P adj/age = 0.38).    Fig 2. In a recessive model, individuals homozygous for the G allele seem to have higher levels of IL-8 (2.14±0.64 pg/mL) compared with individuals (AA +GA 1.18±0.04 pg/mL) bearing a A allele (p = 0.003; P FDR = 0.01) (Fig 2).
Similarly, MIR499A rs3746444 have a tendency on influencing the plasma circulating levels of IL-17 among the HC. GG homozygotes (5.48±1.97 pg/mL) have higher levels of IL-17  Genetic combinations of both variants revealed that individuals bearing three or four G alleles tend to correlate with higher plasma IL-8 (2.24 ± 0.48 pg/mL) and CCL5 (63.4± 29.2 pg/ mL) as shown in S5 Fig.

Discussion
The manifestation of clinical symptoms in Leishmania-infected individuals depends on the capacity of the individual to have a fine regulation of the TH1 response to eliminate the parasite and avoid an exacerbation of expression of pro-inflammatory cytokines. Emerging evidence are showing that miRNAs regulate immune response [19,40,41].
In this study, we showed that male individuals homozygous for the C allele of MIR146A rs2910164 have 30% higher odds of developing Lg-CL compared with individuals bearing a G allele. The MIR146A rs2910164 C allele has been shown to be a risk factor for leprosy in Brazil and psoriasis in South African Indian [42,43]. Homozygous individuals for the C allele have increased risk of developing glioma and decreased survival [44]. A meta-analysis study on autoimmune diseases revealed that individuals bearing the C allele have increased risk of developing the disease [45].
Several studies have also reported that MIR146A rs2910164 is not a risk factor in several diseases. A meta-analysis on ischemic stroke showed that MIR146A rs2910164 is not associated with any risk occurrence of ischemic stroke [46]. A lack of association of MIR146A rs2910164 with the development of RA was also reported [47]. Another study cited that susceptibility to pulmonary tuberculosis is not influenced by the MIR146A rs2910164 [48]. However, a metaanalysis approach on psoriasis case-control studies with rs2910164 revealed that the CC genotype is correlated with decreased risk of psoriasis [28].
Conversely, other studies have reported that the G allele is a risk factor in several diseases. The G allele has been associated with pulmonary TB [49] and ankylosing spondylitis [50]. A meta-analysis study on cancer among Asian patients revealed that GG genotypes are associated with increased risk of cancer [51].
We also showed that in male individuals, the GG genotype of the MIR499A rs3746444 is associated with 50% increased odds of developing CL. The G allele confers 27% elevated odds of developing CL suggesting that the G allele may contribute to the development of Lg-CL.
A recent meta-analysis study showed that the MIR499A rs3746444 G allele is associated with high risk of developing breast cancer [52]. The allele G was also associated with hepatocellular carcinoma [53], and adenocarcinoma [54]. Other studies also showed that the G allele is associated with susceptibility to the development of bronchial asthma [20], RA [29,55], Behcet's disease [56], ankylosing spondylitis [50], myocardial infarction [57] and coronary artery disease [58,59]. Immune response plays critical role in all these diseases. Of note, there is no study of both variants with protozoan infectious diseases to date.
The gender difference observed for both variants in this study may be due to either sexual hormonal interaction with the variant or the small sample size in the female group. Notably, we have 75% and 68% male individuals among patients with CL and HC, respectively. Interestingly, sex-based differences have been shown in L. tropica infected patients. L. tropica-infection manifests commonly in the form of CL. However, females present predominantly CL while male individuals are more incline to develop viscerotropic leishmaniasis [60][61][62]. Furthermore, peripheral blood from patients with CL caused by L. mexicana stimulated or not with lipophosphoglycan demonstrated higher expression of IFNγ and tumour necrosis factor alpha in females than in males [63]. An increased parasite growth has been observed when L. mexicana promastigotes are treated with physiological doses of male dihydrotestosterone [64,65].
The MIR146A is in the cytokine cluster 5q31 and the MIR146A rs2910164 is situated in the stem region of the pre-miR-146a. The change of guanine:uracil pair (G:U) to cytosine:uracil (C:U) in the stem structure affects the production of mature miR146-a [25,66]. The G-allele is reported to be associated with high expression levels of mature miR146-a while the C-allele with low expression [25,67]. Other study cited that the C allele is correlated with higher miR146-a and lower TNFα expression from nerve biopsies of leprosy patients [42]. High expression of miR146-a was also observed in specimens of healthy and tumour tissue from patients with gastric cancer bearing genotype CC compared to GG [68]. In Lupus, CC genotypes was correlated with high miR146-a expression [69]. One study observed that the genotype CC is associated with increased expression of IRAK1 and TRAF6 [70]. miR146-a regulates the IL-17 pathway in human keratinocytes, the first line of defence against pathogens in the skin, to restrain IL17-induced inflammation [71]. Overexpression of miR146-a in keratinocytes inhibits the expression of IL-8 and TNFα, suppressing the chemotactic attraction of neutrophils by keratinocytes [72]. Interestingly, we observed in this study that individuals homozygous for the C allele have lower plasma IL-8. Generation of miRNA regulatory network from Cytoscape revealed that miR146-a play key role in the inflammatory response in Leishmania-infection [73].
In this study, we did not observe any correlation of plasma levels of TNFα, IL-6, IL-1β, IL-2, IL-17, IFNγ and CCL2 by MIR146A rs2910164 genotypes neither among the patients with Lg-CL nor the HC. However, individuals homozygous for the G allele seem to have higher levels of circulating IL-8 and CCL5 compared with individuals bearing a C allele.
Knockout MIR146A -/mice developed exaggerated pro-inflammatory response upon challenging with lipopolysaccharide due to chronic dysregulation of nuclear factor kappa-lightchain enhancer of activated B cells (NFkB) signalling [74,75]. miR146-a is reported to target IRAK and TRAF-6 in the TLRs downstream pathway to downregulate the upregulation of NFkB leading to a decrease in the transcription of pro-inflammatory cytokines IL-6, IL-8, IL-1β, and TNFα [74,[76][77][78][79][80]. A recent study showed that inhibition of miR146-a resulted in a reduced secretion of IL-6 and IL-8 [81], suggesting that miR146-a may downregulate an inflammatory reaction.
miRNA499-a is involved in TLR-signalling [82]. Computational tool analysis suggests that miR499-a may target IL-13 and IL-23 (microRNA.org). MIR499A rs3746444 genotype GG has been suggested to correlate with high expression of miR499-a [83]. Our data showed that individuals homozygous for the G allele have higher plasma IL-8 and a tendency to also have high IL-6, IL-17, and increased risk of developing Lg-CL.
Indeed, in Lb-infected patients with CL, high levels of IL-17 were observed [84] and peripheral blood from these patients released high IL-17 upon stimulation with soluble Leishmania antigen [85]. L. major-infected C57BL/6 mice developed larger lesion size with increased production of IL-17 and neutrophil infiltration at the site of lesion compared with mice treated with anti-IL-17 [85] while L. major-infected BALB/c mice deficient for IL-17 develop smaller lesions [86], suggesting that IL-17 promotes lesions. High levels of IL-6 impaired the cytokineenhanced antileishmanial activity of human macrophages by inhibiting IFNγ and TNFα [87].
Our study has several limitations. Firstly, we considered that our healthy controls from the same endemic areas as the patients are exposed to the Lg-vector and are infected but did not develop the disease. We did not perform DTH to ensure this despite most of the individuals included in the study are farming or agricultural workers. Secondly, the problem of multiple testing of many associations may result in spurious associations but may also discard a true association after correcting for multiple comparisons. Thirdly, our controls are slightly older than our patients with CL and contain slightly more females. Stratification reduces the sample size.
Considering our data, we may hypothesize that MIR146A rs2910164 CC individuals are prone to develop Lg-CL due to an impairment neutrophil migration at the inoculation of the parasite by the sand fly bite. Additionally, these individuals have lower levels of CCL5. CCL5 has been shown to correlate with resistance to L. major infection in animal model [88]. Furthermore, our MIR499A rs3746444 data suggest that individuals with genotype GG are susceptible to develop Lg-CL by attracting too many neutrophils that may pass the parasite to macrophages leading to parasite persistence as they might be high producers of IL-8, IL-6 and IL-17.
Altogether, this is the first study to date to investigate MIR146A rs2910164 and MIR499A rs3746444 in protozoan infectious diseases. Both variants are associated with the development of Lg-CL male individuals exposed to Lg-infection and correlate with plasma IL-8.