A missense variant in DGKG as a recessive functional variant for hepatic fibrinogen storage disease in Wagyu cattle

2.1 Clinical and histological appearance

A 5-month-old male Wagyu calf, weighing 79 kg, was admitted to the Clinic for Cattle of University of Veterinary Medicine Hannover, with a history of recurrent respiratory disease. The calf was born naturally, and was vital and alert after birth. Growth rate and development in the first few weeks of life were reported to be below expectations. A first episode of illness with decreased appetite and obtundation was observed at 4 months of age. The attending veterinarian suspected respiratory disease, which was treated with oxytetracyline (10 mg/kg IM once daily for 3 days) and non-steroidal anti-inflammatory drugs (meloxicam, 0.5 mg/kg SC). Because of lack of treatment response the calf was switched to gamithromycin (6 mg/kg SC as a single dose) and later to procaine penicillin (10 mg/kg IM daily for 3 days) and received a single dose of dexamethasone (0.02 mg/kg IM). The calf was referred to the clinic for diagnostic evaluation because of unsatisfactory response to previous treatment.

On admission, the calf appeared lethargic. The coat was dull and shaggy. Pulse and respiratory rate were 84 beats/min and 20 respirations/min, respectively; rectal temperature was 38.6°C. The calf appeared to be slightly dehydrated. The mucous membranes were pale, and serous nasal discharge was present from both nostrils. Clinical examination of the respiratory and cardiovascular systems identified no abnormalities. The rumen was poorly filled, without stratification and without any motility. The rectum contained small amounts of dark green dry feces partially coated with sticky mucus. Thoracic ultrasonography showed no abnormalities of the interpleural space or lung surface. The rumen appeared empty, and the abomasum and small intestinal loops showed only minimal motility with little content. Arterial blood gas analysis disclosed a marked metabolic acidosis, primarily caused by increased L-lactate and total protein concentrations, consistent with a combined strong ion and A_tot acidosis. Hematology indicated moderate leukocytosis, increased erythrocyte count, and increased PCV. These results, in combination with the clinical examination findings, were attributed to dehydration at the time of admission. Blood biochemistry results showed markedly increased plasma urea and creatinine concentrations, and increased activities of aspartate aminotransferase (AST), glutamate dehydrogenase (GLDH), and creatine kinase (CK). Plasma cholesterol concentration was decreased (Table S1). Based on the clinical findings and blood biochemistry, severe dehydration and acidemia were considered the primary metabolic disturbances to be addressed therapeutically. Intravenous fluid therapy with 5 L of isotonic saline and 500 mL of 8.4% NaHCO₃ solution was administered over a period of 6 hours. Venous blood gas analysis repeated after completion of this initial treatment showed that the acidemia and hyperlactatemia were corrected (Table S1). At this time, the PCV remained slightly above the upper reference limit, whereas the total plasma protein concentration had decreased below the reference range. Isotonic saline (5 L) was administered IV over the next 10 hours because blood urea nitrogen and creatinine concentrations remained markedly increased.

Blood testing performed on the next day, after completion of additional fluid therapy, disclosed a leukocyte count within the reference range and slightly increased red blood cells count (RBC) and PCV. Compared with the time of admission, total bilirubin concentration and activities of AST and GLDH were increased, whereas activities of CK and GGT were decreased. With fluid therapy, plasma total protein and plasma potassium concentrations decreased. Only a moderate decrease in blood urea nitrogen and plasma creatinine concentrations was observed compared with results obtained at the time of admission. To treat persistent azotemia, 10 L of isotonic saline were administered over 12 hours and 20 g of KCl were given PO. Another 5 L of isotonic saline was given IV over 12 hours because plasma creatinine concentration remained increased. On the second day of treatment, the calf was bright, alert, and responsive with normal appetite and thirst. Plasma urea and creatinine concentrations and activity of CK were decreased by approximately 50%, but marked increases in plasma AST, GGT, and GLDH activity were observed. The calf continued to receive 5 L of isotonic saline IV for another 24 hours.

Renal creatinine clearance and fractional excretion of sodium were calculated from spontaneously voided urine and a blood sample, both collected on day 5 and > 48 hours after cessation of IV fluid therapy and were found to be 1.8 mL/min and 0.54%, respectively. Urinary gamma-glutamyl transferase (GGT) activity was 6 U/mL, specific gravity was 1.010 and a pH was 6.5. Urinary GGT activity was below the detection limit. Liver enzyme activities continued to increase, whereas plasma total bilirubin decreased to concentrations within the reference range.

Because the underlying cause of the increased liver enzyme activities and the patient's illness remained unclear, a liver biopsy was performed, and the tissue specimens were fixed in 10% buffered formalin. Histologically, disseminated multiple hepatocytes were moderately swollen and mostly showed a single, sharply demarcated cytoplasmic, slightly eosinophilic inclusion with a ground-glass appearance, with the nucleus displaced to an eccentric position (Figure 1A). Occasionally, small, almost transparent vacuoles were present and single-cell necrosis was seen. The cytoplasmic inclusions did not stain with the periodic acid-Schiff (PAS)-reaction and PAS with diastase (DPAS). Using a rabbit polyclonal antibody against fibrinogen (working dilution 1:500 in phosphate-buffered saline [PBS]; Dako, Germany) with microwave pretreatment in citrate buffer, pH 6.0, a goat anti-rabbit-biotin antibody, and the avidin-biotin peroxidase complex (ABC) method, the cytoplasmic vacuoles in the hepatocytes stained intensely (Figure 1B). Using the pop-off-technique,¹¹ electron microscopy disclosed hepatocytes with membrane-bound cytoplasmic inclusions filled with fine-granular material admixed with a few coarse-grained electron-dense granules (Figure 1C). No fingerprint tubular material was observed.

The animal remained in the hospital 19 days to monitor recovery, and a steady decrease in liver enzyme activities was observed. The calf was discharged after 19 days. Based on the clinical and histopathologic findings, a form of hepatic fibrinogen storage disease (HFSD) was diagnosed. A follow-up discussion with the attending veterinarian indicated that the animal had fully recovered and was still on the farm. All blood results are summarized in Table S1.

2.2 Genetic analysis

We evaluated the hypothesis of a congenital disorder using whole-genome sequencing (WGS; see Data S1 for details of WGS analysis).^12-14 Both the dam and sire of the affected calf had no clinically apparent abnormalities. Therefore, it was difficult to predict whether the disorder was dominant or recessive, so we hypothesized 2 different possible scenarios: a spontaneous, fully penetrant dominant-acting de novo mutation and a recessive mutation that was present in the homozygous state and inherited from both parents. Assuming recessive inheritance in a trio-based approach, filtering the WGS data for homozygous coding variants present in the calf and heterozygous in the parental genomes identified 22 017 variants, of which 81 were protein-changing with a predicted high or moderate effect (Table 1). These 81 variants were further screened for their occurrence in a global control cohort of 4540 genomes from a variety of breeds, identifying 10 remaining protein-changing variants that were exclusively homozygous in the genome of the affected calf and heterozygous in its parents (Tables 1 and S2). Tissue expression of RNA in humans and cattle using publicly available databases (https://www.ncbi.nlm.nih.gov and https://ianimal.pro) for the 9 genes affected by these 10 missense variants showed that they were all widely expressed in different tissues. Seven variants of uncertain importance were classified as neutral (Table 2), whereas 3 missense variants were potentially deleterious according to in silico predictions based on sequence homology using the Protein Variation Effect Analyzer (PROVEAN) web server (https://pubmed.ncbi.nlm.nih.gov/25851949/). Of these 10 remaining private variants, a single variant affected a putative candidate gene for the observed liver phenotype (Figure 2A; Table 2). This homozygous variant at chr1: 81082187C>T represents a missense variant in the DGKG gene, encoding the enzyme diacylglycerol kinase gamma involved in hepatic lipid metabolism (XM_002684869.5: c.2162C>T; Figure 2B). It alters the encoded amino acid of DGKG residue 721 and the threonine-to-isoleucine substitution (XP_002684915.3:p.Thr721Ile) affects an evolutionarily conserved amino acid (Figure 2C,D). The other 2 homozygous missense variants, which are predicted to be deleterious, affect genes with no known role in liver storage diseases. These are TIAM2, which encodes a Rac1-specific guanine nucleotide exchange factor that, when dysregulated, is associated with diseases in humans including cancer, immunological and neurological disorders,¹⁵ and PKD1, which is known to be associated with dominant inherited polycystic kidney disease, which is characterized primarily by kidney cysts and may be accompanied by liver cysts (OMIM 601313). The number of heterozygous carriers in the global control cohort for these 2 variants was higher than the 3 heterozygous DGKG carriers observed in the Wagyu breed alone (Table 3). Furthermore, the occurrence does not seem to be restricted to the Wagyu breed. For TIAM2 there are Angus cattle, and for PKD1 there are Buryat and Hanwoo cattle carrying the variant alleles in addition to Wagyu cattle (Table 3).

Variant filtering identified no private heterozygous protein-changing variants present in the genome of the HFSD-affected calf and absent in both parental genomes and in 5483 controls (Table 1). Thus, a pathogenic de novo mutation that could have occurred in a single parental gamete or during early embryonic development of the calf seems unlikely in the affected animal.