A Combination of Alleles in LMOD2 and a lncRNA is Strongly Associated With Myxomatous Mitral Valve Disease in Cavalier King Charles Spaniels

ABSTRACT

A previous genome‐wide association study identified regions on canine chromosome (cfa) 13 and 14 associated with early onset myxomatous mitral valve disease (MMVD) in Cavalier King Charles Spaniels (CKCS). In the present study, whole genome sequencing (WGS) of 9 CKCS cases (mitral regurgitation (MR) before 4.5 years or congestive heart failure (CHF) at any age due to MMVD) and 10 CKCS controls (no or mild MR after 8 years of age) identified > 2000 genetic variants in the MMVD associated cfa13 and cfa14 regions. Ensembl Variant Effect Predictor (VEP) identified a possible functional impact of 18 variants. These were genotyped in 250 CKCS; 117 cases and 133 controls. The most significantly associated variants were a splice‐site variant in a long noncoding RNA (lncRNA) on cfa13, a nonsynonymous variant in HYAL4, a 39 base‐pair insertion in LMOD2 and a synonymous variant in ENSCAFG00000024436 (p‐values from 2.03E‐08 to 4.20E‐06). Concomitant homozygosity for risk alleles in LMOD2 and the lncRNA gave an odds‐ratio for MMVD of 52.5 compared to homozygosity for the nonrisk alleles (p = 0.00034, 95% CI: 8.8–1023.8). Upon validation of our results in an independent cohort, this gene variant combination in CKCS is expected to enable targeted breeding programs to reduce MMVD prevalence in CKCS.

1. Introduction

Myxomatous mitral valve disease (MMVD) is the most common heart disease in dogs, and certain breeds are predisposed including Cavalier King Charles Spaniels (CKCS) (Egenvall et al. 2006). The disease shares common features with the heart valve disease primary mitral valve prolapse (MVP) in human patients (Pedersen and Häggström 2000). Similar to primary MVP, MMVD in dogs is characterized by valvular changes, including the accumulation of extracellular matrix and fragmentation of collagen fibrils, leading to MVP and/or valvular thickening, and mitral regurgitation, that is, the reverse flow of blood from the left ventricle to the left atrium during ventricular systole (Buchanan 1977; Keene et al. 2019). The disease develops with advancing age and is more frequent in males compared to females at any given age (Pankov 2023). The disease can progress to advanced stages with cardiac enlargement and congestive heart failure (CHF), which may cause premature death. It is estimated that half of the CKCS have a heart murmur at 6–7 years of age (Häggström et al. 1992) and about one fourth of the causes of euthanasia or deaths in CKCS are associated with MMVD (Reimann et al. 2017).

A genetic contribution to the etiology behind MMVD in dogs and primary MVP in human patients has been assumed for decades. In CKCS, a genome‐wide association study identified two loci on canine chromosomes (cfa) 13 and 14 associated with early onset of MMVD (Madsen et al. 2011). Additionally, a whole‐genome sequencing study in CKCS identified gene variants near the NEBL gene on chromosome 2 (Axelsson et al. 2021; Mead et al. 2022) that seemed to be very frequent in this breed predisposed to MMVD (Axelsson et al. 2021). The NEBL gene encodes the actin‐binding myocardial protein nebulette, which may impact the contractile properties of cardiomyocytes. Interestingly, an association between the variants near the NEBL gene and MMVD was found in Dachshunds, and one of the variants was associated with reduced nebulette mRNA expression in the papillary muscle of Dachshunds leading to discussion of possible functional impact (Axelsson et al. 2021).

Identification of causative mutations for MMVD will provide a molecular understanding of the pathology behind MMVD, potentially benefitting human patients with primary MVP as well. Furthermore, a characterization of the genetic background for MMVD will facilitate the development of genetic tests capable of predicting MMVD risk in different dog breeds. Such tests can be used to support breeding strategies aimed at reducing MMVD prevalence and promoting ethical breeding (Lundin and Kvart 2010; Swift et al. 2017; Birkegård et al. 2016). This topic is currently high on the political agenda, as current dog breeding strategies have faced severe criticism for neglecting animal health and welfare, especially in breeds such as CKCS with low genetic diversity. Consequently, breeding of pure bred CKCS has been banned in Norway; however, cross breeding with other breeds is allowed (Supreme Court of Norway 2023).

The present study followed up on results from Madsen et al. (2011), that identified MMVD‐associated regions on cfa 13 and 14 in CKCS. The aim was to identify genetic variants within the two regions and evaluate the association between putative causative variants and MMVD in a larger group of CKCS with and without MMVD. Furthermore, associations between previously identified variants near the NEBL gene and MMVD were investigated in the same dogs. Odds for MMVD were compared between relevant variant genotypes and combinations thereof.

2. Results

In a cohort of CKCS, 10 MMVD cases and 10 healthy controls were selected for whole genome sequencing (WGS). None of the dogs were related at parental level, however two and five dogs in the case and control group respectively were related at grandparental levels, and in addition a case and a control were related at grandparental level. After isolation of genomic DNA, one sample did not meet quality requirements. Hence, nine MMVD cases and 10 healthy controls were sequenced. Genome‐wide depth: Minimum, 40.64; Mean, 46.61; Median, 46.89; Maximum, 51.36; Range, 10.72; SD, 2.54; IQR, 2.65 (Table S6). A total of 2338 variants, 374 in the cfa13 region and 1964 variants in the cfa14 region, were identified, for which all sequenced cases were homozygous for a nonreference allele and the allele frequency among cases were at least 1.5‐fold higher than the allele frequency in the sequenced controls. All analyses allowed for missing data. An analysis using the Ensembl Variant Effect Predictor revealed 19 variants, 13 in the cfa13 regions (in the following named MMVD1‐13) and 6 variants in the cfa14 region (MMVD14‐19) as possible candidate variants (Table 1). This included exon and splice‐site variants in coding and noncoding genes. All these variants plus four variants close to the NEBL gene (Table S4), previously associated with MMVD (Axelsson et al. 2021; Mead et al. 2022), were genotyped in 250 (199 Danish and 51 Swedish) CKCS of which 117 were MMVD cases (46 females, 71 males; 64 with presence of MR at early age and 53 with CHF at any age due to MMVD) and 133 were healthy controls (86 females, 47 males) (Tables S1 and S2). MMVD7 genotypes were excluded from further analyses due to low quality, hence the reported results are based on successful genotyping of 18 variants. Of the 250 genotyped dogs, 145 were also part of the original discovery cohort for the genome‐wide association study (GWAS) study (Madsen et al. 2011) (Table S1). Two variants on cfa14, MMVD15 and MMVD16 obtained genome wide significance (p = 3.66E‐08 and p = 2.03E‐08, respectively, Table 2). The neighboring variant MMVD14 was borderline genome wide significant (p = 7.28E‐08). Linkage disequilibrium (LD) between the three markers were close to 1 (Figure S1). The most significantly associated variants on cfa13 were MMVD6 (p = 3.11E‐06) and MMVD10 (p = 4.20E‐06). None of the previously identified MMVD associated variants close to NEBL (Axelsson et al. 2021), were statistically associated with MMVD status in the present study (Table 3). As also seen in a previous CKCS study (Mead et al. 2022) we found that the frequencies of the reference alleles of the variants near the NEBL gene were low (between 0.04 and 0.07) (Table 3).

TABLE 1.

Position of the 19 genetic markers genotyped on chromosome 13 and 14.

Genetic variant	Position b	Variation reference/alternative	DP c	Genome annotation
MMVD1	Chr13:62537032	A/T	1037	ADAMTS3, NMD a transcript variant
MMVD2	Chr13:62539884	A/G	1184	ADAMTS3, NMD transcript variant
MMVD3	Chr13:62543514	A/C	862	ADAMTS3, NMD transcript variant
MMVD4	Chr13:62545619	G/T	991	ADAMTS3, NMD transcript variant
MMVD5	Chr13:62546218	T/C	1070	ADAMTS3, NMD transcript variant
MMVD6	Chr13:62649778	T/C	729	ENSCAFG00000035673, lncRNA, exon variant
MMVD7	Chr13:62667397	G/A	1241	ENSCAFG00000035673, lncRNA, exon variant d
MMVD8	Chr13:63107245	C/A	704	ENSCAFG00000042273, lncRNA, exon variant
MMVD9	Chr13:63107271	T/C	707	ENSCAFG00000042273, lncRNA, exon variant
MMVD10	Chr13:62740273	A/G	na	ENSCAFG00000045150, lncRNA, splice acceptor variant
MMVD11	Chr13:62740333	A/G	668	ENSCAFG00000045150, lncRNA, exon variant
MMVD12	Chr13:62740357	C/G	689	ENSCAFG00000045150, lncRNA, exon variant
MMVD13	Chr13:62768974	A/G	905	ENSCAFG00000045150, lncRNA, exon variant
MMVD14	Chr14:11689329	A/G	1005	Nonsynonymous SNP in HYAL4
MMVD15	Chr14:11714240	C/T	964	Synonymous SNP in ENSCAFG00000024436, hyaluronidase‐1‐like
MMVD16	Chr14:11870337	T/TTCTCCAGGAGTTTTTGGGACTCCTTTTCCCAATAGGCCA	na	A 39 bp insertion in exon 1 of LMOD2
MMVD17	Chr14:12080417	C/T	1029	Nonsynonymous SNP in ENSCAFG00000001783, vacuolar sorting associated 52
MMVD18	Chr14:12080023	C/T	1048	Synonymous SNP in ENSCAFG00000001783, vacuolar sorting associated 52
MMVD19	Chr14:11935743	T/C	1054	Nonsynonymous SNP in ASB15

Abbreviation: na, no annotation.

^a NMD variant: A variant annotated as associated with nonsense mediated decay.

^b Position refers to the CanFam4 genome assembly.

^c DP SNP number according to Wang et al. (2020) in the UCSC genome browser, CanFam4.

^d MMVD7 was not successfully genotyped in the 250 dogs.

TABLE 2.

Genotype counts and frequencies and associations with myxomatous mitral valve disease (MMVD) of 18 identified gene variants (MMVD1‐6 and 8–19) in chromosome 13 and 14 in 250 Cavalier King Charles Spaniels (CKCS).

Genotypes	Controls (n = 133)			Cases (n = 117)			Missing genotypes	p	LD‐block
	AA	AB	BB	AA	AB	BB
MMVD1	46 35%	58 45%	26 20%	13 11%	59 52%	42 37%	6	1.66E‐05	1
MMVD2	51 39%	54 41%	26 20%	17 15%	57 50%	39 35%	6	7.05E‐05	1
MMVD3	40 32%	24 19%	61 49%	12 10%	26 23%	76 67%	11	0.000223	na
MMVD4	45 34%	62 47%	26 19%	13 11%	61 52%	43 37%	0	2.23E‐05	1
MMVD5	45 34%	61 46%	27 20%	12 10%	61 53%	43 37%	1	1.20E‐05	1
MMVD6	46 35%	58 45%	26 20%	12 10%	60 51%	45 39%	3	3.11E‐06	1
MMVD8	31 23%	65 49%	37 28%	10 8%	50 43%	57 49%	0	0.000297	2
MMVD9	31 23%	65 49%	37 28%	10 8%	50 43%	57 49%	0	0.000297	2
MMVD10	46 35%	60 45%	26 20%	12 10%	61 52%	44 38%	1	4.20E‐06	1
MMVD11	45 34%	61 47%	25 19%	12 11%	63 55%	39 34%	5	1.78E‐05	1
MMVD12	45 34%	61 46%	27 20%	12 10%	61 52%	44 38%	0	9.94E‐06	1
MMVD13	11 8%	45 34%	77 58%	4 3%	34 29%	79 68%	0	0.1496	2
MMVD14	30 23%	55 43%	43 34%	5 4%	35 30%	77 66%	5	7.28E‐08	3
MMVD15	31 23%	58 44%	44 33%	5 4%	35 30%	77 66%	0	3.66E‐08	3
MMVD16	31 23%	58 44%	43 33%	5 4%	34 30%	77 66%	2	2.03E‐08	3
MMVD17	15 11%	47 36%	70 53%	1 1%	24 20%	92 79%	1	7.26E‐06	4
MMVD18	15 11%	47 36%	69 53%	1 1%	24 20%	92 79%	2	5.95E‐06	4
MMVD19	22 17%	51 38%	59 45%	4 3%	26 23%	86 74%	2	3.11E‐06	na

Note: Genotypes are presented as a genotype count and frequency as a percentage of all genotypes for each marker. Associations between MMVD status and genotypes were tested using Fisher's exact test. Bold indicates genetic markers identified by LASSO regression as the most significant predictors for MMVD. The LD block numbers refer to Figure S1.

TABLE 3.

Genotype counts and frequencies and associations with myxomatous mitral valve disease (MMVD) of four gene variants (NEBL 1–4) near the NEBL gene (Axelsson et al. 2021) in 250 Cavalier King Charles Spaniels (CKCS).

Genotypes	Controls (n = 133)			Cases (n = 117)			Missing genotypes	p
	AA	AB	BB	AA	AB	BB
NEBL1	114 91%	11 9%	0 0%	105 92%	9 8%	0 0%	11	0.82
NEBL2	2 2%	15 12%	104 86%	1 1%	10 9%	102 90%	16	0.58
NEBL3	0 0%	9 8%	107 92%	0 0%	9 8%	101 92%	24	1
NEBL4	0 0%	18 15%	106 85%	0 0%	13 11%	101 89%	12	0.56

Note: Genotypes are presented as a genotype count and frequency as a percentage of all genotypes for each marker. Associations between MMVD status and NEBL genotypes tested using Fisher's exact test.

Percentage of genotype similarity between the different variants was calculated. Nine relatively noncorrelated variants (< 90% genotype similarity), MMVD1, MMVD8, MMVD10, MMVD13, MMVD16, MMVD17, MMVD19, NEBL1, and NEBL2, were identified for least absolute shrinkage and selection operator (LASSO) regression analysis (Tibshirani 2018) aimed at selecting a minimal set of variants that could explain the phenotype. The LASSO analysis identified MMVD10 and MMVD16 as most significantly associated with MMVD status (Table 4). It should be noted that several markers were in strong LD with MMVD10 and that MMVD14, MMVD15, and MMVD16 formed an LD block with LD close to 1 (Table 2 and Figure S1). Although the data did not support a statistical test for interaction between MMVD10 and MMVD16, very high odds ratios (OR) were observed for specific allele combinations in the two loci. That is, a test analyzing MMVD16 genotypes alone yielded an OR of 9.31 (3.4–30.2) (95% confidence interval) for MMVD status when comparing homozygosity for the risk associated allele to homozygosity for the wild‐type allele. However, the OR for MMVD increased to 48 (8.5–912.5) or 52 (8.8–1023.8) when the MMVD10 locus simultaneously carried one or two risk alleles, respectively (Table 4). Gene variants near the NEBL gene were not associated with MMVD status in the multivariate logistic regression analysis, neither on their own or in combination with MMVD10 and/or MMVD16 genotypes.

TABLE 4.

Estimated odds ratios (ORs) for myxomatous mitral valve disease (MMVD) based on genotypes of the gene variants MMVD16 and MMVD10 calculated using multivariate logistic regression analysis.

Case vs control	Estimated OR	95% confidence interval	p
Main effect model on MMVD10 alone (compared to AA)
Sex (male)	3.13	1.76–5.68	0.0001
MMVD10 (AB)	3.17	1.46–7.27	0.0047
MMVD10 (BB)	5.00	2.12–12.39	0.0003
Main effect model on MMVD16 alone (compared to AA)
LMOD2 (MMVD16) (AB)	2.82	1.01–9.26	0.063
LMOD2 (MMVD16) (BB)	9.31	3.42–30.26	0.0000
Allele combinations (compared to MMVD10 (BB) + MMVD16 (BB))
Sex (male)	3.12		0.0002
MMVD10 (AA) + MMVD16 (BB)	6.32	1.93–22.98	0.00324
MMVD10 (AB) + MMVD16 (BB)	1.09	0.44–2.81	0.848
MMVD10 (BB) + MMVD16 (AA)	8.63	1.06–185.30	0.0727
MMVD10 (AA) + MMVD16 (AA)	52.54	8.75–1023.85	0.0003
MMVD10 (AB) + MMVD16 (AA)	10.86	2.60–58.38	0.0021
MMVD10 (BB) + MMVD16 (AB)	2.37	0.79–7.23	0.1242
MMVD10 (AA) + MMVD16 (AB)	7.76	2.30–30.07	0.0016
MMVD10 (AB) + MMVD16 (AB)	6.07	2.37–16.57	0.0003

Note: Bold, statistically significant, p < 0.05.

The nonreference variant of MMVD16 was not found in tests of 1168 dogs representing 101 different breeds or breed varieties other than CKCS (Table S3). Furthermore, searching the publicly available WGS dataset with full genome data from 582 dogs and 8 wolves (Jagannathan et al. 2019), MMVD10 and all other variants except MMVD16 were found to be present in several different breeds (Table S5). The MMVD16 variant was only present in CKCS and in a heterozygous state in two Lagotto Romagnolo dogs.

3. Discussion

We identified two variants, MMVD10 and MMVD16, in regions on cfa 13 and 14 previously associated with MMVD. Allelic combinations in these two variants were strongly associated with early onset MMVD in a cohort of Swedish and Danish CKCS. The same variants were identified in publicly available genome data from CKCS in other countries, showing that the variants are not unique for a specific population. It should be noted that MMVD10 was in strong LD with several markers on cfa13 and that MMVD16 was in strong LD with MMVD14 and MMVD15. Any of the variants in strong LD with MMVD10 and MMVD16 can be considered as possible contributors to the etiology behind MMVD.

MMVD10 is a mutation in a splice‐acceptor site in a predicted lncRNA ENSCAFG00000045150 or LOC111098638 (CanFam4, chr13: 62 740 273_A>G). The specific function of this noncoding RNA remains unclear, but generally lncRNAs are recognized for their regulatory functions through interaction with DNA, mRNAs, miRNAs, or proteins. Interestingly, despite their name, some lncRNAs are translated to micropeptides (Nemeth et al. 2024). Recently lncRNAs have been identified as novel regulators in cardiac development and cardiac pathogenesis. For example, lncRNA H19 has been shown to have a regulatory role in cardiac diseases (Wang et al. 2022). The potential influence of the identified MMVD10 variant on MMVD progression, particularly in combination with the MMVD16 locus, remains to be investigated. MMVD10 is in strong LD with variants in two other lncRNA genes (MMVD11 and MMVD12) as well as variants in ADAMTS3 (MMVD1‐6) (Table 1, Figure S1). However, the present analysis remains inconclusive regarding which of these variants may exert functional influence on MMVD development.

MMVD16 is a 39 bp insertion in the LMOD2 gene, which encodes the leiomodin‐2 protein. As reviewed by Henderson et al. 2017 (Henderson et al. 2017), this protein is a member of the tropomodulin family, which contains tropomyosin‐ and actin‐binding domains and plays a crucial role in the elongation of thin filaments in cardiac muscle cells together with nebulin (Henderson et al. 2017; Kiss et al. 2020; Pappas et al. 2015). The length of the thin filaments is important for proper contractile function. LMOD2 is expressed in cardiac muscle and its expression level is positively correlated with the length of the thin filaments (Tsukada et al. 2010).

Our finding of an association between the LMOD2 gene variant and MMVD in CKCS is noteworthy because increased left ventricular systolic dimension and lower fractional shortening predicted survival in a large‐scale study of young CKCS (< 3 years) (Reimann et al. 2017). The study discussed whether CKCS predisposed to MMVD could have myocardial systolic dysfunction at a very early age. As mentioned previously, MMVD in Dachshunds has been associated with gene variants near the NEBL gene encoding the myocardial protein nebulette, and the gene variants were associated with reduced mRNA expression of the nebulette in papillary muscle (Axelsson et al. 2021), which is a part of the mitral valve apparatus.

The 39 bp insertion in LMOD2 is a duplication that leads to a repetition of 13 amino acids, XP_038412591.1:p.Met71‐Glu83dup. The duplicated region coincides with a region of the leiomodin‐2 protein sequence, the ABS1h domain (Pro60‐Leu86), which is essential for its interaction with actin (Boczkowska et al. 2017; Tolkatchev et al. 2020) (Figure 1). We hypothesize that the insertion may affect the length of the thin filaments and hence the contractile function of the heart muscle, leading to disproportionate mechanical stress on the mitral valve apparatus, which could be a primary cause of MMVD (Oyama et al. 2020). Based on the function of the LMOD2 gene, it is possible that MMVD may start as myocardial disease, and the valvular myxomatous changes are secondary changes. Further studies are needed to confirm or reject this hypothesis. In the light of this, it is interesting that other variants in LMOD2 have been associated with severe cardiomyopathies in human patients that often resulted in neonatal death (Ahrens‐Nicklas et al. 2019; Greenway et al. 2021; Sono et al. 2023).

FIGURE 1.

Predicted 3D protein structure of the leiomodin‐2 protein coded by the LMOD2 wild‐type (A) and mutated (B) gene using the public available AI system AlphaFold (www.alphafold.ebi.ac.uk). The identified insertion in the LMOD2 gene (the MMVD16 variant) resulted in a longer alpha‐helix. The helix consists of amino acid 66–85 in the wild‐type but amino acid 66–98 in the mutated version. The helix is indicated with a square in the figures.

The LMOD2 variant is in strong LD with MMVD14 and MMVD15. MMVD14 is a c13T>C nonsynonymous mutation (XM_038556640.1:c13T>C) resulting in a p.Ser5Pro substitution (XP_038412568.1:pSer5Pro) in the HYAL4 encoded hyaluronidase‐4 protein. Hyaluronidase‐4 is an enzyme that degrades glycosaminoglycans, which are known to accumulate in the extracellular matrix during development of MMVD (Grande‐Allen et al. 2003). The mutation was predicted by VEP to have a moderate impact, and the alteration does not affect the functional site, suggesting that the enzymatic capacity should be the same for both versions of the protein. Comparative analysis of HYAL4 amino acid sequences from different species showed that Serine is not conserved, and Proline is found in several species at this position in the protein (data not shown). Moreover, the MMVD14 variant is identified in several dog breeds where a high prevalence of early onset MMVD is not commonly observed (Egenvall et al. 2006). Further studies are needed to elucidate a possible role of this variant for the pathological changes in the mitral valves in dogs.

MMVD15 is a synonymous mutation in a hyaluronidase‐like gene (ENSCAFG00000024436). Compared with MMVD14 and MMVD16, the possible functional impact of this synonymous variation is probably less severe, and we do not consider this variant a strong candidate mutation for MMVD risk.

National CKCS MMVD breeding programs based on cardiac auscultation, echocardiography and/or parental cardiac status are ongoing (Lundin and Kvart 2010; Swift et al. 2017; Birkegård et al. 2016), and an evaluation of a mandatory program showed a significant decrease in the risk (73%) of having a heart murmur due to MMVD over an 8–10 years period of breeding (Birkegård et al. 2016). If our results can be validated in an independent cohort of CKCS, the two identified gene variants may provide a means of early detection of dogs at high risk of early onset of MMVD. Such a test could support more ethical breeding. Moreover, we want to emphasize that breeding decisions in dogs must be balanced and include consideration of effective population size and the risk for inbreeding, especially when strong selection pressure is put on a single trait. Hence, in selection programs designed to decrease MMVD prevalence, high specificity may be preferable compared to high sensitivity.

The study has limitations. Some of the cases were defined as dogs with mild MR before 4.5, but not all dogs with mild MR before 4.5 years will develop CHF due to MMVD. However, mild MR in young CKCS is known as a significant predictor of cardiac death due to CHF caused by MMVD (Reimann et al. 2017). Moreover, the diagnosis in 3 cases and 30 controls was only based on auscultation. It cannot be excluded that some of the control dogs may develop CHF later in life due to early MMVD not detected by auscultation. Moreover, the results from the statistical analyses remained overall the same if the 30 control dogs diagnosed with only auscultation were excluded. Finally, the functional impact of the gene variants identified associated with MMVD in the present study needs to be elucidated in future studies. Another limitation, which should be noted, is that the WGS dataset of short reads and the applied variant discovery method does not support identification of larger structural variations in the DNA sequence. Hence, we cannot rule out the existence of such variations or their possible causal relationships with MMVD.

In conclusion, our analyses show that homozygosity for a 39 bp insertion in LMOD2 (MMVD16) combined with homozygosity or heterozygosity for a lncRNA splice site mutation (MMVD10) is strongly associated with MMVD in this cohort of CKCS, with an odds ratio of 52 (95% confidence interval 8.75–1023.85) depending on the presence of two or one nonreference alleles in the splice‐site mutation (MMVD10), respectively. None of the previously identified variants close to NEBL were associated with MMVD in the present study of CKCS; however, further studies are needed to elucidate the impact of the variants near the NEBL gene since the frequency of the reference allele was low, in accordance with observations in previous CKCS studies. In the present study, we show that two novel genetic markers MMVD16 and MMVD10 identified dogs with high risk of MMVD. However, the ability of the two markers to identify early onset of MMVD needs to be validated in an independent cohort of CKCS. This study supports the possibility of including genetic testing in breeding programs aimed at reducing the prevalence of MMVD in CKCS.

4. Materials and Methods

Venous EDTA stabilized blood samples from MMVD cases and controls were identified among blood samples collected from a cohort of more than 1000 privately owned CKCS participating in MMVD breeding programs or research projects at Department of Veterinary and Animal Sciences, University of Copenhagen, Denmark, or at Department of Clinical Sciences at the Swedish University of Agricultural Sciences from 1999 to 2018. Blood collection had been approved by ethical committees and pet owner consent was obtained. Based on typical findings of cardiac auscultation and echocardiography, MMVD was diagnosed and the severity of the disease was graded by veterinarians with many years of experience within veterinary cardiology. Cases were defined as CKCS with presence of mitral regurgitation (left apical mitral regurgitation murmur intensity > Grade 1 (out of 6) (Levine 1933) and/or echocardiographic mitral regurgitation jet size ≥ 20% of the area of left atrium estimated using color flow Doppler imaging) at an early age (before 4.5 years), equivalent to the criteria used by Madsen et al. (2011). In addition, CKCS with onset of CHF due to MMVD (severe mitral regurgitation, increased left ventricular dimensions and clinical signs of CHF at any age) were considered cases (Keene et al. 2019). Dogs with echocardiographic findings of other cardiac diseases were excluded. Controls were defined as CKCS more than 8 years old with no or mild mitral regurgitation, that is, mitral regurgitation murmur intensity ≤ Grade 2 and/or mitral regurgitation jet size ≤ 50% (Madsen et al. 2011). Echocardiography was performed in 217 CKCS (114 cases and 103 controls), however in 33 CKCS (3 cases and 30 controls) the MMVD diagnosis was based on auscultation.

Ten case and 10 control dogs were selected for WGS. Based on pedigrees, the dogs were selected to be unrelated at the parental level. Sequencing libraries were generated using the NEB Next Ultra DNA Library Prep Kit for Illumina, following the manufacturer's recommendations. After adaptor ligation, DNA fragments 400–500 bp in size were selected for library construction. DNA quality for one case sample proved to be insufficient for proper library construction. 150 bp paired‐end reads were generated for the remaining 19 samples using the Illumina Hiseq 4000 platform. FastQC and trimmomatic‐0.39 (Bolger et al. 2014) were used for quality control and removal of adaptor sequences and low quality sequence. Reads were aligned to the canFam4 dog reference genome using the Burrows‐Wheeler Alignment tool (Li and Durbin 2009) with paired‐end alignment. The GATK software (McKenna et al. 2010) was used for duplication removal, haplotype calling and genotyping following the GATK best practice guidelines adapted to the dog sequence. Hard filtering and VQSR were not performed. Subsequent analyses were focused on the two candidate regions identified by Madsen et al. (2011), that is, chr13:62 500 000–64 300 000 (cfa13‐reg) and chr14:1–12 500 000 (cfa14‐reg). All genetic variations in the two regions, for which a nonreference variant was found in all cases (allowing for missing genotype information) and for which the allele frequency among cases was at least 1.5 times the frequency of the same allele in controls, were extracted. Hereafter, the identified variants were passed through the Ensembl Variant Effect Predictor (https://www.ensembl.org/info/docs/tools/vep/index.html) and 18 single nucleotide variants in exons (synonymous and nonsynonymous) and splice sites and one 39 bp insertion in LMOD2 were selected for genotyping. Seventeen SNPs were successfully genotyped in the CKCS cohort described above using TaqMan assays. Genotypes for one SNP (MMVD7) were discarded due to low quality. The insertion, MMVD16, was genotyped by PCR followed by capillary gel‐electrophoresis (Primers: Forward: FAM‐5′‐AGCCTACCTTTCCACACTCC, Reverse: 5′‐AACTTCTGGCTTCCCTGTCA). Furthermore, previously identified variants near the NEBL gene (NEBL 1–4) (Axelsson et al. 2021) were genotyped by TaqMan assays in the same dogs.

Data was analyzed using the statistical software R version 4.2.3. For each variant, a Fisher's exact test was performed to examine associations between genotypes (homozygous AA, heterozygous AB, or homozygous BB) and MMVD status (case or control). A genome‐wide significance threshold of p = 5E‐08 was applied. Subsequently, LASSO multivariate logistic regression analysis (Tibshirani 2018) was applied to identify a minimal set of candidate variants that were most significantly associated with MMVD status. Multiple logistic regression was used to quantify the effect of identified variants. Sex was included as a covariate. Combinations of alleles of the most significant gene variants were tested and odds ratios were estimated accordingly. Linkage disequilibrium between variants was calculated using Plink v1.9 software (Purcell et al. 2007) and blocks with LD > 0.8 were identified.

Finally, 1174 dogs representing 101 different breeds or breed varieties (sample left over from genetic tests service or previous non‐MMVD related research studies) were genotyped for the MMVD10 and MMVD16 variations to study the presence of the most significant candidate variant associated with MMVD in dog breeds other than CKCS. The same TaqMan and PCR‐based test as described above was used. In addition, genotypes of all variants were examined in a publicly available WGS dataset with full genome data from 582 dogs and 8 wolves (Jagannathan et al. 2019).

Author Contributions

Lisbeth Høier Olsen: conceptualization, investigation, writing – original draft, methodology, validation, writing – review and editing, data curation, resources. Isabella Jane Larsen: investigation, writing – review and editing, formal analysis, data curation. Ingrid Ljungvall: investigation, writing – review and editing, resources. Jens Häggström: investigation, resources, writing – review and editing. Torkel Falk: investigation, writing – review and editing. Maria Josefine Ubbe Reimann: investigation, writing – review and editing, methodology, formal analysis, data curation. Arney Eva Gunnlaugsdóttir: investigation, writing – review and editing, formal analysis, data curation. Torben Martinussen: investigation, formal analysis. Foteini Papadaki: investigation, writing – review and editing, formal analysis, data curation. Majbritt Busk Madsen: investigation, writing – review and editing, formal analysis, data curation. Henrik Duelund Pedersen: investigation, formal analysis. Merete Fredholm: conceptualization, investigation, writing – review and editing, resources. Peter Karlskov‐Mortensen: conceptualization, investigation, funding acquisition, methodology, validation, writing – original draft, writing – review and editing, formal analysis, data curation, supervision, resources.

Funding

This work was supported by Agria Djurförsäkring (N2014‐0018, N2019‐0015) and Danmarks Frie Forskningsfond (7017‐00131B).

Ethics Statement

The Danish Animal Experiments Inspectorate (licenses no. 2006/561‐1145, 2011/561‐71 and 2016‐15‐0201‐01074) and the Swedish “Uppsala djurförsöksetiska nämnd” (license no. C12/15, C318/9 and C139/9). Data in Table S3 were based on left over blood samples from the genetic dog service at the Department of Veterinary and Animal Sciences, Animal Genetics & Breeding by written consent from the dog owners and previous ethically approved non‐MMVD research projects. Authors declare human ethics approval was not needed for this study.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Heatmaps illustrate LD between genotyped markers on cfa13 (A) and cfa14 (B). LD‐blocks are indicated by blue squares and numbers 1, 2, 3 and 4.

Table S1: Sex, examiner and age at diagnosis in the 250 included cavalier King Charles spaniels (CKCS).

Table S3: Different dog breeds and breed varieties genotyped for MMVD16 genotypes. None of these 1168 dogs had the 39 bp insertion in LMOD2, that is, the nonreference variant of MMVD16.

Table S4: Position of NEBL variants in CanFam3 and CanFam4 (from Axelsson et al. 2021).

Table S6: Sequencing depth per sample.

Table S7: Allele combination genotype counts in cases and controls.

Table S2: Genotypes for all markers in all tested dogs.

Table S5: MMVD marker genotypes in different dog breeds, WGS sequence (Jagannathan et al. 2019).

Data Availability Statement

All relevant data are available from the corresponding author upon reasonable request. Sequencing data have been deposited at the public repository European Nucleotide Archive under project number PRJEB90338.