(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



An ancient haplotype containing antimicrobial peptide gene variants is associated with severe fungal skin disease in Persian cats

['Alexandra N. Myers', 'Department Of Veterinary Pathobiology', 'College Of Veterinary Medicine', 'Biomedical Sciences', 'Texas A M University', 'College Station', 'Texas', 'Unites States Of America', 'Sara D. Lawhon', 'Alison B. Diesel']

Date: 2022-04

Dermatophytosis, also known as ringworm, is a contagious fungal skin disease affecting humans and animals worldwide. Persian cats exhibit severe forms of the disease more commonly than other breeds of cat, including other long-haired breeds. Certain types of severe dermatophytosis in humans are reportedly caused by monogenic inborn errors of immunity. The goal of this study was to identify genetic variants in Persian cats contributing to the phenotype of severe dermatophytosis. Whole-genome sequencing of case and control Persian cats followed by a genome-wide association study identified a highly divergent, disease-associated haplotype on chromosome F1 containing the S100 family of genes. S100 calcium binding protein A9 (S100A9), which encodes a subunit of the antimicrobial heterodimer known as calprotectin, contained 13 nonsynonymous variants between cases and controls. Evolutionary analysis of S100A9 haplotypes comparing cases, controls, and wild felids suggested the divergent disease-associated haplotype was likely introgressed into the domestic cat lineage and maintained via balancing selection. We demonstrated marked upregulation of calprotectin expression in the feline epidermis during dermatophytosis, suggesting involvement in disease pathogenesis. Given this divergent allele has been maintained in domestic cat and wildcat populations, this haplotype may have beneficial effects against other pathogens. The pathogen specificity of this altered protein should be investigated before attempting to reduce the allele frequency in the Persian cat breed. Further work is needed to clarify if severe Persian dermatophytosis is a monogenic disease or if hidden disease-susceptibility loci remain to be discovered. Consideration should be given to engineering antimicrobial peptides such as calprotectin for topical treatment of dermatophytosis in humans and animals.

Fungal skin infections known as ringworm or dermatophytosis affect billions of humans and animals worldwide. Normally the disease is self-limiting in affected individuals. The Persian cat breed is a popular breed known for its long hair coat and short nose as well as its propensity to develop severe, chronic dermatophytosis. By examining the genomes of Persian cats, we discovered that a specific region of DNA is highly altered between cats with and without severe dermatophytosis. The DNA sequence in this region is particularly divergent within a cluster of genes involved in immune defense against pathogens. Notably, alterations to the DNA sequence cause several changes in the antimicrobial protein known as calprotectin, which defends against pathogens in the skin of cats. Persian cats with severe dermatophytosis have a version of calprotectin similar to a version maintained by certain desert-dwelling wild felids such as sand cats and Asiatic wildcats. Therefore, we think this version of the protein is beneficial in some environments or against certain pathogens but not against the fungus that causes ringworm in cats. Our findings suggest changes to calprotectin may affect pathogen specificity and engineered calprotectin could be considered as a novel therapy for dermatophytosis in humans and animals.

Persian cats are more likely than other breeds to develop dermatophytosis caused by M. canis, including chronic, extensive, and/or deep infections ( Fig 1 ) [ 9 , 17 ]. A unique form of deep dermal to subcutaneous dermatophytosis, termed dermatophytic pseudomycetoma, is seen almost exclusively in Persian cats. A genetic component to either susceptibility or severity of disease has long been suspected in this breed, supported by observations that particular Persian catteries housing genetically related cats have significantly higher prevalence of chronic disease than others [ 18 ]. A recessive mode of inheritance is suspected given severe, deep infections are relatively isolated to the Persian breed despite the Persian being used in the development of many other breeds of cat. Elucidating the genetic underpinnings of severe dermatophytosis in the Persian cat is likely to translate into better understanding of the pathogenesis and treatment of dermatophytosis in other species, including humans. The objective of this study was to identify genetic variants associated with severe dermatophytosis in Persian cats using whole-genome sequencing (WGS). We further investigated the expression and evolution of a key, skin-expressed antimicrobial peptide gene found within a divergent and disease-associated haplotype.

Most dermatophyte infections are mild; however, severe and even fatal cases are occasionally reported [ 10 , 11 ]. In severe cases, the infection may become chronic, cover a large area of the body (extensive infection), and/or penetrate beyond the epidermis into the dermis and other tissues (deep infection). In humans, monogenic inborn errors of immunity can cause severe dermatophytosis, including autosomal recessive mutations in caspase recruitment domain family member 9 (CARD9) and autosomal dominant mutations in signal transducer and activator of transcription 1 (STAT1) [ 11 – 14 ]. Reports of genetic variation contributing to disease susceptibility rather than severity also exist, although full functional validation of these variants is lacking [ 15 , 16 ].

Dermatophytes rely upon keratin as a source of nutrition. Ancestral dermatophytes gleaned this keratin primarily from soil, but recent adaptive radiation of this lineage has resulted in fungi that are adapted to keratins of specific host species [ 5 – 7 ]. Anthropophilic dermatophytes, such as those causing diseases commonly known as “athlete’s foot” and “jock itch”, are adapted to human keratins, while zoophilic dermatophytes are adapted to specific animal keratins [ 8 ]. Microsporum canis prefers the dog and cat as its primary host species but is commonly transmitted to humans [ 9 ].

Superficial fungal infections of the hair, skin, and nails affect an estimated one billion people globally and are the most common type of fungal infection in the world. These infections are most often caused by fungi known as dermatophytes, and the infection itself may be referred to as dermatophytosis, tinea, or a more commonly known name—ringworm [ 1 ]. While the disease is generally mild and self-limiting, its impact on human and animal welfare can be enormous [ 2 ]. A large, ongoing epidemic of treatment-resistant dermatophytosis in India has called attention to the serious economic and welfare implications of this disease and stimulated more research on pathogenesis and treatment [ 2 – 4 ].

S100A9 and S100A8 dimerize in vivo to form the antimicrobial peptide known as calprotectin. Zero of five non-lesional domestic shorthair cats exhibited epidermal immunolabeling for S100A8/S100A9 (calprotectin) while ten of ten cats (5 domestic shorthair cats and 5 Persian cats) with dermatophytosis exhibited moderate to strong, multifocal to diffuse epidermal immunolabeling ( Fig 5 ). These findings demonstrate a clear association between dermatophyte infection and increased expression of calprotectin in feline keratinocytes. This provides indirect evidence that calprotectin plays a role in defense against dermatophytes at the skin surface. We were unable to obtain sufficient concentrations of DNA from these FFPE skin samples to perform S100A9 genotyping.

Alignment of raw reads from a case and control cat to a new, single-haplotype Felis catus genome assembly (BioProject ID PRJNA670214) revealed no differences in alignment or haplotypes from the standard felCat9 assembly ( S5 Fig ). Further, CNVnator analysis of chromosome F1 did not reveal gene duplications within the disease-associated locus. These findings indicate that genome assembly errors and copy number variation did not interfere with read mapping at this locus or lead to an artifactual increase in non-synonymous variation. Additionally, many cats are homozygous for the case or control haplotype, providing evidence that paralogous genes are not leading to mis-mapping of reads at this locus as this phenomenon should result in heterozygous sites.

To explore the origin of the divergent alleles, a maximum-likelihood phylogenetic tree was inferred from the phased S100A9 nucleotide sequences of 22 wild felids (publicly available sequence data) and six domestic cat S100A9 haplotypes ( Fig 4B ) . The common Persian control haplotype (domestic cat H2) was found in its expected phylogenetic position with haplotypes from other species of the domestic cat lineage [ 22 , 23 ]. However, the Persian case haplotype (domestic cat H1) and the Felis margarita (sand cat) haplotypes formed a unique and highly divergent cluster from other Felis haplotypes as well as all other Felidae lineages. An approximately unbiased (AU) test confirmed this placement deviates significantly from the Felidae species tree topology (p-value = 6.19x10 -7 ). While the domestic cat H1 S100A9 nucleotide sequence was only 97.91% identical to the domestic cat H2 haplotype, it shared 99.27% identity with the sand cat sequence. The amino acid sequence of the domestic cat H1 haplotype shared 90.4% identity with the domestic cat H2 haplotype but 95.6% identity with the sand cat sequence. Overall, the Persian case haplotype is more diverged from the Persian control haplotype than from the sand cat haplotype [ 22 ]. Elsewhere in the phylogeny, there is little deviation from the species tree as would be expected with incomplete lineage sorting; therefore, ancient hybridization and introgression of this allele into ancestors of the sand cat and domestic cat populations is a more plausible explanation.

(A) Amino acid sequence alignment and secondary structure depiction of the divergent Persian cat case and control S100A9 haplotypes, created with ESPript 3.0. Conserved residues are boxed in red while similar residues are in bold text and boxed in yellow. (B) Maximum-likelihood phylogenetic tree inferred from the phased S100A9 nucleotide sequences of 22 wild felids and the domestic cat haplotypes. Bootstrap support values are given on nodes. The domestic cat lineage is split across the tree: the Persian cat case haplotype (Domestic cat H1) clusters with the sand cat near the root of the tree while the control haplotype (Domestic cat H2) is found in the expected location for the domestic cat lineage.

(A) Maximum-likelihood phylogenetic tree inferred from the amino acid sequence of six distinct domestic cat S100A9 haplotypes with the lion as the outgroup. Bootstrap support values are given on nodes. (B) Overall distribution of S100A9 haplotypes in 161 domestic cats of varied breed and background. Haplotypes that cluster together with the Persian case haplotype (H1) are shaded in oranges while haplotypes that cluster with the Persian control haplotype (H2) are shaded in blues. (C) Relative frequency of haplotypes in various cat breeds. Labels for ‘Persian_Controls’ and ‘Persian_Cases’ represent the haplotype frequencies for the Persian cats enrolled in this study while the label for ‘Persian’ represents Persian cats in the general cat population. Haplotypes H3 and H4 predominate in cat breeds previously determined to be of Eastern origin. Breeds were included only if represented by at least 3 individuals (6 haplotypes).

Of all genes in the disease-associated locus, S100A9 (F1:71,653,735–71,656,597) exhibited the highest sequence divergence between the common Persian case haplotype (domestic cat H1) and the common Persian control haplotype (domestic cat H2). Given this high level of sequence divergence, numerous amino acid alterations, its well-described role in antifungal immunity, and its reported epidermal expression in other species, S100A9 was considered the most compelling candidate gene [ 19 – 21 ]. We therefore chose to focus on S100A9 for further analysis. Upon examination of publicly available variant data from domestic cats of various breeds and backgrounds, six distinct S100A9 haplotypes were identified in the domestic cat population, including domestic cat H1 and H2 as well as domestic cat H3, H4, H5, and H6 ( Fig 3A–3C ). Of note, the domestic cat H1 haplotype, which is homozygous in most Persian cases, had an allele frequency of 13% in the general domestic cat population, and 6.8% of this population are homozygous H1/H1. Both the domestic cat H1 and H2 haplotypes were also identified in a wild-caught Asiatic wildcat (Felis silvestris ornata) from Tajikistan. The domestic cat H1 and H2 S100A9 haplotypes are highly divergent for two populations of the same species, sharing 97.91% nucleotide sequence identity across 2,864 base pairs and 90.4% identity across 135 amino acids. This protein, though relatively small, exhibited 13 amino acid substitutions between the common case (domestic cat H1) and control (domestic cat H2) haplotypes ( Figs 4A and S4 ).

Weir and Cockerham’s Fst and Tajima’s D were higher across the 1 Mb disease-associated locus than anywhere else on chromosome F1, indicating sequence divergence between cases and controls and an abundance of intermediate frequency alleles, respectively ( Fig 2C ). The beta statistic (β), a measure of long-term balancing selection, is also higher within the disease-associated locus than anywhere else on the chromosome ( Fig 2C ), but unlike Tajima’s D and Fst, the elevation is confined to a narrow region within the locus: the S100 antimicrobial peptide gene region ( Fig 2D ) . Together these findings suggest balancing selection is acting to maintain divergent haplotypes in this region.

(A) Manhattan plot depicting single locus linear mixed model output for 10 Persian cat cases of severe dermatophytosis and 16 Persian cat controls. A locus on chromosome F1 achieves genome wide significance of p-value = 1.51x10 -10 . The Bonferroni-corrected significance threshold of p-value = 3.30 x 10 −9 is shown in red. An additional Manhattan plot focused on the disease-associated locus can be found in S2 Fig . (B) Table depicting the genotypes of cases and controls for the disease-associated locus on F1. Two heterozygous controls are missing from the table as they were heterozygous for additional haplotypes (H3/H4 and H2/H6). (C) A Manhattan plot depicting the entirety of chromosome F1 is shown along with Tajima’s D, Weir and Cockerham’s Fst, and the β statistic for detection of long-term balancing selection. (D) A regional Manhattan plot depicting the last ~100 kb of chromosome F1 is shown along with β and the gene annotation. Within the disease associated locus, β is most elevated over the S100A antimicrobial peptide genes, S100A9 (A9) and S100A15 (A15).

All Persian cat whole genome sequencing data was uploaded to the NCBI Short Read Archive under BioProject PRJNA704916. Information on each Persian cat sampled for this study is listed in S1 Table . A genome-wide association study (GWAS) performed on variants called from the whole-genome sequencing data revealed a single peak of markers surpassing the Bonferroni correction threshold at the distal end of the acrocentric chromosome F1, with the most significant markers having a p-value of 1.51x10 -10 ( Fig 2A ). A list of all significant markers and associated p-values can be found in S2 Table , and the GWAS quantile-quantile (QQ) plot and genomic inflation factor can be found in S1 Fig . Upon closer inspection, this peak marked the location of a large, ~1 Mb locus (chrF1:70,630,048–71,656,506) where eight of ten Persian cases (including the two cats with concurrent pseudomycetoma) were homozygous for a shared haplotype, henceforth referred to as domestic cat H1 ( Fig 2B ) . None of the 16 Persian control cats were homozygous for domestic cat H1, though 6 controls were heterozygous for this haplotype. This peak remained the only significant association whether the GWAS was conducted using all control cats (including those with unknown exposure to dermatophytes) or only the eight control cats with known previous exposure or mild/transient disease ( S3 Fig ). The ~1 Mb disease-associated locus contained 42 genes, eight of which contain disease-associated non-synonymous single nucleotide variants (SNVs) ( Tables 1 , S3 and S4 ). Half of these genes are well-characterized immune genes, including a cytokine receptor gene and antimicrobial peptide genes: interleukin 6 receptor (IL6R), S100 calcium binding protein A15 (S100A15), S100A12, and S100A9. The most significant non-synonymous SNV (p-value 1.51x10 -10 ) is a missense mutation within S100A9 predicted to cause a glycine to glutamic acid substitution: XP_003999832.3:p.Gly49Glu ( S3 Table ). Within the disease-associated locus, structural variant (SV) callers identified two deletions within the domestic cat H1 haplotype relative to the domestic cat H2 haplotype: a 1,321 bp intergenic deletion between S100A6 and S100A15 (F1:71,605,113) and a 255 bp intergenic SINE deletion between S100A8 and S100A9 (F1:71,648981). No SVs were found within coding regions or at the boundary of the disease-associated locus.

Discussion

Using whole-genome sequencing, we identified a highly divergent haplotype containing antimicrobial peptide genes that was significantly associated with severe dermatophytosis in Persian cats. In particular, the S100A9 gene encoding a subunit of the antimicrobial peptide (AMP) known as calprotectin included numerous amino acid substitutions between case (domestic cat H1) and control (domestic cat H2) haplotypes. We demonstrate that calprotectin (S100A8/A9) is likely to be an important player in anti-dermatophyte defense at the skin surface where this protein is highly expressed during infection.

Other genes within the disease-associated locus contained nonsynonymous mutations, but there are several reasons we chose to focus on S100A9 as the most compelling candidate gene. First, the level of divergence between domestic cat H1 and H2 within S100A9 was far higher than that of other genes with alteration of nearly 10% of amino acids between the haplotypes (13/135 amino acids). This striking level of divergence suggests some impact on function. Second, calprotectin has known antifungal activity and epidermal localization [19,24–26]. Other genes with multiple non-synonymous mutations such as S100A12 and NUP210L are not reported to be skin-expressed; in humans, expression of NUP210L is reported to be almost exclusively testis-specific, making it an unlikely candidate gene [27]. As dermatophytes cause infection by adhering to and penetrating hair and epithelial cells at the skin surface, candidate genes for severe dermatophytosis are likely to be expressed in skin and/or hair. Certainly, mutations in other immune genes such as S100A15 or IL6R also have the potential to play a role in severity and chronicity of infection and modulation of the immune response to various pathogens, as do non-coding mutations that may be impacting gene regulatory elements. The expression and regulation of multiple genes within this disease-associated locus are of interest and bear further investigation in the future.

AMPs such as calprotectin are an important component of the immune defenses of multicellular eukaryotes [28]. They are a highly diverse group of small proteins with activity against bacteria, fungi, and viruses [28]. Interestingly, AMPs were previously thought to have broad-spectrum activity against various pathogens; however, recent studies suggest a higher degree of specificity for particular pathogens and show that single amino acid substitutions can alter specificities [28]. The AMP calprotectin has been shown to exhibit activity against bacteria and fungi such as Candida sp. and Aspergillus sp., but studies evaluating its efficacy against dermatophytes are lacking [19,24–26]. Calprotectin exerts antimicrobial effects by chelating transition metals such as zinc and manganese needed for microbial growth—this is termed nutritional immunity [29]. Recently, however, Besold et al. demonstrated that calprotectin also inhibits bacterial growth through direct physical interaction that does not involve metal withholding [30]. The full mechanism for calprotectin’s antimicrobial action is still being elucidated.

Although the Persian cat case allele of calprotectin appears to render these cats susceptible to severe Microsporum canis infections, we suspect that this allele would be beneficial against other pathogens, given its relatively high frequency in the domestic cat population. AMPs of many species have been described to evolve under pathogen-driven balancing selection resulting from an evolutionary arms race between pathogens and the host immune proteins they directly interact with [31–34]. Maintenance of divergent alleles within a population can serve as a defense mechanism for a population by providing a heterozygote advantage and/or a reservoir of alleles that may be useful in specific conditions/environments or against particular pathogens [31]. Interestingly, high levels of non-synonymous variation have been previously identified within S100A antimicrobial peptide genes of taurine cattle (Bos taurus) and yak (Bos grunniens), suggesting a similar scenario of divergent haplotypes as observed in cats [35].

Previous studies evaluating signatures of selection in the Persian cat breed did not find evidence of selection on chromosome F1; however, these studies focused on selection signatures specific to the Persian breed, not other breeds or species [36,37]. We identified the divergent S100A9 case and control alleles not only in various domestic cat breeds but also in an Asiatic wildcat (Felis silvestris ornata) from Tajikistan. Further, the case allele is similar to the sand cat (Felis margarita) allele, and together these alleles cluster much closer to the root of Felidae than expected. These findings indicate either introgression of the case allele into Felis species after an ancient hybridization event, long-term balancing selection, or a combination of the two (selection acting upon introgressed variation). We speculate that the haplotypes clustering in the clade with the sand cat and Persian case haplotypes may have been adaptive in the arid desert environments inhabited by the sand cat and many of the Felis silvestris subspecies thought to be the predecessors of modern domestic cats [38]. Arid regions harbor a unique repertoire of pathogens, and studies have shown dermatophytes are more prevalent in humid environments rather than arid conditions [39–41]. In modern long-haired Persian cats not confined to the desert, the divergent S100A9 allele may be maladaptive. Further tests to assess effectiveness of the divergent calprotectin protein against other common feline pathogens should be undertaken to clarify any potential benefit of this allele.

Other explanations exist for why the case allele (domestic cat H1) is not as rare in the domestic cat population as might be expected (allele frequency of 13%) and why it is seen in breeds without a known predisposition to dermatophytosis. One explanation is that Persian cats possess a hidden disease susceptibility locus (or loci) not found in other breeds that we simply don’t have the power to detect with this study design. Alternately, it is possible that there are cases of severe dermatophytosis occurring in these other breeds that are not as well-recognized or reported. Clinical history regarding potential dermatophytosis is not readily available for the domestic cats in the publicly available dataset used here—some of the cats homozygous for H1/H1 in this dataset may have experienced severe disease. Additional work is suggested to characterize the frequency and severity of dermatophytosis in other cat breeds that carry domestic cat H1 or those haplotypes that cluster closely with it.

Regarding the two Persian cases that were not homozygous H1/H1 (both were H1/H2): we were unable to confirm that these cats had received adequate treatment for dermatophytosis, and they may have lived in highly contaminated environments. Indeed, one of these cats was the only cat included in the study known to have lived in a cattery with many other infected cats. Regardless of genotype, a highly contaminated environment may lead to chronic and eventually severe disease. Ideally, such environmental variables would be avoided in future studies by adhering to strict enrollment criteria including proper environmental decontamination.

The phenotype assessed in this study was that of chronic and extensive dermatophytosis in Persian cats despite appropriate veterinary-directed antifungal treatment. We were unable to specifically assess the phenotype of dermatophytic pseudomycetoma—a particular form of deep dermatophytosis that is also reported in humans—due to limited sample size [42,43]. Only two Persian cat cases included in this study had concurrent pseudomycetoma. These cats were homozygous for the case haplotype, but a larger sample size is needed. The clinical and histologic appearance of pseudomycetoma is quite different from that of superficial dermatophytosis and primarily involves infection of deeper layers of the skin and subcutaneous tissues rather than the epidermis; therefore, the immune genes and pathways involved may be different as well. Additional genotyping for pseudomycetoma cases is needed to determine if the S100A9 case haplotype or closely-related haplotypes are associated with this specific phenotype. Overall, the small sample size of Persian cats enrolled for sequencing and GWAS was primarily due to our stringent enrollment criteria that 1) each Persian be diagnosed, sampled, and treated by a veterinarian (many owners and breeders attempt to diagnose and treat this disease on their own) and 2) that a detailed past history be known for case and control cats. However, several recent articles describe why small sample sizes can be used to successfully map disease loci in canine and feline GWAS [44,45]. In particular, dogs and cats have higher linkage disequilibrium than many other species due to inbreeding. Our findings here should still be validated and further explored in a larger cohort of cats while controlling for environmental variables as much as possible.

In summary, a divergent haplotype on chromosome F1 containing S100A antimicrobial peptide genes is associated with development of severe dermatophytosis in Persian cats. If this haplotype is not found to confer any significant benefit against other pathogens, a genetic test could be easily created to identify cats with the case haplotype and to help adjust breeding programs. These findings open the door for future investigation into AMPs as treatment options for dermatophytosis in humans and animals. Numerous AMPs are already in clinical trials for treatment of other diseases in humans [28,46,47]. Interest in engineering these peptides to increase specificity against particular pathogens is burgeoning, as AMPs are expected to have less trouble with development of antimicrobial resistance [48]. The application of certain forms of calprotectin as a topical treatment for dermatophytosis may be of value. Alternate treatments are becoming more important as the incidence of dermatophytosis continues to rise [41].

[END]

[1] Url: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1010062

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.