Genetic variants in the PSMA 6 , PSMC 6 and PSMA 3 genes associated with childhood asthma in Latvian and Taiwanese populations

Genomics and Bioinformatics, Institute of Biology, University of Latvia 3, Miera Str., Salaspils, Latvia, LV-2169 Institute of Biomedicine, Tzu-Chi University 701, Zhongyang Rd., Sec 3., Hualien 97004, Taiwan Faculty of Medicine, University of Latvia 1a, Sharlotes Str., Riga, Latvia, LV-1001 University Children Hospital 20, Juglas Str., Riga, Latvia, LV-1079 Division of Allergy and Clinical Immunology, College of Medicine, National Cheng Kung University 1, University Road, Tainan City 701, Tainan, Taiwan Latvian Institute of Organic Synthesis 21, Aizkraukles Str., Riga, Latvia, LV-1006


Introduction.
Asthma is a chronic inflammatory disease caused by complex gene-gene and gene-environment interactions with hyper-responsiveness to various nonspecific stimuli [1][2][3] being to a large extent geneti-cally heterogeneous between human populations [4].A number of genes implicated in asthma encode various signaling proteins and transcription factors including those driven by NF-kB signaling pathways [5][6][7].
However, interplay of multiple risk alleles and/or genotypes and primary driver of the disease remains still unclear.
In eukaryotes, processing and degradation of vast majority of regulatory proteins are mediated by ubiquitin-proteasome system (UPS).Proteasomes, key UPS enzymatic complex possess several types of peptidase, endoribonuclease, protein-chaperone and DNA-helicase activities [8][9][10] allowing strict control and coordination of all steps of gene expression, genes and proteins networks and processes of genome-environment interaction.Insufficient proteasome function was implicated in pathophysiology of various acute and chronic lung diseases and their complications [11][12][13][14] and potentially could be a consequence of particular proteasomal genes structural variations.
Aim of the current study was to genotype five single nucleotide polymorphisms (SNPs) belonging to the PSMA6 (rs2277460 and rs1048990), PSMA3 (rs2348071), and PSMC6 (rs2295826 and rs2295827) proteasomal genes and evaluate main and sex-specific association between variations of these genes and asthma in Latvians and Taiwanese.
Materials and methods.One hundred two children (28 girls) aged under five and 159 (69 girls) aged under three represented Latvian (LV) and Taiwanese (TW) asthma groups respectively.LV BA patients were enrolled from the outpatient clinic of P. Stradins Clinical University Hospital and Children Clinical University Hospital «Gailezers» in Riga, Latvia.
TW study subjects were enrolled from elementary school for allergy diseases screen Taoyuan General Hospital, Taiwan.All patients were diagnosed with mild or moderate persistent asthma according to the guidelines of the Global Initiative for Asthma (GINA; http://www.ginasthma.org/local/uploads/files/GINA_Under5_Pocket_20091_1.pdf).The studies were approved by the Central Medical Ethics Commission of the Republic of Latvia Ministry of Health and Ethics Committee of Taoyuan General Hospital, Taiwan.Informed consent was obtained from parents of study participants.
Latvian and Taiwanese control groups of 191 (age = = 54.8 ± 18.6; 117 women) and 1097 (aged under five; 558 girls) participants respectively, were described and genetic diversity of SNPs of interest was studied previously [28] providing primary genotyping data to be used in current study to evaluate asthma main effects for each particular SNP, construct multi locus genotypes and stratify controls by sex to reveal asthma sex specific associations in single-and multi-locus models.
DNA extraction and genotyping technologies were the same as in [28].
For quality control, of the 16 randomly chosen samples per each marker were genotyped in duplicate in different experiments for asthma samples from both Latvian and Taiwanese collections.The concordance of the genotyping was 100 %.The chromosome 14 GRCh37.p5assembly (NCBI reference sequence: NC_ 000014.8)sequence information was used for loci description.
Personalised genotyping data documentation resulted in knowledge of 5 locus genotype (5-LG: rs2277460/ rs1048990/rs2295826/rs2295827/rs2348071) of each individual participant of the study.The 5-LGs, single locus genotypes (SLGs) and alleles frequencies were estimated by direct gene counting.DnaSP version 5 (http:// www.ub.es/dnasp/ [29] was used to reconstruct the haplotypes from un-phased genotypes, evaluate the nucleotide and haplotype genetic diversity and pairwise linkage disequilibrium (LD) between the loci (D' and r 2 ).Both the two-tailed Fisher's exact test and the c 2 test were applied to evaluate the linkage between the rs2295826 and rs2295827 polymorphic sites at three p-value levels (p < 0.05; p < 0.01; p < 0.001).The Bonferroni correction included in DnaSP analysis was taken into account to support the significance of the revealed disequilibrium (a' = = 0.05).
Deviation from the Hardy-Weinberg equilibrium and differences between case and control groups in al-lele, genotype and haplotype frequencies as well as permutation test (Monte Carlo method/number of simulations = 10000) were evaluated by c 2 using XLSTAT 2013 software for Windows.Dominant, recessive, over dominant and multiplicative genetic models for every individual locus were designed according to Lewis [30] and analysed by using 2 ´2 contingency tables.Odds ratio (OR) more than 2 and less than 0.5 was considered to be clinically significant.Stratification was performed by sex.
Results and discussion.In both Latvian and Taiwanese sample collections the genotyping call rate was 100 % for all markers; alleles and genotypes frequencies are given in Tables 1 and 2. The rs2348071 being in Latvian patients in HWE, significantly (P < 0.001) deviated from equilibrium in Taiwanese.Other markers were found to be in HWE in both LV and TW patients.The rs2295826 and rs2295827 were observed in complete (D' = 1, r 2 = 1) and slightly disrupted (D' = 1, r 2 = 0.896) LD in Latvians and Taiwanese respectively.
The distributions of alleles and genotypes in case groups were compared with those previously identified in the populations [28] and the data on single-locus association are summarized in Table 1 and Table 2.In Latvians all five loci showed the asthma main effect for rare alleles and heterozygous genotypes.The rs1048990 was associated with the disease in both females and males.The asthma susceptibility of resting loci was characterized by nonadditivity; that was, the rs2277460 and rs2348071 were associated with asthma in females, and the rs2295826 and rs2295827 were associated with asthma in males.In Taiwanese, the rs2277460 appears to be asthma neutral and resting loci showed the disease susceptibility only in females.The asthma risk effect was observed for the rs1048990 and rs2348071 GG genotypes and rs2348071 allele G. Rare alleles and heterozygous genotypes of the rs2295826 and rs2295827 showed female-specific asthma protective effect.
The multi-locus genotypes showing asthma risk or protective effect in any of our populations are listed in Table 3 (see suppl.).In Latvians a statistically significant protective effect was observed for all variants of multi locus genotypes being homozygous for the alleles common in the population.The 5LG of CC/CC/AA/CC/ GG configuration being protective in Latvians (OR = = 0.322 [0.196 -0.651]), showed the asthma risk effect in Taiwanese females (OR = 2.911 [1.327-6.387]).Similarly, being protective in Latvians the AA/CC/GG (rs2295826/rs2295827/rs2348071) genotype appears to be the disease susceptible in Taiwanese.In Latvians, the risk effect was observed for the rs2277460/rs2348071, rs1048990/rs2348071, rs1048990/rs2295826/rs2295827 and rs2295826/rs2295827/rs2348071 genotypes being simultaneously heterozygous at all loci involved.All mentioned genotypes were neutral in Taiwanese.In contrast, the rs2295826/rs2295827/rs2348071 genotype of AG/CT/AA configuration being neutral in Latvians, showed the protective effect in Taiwanese females.
The data of haplotype analysis are given in Table 4 (see suppl.).Haplotype diversity was higher in Latvian patients than in the population and did not differ between the cases and population in Taiwanese.The most frequent in Latvians the Hap1 (CCACG) showed male specific asthma protective (OR = 0.633 [0.416-0.963])effect in this population.The Hap6 (CGACG) showed strong (P < 0.0001) association with the asthma risk in both Latvians (OR = 4.525 [2.286-8.958])and Taiwanese (OR = 2.448 [1.763-3.399])for both females and males.Minor in both populations and Taiwanese cases the Hap9 (CCGTA) was strongly (P < 0.0001) associated with the asthma phenotype in Latvians.
In this paper we provide for the first time the evidence that polymorphism in the PSMA6, PSMC6, and PSMA3 proteasomal genes may contribute to the risk of childhood asthma in both Latvian and Taiwanese populations.The most remarkable finding of our study is that haplotype CGACG was revealed to be a strong (P < 0.0001) asthma risk factor in both Latvians and Taiwanese.Other identified asthma risk and protective single-and multi-locus genetic variants are different between two populations.The difference between human  populations in asthma genetics is a well-known phenomenon described for a number of asthma susceptible loci having mainly ethnos specific differences in genetic diversity [4].The Latvian and Taiwanese populations significantly differ in genetic diversity of loci studied here [28].This suggests involvement of these loci in the processes of evolutional and/or geographical adaptation to environment and a potential for allele substitutions to have different ethnic specific influence on the human health and population morbidity [28].
Several associations revealed in our study showed non-additivity between sexes.In Latvians the rs2277460 and rs2348071 were associated with asthma in females, and the rs2295826 and rs2295827 were disease susceptible in males; in Taiwanese all asthma susceptible loci were limited to females.Sex specific differences in incidence, prevalence, and severity are also well known features of asthma epidemiology.Sex-specific associations with the disease have been recently reported for SNPs of several genes-candidates including the IFNG [31], IL17F [32], TSLP [33], VDR [34], and KCNB1 [35] genes.Our analysis of the BA main effect in Latvian population is a subject to some limitation as sexes were not equally presented in both BA and control groups.Although a significant asthma main effect was detected for all five loci studied, only the rs1048990 showed an additive effect that was an association in both females and males.The replication study in additional larger cohorts represented by sexes equally is required to validate the results found in the current study for Latvian population.
Due to the pleiotropic effect, a frequent phenomenon in human complex traits and diseases [36], some loci of susceptibility may be shared among many autoimmune and other immune-mediated diseases [37,38].Earlier the genetic pleiotropic effect has been reported for asthma and obesity [39,40] and for asthma and juvenile rheumatoid arthritis [41,42].Similarly, SNPs associated with asthma in our current study, previously have been found to be susceptible in Latvians to other immune-mediated pathologies including juvenile idiopathic arthritis [24,43], children obesity [44] and multiple sclerosis [45].The rs1048990 was widely genotyped in many human populations and reported as an ethnic specific risk factor for inflammation within the cardio-vascular system [27,28].
All loci we have studied here belong to the non-coding regions of corresponding genes and nucleotide substitutions potentially may influence the gene expression through allele specific targeting of different regulatory elements.Among the allele-specific targets described by Sjakste with co-authors earlier [24,28], several sites showed affinity to transcription factors and splicing signals implicated previously in immunity, lung function and lung pathology.The targeting of these regulatory proteins may influence asthma pathogenesis and needs to be mentioned in respect of current study.The rs2277460 ancestral allele C, the major allele in human populations over the world, appears to be functionally neutral.Substitution to A generates a target to hnRNP A1, a multifunctional protein implicated in the association with multiple promoter sequences and modulation of a number of transcriptional events [46].The hnRNP A1 has been shown to play a key role in many human pathologies including lung cancer and response to viral pathogens [46,47].It may influence protein-protein interactions including those with participation of NF-kB [46] playing in turn a significant role in the asthma development and progression [5][6][7].Additionally, it is involved in crosstalk with ubiquitin proteasome system at different levels of NF-kB and other regulatory proteins signaling pathways [48].Allele A also assists to sequence affinity to the BARBIE box proteins found to be involved in inflammatory response of alveolar macrophages [49].Substitution C ® G at the rs1048990 was shown to influence the gene expression in vivo and in vitro [27,50] and significantly change the sequence capacity to bind a number of splicing signals and transcription factors [28].Rare allele G generates binding sites for the multifunctional proteins of p53 and DMRT families implicated in the processes of climatic [51] and evolutional [52] adaptation.The targeting of these proteins potentially could be involved in the mechanisms of natural selection and ethnos specific susceptibility to inflammation [27,28].
Common allele A of the rs2295826 (first intron of the PSMC6 gene) generates the targets for the mentioned above hnRNP A1 regulatory protein and for the transcription factor of CREB family involved in transcriptional control of many pro-inflammatory genes [53,54] and implicated in asthma pathogenesis [55], asthma phenotypes and response to therapy [56].
The rs2348071 SNP strongly discriminates Latvians having a major allele G (about 70 %) and Taiwanese having a major ancestral allele A (about 70 %).Previously we have suggested [28] that transition A ® G happened in Caucasians about 15,000 years ago was supported by positive selection.This mutation eliminates potential targets for hnRNP A1 and the transcription factors of CART family shown to be an essential participant of signaling respiratory network [57] and the MEF2 family implicated in transcriptional switch between metabolism and immunity [58].
Summarizing mentioned results we suggest that the nucleotide substitutions we have studied may significantly modulate the transcription of related genes and gene network in response to the inflammation and other environmental stimuli and influence the asthma susceptibility.
Conclusions.Our findings provide an evidence that single-and multi locus variations in the 14q13-23 PSMA6/ PSMC6/PSMA3 proteasomal genes cluster are associated with childhood asthma in Latvian and Taiwanese populations and could play an important role in asthma and other immune-mediated pathologies in both Caucasians and Asians, as either the risk or protective ethnicand sex-specific genetic factors.
Identification of genetic variants susceptible to asthma and other immune-mediated pathologies, both common and different across populations, is important in understanding pathogenesis and phenotype variability of these multifactorial diseases.It might be a subject of thorough investigation in the nearest future. Funding.

Table 1
SNP allele and genotype distribution and data on association with paediatric asthma in Latvian population [24,28]n an allele and genotype presentation in control group are given according to Sjakste et al.[24,28]; P -probability calculated by c 2 test; c -corrected probability calculated by Monte Carlo method with 10000 simulations.