Leprosy Association with Low MASP-2 Levels Generated by MASP2 Haplotypes and Polymorphisms Flanking MAp19 Exon 5

Background The gene MASP2 (mannan-binding lectin (MBL)-associated serine protease 2) encodes two proteins, MASP-2 and MAp19 (MBL-associated protein of 19 kDa), bound in plasma to MBL and ficolins. The binding of MBL/MASP-2 and ficolin/MASP-2 complexes to microorganisms activates the lectin pathway of complement and may increase the ingestion of intracellular pathogens such as Mycobacterium leprae. Methods We haplotyped 11 MASP2 polymorphisms with multiplex sequence-specific PCR in 219 Brazilian leprosy patients (131 lepromatous, 29 borderline, 21 tuberculoid, 14 undetermined, 24 unspecified), 405 healthy Brazilians and 291 Danish blood donors with previously determined MASP-2 and MAp19 levels. We also evaluated MASP-2 levels in further 46 leprosy patients and 69 Brazilian controls. Results Two polymorphisms flanking exon 5 of MASP2 were associated with a dominant effect on high MASP-2 levels and an additive effect on low MAp19 levels. Patients presented lower MASP-2 levels (P = 0.0012) than controls. The frequency of the p.126L variant, associated with low MASP-2 levels (below 200 ng/mL), was higher in the patients (P = 0.0002, OR = 4.92), as was the frequency of genotypes with p.126L (P = 0.00006, OR = 5.96). The *1C2-l [AG] haplotype, which harbors p.126L and the deficiency-causing p.439H variant, has a dominant effect on the susceptibility to the disease (P = 0.007, OR = 4.15). Genotypes composed of the *2B1-i and/or *2B2A-i haplotypes, both associated with intermediate MASP-2 levels (200–600 ng/mL), were found to be protective against the disease (P = 0.0014, OR = 0.6). Low MASP-2 levels (P = 0.022), as well as corresponding genotypes with *1C2-l and/or *2A2-l but without *1B1-h or *1B2-h, were more frequent in the lepromatous than in other patients (P = 0.008, OR = 8.8). Conclusions In contrast with MBL, low MASP-2 levels increase the susceptibility to leprosy in general and to lepromatous leprosy in particular. MASP2 genotypes and MASP-2 levels might thus be of prognostic value for leprosy progression.


Introduction
Mycobacterium leprae causes leprosy (Hansen's disease) by invading macrophages and Schwann cells. This leads to progressive and irreversible disabilities associated with heavy social stigma. Leprosy has been eliminated from 119 countries out of 122 since 1985, but Brazil still presents the second highest prevalence rate in the world with 1.52/10,000 inhabitants, and the highest detection rate of new cases with 17.86/100,000 inhabitants [1].
On exposure to M. leprae, only a small group of infected individuals develop the disease, which is characterized by a wide spectrum of immunologically-defined clinical manifestations. The lepromatous form is the result of a Th2 response and characterizes the low resistance pole of Hansen disease. The other pole is represented by the tuberculoid form and is defined by a strong Th1 response controlling bacillus dissemination [2]. The innate immune response interacts with the adaptive Th1 and Th2 responses, e.g. by cells secreting different key cytokines and by generating anaphylatoxins through activation of the complement system. The lectin pathway of complement becomes activated when the pattern recognition molecules mannan-binding lectin (MBL) or ficolins recognize foreign structures. This recognition leads to the activation of MBL-associated serine proteases 1 and 2 (MASP-1 and-2). Recently, it was found that MASP-1 is the first enzyme to get activated leading thereafter to the activation of MASP-2 and culminating with the activation of other complement system components [3,4]. As a result, fragments of complement components deposit on the pathogens surface, leading to opsonization and phagocytic ingestion, a mechanism subverted by M. leprae to allow intracellular infection [5].
In this work, we investigated if MASP2 polymorphisms and MASP-2 or MAp19 levels in serum can play a role in the susceptibility to leprosy. We found an association between nine previously described and two formerly not investigated MASP2 polymorphisms and MASP-2 and MAp19 levels in healthy individuals and in patients with leprosy disease, as well as an association between some of these SNPs and the susceptibility to the disease and to the lepromatous condition.  . To reduce the potential for selection bias, we selected controls from different sources: 325 blood donors from three different blood banks in Curitiba (122 from the Hospital de Clínicas -UFPR, 164 from the Hemepar and 64 from the Biobanco of the Hospital Evangélico) and 80 volunteers without leprosy from Curitiba and surrounding cities. Patients and control subjects had the same socioeconomic status and were from the same geographical area. The patients and controls were defined as being from major European or African ascendency based on physical characteristics. Based on HLA allelic frequencies of South Brazilian population samples identified in the same way, Euro-Brazilians have an average sub-Saharan African component of 9% and an average Amerindian component of 5%, whereas Afro-Brazilians have at least 40% of African and 6% of Amerindian ancestry [23,24]. Furthermore, the genotype distribution of MBL2 haplotypes of Euro-Brazilians of Southern Brazil, with ascendency defined in the same manner, is homogeneous with the MBL2 genotype distribution of most European populations, whereas Afro-Brazilians are similar to Mozambicans, resulting from the large eastern African input in Brazil during the first half of the 19th century [25,26]. All patients and control subjects provided written informed consent. We also obtained written informed consent from the next of kin, caretakers, or guardians on the behalf of the minors/children participants involved. The study was approved by the local medical ethics committee (Comitê de É tica em Pesquisa-Hospital de Clínicas 497.079/2002-06). In order to verify a possible correlation between MASP2 polymorphisms and protein levels, we also genotyped 291 Danish individuals with Figure 1. Approximate gene location and association with plasma levels of the investigated MASP2 polymorphisms. SNPs flanking the alternative exon 5 were investigated for the first time in this study and are gray-shadowed. Arrows pointing downward: SNPs reportedly associated with low MASP-2 levels (400-200 ng/mL) or very low MASP-2 levels (under 200 ng/mL, shown as a double arrow) [9,10,12]. Arrows pointing upward: SNPs reportedly associated with high MASP-2 levels (above 600 ng/mL) [12]. SNPs with unknown or neutral effect are indicated without an arrow. Nucleotide substitutions that correspond to the amino acid changes in the NG007289.1 reference sequence are: g.5557G.A (p.R99Q), g.5620A.G (p.D120G), g.5638C.T (p.P126L), g.21370G.T (p.Y371D), g.21389T.C (p.V377A) and g. 24599G

MASP2 Genotyping
We optimized a multiplex sequence-specific amplification method (multiplex PCR-SSP) to rapidly identify nine single nucleotide polymorphisms (SNPs rs61735600, rs7548659, rs72550870, rs56392418, rs17409276, rs12711521, rs2273346, rs12085877 and rs1782455) and ten corresponding haplotypes (*1A, *1B1-h, *1B2-h, *1C1-l, *1C2-l, *2A1, *2A2-l, *2B1-i, *2B2A-i, *2B2B-l) at low cost and with high accuracy, as previously published [12]. We further investigated the SNPs rs2273344 and rs9430347 flanking the alternative exon 5. These SNPs could be responsible for a hitch-hiking effect with the intron 9 g.21081T polymorphism, explaining the previous association found by our group with this variant and high MASP-2 as well as low MAp19 levels [12]. To this end, we amplified a 316 bp fragment specific for NG_007289.1:g.7164G.A (rs2273344) in intron 4 and NG_007289.1:g.7441A.G (rs9430347) in intron 5 using the MASP2 Intr4_rs2273344Gf (59 GTTCCCTGCACTGTGGGACG 39) and MASP2 In-tr4_rs2273344Af (59 GTTCCCTGCACTGTGGGACA 39) forward sequence-specific primers (SSPs) conjugated with the MASP2 Intr5_ rs9430347Ar (59 CTCCCACCCCAGAGACACGT 39) and MASP2 Intr5_rs9430347Gr (59 CTCCCACCCCAGAGA-CACGC 39) reverse SSPs. In order to control for the quality of the reactions, we added two generic primers (HGHf: 59 TGCCTTCCCAACCATTCCCTTA 39 and HGHr: 59 CCACT-CACGGATTTCTGTTGTGTTTC 39) to amplify a fragment of 431 bp of the human growth hormone gene (HGH). The amplification protocol starts with a 3 min denaturation step at 96uC, followed by 30 cycles of 15 sec at 94uC, 30 sec at the specific annealing temperature and 30 sec at 72uC, concluding with 5 min at 72uC in the final DNA extension step. Annealing temperature decreased every 10 cycles (69uC, 67uC and 65uC), according to a previously published ''touch-down'' strategy which assures higher specificity to the amplification, while providing a larger amount of the desired PCR product [29]. The haplotypes defined by two SNPs, amplified by a pair of SSPs, were identified by the presence or absence of specific bands after agarose gel electrophoresis. Control bands informed on the quality of the reactions.

MASP-2 Concentration Assay
We measured MASP-2 concentrations in 1:40 diluted sera of 45 patients and 82 controls with the same proportion of selected MASP2 genotypes, using the enzyme-linked immunosorbent assay HK326 (Hycult Biotechnology, Uden, The Netherlands).

Statistics
Genotype, allele and haplotype frequencies were obtained by direct counting. Neighbor SNPs distributed from the promoter to exon 12 were phased with the SSP primers. In most cases, the phase between distantly situated SNPs could be deduced due to strong linkage disequilibrium between the variants. The hypothesis of Hardy-Weinberg equilibrium and of homogeneity between genotype distributions (exact test of population differentiation of Raymond and Rousset) were verified with the ARLEQUIN software package version 3.1 (http://anthro.unige.ch/arlequin/). Tests of independence between patients and controls, as well as between patients with the lepromatous and non-lepromatous forms, were performed using Fisher exact test. Corresponding false discovery rate q-values were calculated with the Microsoft research false discovery rate tool for 262 tables. MASP-2 levels were compared between the groups using nonparametric Mann-Whitney/Kruskal-Wallis tests (or ANOVA if data passed normality tests). The statistical analysis was undertaken using the GraphPad Prism 3.0 software package. Two-tailed P-values less than 5% were considered significant. They were additionally corrected with the Bonferroni method (shown as P Bf ). HAPSTAT software was used to evaluate associations between SNPs or haplotypes and leprosy per se as well as lepromatous leprosy, adjusted for age, gender and ethnic group. HAPSTAT software reconstructs haplotypes using an expectation maximization (EM) algorithm to calculate maximum likelihood estimates of haplotype frequencies, while taking into account phase ambiguity. It evaluates association within the additive, recessive, dominant, codominant additive and codominant recessive models [30]. EMgenerated haplotype frequencies were somewhat different to those generated by physical PCR-SSP haplotyping, with 2.3-3.4% discrepancy, which is comparable to the percentage previously reported for MBL2 in the Gabonese population [26]. We choose the best-fitting model between significant additive, recessive,  Data is shown with medians and interquartile ranges, differences were calculated with Mann-Whitney tests (P Bf = 0.02 between AG/AG and P Bf = 0.1 between AG/GA groups). Pat: Patients; Co: controls; diamonds: genotypes with p.126L (*1C1-l and *1C2-l); closed circles: genotypes with *2B1-i. B) Differences between MASP-2 levels of lepromatous and non-lepromatous patients. Data is shown with means and standard error of the means; differences between groups were calculated with an unpaired t test. Diamonds: genotypes with *1C2-l and/or *2A2-l but without *1B1-h or *1B2-h. doi:10.1371/journal.pone.0069054.g003 dominant, codominant additive and codominant recessive models adjusted for age, ethnic group and gender distributions, based on the lowest information criterion of Akaike [31]. We also used the default parameters of the Human Splicing Finder program to evaluate in silico possible effects of the intron 4 and 5 variants [32]. Logistic regression models were used to adjust results for age, gender and ethnic group, as well as for previously published MBL2, FCN1 and FCN2 genotypes [33][34][35], using STATA v.9.2 (Statacorp, USA).

MASP2 Polymorphisms and Haplotypes
Genotype distribution was in Hardy-Weinberg equilibrium. Different combinations of nine out of 11 investigated polymorphisms (g.4847A.C in the promoter, p.R99Q, p.D120G and p.P126L in exon 3, g.21081C.T in intron 9, p.Y371D and p.V377A in exon 10, p.R439H and g.24762T.C in exon 12) resulted in 11 haplotypes, one of them being a possible recombinant between the promoter-exon 3 block of *2B2A-i and the exon 9-exon 12 block of *1B1-h, thus named *2B2A-i.1B1-h [NG_007289.1:g.4847A, p.99R, p.120D, p.126P6g.21081T, p.371D, p.377V, g.24762C]. Further six haplotypes presented recombination of the region surrounding the MAp19 alternative exon 5, as defined by the SNPs rs2273344 and rs9430347. The SNPs occurred in three different combinations in cis, AG being the most common (around 85%), followed by GA (around 15%) and GG, which appeared only twice in the patient group, clustered in the *2B2A-i haplotype. Despite being rare in the human, GG is conserved in the chimpanzee (BAC clone CH251-686E20 from chromosome 1). The phylogenetic nomenclature adopted for MASP2 cannot accommodate recombinant haplotypes whose parental haplotypes are not clear, thus we decided to add [AG], [GA] or [GG] to the original names to account for the information regarding these polymorphisms.

MASP2 Association with MASP-2 and MAp19 Levels
The AG variant combination flanking exon 5 occurred in all haplotypes excepting *1B2-h, which is associated with high MASP-2 levels. On the other hand, GA occurred in all haplotypes except in *1C1-l, *1C2-l, *2A2-l and *2B2B-l, which are all known to be associated with low MASP-2 levels [9,10,12]. In the Danish population, where we did not find (and thus expect no interference from) the *1B2-h, *1C1-l and *1C2-l haplotypes, AG/AG genotypes presented lower MASP-2 levels than AG/GA and GA/GA genotypes (medians: 378,492,686 ng/mL, respectively, Kruskal-Wallis P,0.0001). The opposite occurred with MAp19 levels: AG/AG presented higher values than the other genotypes (medians: 238.1.184.4 and 138.6 ng/mL, respectively, Kruskal-Wallis P,0.0001) (Figure 2). In silico, the g.7164G (rs2273344) of the GA allelic combination creates the silencer motif 59-GGGCCA-39, which could be responsible for skipping the alternative exon 5, increasing MASP-2 and reducing MAp19 levels. The GA combination also shares a highly significant, strong dominant effect on generating MASP-2 levels higher than 600 ng/ mL with four other minor alleles: g.4847C in the promoter (rs7548659), g.21081T in intron 9 (rs17409276), g.21370T or p.371D in exon 10 (rs12711521) and g.24762C or p.S493 in exon 12 (rs1782455) ( Table 1). The same SNPs share a smaller, additive effect on the lowest MAp19 levels (less than 100 ng/mL) ( Table 1). They constitute the second most common *1B1-h [GA] haplotype, whereas the major alleles at these SNPs define the most frequent *2B2A-i [AG] haplotype. As expected, these haplotypes also present opposite effects on MASP-2 and MAp19 levels ( Table 2). Furthermore, the p.120G variant and its corresponding *2B2B-l haplotype present a strong additive effect on very low MASP-2 levels (less than 200 ng/mL) (Tables 1 and 2).
We performed multivariate analysis to account for any discrepancies in demographic factor distribution and genotype interaction. As expected, genotypes with the p.126L and/or p.439H variants were not independently associated with the disease, since both can occur in the same haplotype (*1C2-l). Each remained significantly associated with the disease after correction for age, gender, ethnic group and genotypes with the *2B1-i and/ or *2B2A-i haplotypes, which did also remain associated with protection against leprosy development (p.126L P = 0.001, p.439H P = 0.034, *2B1-i and/or *2B2A-i P = 0.049). Significance was lost after the inclusion of genes previously associated with the disease for the same set of patients and controls (only genotypes with FCN2*AGA remained significant-P = 0.026 among genotypes with MBL2*LYPA, FCN1*-542A-144C and MASP2*126L). MASP-2 levels were also associated with the disease after correction for demographic factors (P = 0.001).

Discussion
Due to extreme difficulties in cultivating M. leprae in vitro and its slow growth in animal models, the molecular mechanisms of infection and immune evasion by the bacillus are still poorly known [36]. Nevertheless the importance of complement activation and opsonophagocytosis in mycobacterial monocyte/macrophage invasion has been increasingly underlined by several different polymorphisms of pattern recognition receptors that modulate phagocyte invasion in the host [2]. Among them, genetically-defined low MBL and M-ficolin levels have been found protective by us and by others, against leprosy per se and/or the lepromatous disease. In contrast, haplotypes associated with normal L-ficolin levels were protective against clinical leprosy [34] and C4B deficiency (C4B*Q0) increases the risk to pathological immune reactivity in lepromatous disease [37].
Using multiplex sequence-specific amplification for physical MASP2 haplotyping [12] and the same genotype-phenotype approach that proved to be informative in our previous investigations with MBL and M-ficolin, as well as with MASP-2 in Chagas disease [14], we were able to go deeper in our investigation of lectin pathway components in the susceptibility to leprosy. To this end, we refined our haplotyping strategy to include two common polymorphisms flanking the alternative exon 5, which is exclusive of the truncated MAp19 protein [38]. In fact, the NG_007289.1:g.7164G variant in intron 4 and g.7441A in intron 5 were as strongly associated with the generation of high MASP-2 and low MAp19 levels, as the formerly reported g.21081T polymorphism in intron 9 [12]. The association with the g.21081T variant is thus most probably a hitch-hiking effect, since all three intronic variants are strongly linked. In fact, we observed the g.7164G variant in silico to create a silencer motif which might cause the alternative exon 5 to be skipped in the pre-mRNA splicing process. Interestingly, recombination has shuffled the intron 4 and 5 variants in a manner that they happen to occur in their two common AG and GA combinations in most of the clade 1 and clade 2 MASP2 haplotypes with wild-type amino acid sequence (excluding only p.Y371D). The modulation of MASP-2/ MAp19 levels may thus be of ancient selective advantage, similarly to the splicing variants of endoplasmic reticulum amino peptidase 2 (ERAP2), which has a fundamental role in MHC class I presentation [39]. This balance of alternative mRNA processing may differ for different cell lineages, since it is known that MAp19 is mainly expressed in Kupffer cells, whereas MASP-2 is mainly synthesized in hepatocytes [38].
Surprisingly, we found that genetically-defined low MASP-2 levels do not mimic the protective association found between MBL and M-ficolin deficiencies and leprosy. Instead, low MASP-2 levels generated by p.126L, p.377A and/or p.439H combined with the exon 5 flanking AG variants were associated with susceptibility to leprosy per se as well as with lepromatous disease. Among these variants, p.439H is the only one known to hinder complement activation. This mutation disrupts the activation peptide. In contrast to p.120G, which is also known to cause MASP-2 deficiency [8][9][10], the resulting protein binds to MBL. Nevertheless it cannot autoactivate [10]. Thus, in addition to being unable to activate complement, p.439H probably functions as a ''sink'' for MBL, reducing binding opportunities with full-working MASP-2 homodimers in heterozygotes. This could also be the underlying reason for its dominant effect on the susceptibility to the disease. Importantly, the distribution of p.120G did not differ between patients and controls, as well as between lepromatous and nonlepromatous patients. One might argue that, beside functional differences, the contrasting results between p.439H and p.120G could be the result of population substructure, since p.439H are most probably of African, and p.120G of European origin [9][10][11]. Nevertheless the association of the *1C2-l haplotype (harboring both p.126L and p.439H) with leprosy per se was especially evident after the exclusion of Afro-Brazilian patients, which present at least 40% of sub-Saharan African admixture [23,24], whereas the frequencies of p.120G in patients and controls remained similar.
Both leprosy and malaria are intracellular infections associated with the same MBL and MASP-2 variants that disturb the lectin pathway of complement activation, nevertheless with opposite effects on the disease susceptibility. In sharp contrast with leprosy, where susceptibility to the disease is associated with haplotypes generating high MBL levels [35] and low MASP-2 levels (this study), low MBL levels generated by the MBL2*LYQC haplotype increase susceptibility to malaria [16,[40][41][42], whereas low MASP-2 levels generated by the p.439H variant are protective [16]. These opposite effects cannot be explained solely on the basis of disturbed complement activation, since MBL and MASP-2 are known to act synergistically in this process. A possible explanation could be the ability of MASPs to block the interaction site of MBL with the CD91/calreticulin complex on the macrophage cell membrane, thereby preventing MBL-driven phagocytosis [43]. If this mechanism is operating in mycobacterial cell invasion, low MASP-2 levels would be insufficient to avoid subsequent infection. Surely this hypothesis needs to be tested in a functional setting, maybe using the recently achieved bioengineered BCG bacillus [36]. Otherwise in malaria, the ingestion of hemozoin could be beneficial to control for dyserithropoiesis [44]. It should also be kept in mind that MASP-2 is a pleiotropic molecule, being involved in the coagulation cascade through the cleavage of prothrombin, generating cross-linked fibrin covalently bound on bacterial surfaces [45,46]. Disruption of this pathway could modulate susceptibility to infections, as well. In addition to low MASP-2, high MAp19 levels could play a role in leprosy pathogenesis, but the strong association with p.439H, which does not occur in the truncated protein, disfavors an important role for it. Finally, it should be kept in mind that associations of genes of the lectin pathway of complement with the disease are not independent of each other. Multivariate analysis predicted a more important role for the FCN2*AGA haplotype [34], known to be associated with normal L-ficolin levels, but whose functional relevance awaits further investigation.

Conclusion
The results lead us to suggest that, in contrast to MBL, low MASP-2 levels increase the susceptibility to leprosy in general and to lepromatous leprosy in particular. MASP2 genotypes and MASP-2 levels might be used as biomarkers to predict disease progression in leprosy.