Identification of an F13A1 frameshift variant associated with factor XIII deficiency in a Coonhound dog with severe coagulopathy

Abstract

Background

Factor XIII (FXIII) deficiency is a rare autosomal recessive bleeding disorder characterized by unstable fibrin clots and severe hemorrhagic complications. In humans, pathogenic variants have been described in F13A1 and F13B, which encode the subunits comprising the FXIII heterotetramer. However, cases in animals are exceedingly rare.

Hypothesis/Objectives

The objective of this work was to characterize a naturally occurring FXIII deficiency in a dog.

Animals

A 4-month-old male Black and Tan Coonhound presented with spontaneous hemoperitoneum, thrombocytopenia, and persistent bleeding after surgical procedures.

Methods

Hemostasis testing and whole genome sequencing were performed to characterize the phenotypic and molecular genetic basis of the bleeding disorder.

Results

A functional FXIII deficiency was identified, and a private, homozygous variant (c.1234_1239delinsTCAA) was found in exon 11 of F13A1 that predicts a frameshift and premature stop codon.

Conclusions and clinical importance

This report represents only the second clinical description of FXIII deficiency in dogs and the first genetic characterization of this disorder in companion animals. The identified F13A1 variant provides a molecular diagnosis and enables genetic testing for this bleeding disorder in Black and Tan Coonhounds should additional cases arise.

Introduction

Factor XIII (FXIII) deficiency is a rare and potentially fatal bleeding disorder in humans caused by functional or quantitative defects of FXIII, a transglutaminase that cross-links fibrin fibers and thus promotes fibrin clot stability. This coagulopathy is typically characterized as a severe bleeding diathesis that can manifest as intracranial hemorrhage, the leading cause of mortality and disability in affected individuals, and is also associated with poor wound healing and miscarriage.^1-3 Inherited FXIII deficiency in humans (OMIM#613225, #613235) is an autosomal recessive trait with an estimated worldwide incidence of 1 in 2 million.^3,4 The FXIII protein is a heterotetramer composed of 2 catalytic A subunits and 2 carrier B subunits. More than 200 variants have been associated with FXIII deficiency in humans, affecting the F13A1 and F13B genes that encode the A and B subunits, respectively.⁵ Defects in the B subunit are associated with milder bleeding symptoms presumably due to residual FXIII catalytic activity that remains despite dysfunctional or absent carrier subunits.⁶

The heterotetramer (FXIIIA₂B₂) circulates in plasma as an inactive zymogen pro-γ-transglutaminase (pFXIII). In its main activation pathway, thrombin cleaves the N-terminal activation peptide of the pFXIII A subunit, followed by Ca²⁺-induced subunit dissociation and conformational changes. Activated FXIII covalently cross-links fibrin polymers through glutamine and lysine sidechains, stabilizing the nascent clot and protecting it from shear stress. In addition, FXIII cross-links fibrin to the fibrinolysis inhibitor, α2-antiplasmin, thereby slowing the rate of plasmin-mediated clot degradation.⁷

Acquired FXIII deficiency caused by protein activation, consumption, and depletion has been described in human patients with severe systemic diseases associated with disseminated intravascular coagulation⁸ and after extracorporeal perfusion procedures.⁹ Mouse F13A1 knockout models demonstrate a severe bleeding phenotype with pregnancy loss due to uterine hemorrhage, comparable to the human disease counterpart.^10,11 In contrast, targeted inhibition of the FXIII B subunit results in a mild bleeding phenotype characterized by rebleeding at sites of injury.¹² To date, naturally occurring FXIII deficiency in animals has been described only in Japanese Black cattle and a domestic dog. The trait in Japanese Black cattle and the related Kumamoto sub-breed (OMIA:001818-9913) is characterized as a clinically severe and potentially lethal bleeding disorder caused by a point mutation in F13A1 (c.248T > C). The canine case report described a male Toy Poodle with recurrent bleeding after surgery, abnormal clot stability in viscoelastic assays and in a FXIII deficiency screening test.¹³ Herein, we report the clinicopathologic features and genetic characterization of a severe coagulopathy in a Black and Tan Coonhound dog with dysfunctional FXIII.

Materials and methods

Animals

The proband was a 4-month-old, male intact purebred Black and Tan Coonhound evaluated for persistent bleeding from an incisional biopsy site. Before the bleeding episode that prompted hemostatic testing, the dog had a history of shifting limb lameness, bleeding from an oral lesion, and spontaneous hemoperitoneum secondary to a splenic laceration of unknown cause.

Hemostasis testing

Citrate plasma and EDTA blood samples were submitted to an AAVLD-accredited laboratory (Comparative Coagulation Laboratory, Animal Health Diagnostic Center, Cornell University) to further characterize the bleeding disorder. Analyses included coagulation screening tests (activated partial thromboplastin time [aPTT], prothrombin time [PT], and thrombin time), measurement of clottable (Clauss) fibrinogen and VWF:Ag, and procoagulant factor activity assays (Factor II:C, V:C, VII:C, VIII:C, IX:C, X:C, XI:C, XII:C). Platelet membrane marker expression (CD61, and poststimulation annexin V binding) was evaluated by flow cytometry. Screening tests to detect FXIII deficiency were performed by measuring clot solubility in a urea solution, using samples collected and submitted at 2 different time points. Plasma FXIII activity was measured using a commercial chromogenic assay designed for use with human samples. See the Supplementary Appendix for complete hemostasis test descriptions.

Genetic analyses

DNA extraction from EDTA whole blood was performed using standard procedures. A TruSeq PCR-free library was prepared and WGS data targeting 30X coverage were generated with 2 × 150 bp paired-end reads on an Illumina NovaSeq X. The resultant sequence files were processed using the Whole Animal Genome Sequencing (WAGS) pipeline¹⁴ for alignment with the canFam4 reference genome. Given the rare nature of the disease, and the fact that the clinically unaffected parents shared a sire, the disease allele was expected to be private to the affected dog. Sequencing data from the affected dog were compared to 3023 genomes (Table S1), including 2628 purebred dogs of 362 breeds, using the private variant pipeline OnlyWAGS to identify unique changes occurring only in the affected Black and Tan Coonhound. Private variants were filtered by genotype quality, retaining those with GQ ≥ 20. The Ensembl Variant Effect Predictor was used to predict the impact of the private alleles present on autosomes and sex chromosomes. All characterized genes with identified protein-changing variants were input into VarElect, a program that ranks genes based on their association with phenotypes, with the keywords “factor XIII deficiency,” “bleeding diathesis,” and “coagulation factor” (Table S2). F13A1 received the highest score of 130.20 (Table S3). No other genes received a score over 3.5 (Tables S3 and S4). Candidate genes, F13A1 and F13B, were also visually inspected using the Integrative Genomics Viewer.¹⁵ The variant was confirmed by PCR and Sanger sequencing. See the Supplementary Appendix for primer sequences and thermal cycling conditions.

Results

Clinical presentation

The proband, a 4-month-old male Black and Tan Coonhound with a history of intermittent shifting limb lameness, and prolonged bleeding from an oral lesion, developed a hemoperitoneum secondary to a splenic laceration of unknown cause. At that time the dog was thrombocytopenic (34 000/μL, reference interval 148-484 K/uL), confirmed as accurate by a manual platelet estimate on a thin blood film evaluation, with normal PT (14 s [11-17 s]), a slight prolongation of aPTT (108 s [72-102 s]), and normal plasma von Willebrand factor concentration (VWF:Ag). The dog’s hemoperitoneum was managed by autotransfusion of washed, recovered shed blood, allogeneic red blood cell transfusion, and splenectomy. There was excessive postoperative bleeding at the incision site that was treated with a transfusion of fresh-frozen plasma (FFP) and the administration of ε-aminocaproic acid (ε-ACA).

Upon re-evaluation 2 weeks after discharge, the dog had recovered from surgery, and the platelet count was 511 000/μL. The dog’s rostral mandible was swollen and appeared painful. There were persistent deciduous teeth and a fluctuant swelling with focal mucosal ulceration. Cross-sectional imaging demonstrated abnormal mandibular dentition and a well-demarcated cavitated intraosseous mass with associated osteolysis (Figure S1). Incisional biopsies revealed a papillomatous squamous cell carcinoma. The dog developed persistent bleeding associated with the incisional biopsy site prompting referral to a veterinary emergency and critical care specialist. At that time the dog had melena, and the rostral mandible was markedly swollen with organized but easily disrupted clots at the biopsy site. The dog was anemic (HCT 23%) but not thrombocytopenic (248 000/μL), and clotting times were normal. The cause of the dog’s bleeding was unknown, thus an untargeted therapeutic approach was taken involving a crossmatched packed red blood cell transfusion, ε-ACA, FFP, and desmopressin. This strategy successfully controlled the dog’s bleeding and enabled a subsequent elective bilateral rostral hemimandibulectomy to remove the oral tumor. At that time, the dog received 325 mg tranexamic acid (TXA, 13 mg/kg, PO q6h) and 25 mL/kg FFP total over 24 h preoperatively. Intraoperatively the dog was transfused with cryoprecipitate (15 mL/kg over 5 h). Postoperatively, he received 325 mg TXA (13 mg/kg, PO, q6h then q8h). Point-of-care viscoelastic coagulation testing (VCM Vet, Entegrion, Durham, NC) was performed serially during the procedure (Figure 1). Preoperatively the tracing appeared hypocoagulable based on low maximum clot firmness (MCF) and hyperfibrinolytic based on residual clot firmness at 30- and 45-min post-MCF (LI30 and LI45), becoming normal with the administration of plasma products (Table 1).

Figure 1.

Viscoelastic test profiles with derived parameters from the dog from before the bilateral rostral hemimandibulectomy was performed; in the postoperative period; and before hospital discharge. The X-axes represent time in seconds (0-3600); while the Y-axes represent tracing amplitude in mm (−60 to 60). The preoperative tracing performed before blood product transfusion or tranexamic acid administration suggests abnormally low clot firmness and hyperfibrinolysis. The postoperative tracing performed after surgery and after administration of fresh-frozen plasma and tranexamic acid suggests improved clot firmness and a reduction in the speed of clot lysis. At the time of hospital discharge and subsequent to additional administration of cryoprecipitate, the clot firmness has improved further and there was no indication of hyperfibrinolysis.

Table 1.

Summary of whole blood viscoelastic coagulometer test variables indicating the response to treatment with blood product transfusion and antifibrinolytic drug administration.

Variable	Reference interval	Baseline	After FFP + TXA	After FFP + TXA + Cryo
CT (s)	241-470	231	399	214
CFT (s)	104-266	156	153	180
Alpha (°)	43-64	60	56	48
MCF (mm)	29-44	20	32	44
A10 (mm)	16-30	19	28	29
A20 (mm)	22-38	20	32	39
LI30 (%)	99-100	92	97	100
LI45 (%)	98-100	69	94	100

Abbreviations: A10 = tracing amplitude at 10 min; A20 = tracing amplitude at 20 min; Alpha = clot formation angle; CT = clot time; CFT = clot formation time; Cryo = cryoprecipitate; FFP = fresh-frozen plasma; LI30 = lysis index at 30 min; LI45 = lysis index at 45 min; MCF = maximum clot firmness; TXA = tranexamic acid.

The hemimandibulectomy procedure was routine, and the dog’s postoperative recovery was uneventful with no bleeding complications noted. The dog was subsequently hospitalized 3 times to manage a subcutaneous hematoma resulting from a fall, for the elective resection of an aural mass, and for management of a hematoma after microchip implantation. All bleeding episodes were successfully managed with transfusions of allogeneic frozen plasma.

Hemostasis testing

Results of coagulation screening tests, fibrinogen, platelet flow cytometry assessment of CD61 (GPIIIa) expression and poststimulation annexin V binding, and individual factor assays were within reference intervals, with the exception of Factor X:C and VWF:Ag (Table 2). A mild reduction in Factor X:C (69%) was interpreted as clinically irrelevant, since values > 60% are considered sufficient for normal hemostasis and are not associated with abnormal bleeding.¹⁶ Plasma VWF:Ag of 193% was slightly above the reference interval, ruling out von Willebrand disease. Results of FXIII deficiency screening tests revealed abnormally rapid clot dissolution, with complete clot lysis within 15 min of incubation in a 5 M urea solution. This rapid lysis suggests a functional FXIII deficiency (typically < 5% activity) and was a reproducible finding on a second sample collected approximately 3 weeks after the first submission. Plasma FXIII activity measured by chromogenic assay was 5% relative to a pooled normal canine plasma standard assigned a value of 100%.

Table 2.

Summary of coagulation, factor assays, and platelet flow cytometry testing.

Test (unit)	Result	Reference interval
Activated partial thromboplastin time (s)	16.9	11.0-17.5
Prothrombin time (s)	13.4	11.0-15.5
Thrombin time (s)	7.1	5.0-9.0
Fibrinogen (mg/dL)	200	150-490
Factor II:C (%)	74	50-150
Factor V:C (%)	50	50-150
Factor VII:C (%)	104	50-150
Factor VIII:C (%)	84	50-200
Factor IX:C (%)	56	50-150
Factor X:C (%)	69	80-175
Factor XI:C (%)	101	60-150
Factor XII:C (%)	85	60-150
Factor XIII transglutaminase activity (%)	5	–
VWF:Ag (%)	193	70-180
Platelet integrin αIIbβ3 expression (CD61, GPIIIa)	Normal	–
Platelet poststimulation annexin V binding	Normal	–

Abbreviation: VWF:Ag = von Willebrand factor concentration.

Genetic characterization

Pedigree analysis revealed that the parents of the proband were half-siblings through their sire, suggesting that a causal variant was inherited identical by descent from each parent. Whole genome sequence data generated from the affected proband, achieved 31.14X average coverage. Comparison to canine WGS (n = 3023), comprising 326 breeds, revealed 25 572 unique variants on canine chromosomes 1-38, X, and Y in the proband. Variant consequences on protein sequence revealed a total of 79 variants that were protein-coding or within splice sites (Table S5). Of these, 9 were homozygous, consistent with inheritance identical by descent. Three variants occurred nearly consecutively in exon 11 of the candidate gene F13A1 located on canine chromosome 35 (Figure 2A). Upon inspection, these separately called variants represent a deletion–insertion (delins) mutational event, XM_038584097.1:c.1234_1239delinsTCAA (Figure 1B), which was confirmed by Sanger sequencing (Figure S2). The delins predicts a frameshift from amino acid 412 and a premature stop codon 7 amino acids downstream of the delins (Figure 1C). The remaining 76 protein-changing candidate variants, inclusive of 67 heterozygous variants, occurred in genes with no known role in coagulation.

Figure 2.

Visualization of the identified functional candidate variant in F13A1 in the Black and Tan Coonhound with factor XIII deficiency. (A) Integrative Genomics Viewer display of the homozygous frameshift variant in the affected dog. (B) Illustration of the F13A1 transcript with the identified variant in exon 11 of F13A1. (C) Wild-type and mutant cDNA and protein sequences showing the impact of the variant, with the frameshift causing a premature stop codon after the inclusion of seven aberrant amino acids. The deleted sequence is highlighted.

Discussion

We investigated clinically severe and recurrent abnormal bleeding in a 4-month-old Black and Tan Coonhound. Hemostasis testing ruled out common coagulopathies and confirmed a clot stability defect, compatible with a diagnosis of FXIII deficiency. Whole genome sequencing of the affected dog revealed a deletion–insertion variant predicting a frameshift and premature stop codon in F13A1, the disease gene most often associated with autosomal recessive FXIII deficiency in humans.¹⁷

The affected dog’s reported bleeding history of shifting leg lameness and severe hemorrhage after injury or surgery was suggestive of a secondary coagulopathy (ie, a plasma protein defect, deficiency or inhibition) resulting in impaired formation of a stable fibrin clot. Hemostasis testing to identify a coagulation factor deficiency was initially prioritized due to the relatively high incidence of these defects in dogs.¹⁸ In particular, the X-linked recessive traits, hemophilia A and B, were considered possible differentials given the signalment of a young male dog. Specific measurement of VWF:Ag was included in this preliminary diagnostic evaluation due to the many described von Willebrand disease breed-variants in dogs and observations that new cases can arise in previously unaffected breeds (OMIA 001057; 001058; 001339). Results of coagulation screening tests and factor assays, however, ruled out specific coagulation factor or VWF deficiency as the cause of clinical signs. Flow cytometric screening tests also ruled out Glanzmann thrombasthenia and canine Scott syndrome on the basis of normal platelet fibrinogen receptor expression (CD61/GPIIIa) and normal procoagulant surface expression (poststimulation annexin V binding). In contrast, the FXIII deficiency screen revealed markedly abnormal clot stability, consistent with the measured reduction in FXIII activity. Unlike other procoagulant factors, FXIII is not required for the transformation of plasma fibrinogen into the polymerized fibrin fibrils that form a fibrin clot. Thus, coagulation screening tests based on detection of a fibrin clot endpoint are not useful for diagnosing FXIII defects. The transglutaminase activity of FXIII results in fibrin cross-linkage and stabilization that delays fibrinolysis and renders fibrin clots insoluble in urea solutions. The findings of unstable clot formation and premature clot lysis in a viscoelastic assay combined with complete clot dissolution in urea indicated that FXIII deficiency should be considered high on the differential list as the underlying cause of abnormal bleeding for this dog. Assessment of the transglutaminase activity of the dog’s plasma confirmed a marked reduction in FXIII activity.

Genetic analysis of the affected dog supported the diagnosis of FXIII deficiency. A novel homozygous delins variant was identified in exon 11, out of 16, in F13A1 that introduced a frameshift and premature stop codon within 7 amino acids. The candidate causal variant was absent in the genomes of 3023 other dogs, and no other protein-changing variants were called within the region impacted by the delins. The premature stop codon predicts a loss of 43% of the amino acid sequence; thus, the transcript likely undergoes nonsense-mediated decay. This mechanism is consistent with the FXIII deficiency observed in the affected dog.

F13A1 encodes a 733 amino acid protein in dogs (XP_038440025.1), with 96% sequence similarity (88% sequence identity) to the human protein (NP_000120.2). In humans, variants in F13A1 are responsible for approximately 90% of FXIII deficiency cases.¹⁹ A missense variant, p.Gly411Cys, overlapping the delins site and within the core catalytic domain has been reported in one human FXIII deficiency patient, as well as frameshift variants affecting downstream exons.¹⁹

In conclusion, we describe the clinical diagnostic features of FXIII deficiency with a likely autosomal recessive inheritance in a Black and Tan Coonhound. We identified a homozygous deletion–insertion variant in F13A1 that results in a frameshift and premature stop codon and is the likely causative variant. This work genetically characterizes FXIII deficiency in dogs and describes the second reported case.

Abbreviations

AAVLD

American Association of Veterinary Laboratory Diagnosticians

aPTT

activated partial thromboplastin time

FXIII

factor XIII

FFP

fresh-frozen plasma

MCF

maximum clot firmness

TXA

tranexamic acid

VWF:Ag

von Willebrand factor concentration

WAGS

whole animal genome sequencing

WGS

whole genome sequencing

ε-ACA

ε-aminocaproic acid

Author contributions

Leo Pieples (Investigation, Writing—original draft, Writing—review & editing), Shawna Cook (Data curation, Investigation, Visualization, Writing—original draft, Writing—review & editing), Audrey Tinsman (Investigation, Resources, Writing—original draft, Writing—review & editing), Marjory B. Brooks (Conceptualization, Investigation, Resources, Writing—review & editing), Robert Goggs (Conceptualization, Investigation, Resources, Writing—original draft, Writing—review & editing), and Jacquelyn Evans (Conceptualization, Investigation, Resources, Writing—original draft, Writing—review & editing)

Conflicts of interest

The authors declare no conflicts of interest.

Funding

This work was supported in part by the Hunter R. Rawlings III Cornell Presidential Research Scholars program (LAP) and The Hartwell Foundation (SRC).

Data availability

Whole genome sequence data generated herein are deposited in the Sequence Read Archive under accession # SRR35980927.

Off-label antimicrobial declaration

The authors declare no off-label use of antimicrobials.

Institutional animal care and use committee or other approval declaration

An EDTA-anticoagulated whole blood sample was collected from the affected dog for DNA extraction under an approved Cornell University institutional animal care and use protocol (2005-0151, approved February 19, 2015).

Human ethics approval declaration

The authors declare that human ethics approval was not needed.