Assessment of Toll-like receptor 2, 4 and 9 SNP genotypes in canine sino-nasal aspergillosis

Background The exact aetiology of canine sino-nasal aspergillosis (SNA) is unknown. In man, dysfunction in innate immunity, particularly in the function of pattern recognition receptors, is implicated in the pathogenesis of inflammatory sino-nasal disease and in fungal diseases. Associations between single nucleotide polymorphisms (SNPs) in Toll-like receptors (TLRs) and these diseases have been identified. Similarly, in dogs SNPs in genes encoding TLRs may be important in the pathogenesis of SNA. The aims of the present study were (1) to identify the presence of non-synonymous SNPs in the coding regions of the TLR2, 4 and 9 genes in dogs suffering from SNA, and (2) to investigate the SNP genotypes in dogs with SNA compared with a control population. Results Direct sequencing of nine dogs of various breeds with SNA revealed two non-synonymous SNPs in the coding region of TLR2, eight in TLR4 and four in TLR9. These non-synonymous SNPs were further evaluated in a case-control study of affected Golden Retrievers, Labrador Retrievers, Rottweilers and Beaucerons. Genotyping was performed using a combination of allele-specific primers and hydrolysis probe assays in 31 dogs with SNA and 31controls. No significant difference in minor allele frequency was identified between these groups, for all studied SNPs, in any of the four breeds. Conclusions These findings do not support a role for non-synonymous SNPs in the TLR 2, 4 and 9 coding regions in the pathogenesis of canine SNA, but do not exclude a role for innate immunity in the pathogenesis of the disease.


Background
The sino-nasal mucosa is an important interface with the environment and represents a dynamic system for innate host defence [1]. In addition to providing a barrier to the entry of pathogens, epithelial cells lining the airway are able to recognise pathogen-associated molecular patterns (PAMPs) [2]. Toll-like receptors (TLRs) are situated on these cells and on dendritic cell populations in this location and appear to play crucial roles in the distinction between commensal flora and PAMPs. Single nucleotide polymorphisms (SNPs) within the TLR genes have been reported to impact the immune reaction to microbial antigens, and to be associated with the incidence and outcome of infections [3]. Collectively, the TLRs cover the recognition of a wide variety of pathogens (i.e. viruses, bacteria, parasites and fungi). TLR2, TLR4 and TLR9 are particularly associated with recognition of fungi (e.g. Candida albicans and Aspergillus fumigatus) [4][5][6][7][8][9], and SNPs in these TLRs have been associated with higher risk of fungal disease in man [10].
In dogs, sino-nasal aspergillosis (SNA) is a cause of chronic rhinitis, and the disease is most often caused by A. fumigatus [11]. SNA is characterized by the formation of superficial mucosal fungal plaques within the nasal cavity and/or frontal sinus, associated with a severe mucosal inflammatory response and a local destruction of nasal bone [12]. As the disease usually affects systemically healthy dogs, and A. fumigatus is a ubiquitous fungus, a local immune deficiency in affected dogs has been suggested [13].
The clinicopathological features of SNA are similar to those of human chronic erosive non-invasive fungal sinusitis [14]. Although TLRs have not been studied in this type of sinusitis in people, they have been better characterized in chronic rhinosinusitis (CRS) including fungal CRS, and some forms of aspergillosis. Pathogenesis of the CRS complex remains unclear, but association of this disease with TLR polymorphisms, particularly TLR2 SNPs, has been described [15].
Association between TLR SNPs and incidence of various forms of human Aspergillus infection have been reported. A significant association between the incidence of invasive aspergillosis (IA) in recipients of transplanted allogeneic stem cells and the presence of some TLR4 SNPs in donors has been reported [16,17]. Hence, Carvalho et al. [10] have described an increased frequency of chronic pulmonary aspergillosis among TLR4 SNP carriers. In the same study, a SNP in TLR9 was shown to predispose to the allergic form of pulmonary aspergillosis.
Polymorphisms in TLR genes have not yet been described in dogs with SNA, but alteration in these receptors could hamper their function, therefore disrupting the recognition of fungi, eventually leading to fungal escape from immune surveillance. Recently, up-regulation of TLR2, 4 and 9 mRNA expression has been reported in the nasal mucosa of affected dogs [18]. Comparison of the canine and human TLR genetic sequences show a high degree of homology [19]; suggesting the polymorphisms in TLR genes found in people with aspergillosis and/or fungal CRS might also be observed in dogs.
A retrospective analysis of the medical records at the Veterinary Clinic of the University (CVU) of Liège (Belgium) over the last 5 years revealed four predominant canine breeds diagnosed with SNA: Golden Retriever, Labrador Retriever, Rottweiler and Beauceron Sheepdog. A candidate gene approach was used to test the hypothesis that SNPs in TLR2, 4 and 9 genes were associated with canine SNA. The presence of non-synonymous SNPs in the coding regions of the TLR2, 4 and 9 genes was investigated by direct sequencing of PCR-amplified products from these regions, using DNA from nine dogs with SNA belonging to various breeds (including the four aforementioned breeds). Secondly, a combination of allele-specific primers and Taq-man probe-based PCR genotyping was applied to samples from affected and control dogs derived from these four breeds in order to assess the significance of these SNPs in canine SNA.

Selection of single nucleotide polymorphisms
The initial part of the study focused on a mutational analysis of TLR2, 4 and 9 genes in nine dogs with SNA belonging to various breeds. This analysis revealed two non-synonymous SNPs in the coding region of TLR2: C137A (rs22410121) and C1547T (not reported in the canine genome database). Eight non-synonymous SNPs were identified in the TLR4 exons: T23C (not reported in the canine genome database but described by Kathrani et al. [20]), C107T (not reported previously), T500C (rs22145736), A600C (rs22189454), A688G (rs22189456), G1039A (not reported in the canine genome database but described by Kathrani et al. [20]), A1571T (rs22124023) and G1807A (rs22123995). Four non-synonymous SNPs were identified in TLR9 exon: G1138A (not reported previously), A1372C (not reported previously), A2158G (rs22882109) and G2927A (not reported previously).
All non-synonymous SNPs identified resulted in a change in the class of amino acid coded (Table 1).

SNP genotyping and association study
To assess the significance of the identified SNPs, the allelic frequencies in a group of dogs with SNA belonging to four breeds commonly affected by SNA (Golden Retriever, Labrador Retriever, Rottweiler and Beauceron Sheepdog) were compared with those found in control dogs of the same breeds (Table 2), and no significant difference was found. Two SNPs (TLR2 C1547T and TLR4 T500C) were found monomorphic between and within samples from the four selected breeds.
The same analysis was conducted within the four breed groups separately. Again, the allelic differences, although variable across breeds, were not related to the disease phenotype (Tables 3, 4, 5 and 6). In Golden Retrievers, the TLR4 C107T SNP was not found, and the least common alleles for all breeds taken together for TLR9 A1372C and TLR9 A2158G were the most common in this breed. In Golden Retrievers, the association of the C137A SNP with the disease almost reached significance. In the Labrador Retriever dogs, the least common alleles

G2927A
Arginine Histidine for all breeds taken together for TLR2 C137A, TLR4 A600C, TLR4 A688G, TLR4 and G1039A were most common. In Rottweiler dogs, TLR4 C107T, TLR4 A1571T, TLR4 G1807A, TLR9 G1138A and TLR9 G2927A SNPs were absent, and the least common allele for all breeds taken together for TLR4 A600C was the most common allele. In the Beauceron Sheepdogs, TLR2 C137A, TLR9 G1138A, TLR9 A2158G and TLR9 G2927A SNPs were absent, and the least common allele in all breeds taken together for TLR9 A1372C was the most common allele in this breed.
Statistical significance was set at P < 0.05.
Statistical significance was set at P < 0.05.
Statistical significance was set at P < 0.05.

Discussion
The present study investigated the association between SNPs in TLR2, 4 and 9 genes, and the risk of developing SNA in dogs. In the first part of the study, sequencing of the coding regions of these genes in nine dogs with SNA (from various breeds) revealed two, eight and four nonsynonymous SNPs in the coding region of TLR2, TLR4, and TLR9 genes, respectively. In the second part of the study, no association was found between these SNPs and the presence of SNA. Some of the SNPs identified in the first part of the study were not found in samples from the four breeds of dog genotyped in the second part, and minor allele frequencies varied extensively between breeds, with some uncommon alleles in one breed becoming the most common alleles in the other, and vice versa. This supports a wide genetic variations between canine breeds [21].
The only result close to significance was found for TLR2 C137A SNP in the Golden Retrievers. Due to the unpredictable direction of allelic frequency shift between dogs with SNA and controls, PLINK software was used to compute two-tailed Fisher's exact test p-values. Using a one tailed test would have led to half the actual p-value (0.04; data not shown), thereby reaching significance. Although this might be interpreted as a significant result, the rationale for choosing a one-tailed test might be debated. Furthermore, multiple testing issues (i.e. performing several tests) could be considered, leading to a lowered significance threshold that would make this result not significant. On the other hand, the power of the test might be low due to the relatively small size of the sample. Therefore, this SNP could merit further investigation with a larger number of dogs.
Although some researchers have studied the association between TLR SNPs in human CRS and some fungal diseases, the results are heterogeneous and sometimes contradictory. Several researchers have shown a significant association between TLR4 and/or TLR9 SNPs and the incidence of several forms of aspergillosis [10,16,17]; however, another study failed to associate these polymorphisms with that disease. Recently, Carvalho et al. noted a positive association between TLR4 SNPs and fungal colonization, but not susceptibility to fungal infection [22]. Kesh et al. did not report any positive association of recipient or donor TLR4 SNPs with the incidence of IA in patients undergoing allogeneic stem cell transplantation [23]. In human CRS, there are also discrepancies between studies regarding the role of SNPs in TLR genes. Park et al. reported an association between TLR2 SNPs in a Korean population [15], but Tewfik et al. could not confirm these results in Quebec [24].
The effect of the environment on susceptibility to certain diseases is another important aspect that should be considered together with genetic background. Vercelli [25] reported that genes and environment are intertwined in complex nonlinear relationships. Therefore, the same genetic background may result in the expression of different phenotypes when exposed to different environments. Consequently, in a disease with a complex pathogenesis, the genotype of multiple contributing genes, together with environmental factors, may modify the likelihood of an individual developing the disease. In such instances, individual gene mutations are insufficient to cause the disease on their own [26].
There are several limitations to the present study. Firstly, although we cannot confirm an association between polymorphism in the TLR2, 4 and 9 genes with SNA, it is possible that because of the small size of the study population, a slight/weak but real contribution of TLR2, 4 or 9 SNPs has been missed. Another potential limitation is the selection of control dogs. Although efforts were made to exclude dogs with respiratory, inflammatory, immune-mediated, neoplastic or infectious diseases, the possibility cannot be excluded that these dogs might develop such disease later in life. Nevertheless, to try to overcome this problem, we recruited only dogs older than 6 years. Another limitation was the use of a candidate gene approach, which permits comparison of allele frequencies of SNPs in genes suspected to be involved in SNA among dogs with the disease and others without (controls), but this method precludes finding other genes of interest. Genome-wide approaches would overcome this limitation, but the need for a large number of affected dogs and the high associated expense renders this difficult.

Conclusions
The findings of the present study suggest that TLR2, 4 and 9 are not implicated in the pathogenesis of SNA in dogs. However, these results are limited by the size of the study population. This does not rule out the possibility of innate immune defects contributing to the development of the disease. Other TLR genes and/or other genes involved in mucosal immunity may be important and require further investigation.

Selection of single nucleotide polymorphisms
A mutational analysis of the TLR2, 4 and 9 exons was carried out in nine dogs diagnosed with SNA in order to determine the presence of non-synonymous SNPs in the coding regions of the genes and to investigate these further in a case-control association study to determine their significance in SNA. The nine selected dogs (aged 10 months to 14 years; median 6.7 years; 3 females and 6 males) were from various breeds: Golden Retriever (n = 2), Labrador Retriever (n = 2), Rottweiler (n = 2), and one each of Beauceron, German Shepherd Dog and crossbreed. Diagnosis of SNA was made on the basis of physical examination, observation of intranasal fungal plaques associated with severe turbinate destruction on rhinoscopy, presence of serum antibody specific for A. fumigatus and culture of A. fumigatus from suspected lesions. A computed tomography examination of the head was performed in some dogs.
Blood stored in ethylenediaminetetraacetic acid (EDTA) or biopsy tissue from the nasal mucosa were available for each dog and were used for extraction of genomic DNA using the Macherey-Nagel NucleoSpin blood (or tissue) isolation kit (ABgene, Epsom, UK). All biological material used to perform the present study was collected for diagnostic purpose in all dogs. Nevertheless, the right to use it was obtained by owner consent.
Conventional polymerase chain reaction (PCR) was performed using Taq Platinum® PCR master mix (Invitrogen, California, USA), with reactions heated to 94°C for 2 min, followed by 35 cycles of 94°C for 25 sec, 60°C for 25 sec and 68°C for 4 min then a further 72°C for 7 min. The PCR products obtained were separated by 2% w/v agarose gel electrophoresis and appropriately sized products were excised with a clean scalpel blade and purified using the Nucleospin® Extract II (Macherey-Nagel, Hoerdt, France) as per the manufacturer's instructions. Gel-purified PCR products were sent to GIGA research centre in Liège, Belgium for sequencing using sequencing primers specifically designed to cover the entire sequence of the exons ( Table 8).
The sequence data were compared with the caninespecific GenBank sequences using GeneStudio contig assembly editor (http://www.genestudio.com/), and nonsynonymous SNPs were identified.

SNP genotyping
The non-synonymous SNPs identified in the first part of the study were further analysed in case and control populations by genotyping.
The case group consisted of 31 dogs diagnosed with SNA (aged 10 months to 12 years; median 6.5 years; 7 females and 24 males) from four different breeds: Golden Retriever (n = 10; minimum age 10 months; maximum age 12 years; median 7 years; 1 female and 9 males), Labrador Retriever (n = 8; minimum age 1.3 years; maximum age 12 years; median 7.1 years; 2 females and 6 males), Rottweiler (n = 9; minimum age 3 years; maximum age 9 years; median 5 years; 2 females and 7 males) and Beauceron (n = 4; minimum age 10 months; maximum age 10.5 years; median 2.5 years; 2 females and 2 males). Diagnosis of SNA was based on the criteria given above. These dogs were presented at the Veterinary Clinic of the University (CVU) of Liège (Belgium) between July 2007 and June 2013. During that period, a total of 72 dogs from various breeds were diagnosed with SNA at the same institution. The breeds of dogs included in the present study were the 4 most frequently encountered breeds in these 72 dogs. A Fisher's exact test was used to assess whether those breeds were predisposed to the disease as compared with the general population of dogs presented to the CVU during the same period of time. The four canine breeds were proved to be predisposed to the disease (p < 0,01 in the Golden Retriever and the Rottweiler; p < 0,05 in the Labrador Retriever and the Beauceron).
The control group consisted of 31 dogs of the same breeds: Golden Retriever (n = 10; minimum age 6 years; maximum age 14 years; median 9 years; 4 females and 6 males), Labrador Retriever (n = 8; minimum age 9 years; maximum age 14 years; median 12 years; 3 females and 5 males), Rottweiler (n = 9; minimum age 6 years; maximum age 8 years; median 7.8 years; 4 females and 5 males) and Beauceron (n = 4; minimum age 8 years; maximum age 12 years; median 11 years; 1 female and 3 males), with no history or signs of respiratory disease. Inflammatory, immune-mediated, neoplastic or infectious diseases were also excluded. Twelve dogs were healthy; five were presented for laryngeal paralysis, four for degenerative osteoarthritis, three for a lipoma and two for perineal hernia. One dog each had vaginal haemorrhage, arrhythmia, anal sac impaction, hyperadrenocorticism or hypoadrenocorticism. All dogs in this group were older than 6 years of age (6 -14 years; median 9 years; 15 females and 16 males). Owner consent was again obtained for each dog.
Blood stored in ethylenediaminetetraacetic acid (EDTA) was used for genomic DNA extraction using the Macherey-Nagel NucleoSpin Blood Isolation Kit (ABgene, Epsom, UK).
The allele-specific primer-based PCR was used to genotype the SNPs [28,29]. Briefly, allele-specific PCR primers were designed so that the 3′ terminal nucleotide of a primer corresponded to the site of the SNPs (Table 9). An allele-specific primer matches perfectly with one allele (the specific allele) or has a 3′ mismatch with a non-specific allele. Because mismatched 3′ termini are extended by Taq polymerases with much lower efficiency than correctly matched termini, the allele-specific primer preferentially amplified the specific allele. To improve the specificity of traditional allelespecific primers, we followed a modified method [30,31] in which an artificial base pair mismatch was introduced between the third and the fifth base from the 3′ terminus to each primer (Table 9). Two positive controls were designed for each SNP, to include allele-specific target sequence (without the addition mismatch) and the common primer (forward or reverse as appropriate), to test the ability of the assay to identify the SNP (Table 9). Primers and positive controls were synthesized by Eurogentec S. A. (Seraing, Belgium) and reconstituted in TE buffer (10 mM Tris, 1 mM EDTA, pH 8) before use.
The allele-specific reactions were performed by quantitative PCR (qPCR) using SYBR Green I with separate reactions for each primer set (two sets per SNP). SNP genotyping was performed by comparison of Ct values ( Figure 1) with a homozygous sample having a lower Ct with the specific primer set and a higher Ct with the non-specific one, and a heterozygous sample having approximately equal results from both assays. Each assay was optimized with the synthetic positive controls and residual DNA from the samples used in the sequencing study. The qPCR was performed using GoTaq Colourless Master Mix (Promega Corporation, Southampton, UK) in a Stratagene Mx3005P (Agilent Technologies, Stockport, Cheshire, UK) with results analysed using the MxPro Software (ver. 4.1; Agilent Technologies). Optimisation involved varying the annealing temperature (55, 60 and 65°C), with and without additional MgCl 2 (final concentration: 3.0 and 4.5 mM), to balance the Ct value difference between the target and non-target allele-specific PCR, and non-specific product formation with genomic DNA and water negative controls. The final reactions consisted of 200nM of each primer, 4.5 mM Mg and 1:50,000 SYBR Green I (Invitrogen) with a thermocycling protocol of incubation at 95°C for 2 min for Taq activation and then 35 cycles of 95°C for 15 sec and annealing (55 or 60°C: Table 9) for 30 sec during which the fluorescence data were collected. The thermocycling protocol was extended by heating samples from 75°C to 95°C in 0.5°C increments for 15 sec, during which time the fluorescence data were collected in order to create a melt curve. For each SNP, the two genotyping assays were run for a single DNA sample on the same plate which also included a number of additional controls including the target and non-target synthetic positive control, and AE buffer (negative control).
For three SNPs (TLR9 A1372C, TLR9 C2607T, TLR9 G2927A) it was not possible to genotype on the basis of Ct value alone because Ct values were too similar with both homozygous and heterozygous animals due to the presence of additional amplified products in both reactions ( Figure 2). Therefore, capillary electrophoresis (QIAxcel Advanced System, QIAGEN Group, UK) using the QIAxcel DNA High Resolution Kit (QIAGEN), QX Alignment Marker 15 bp/500 bp (QIAGEN) and QX DNA Size Marker 25-500 bp (QIAGEN) was used to

AGCCGCACATAGCGGGtGT AGCCGCACATAGCGGGAGTCTGTTGGGCCAGCAGGAAGCTGGCACGCAGGAG
The polymorphic base is in bold, the additional inserted 3′ mismatch is in lowercase, the amplicon size in base pairs (bp) and annealing temperature [°C] used for the qPCT are shown below the SNP name.

GG
AA GA A C B Figure 1 Real time PCR results for TLR4 A688G SNP genotyping. Figures 1A and B show the results for homozygous dogs; a lower Ct is obtained with the specific primer set and a higher Ct with the non-specific primer set. Figure 1C shows the results for a heterozygous dog; both assays give approximately equal Ct results.  (Figure 2A and B) and heterozygous ( Figure 2C) dogs. Capillary electrophoresis of PCR products permits differentiation of the products according to their size.