Independent COL5A1 Variants in Cats with Ehlers-Danlos Syndrome

We investigated four cats with similar clinical skin-related signs strongly suggestive of Ehlers-Danlos syndrome (EDS). Cases no. 1 and 4 were unrelated and the remaining two cases, no. 2 and 3, were reportedly siblings. Histopathological changes were characterized by severely altered dermal collagen fibers. Transmission electron microscopy in one case demonstrated abnormalities in the collagen fibril organization and structure. The genomes of the two unrelated affected cats and one of the affected siblings were sequenced and individually compared to 54 feline control genomes. We searched for private protein changing variants in known human EDS candidate genes and identified three independent heterozygous COL5A1 variants. COL5A1 is a well-characterized candidate gene for classical EDS. It encodes the proα1 chain of type V collagen, which is needed for correct collagen fibril formation and the integrity of the skin. The identified variants in COL5A1 are c.112_118+15del or r.spl?, c.3514A>T or p.(Lys1172*), and c.3066del or p.(Gly1023Valfs*50) for cases no. 1, 2&3, and 4, respectively. They presumably all lead to nonsense-mediated mRNA decay, which results in haploinsufficiency of COL5A1 and causes the alterations of the connective tissue. The whole genome sequencing approach used in this study enables a refinement of the diagnosis for the affected cats as classical EDS. It further illustrates the potential of such experiments as a precision medicine approach in animals with inherited diseases.


Introduction
Ehlers-Danlos syndromes (EDS) are heritable disorders of connective tissue that share several clinical features, such as skin hyperextensibility, joint hypermobility, abnormal wound healing, and easy bruising [1]. Based on clinical findings and the mode of inheritance, different subtypes are recognized in humans, caused by alterations in 20 different genes [1][2][3] (Supplementary Table S1). One subtype is classical EDS, which is caused by defects in type V collagen and results from pathogenic variants in COL5A1 or COL5A2 [2,[4][5][6][7][8][9]. Type V collagen is a major regulator of collagen fibril diameter and has a critical role during the early process of collagen fibril nucleation [1]. Targeted and spontaneous animal models for EDS have recently been reviewed [10].
In this study, we characterized the clinical and histopathological phenotype of four cats with clinical signs of classical EDS and performed genetic analyses to search for the causative genetic variants.

Animal Selection
This study included four EDS-affected cats, one Bengal cat (case no. 1), two presumably related shelter cats reported to be Bombay with unknown lineage (cases no. 2 and 3), and one domestic shorthair cat (case no. 4). For case no. 4, we additionally had samples of two unaffected siblings and both parents. Samples from 54 additional genetically diverse control cats of the Vetsuisse Biobank were used for the whole genome sequencing data analysis (Supplementary Table S2).

Clinical and Histopathological Examinations
All cats underwent clinical examination. For case no. 4, hematological and biochemical parameters were measured, and ELISA tests for Feline Leukemia Virus (antigens), Feline Immuno-deficiency Virus (antibodies), and Feline Coronavirus (antibodies) were performed (IDEXX Laboratories, Kornwestheim, Germany). Furthermore, a skin intradermal allergy test (Artuvetrin Haut Test, Nextmune, Leipzig, Germany) and a serum allergy test (IDEXX Laboratories, Kornwestheim, Germany) were carried out.
Skin biopsies were obtained from cases no. 1, 2, and 4. The biopsies were fixed in 10% buffered formalin and processed routinely. The slides were stained with hematoxylin and eosin (H&E), periodic acid-Schiff (PAS), and Masson's Trichrome (MT), prior to histopathological examination.

DNA Extraction
Genomic DNA was isolated from EDTA blood using the Maxwell ® RSC Whole Blood DNA Kit with the Maxwell ® RSC instrument (Promega, Dübendorf, Switzerland).

Whole-Genome Sequencing and Variant Calling
Illumina TruSeq PCR-free DNA libraries were prepared for cases no. 1, 2, and 4 and sequenced on a NovaSeq 6000 instrument. The sequence data were submitted to the European Nucleotide Archive with the study accession PRJEB7401 and sample accessions SAMEA5885921 (case no. 1), SAMEA5885927 (case no. 2), and SAMEA7853381 (case no. 4). Mapping and alignment were performed as described [25]. Variant calling was performed using GATK HaplotypeCaller [26] in gVCF mode as described [25]. To predict the functional effects of the called variants, SnpEff version 4.3t software [27] together with NCBI annotation release 104 for the felCat9.0 genome reference assembly was used. For variant filtering, we used 54 control genomes (Supplementary Table S2).

Gene Analysis
We used the felCat9.0 reference genome assembly and NCBI annotation release 104. Numbering within the feline COL5A1 gene corresponds to the NCBI RefSeq accession numbers XM_023242950.1 (mRNA) and XP_023098718.1 (protein).

PCR and Sanger Sequencing
Candidate variants were confirmed and genotyped either by direct Sanger sequencing of PCR amplicons or fragment length analysis of PCR products on a 5200 Fragment Analyzer (Agilent, Basel, Switzerland). PCR products were amplified from genomic DNA using AmpliTaq Gold 360 Mastermix (Thermo Fisher Scientific, Reinach, Switzerland) together with a forward and reverse primer (Supplementary Table S3). After treatment with exonuclease I and alkaline phosphatase, amplicons were sequenced on an ABI 3730 DNA Analyzer (Thermo Fisher Scientific). Sanger sequences were analyzed using the Sequencher 5.1 software (GeneCodes, Ann Arbor, MI, USA).

Clinical History and Examination
Case no. 1 was a 1-year-old purebred Bengal cat with a history of multiple repetitive skin lesions after minimal trauma occurring since the age of 6 months. At the time of presentation, a scarring alopecic lesion of about 5 × 5 cm was visible on the dorsal neck, which resulted from a previous trauma. Furthermore, the skin was diffusely thinner and much more elastic compared to the skin of a healthy young Bengal cat. In addition, the patient presented with multiple small skin tears distributed over the entire trunk/back area ( Figure 1). The cat had no pruritus and there were no signs of active inflammation. Skin cytology, skin scrapings, and the trichogram were unremarkable. Cases no. 2 and 3, presumably related female spayed Bombay cats, were presented for multiple wounds in the hyperextensible skin predominantly at the neck, base of ear, and shoulders ( Figure 2a). The wounds occurred after self-trauma or exuberant play. Case no. 2 additionally showed spasms of the cutaneous trunci followed by frantic self-mutilation since adoption at 8 weeks of age. At 11 months of age, case no. 2 also experienced multiple partial and grand mal seizures necessitating phenobarbital therapy. Case no. 3 experienced a grade 3 patellar luxation at 18 months of age presumed to be related to joint hypermobility due to EDS. Excessive facial skin folds were apparent (Figure 2b). Cases no. 2 and 3, presumably related female spayed Bombay cats, were presented for multiple wounds in the hyperextensible skin predominantly at the neck, base of ear, and shoulders ( Figure 2a). The wounds occurred after self-trauma or exuberant play. Case no. 2 additionally showed spasms of the cutaneous trunci followed by frantic self-mutilation since adoption at 8 weeks of age. At 11 months of age, case no. 2 also experienced multiple partial and grand mal seizures necessitating phenobarbital therapy. Case no. 3 experienced a grade 3 patellar luxation at 18 months of age presumed to be related to joint hypermobility due to EDS. Excessive facial skin folds were apparent (Figure 2b). Cases no. 2 and 3, presumably related female spayed Bombay cats, were presented for multiple wounds in the hyperextensible skin predominantly at the neck, base of ear, and shoulders ( Figure 2a). The wounds occurred after self-trauma or exuberant play. Case no. 2 additionally showed spasms of the cutaneous trunci followed by frantic self-mutilation since adoption at 8 weeks of age. At 11 months of age, case no. 2 also experienced multiple partial and grand mal seizures necessitating phenobarbital therapy. Case no. 3 experienced a grade 3 patellar luxation at 18 months of age presumed to be related to joint hypermobility due to EDS. Excessive facial skin folds were apparent (Figure 2b). Case no. 4 was a 2-year-old male castrated indoor DSH cat that was examined for recurrent laceration wounds on the neck and shoulders since adoption (one year before), and moderate head and neck pruritus (observed all year-round with exacerbation in June and August). Hematology and biochemistry (including basal cortisol) were normal, and Feline Leukemia Virus, Feline Immunodeficiency Virus, and Feline Coronavirus were negative one year before referral. On physical examination, a 1 cm scar on the ventral neck and a 0.5 cm crust above the right eye were noted. The skin extensibility index was 22% ( Figure 3a). No joint laxity, hernias, ocular, or heart abnormalities were found. Neither ecto-and endoparasite treatment of both cats, nor an eight-week elimination diet improved the pruritus. Over the course of the disease, the cat developed new lesions due to pruritus when treated only with oral antihistamines (Fenistil, GlaxoSmithKline Consumer Healthcare GmbH & Co. KG, Munich, Germany) (Figure 3b,c). With short courses of systemic prednisolone (injections of 2 mg/kg, or orally 0.6 mg/kg once daily, as needed), the scratching stopped, and the lesions healed. Systemic medication was withdrawn as indicated prior to biopsy sampling. A welfare score (8; 0-21) was not suggestive of a psychogenic induced pruritus [28]. Based on all the above and since the patient fulfilled Favrot's Case no. 4 was a 2-year-old male castrated indoor DSH cat that was examined for recurrent laceration wounds on the neck and shoulders since adoption (one year before), and moderate head and neck pruritus (observed all year-round with exacerbation in June and August). Hematology and biochemistry (including basal cortisol) were normal, and Feline Leukemia Virus, Feline Immunodeficiency Virus, and Feline Coronavirus were negative one year before referral. On physical examination, a 1 cm scar on the ventral neck and a 0.5 cm crust above the right eye were noted. The skin extensibility index was 22% (Figure 3a). No joint laxity, hernias, ocular, or heart abnormalities were found. Neither ecto-and endoparasite treatment of both cats, nor an eight-week elimination diet improved the pruritus. Over the course of the disease, the cat developed new lesions due to pruritus when treated only with oral antihistamines (Fenistil, GlaxoSmithKline Consumer Healthcare GmbH & Co. KG, Munich, Germany) (Figure 3b,c). With short courses of systemic prednisolone (injections of 2 mg/kg, or orally 0.6 mg/kg once daily, as needed), the scratching stopped, and the lesions healed. Systemic medication was withdrawn as indicated prior to biopsy sampling. A welfare score (8; 0-21) was not suggestive of a psychogenic induced pruritus [28]. Based on all the above and since the patient fulfilled Favrot's criteria [29], which help to establish the diagnosis of atopic dermatitis, a feline atopic skin syndrome (FASS) was diagnosed and a concomitant collagen disorder leading to lacerations due to pruritus (traumatic splitting) was suspected [30]. A skin intradermal allergy test as well as a serum allergy test were negative and could not identify the triggering agents of FASS. The owner declined antiallergic therapies other than antihistamines, and the cat was euthanized 2.5 years after the first presentation due to a very large laceration wound and deterioration, despite surgical and emergency treatment. criteria [29], which help to establish the diagnosis of atopic dermatitis, a feline atopic skin syndrome (FASS) was diagnosed and a concomitant collagen disorder leading to lacerations due to pruritus (traumatic splitting) was suspected [30]. A skin intradermal allergy test as well as a serum allergy test were negative and could not identify the triggering agents of FASS. The owner declined antiallergic therapies other than antihistamines, and the cat was euthanized 2.5 years after the first presentation due to a very large laceration wound and deterioration, despite surgical and emergency treatment. Two unaffected siblings and both parents of the patient were also clinically evaluated. They did not have any history compatible with allergy and no abnormality was noted. Their skin extensibility indices were normal (father: 15.6%, mother: 6.6%, brother 17.6%, sister: 10.7%).

Histopathological Examination
The pathological changes in the skin biopsies were characterized by severely altered dermal collagen fibers and a slightly thinner epidermis. Changes were more severe in the superficial and mid dermis but were also present in the deep dermis. The abnormal collagen fibers were haphazardly arranged, shortened, and sometimes curled. They were uneven in length and width. Many fibers were very wispy and the interfibrillar spaces were widened. Multifocally small areas of hemorrhage were seen (Figure 4). Two unaffected siblings and both parents of the patient were also clinically evaluated. They did not have any history compatible with allergy and no abnormality was noted. Their skin extensibility indices were normal (father: 15.6%, mother: 6.6%, brother 17.6%, sister: 10.7%).

Histopathological Examination
The pathological changes in the skin biopsies were characterized by severely altered dermal collagen fibers and a slightly thinner epidermis. Changes were more severe in the superficial and mid dermis but were also present in the deep dermis. The abnormal collagen fibers were haphazardly arranged, shortened, and sometimes curled. They were uneven in length and width. Many fibers were very wispy and the interfibrillar spaces were widened. Multifocally small areas of hemorrhage were seen (Figure 4).

Ultrastructural Examination
TEM analysis of case no. 4 skin connective tissue revealed abnormalities in the collagen fibril organization and structure. Abnormal findings from the reticular part of the dermis were loosely packed and disorganized collagen fibers (Figure 5b,d). Individual collagen fibers contained curled fibrils (Figure 5b,d). Higher magnifications showed crosssections of collagen fibrils with different diameters and irregularly outlined fibrils within a fiber (Figure 5f,h). In contrast, the connective tissue of an age-matched control cat contained densely packed collagen fibers with parallel-aligned fibril organization ( Figure  5a,c). Cross-sections of fibrils showed uniform diameters and regularly shaped, almost round outlines (Figure 5e

Ultrastructural Examination
TEM analysis of case no. 4 skin connective tissue revealed abnormalities in the collagen fibril organization and structure. Abnormal findings from the reticular part of the dermis were loosely packed and disorganized collagen fibers (Figure 5b,d). Individual collagen fibers contained curled fibrils (Figure 5b,d). Higher magnifications showed cross-sections of collagen fibrils with different diameters and irregularly outlined fibrils within a fiber (Figure 5f,h). In contrast, the connective tissue of an age-matched control cat contained densely packed collagen fibers with parallel-aligned fibril organization (Figure 5a,c). Cross-sections of fibrils showed uniform diameters and regularly shaped, almost round outlines (Figure 5e,g).

Genetic Analysis
To identify the causative genetic variants, the genomes of three affected cats (cases no. 1, 2, and 4) were sequenced. Case no. 3 was presumably related to case no. 2 and assumed to share the same pathogenic variant. Since the other cats were not known to be related, we hypothesized that their phenotypes were due to independent pathogenic variants. We therefore performed an individual variant filtering in each of the affected cats by comparing them to the genomes of 54 control cats. We focused our search on private protein-changing variants in the 20 known EDS candidate genes (Tables 1, S1, S2 and S4). In case no. 1, we identified one heterozygous private protein-changing variant within an EDS candidate gene (Figure 6a). The variant, COL5A1:c.112_118+15del, is a 22 bp deletion that removes the exon1/intron1 boundary (Table 2, Figure 6b). The presence of the variant was confirmed by PCR and fragment length analysis.

Discussion
In this study, we describe four cats with classical EDS. All cats were characterized by similar clinical skin-related symptoms strongly suggestive of EDS [2]. Histopathology of the examined skin biopsies was also suggestive of EDS. In addition, TEM on case no. 4 identified abnormal morphological features indicative of EDS, such as variable diameters of collagen fibrils, disorganized, distorted collagen fibers, and irregularly outlined fibrils, which have been reported in the literature to be present in most patients with a genetically confirmed diagnosis of EDS [31][32][33].
Whole genome sequencing was performed to confirm the suspected clinical and his-  In case no. 2, we identified another single candidate variant, a heterozygous nonsense variant in COL5A1, predicted to truncate 36% of the coding sequence ( Table 2). We confirmed identical heterozygous genotypes in cases no. 2 and 3 by Sanger sequencing (Figure 6a,c).
In case no. 4, we identified three heterozygous variants in known EDS candidate genes, a frameshift deletion in COL5A1 and two missense variants in TNXB. The targeted genotyping of all three variants in the complete family comprising both parents and two additional unaffected littermates excluded the TNXB variants (Supplementary Figure S1). The remaining variant, COL5A1:c.3066del, is a one base pair deletion, resulting in a frameshift and predicted to truncate 44% of the open reading frame (Table 2, Figure 6a,d). The genotyping results in the family further revealed that COL5A1:c.3066del represented a de novo variant as the mutant allele was absent from the leukocyte DNA of both parents.

Discussion
In this study, we describe four cats with classical EDS. All cats were characterized by similar clinical skin-related symptoms strongly suggestive of EDS [2]. Histopathology of the examined skin biopsies was also suggestive of EDS. In addition, TEM on case no. 4 identified abnormal morphological features indicative of EDS, such as variable diameters of collagen fibrils, disorganized, distorted collagen fibers, and irregularly outlined fibrils, which have been reported in the literature to be present in most patients with a genetically confirmed diagnosis of EDS [31][32][33].
Whole genome sequencing was performed to confirm the suspected clinical and histological diagnosis and to identify the underlying causative genetic variants in all four cats. We identified three independent heterozygous private protein-changing variants in COL5A1 in the four investigated affected cats. A de novo mutation event was experimentally confirmed for case no. 4. We assume that the pathogenic COL5A1 variants in the other affected cats were also due to recent de novo mutation events. For cases no. 2 and 3, the most likely scenario was a germline mutation in one of their parents. These results resemble previously described findings in a cat and three dogs affected by EDS due to de novo mutations in COL5A1 [24,34].
COL5A1 encodes the proα1 chain of type V collagen, which together with type I collagen forms heterotypical fibrils. Type V collagen is a major regulator of collagen fibrillogenesis, and the reduction in collagen V expression results in fewer collagen I fibrils with increased diameters [35,36]. Correct fibril formation is required for the integrity of the skin [1].
In humans, the functional loss of one COL5A1 allele is the most commonly reported molecular mechanism in classical EDS [36,37]. As type V procollagen molecules cannot accommodate more than a single proα2(V) chain, the reduction in available proα1(V) chains results in the production of about half the normal amount of type V collagen [38]. We assume that all three newly reported feline variants lead to nonsense-mediated mRNA decay, which results in a haploinsufficiency of COL5A1 and causes the alterations of the connective tissue.
Case no. 2 experienced epilepsy and presumptive hyperesthesia signs. Notably, an EDS-affected Burmese cat was described as experiencing prolonged episodes of disturbing vocalizations that were considered to be partial seizures. Upon necropsy, this cat was found to have flattened cerebral gyri [14]. Epilepsy has been noted in human patients with EDS and is associated with many structural brain defects, including periventricular heterotopia in patients with COL5A1 variants [39,40]. EDS-associated small fiber polyneuropathy in human patients manifests as acute, unpredictable burning pain, muscle cramping, and also compulsive scratching due to severe episodic itch [41]. Given comparable observations in human EDS patients and the earlier reported EDS-affected Burmese cat, it is reasonable to suspect that epilepsy and hyperesthesia were likely part of the EDS phenotype in case no 2.
Case no. 4 was diagnosed with EDS and concurrent FASS. Diagnosis of FASS can be difficult since it is a clinical diagnosis, and no diagnostic tests exist to diagnose or exclude FASS. Allergy tests such as serology and intradermal testing only help to identify the triggering agents to perform allergen specific immunotherapy. However, FASS cases have sometimes been described to have negative allergy testing results [30]. Furthermore, the cat was presented with head and neck pruritus. This is a skin reaction pattern that is consistent with the diagnosis of FASS after excluding other possible causes, as performed in our case [30]. Therefore, based on the clinical presentation and the fact that the cat showed a seasonal exacerbation, as well as the response to histamines (partial) and steroids (complete), we concluded that the itchiness has an underlying allergic cause. A neurological component of the itch therefore seemed unlikely, but it could not be completely excluded without further diagnostic imaging or necropsy, which was declined by the owner.
EDS-affected pets with other concurrent diseases inducing self-trauma (such as itchiness in case no. 4 due to FASS) have exacerbated symptoms that might lead to euthanasia. A faster whole genome sequencing-based diagnosis of the primary collagen genetic disease might lead to a better overall management of the patients, with the potential of decreasing the euthanasia rates of such complicated cases. A precise classification of EDS relies on molecular confirmation with identification of a causative variant [2]. This study illustrates the potential of whole genome sequencing as a precision medicine approach in animals with inherited diseases.
Consensus criteria for the interpretation of genetic variants in domestic animals have not yet been defined. However, such criteria have been established for human patients and they have proven highly useful to provide a standardized classification reflecting the strength of evidence for pathogenicity [42]. When we apply the human guidelines to the investigated cats, all three reported COL5A1 variants would be classified as pathogenic according to the ACMG/AMP criteria [42]. All three reported variants represent null variants (very strong evidence of pathogenicity, PVS1), are absent from controls (moderate evidence of pathogenicity, PM2), and the patients' phenotypes are highly specific for a disease with a single genetic etiology (supporting evidence of pathogenicity, PP4). For case no. 4 we have the confirmed de novo mutation event and family segregation data as additional criteria for pathogenicity.

Conclusions
We characterized four cats with EDS and identified three independent causal variants in the COL5A1 gene. De novo mutation events leading to haploinsufficiency of COL5A1 appear to represent an important molecular etiology in feline EDS.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/genes13050797/s1, Figure S1: Pedigree with genotypes from the family of case no. 4.   Institutional Review Board Statement: All animal procedures were performed according to local regulations. The "Cantonal Committee for Animal Experiments" approved the collection of blood samples from cats (Canton of Bern; permit 71/19).

Informed Consent Statement:
The owners of the cats in this study gave informed consent to use samples and data for research.

Data Availability Statement:
The accessions for the sequence data reported in this study are listed in Table S2.