Association of Ficolin-2 Serum Levels and FCN2 Genetic Variants with Indian Visceral Leishmaniasis

Background Visceral leishmaniasis (VL), one of the neglected tropical diseases, is endemic in the Indian subcontinent. Ficolins are circulating serum proteins of the lectin complement system and involved in innate immunity. Methods We have estimated ficolin-2 serum levels and analyzed the functional variants of the encoding gene FCN2 in 218 cases of VL and in 225 controls from an endemic region of India. Results Elevated levels of serum ficolin-2 were observed in VL cases compared to the controls (adjusted P<0.0001). The genetic analysis revealed that the FCN2 structural variant +6359 C>T (p.T236M) was associated with VL (OR=2.2, 95% CI=1.23-7.25, P=0.008) and with high ficolin-2 serum levels. We also found that the FCN2*AAAC haplotype occurred more frequently among healthy controls when compared to cases (OR=0.59, 95%CI=0.37-0.94, P=0.023). Conclusions Our findings indicate that the FCN2 variant +6359C>T is associated with the occurrence of VL and that ficolin-2 serum levels are elevated in Leishmania infections.


Introduction
Visceral leishmaniasis (VL; Kala-Azar), a neglected tropical disease strongly associated with poverty, claims 400.000 new cases and 40.000 deaths annually [1]. VL leads to a loss of about 2 million disability adjusted life years (DALYs) every year [2]. The vector-borne infection occurs in the four distinct clinical manifestations as cutaneous leishmaniasis, muco-cutaneous leishmaniasis, VL and post-kala-azar dermal leishmaniasis [3]. VL is the severest form and severely affects visceral organs including the spleen, liver and lymph nodes [4]. Although transmission of VL has been reported in 66 countries, more than 90% of the disease burden are observed in six countries only, viz. Bangladesh, India, Nepal, Sudan, Ethiopia and Brazil [3]. Among these countries, the Indian sub-continent (India, Nepal and Bangladesh) harbours 67% of the global VL disease burden [5]. In particular, the Bihar state of India shares 50% of VL and is considered a "hot spot" of VL [6]. Inadequate vector control practice and disease management have been claimed to be responsible for the increased incidence of VL and associated mortality in India [7].
Leishmania donovani is the causative agent of VL in India. The organism is transmitted to mammalian hosts by infective bites of the sandfly Phlebotomus argentipes. L. donovani is a unicellular trypanosomatid protozoan parasite with a dimorphic life cycle between the sandfly vector (extracellular promastigotes) and the human host (intracellular amastigotes) [8]. Both developmental stages of L. donovani are coated with various secreted and membrane bound phosphoglycans. During the promastigote stage, abundant lipohosphoglycan (LPG) and gp63 are expressed, which aid immune evasion of the parasite by inhibiting the phagolysosome biogenesis in phagocytes [9]. Further, these glycoconjugates facilitate the parasite´s survival in the hostile macrophage environment [10]. However, LPG and gp36 may also serve as pathogen-associated molecular patterns (PAMPs) which are recognized by pattern recognition molecules (PRMs) of the innate system such as complement serum proteins, mannose-binding lectin (MBL), ficolins (FCN), other soluble C-type lectins and toll-like receptors [11]. Serum complement activating pattern recognition molecules act in a first-line innate defense against promastigotes inoculated by the sandfly bite. Leishmania parasites have developed various evasion strategies to avoid the lytic action of the complement system. The parasites use host complement proteins to escape the immune attack by entering into macrophages [12]. Mannosebinding lectin (MBL), a circulating serum protein, recognizes the carbohydrate domain of L. major, L. mexicana, and L. braziliensis. MBL binds to the surface of Leishmania promastigotes to opsonize the parasites. Upon binding to parasites, MBL initiates the complement cascade and provides an additional uptake mechanism of parasites by enhancing opsonophagocytosis and protects them from the immune attack [13,14] and, thus, modulates the clinical outcome of VL [15].
Ficolins are serum complement lectins that are structurally and functionally analogous to MBL [16] and, hence, expected to modify the clinical outcome of VL due to their involvement in innate immunity. Interestingly, a significant association of a distinct FCN2 haplotype with cutaneous leishmaniasis has been reported from a Syrian population [17]. Ficolins are a group of complement activating pattern recognition molecules consisting of a collagen-like tail region and a fibrinogen-like domain (FBG) [18]. Three types of ficolins (Ficolin-1, -2, -3) of similar structure exist in humans. These types have differential tissue expression patterns and functions [19]. The role of ficolin-2, as an innate immunity component, has been studied in several infectious diseases including Hepatitis B, schistosomiasis, Chagas disease and others [16,[20][21][22]. Ficolin-2 recognizes superficial acetylated compounds of invading pathogens by their FBG domain and initiates the lectin complement cascade [23]. The FCN2 gene localizes to chromosome 9q34.3 (OMIM 601624) and hepatic cells predominantly express the corresponding protein. The variants in the promoter region of FCN2 gene at positions -986A>G, -602G>A and -4A>G have been observed to modulate the circulating ficolin-2 concentration in a dosedependent manner. The non-synonymous exon-8 variant alleles at positions +6359C>T and +6424G>T were shown to exhibit differential binding affinities to acetylated compounds when compared to the wildtype reference alleles [24]. Studies have shown that inter-individual variation of circulating ficolin-2 concentration are correlated with polymorphisms in the promoter and exon-8 regions [25].
Although it has been showed that FCN2 gene polymorphisms and haplotypes are associated with cutaneous leishmaniasis [17], no investigations of ficolins have so far focused on VL. Moreover, we recently observed that functional MBL2 polymorphisms and lower MBL levels confer relative protection against VL (unpublished). As ficolin-2 shares similarities both in structure and function with MBL [26], we aimed to explore the role of potentially important FCN2 gene variants and circulating ficolin-2 levels in VL in our Indian study group. Three promoter SNPs (-986A>G, -602G>A and -4A>G) and two structural SNPs in exon 8 (+-6359C>T and +6424G>T) were genotyped and studied.

Ethics statement
Informed written consent was obtained either from the participating individual or from the parents/guardians if an individual was less than 18 years old. The study was approved by the Institutional Ethical Committee (IEC) of the CSIR-Centre for Cellular and Molecular Biology (CCMB), Hyderabad, India. Permission was also sought for and obtained from district government hospitals.

Study design and sample collection
This is a case-control study matched for ethnicity, sex and geographical location. All cases and controls were recruited through multiple field visits from villages located within a radius of 120 km from the city of Muzaffarpur in the Bihar state of India. Previous epidemiological studies of VL have indicated that the Bihar state is a hot spot for VL with an average annual incidence of 2.49/1000 individuals [6]. The sample size was calculated prior to recruitment using the Open Epi platform (http://www.openepi.com/) based on the incidence rate and the risk of VL in the study area. A total of 443 unrelated subjects (218 cases and 225 healthy controls) were recruited. The mean age of VL cases was 28.7±16.7 and 35.3±16.2 in healthy controls (P = 0.001). No significant difference in the male/female ratio was observed in cases (125:95) and controls (122:93). The cases were determined based on the clinical features of VL in medical records issued by government hospitals in the study region. Typical clinical features of the cases included fever with rigors and chills and significant splenomegaly. Cases were tested with the rk39 leishmanin antigen by nitrocellulose dipstick tests (InBios International, Seattle, USA). The control subjects were free of any relevant infectious disease. Pregnant women, cases with other infections, healthy controls with a family history of VL and relatives of cases were excluded from the study. About 5.0 mls of full venous blood were collected from study subjects for serological and genetic studies. The samples were immediately transported to the lab and the serum samples were separated from whole blood and stored in the same type of tubes at -20°C until further use.

DNA isolation and FCN2 genotyping
Genomic DNA was extracted from peripheral blood leukocytes using the protocol described previously [27]. The reference genomic sequence was retrieved from the Ensembl database (www.ensembl.org). The five FCN2 variants studied were PCR amplified from two genomic regions. The three promoter variants -986A>G, -602G>A and -4A>G were amplified by the primer pairs PromF-5'-ATTGAAGGAAAATCCGATGGG-3' and PromR-5'-GAAGCCACC AATCACGAAG-3', and the two exon-8 variants +6359 C>T and +6424 G>T were amplified using the primer pairs Exon8F-5'-CCAGCTCCCATGTCTAAAGG-3' and Exon8R-5'-TTACAAACCGTAGGGCCAAG-3'. Primers were designed by Primer-BLAST (http://www. ncbi.nlm.nih.gov/tools/primer-blast) and synthesized commercially (Eurofins, Bangalore, India). The target regions were amplified using an Emerald PCR master mix (TaKaRa, Shiga, Japan) and reactions were carried out in the ABI GeneAmp PCR system 9700. The thermal cycling parameters for both amplicons were: initial denaturation at 95°C for 5 minutes, followed by 35 cycles of denaturation at 94°C for 1 minute, annealing at 60°C for 30 seconds and elongation at 72°C for 1 minute. PCR products were purified using Exo-SAP-IT (USB-Affymetrix, Santa Clara, USA) and 1.0 μl of the products were directly used as templates for sequencing using the BigDye terminator (v.3.1) cycle sequencing kit (Applied Biosystems, Texas, USA) on an ABI 3730XL DNA Analyzer. Variations were identified by assembling DNA sequences with the reference sequence using AutoAssembler software (Applied Biosystems, Texas, USA) and were reconfirmed visually from their electropherograms.

Ficolin-2 serological assay
Ficolin-2 levels were measured in the sera of VL cases (n = 166) and healthy controls (n = 85) using the human Ficolin-2 ELISA kit following manufacturer's instructions (Hycult Biotech, Uden, The Netherlands). The detection limit of the assay was 16 ng/mL.

Statistical analysis
Data were analyzed using the STATA software (Intercooled STATA, STATA Corp., College Station, TX, USA) and the level of significance was set to a P value of <0.05. Genotype or haplotype frequencies were calculated by simple gene counting and by expectation-maximum (EM) algorithm and the deviations from Hardy-Weinberg equilibrium were tested using the random-permutation procedure as implemented in the Arlequin v.3.5.1.2 software (http://lgb. unige.ch/arlequin). The linkage disequilibrium (LD) analysis was performed using Haploview v.3.2 (http://broadinstitute.org/haploview). Multivariate analysis was performed after adjustment with the confounding factors such as age, ethnicity and gender using the STATA software. In all comparisons, P values <0.05 were considered significant. Kruskal-Wallis or Wilcoxon-Mann-Whitney rank sum tests were applied wherever appropriate to analyze the correlation of serum ficolin-2 levels with FCN2 variants and haplotypes by using the Kaleidagraph software (www.synergy.com).

Association of FCN2 variants with the risk of VL
The genotype and allele frequencies for the variants -986G>A, -602A>G, -4A>G and +-6359C>T) in VL cases and controls were in Hardy-Weinberg equilibrium (P>0.05). This did not apply to the variant +6424G>T in VL cases. This variant was excluded from further analyses. The LD patterns of the FCN2 variants are given in Fig 1. The LD plot indicates that the promoter variants -986G>A, -4A>G and the exon 8 variant +6359C>T were in strong LD with each other both in cases and controls. The variant -602A>G was in LD with +6359C>T only in controls. Significant differences were observed both in genotype and allele distributions between cases and controls for the non-synonymous variant +6359C>T (p.Thr236Met). The homozygous genotype +6359TT occurred more frequently among VL cases compared to controls after adjusting for age, sex and ethnicity (OR = 2.2, 95%CI = 1.23-7.25, P = 0.008), indicating that this variant was associated with an increased risk for L. donovani infection (Table 1) Table 1). The different genetic models indicate that the +6359T minor allele increases the susceptibility of L. donovani infection. The other investigated FCN2 variants were not significantly associated with VL.

Discussion
Visceral leishmaniasis develops when L. donovani parasites are successfully inoculated and survive the first-line attack of innate immune components such as phagocytes and the complement system. Indeed, these innate immune components play a major role both in the control and establishment of L. donovani infections [28]. Complement components including lectins are the primary molecules of the innate immune system to encounter inoculated metacyclic promastigotes. The early activation of the complement system during pathogen invasion occurs predominantly by the lectin pathway, as it is independent of a specific antibody response. Moreover, it prompts the activation of the alternative pathway [29]. The lectin pathway protein MBL induces opsonophagocytosis by depositing C3b on the surface of Leishmania which is crucial for parasite survival and multiplication [13,14]. We assume that, as ficolins are functionally similar to MBL, they equally influence the outcome of VL. No study so far, however, has focused on the role of ficolin-2 in VL. We studied the contribution of ficolin-2 serum levels and of FCN2 functional variants in VL. The structural variant +6359C>T (p.T236M) in the fibrinogen-like domain of the FCN2 gene confers relative susceptibility to VL. The finding remains consistent in recessive and allelic genetic models. The computational prediction revealed that the T236M substitution has a major impact on the physiochemical property of ficolin-2 [30]. In addition, the +6359T allele was found associated with higher ficolin-2 serum levels [31] and the observation was reconfirmed in a cohort of neonates [32]. We also observed a similar effect of the +6359T allele in controls, but not in cases. Our results inferred that ficolin-2 serum levels were modulated significantly by the infection in VL cases rather than by FCN2 variants. Moreover, the ficolin-2 protein with T236M substitution had a markedly decreased binding capacity to acetylated agarose beads. Therefore, this structural variant is believed to alter the binding properties of the protein to recognize invading pathogens [24,33,34]. These reports indicate that individuals with higher ficolin-2 serum levels and altered binding capacities might favor L. donovani invasion into macrophages and the development of VL. Previous studies have also reported that +6359C>T in the FCN2 gene is a risk factor for staphylococcal peritonitis in continuous ambulatory peritoneal dialysis cases [35] and for bloodstream infections in kidney transplant recipients [36].
In FCN2 gene-association studies, haplotype analyses should be taken into account as they may influence disease susceptibility [37]. Our FCN2 haplotypes revealed that the FCN2 Ã AAAC haplotype frequency was higher among controls than in VL cases, indicating that individuals with this haplotype had a diminished probability to develop VL. The FCN2 Ã AAAC haplotype harbors the +6359C major allele, which accounts for reduced ficolin-2 levels [31,32]. In light of Ficolin-2 and FCN2 Variants and Visceral Leishmaniasis these observations, it is evident that the FCN2 genetic factors that contribute to low ficolin-2 level decrease the risk of VL. FCN2 promoter haplotypes did not show any differences among groups, suggesting the relative contribution of the +6359C>T genotype in Indian VL.
Ficolin-2 serum levels were elevated in VL cases compared to controls, indicating that ficolin-2 is a susceptibility factor. The result is in accordance with a study published previously [15]. Corresponding results were also observed in infections with Mycobacterium spp., where higher MBL serum levels increased the risk of infection [38][39][40]. The proposed mechanism may be that intracellular parasites abuse C3 opsonization and enhance opsonaophagocytosis by monocytes/macrophages to avoid complement attacks. Any increase in the MBL and ficolin levels in turn may enhance complement activation and, thus, the probability of parasitization by depositing C3b on parasite surfaces [41]. Our observation supports this notion as cases with VL had higher ficolin-2 levels than uninfected controls. Nevertheless, discordant results were reported for ficolin-2 in tuberculosis and Chagas disease, where cases presented lower ficolin-2 plasma levels than did controls [22,42]. No clear mechanism is proposed to address the conflictive observations of functionally similar proteins in intracellular habitant infections. In addition, the recognition and interaction of mannose binding lectin (MBL) with Leishmania parasites are well established [13,14] and ficolins were shown to be functional analogous to MBL [16]. However, a limitation of our study is that there is a lack of data showing the interaction of ficolin-2 with L. donovani. Nevertheless, our earlier study demonstrated the genetic association of FCN2 polymorphism with cutaneous leishmaniasis in Syrian population [17].
In conclusion, our results show that the FCN2 +6359C>T variant is associated with increased susceptibility to VL and that the FCN2 Ã AAAC haplotype is associated with relative protection. Higher serum ficolin-2 levels were observed in cases with VL than among controls.