Gap junction protein gamma 2 (GJC2) gene frameshift deletion in Toy Poodles with hypomyelinating leukodystrophy resembling human Pelizaeus–Merzbacher–like disease 1

Abstract

Pelizaeus–Merzbacher-like disease type 1 (PMLD1), also known as hypomyelinating dystrophy 2, is a typical slowly progressive hypomyelinating leukodystrophy (HLD) that manifests in the neonatal period or early infancy. In humans, PMLD1 is inherited as an autosomal recessive trait, with genetic variants identified in the gap junction protein gamma 2 (GJC2) gene. Six juvenile-age Toy Poodle dogs exhibited coarse, vertical head and trunk tremors from immediately after birth. Conventional magnetic resonance imaging revealed uniform signal hyperintensities throughout the subcortical white matter in all dogs. All dogs were euthanized. Subsequent histopathological examination revealed vacuolar changes and demyelination in the cerebral and cerebellar white matter, leading to a diagnosis of HLD. Genetic analysis of five affected dogs identified a shared homozygous 14-base pair deletion in the GJC2 gene, XM_038556433.1:c.920_933del. To the best of our knowledge, this is the first report of a pathogenic GJC2 variant in dogs or any other domestic animal species. The results will enable genetic testing and provide a spontaneous large animal model for the homologous human disease.

INTRODUCTION

Pelizaeus–Merzbacher-like disease 1 (PMLD1) is a typical slowly progressive hypomyelinating leukodystrophy (HLD) that manifests in the neonatal period or early infancy [11]. PMLD1 arises in patients with variants in the gap junction protein gamma 2 (GJC2) gene, encoding connexin 47 (Cx47) [4]. Cx47 contributes to gap junction communication channels between oligodendrocytes, the myelinating cells in the central nervous system (CNS), and between oligodendrocytes and astrocytes. Prior cell-line studies have shown that PMLD1 mutants such as p.P87S display defective protein trafficking, intracellular retention in the endoplasmic reticulum and loss of function [4]. In humans with PMLD1, cerebellar clinical signs (ataxia, dysmetria, intention tremor, nystagmus, and dyspraxia) frequently appear during childhood. PMLD1 is inherited in humans as an autosomal recessive trait [11].

HLD in dogs has been comprehensively summarized for hereditary diseases with identified causative gene variants [1]. Specifically, reports have described Pelizaeus–Merzbacher disease (PMD) based on proteolipid protein 1 (PLP1) gene variants [10], hypomyelination based on folliculin-interacting protein 2 (FNIP2) gene variants [12], pontocerebellar hypoplasia based on tRNA splicing endonuclease subunit 54 (TSEN54) gene variants [17], and canine spongiform leukoencephalo-myelopathy based on cytochrome b (CYTB) gene variants [7]. On the other hand, no reports to date have described PMLD1 based on GJC2 gene variants in not only dogs, but also other animals.

Six young Toy Poodles (age range, 87–147 days old) exhibited severe tremors of the head and trunk and ataxia from shortly after birth. Postmortem histopathology following euthanasia confirmed vacuolar degeneration and demyelination in the cerebral and cerebellar white matter, leading to a diagnosis of HLD. Genetic analyses of five of the affected dogs identified a shared homozygous 14-base pair deletion in GJC2, as XM_038556433.1:c.920_933del. The fact that deletions were observed in similar regions in both humans and dogs suggests that the results of this study may provide important insights not only for dogs, but also for congenital HLD in general.

MATERIALS AND METHODS

Animals

A Toy Poodle breeder residing in Japan consulted clinical veterinarians to determine the cause of neurological signs exhibited by two puppies. Subsequently, four additional cases presenting with similar neurological signs were investigated (Table 1). Except that Cases 3 and 4 were littermates, no information regarding blood relationships was disclosed by the breeder or keepers. All six dogs underwent clinical examinations and magnetic resonance imaging (MRI) of the head. Neurological examination could not be performed due to excessive generalized tremors.

Table 1. Case summary.

No.	Sex	Age at MRI (days)	Clinical symptoms	Abnormal blood test results†	Special notes
1	F	147	Tremors throughout the body since shortly after birth Difficulty maintaining standing position Ataxia	WBC: 22,110/μL IP: 7.4 mg/dL CRP: 2.4 mg/dL	CSF examination performed
2	M	142	Tremors throughout the body since shortly after birth Difficulty maintaining standing position Ataxia	Hemoglobin: 9.3 g/dL PCV: 27.9%	Genetic testing not performed
3*	M	109	Tremors throughout the body since shortly after birth Difficulty maintaining standing position Ataxia	IP: 7.9 mg/dL CRP: 3.1 mg/dL
4*	F	109	Tremors throughout the body since shortly after birth Difficulty maintaining standing position Ataxia	Hemoglobin: 12.5 g/dL IP: 7.3 mg/dL CRP: 2.8 mg/dL
5	M	108	Tremors throughout the body since shortly after birth Ataxia	Hemoglobin: 12.4 g/dL PCV: 37.1%	CSF examination performed
6	F	87	Tremors throughout the body since shortly after birth Difficulty maintaining standing position Ataxia	WBC: 26,740/μL Hemoglobin: 12.7 g/dL IP: 8.3 mg/dL CRP: 1.2 mg/dL

* Same womb; ^† WBC, white blood cell count (reference range: 5,050–16,760/μL); IP, inorganic phosphorus (reference range: 2.5–6.8 mg/dL); CRP; C-reactive protein (reference range: 0–1.0 mg/dL); hemoglobin (reference range: 13.1–20.5 g/dL); PCV, packed cell volume (reference range: 37.3–61.7%).

Clinical examinations

Clinical examinations included observation, routine blood tests, serum biochemical profiles and serum electrolyte concentrations.

MRI examination

Routine brain MRI (1.5T HITACHI Echelon Smart; HITACHI, Tokyo, Japan) included the following sequences: T1-weighted imaging (T1WI); T2WI; contrast-enhanced T1WI; and T2-weighted fluid-attenuated inversion recovery (FLAIR). Magnescope^® intravenous (IV) injection 38% syringes (0.3 mL/kg IV; Guerbet, Tokyo, Japan) were used as the contrast agent. MRI was performed under general anesthesia induced with propofol (1 mg/kg repeated injection of “VTRS” propofol 1% IV injection 50 mL; Viatris Healthcare LLC, Tokyo, Japan) and maintained with isoflurane (“VTRS” isoflurane inhalation anesthetic solution; Viatris Healthcare LLC) in >95% oxygen.

Euthanasia

Based on the clinical history, laboratory test results, and MRI findings, the veterinarians clinically diagnosed all cases with HLD. Considering the clinical history, no scope for improvement in clinical signs was expected for any of the animals, and prognosis was determined to be poor from an animal welfare perspective. The breeders and keepers did not wish to have the animals awakened and requested euthanasia, which was performed by administering an excessive dose of propofol while maintaining general anesthesia with isoflurane.

Cerebrospinal fluid (CSF) examination

CSF testing was performed on Dogs 1 and 5 after obtaining informed consent for post-euthanasia CSF collection. A 23-gauge needle attached to a 2.5-mL syringe was used to perform puncture aspiration through the foramen magnum. The CSF was examined for cell count, protein concentration, glucose concentration, and electrolytes.

Pathological examination

Only brain sampling was permitted by the breeder and keeper. Brains were immediately immersed in 10% neutral-buffered formalin, and routine formalin-fixed paraffin-embedded histological examination was performed. Tissue sections were prepared as 4-μm-thick slices following automated processing and paraffin embedding. Sections were stained with hematoxylin and eosin (HE) and luxol fast blue (LFB). Immunohistochemical staining was also performed using anti-glial fibrillary acidic protein (GFAP) antibody (rabbit polyclonal; Dako, Glostrup, Denmark), anti-Olig2 antibody (mouse clone 211F1.1; Merck, Rahway, NJ, USA), and anti-lba-1 antibody (rabbit polyclonal; Wako, Osaka, Japan).

Genetic analyses

Genetic analyses were performed using peripheral blood samples from five individuals, excluding Dog 2, for whom the samples were lost.

We investigated the PLP1, FNIP2, TSEN54, CYTB, and GJC2 genes, which are considered the primary causative genes for congenital HLD. Genomic DNA was extracted using standard methods. For the PLP1 gene, we used primers from Nadon et al. [10] and performed alignment analyses using NM_001013834.2 as the reference sequence. For the other genes, we designed primers using Primer BLAST with the genome assemblies UU_Cfam_GSD_1.0 and Dog10K_Boxer_Tasha (Table 2). Furthermore, alignment analyses were performed using XM_038559114.1 as the reference sequence for the FNIP2 gene, XM_038546615.1 for the TSEN54 gene, XM_038556433.1 for the GJC2 gene, and NC_002008.4 for CYTB. For five individuals with white matter dystrophy and a non-affected control individual (Toy Poodle, female, 11 years old; one example of a blood sample obtained at the veterinary clinic as described below), polymerase chain reaction amplification and direct sequencing were performed on genomic DNA for the coding sequence (CDS) of the PLP1, FNIP2, TSEN54, CYTB, and GJC2 genes. In addition, for GJC2, we compared sequences from wolf (NC_132855.1), cat (NC_058368.1), horse (NC_091697), cow (NC_037334.1), sheep (NC_056058.1), and goat (NC_030814.1).

Table 2. Primers used for analysis of folliculin-interacting protein 2, tRNA splicing endonuclease subunit 54, cytochrome B, and gap junction protein gamma 2 genes.

No.	Primer name	Sequence (5′-3′)	Reference assembly	Chr	Start	End	Length (bp)	Tm (°C)	Amplicon (bp)
Folliculin-interacting protein 2 (FNIP2)
1	FNIP-1_F	CAGCATCATGGCCCCGAC	UU_Cfam_GSD_1.0	15	56,225,191	56,225,208	18	60.9	129
2	FNIP-1_R	GCTCACCTAAACGCGGCTC	UU_Cfam_GSD_1.0	15	56,225,301	56,225,319	19	61.5
3	FNIP-2_F	GACCCAGTTTGACCGCAGGATCCAG	UU_Cfam_GSD_1.0	15	56,226,080	56,226,104	25	67.3	145
4	FNIP-2_R	GGACAGGGCTTAAACCTCTCCAC	UU_Cfam_GSD_1.0	15	56,226,202	56,226,224	23	62.8
5	FNIP-3_F	GTCCCGAGTCCATGTTGGTAGTCAGG	UU_Cfam_GSD_1.0	15	56,266,801	56,266,826	26	66.3	376
6	FNIP-3_R	GCTACACACCCACCAACCACCCCATG	UU_Cfam_GSD_1.0	15	56,267,151	56,267,176	26	69.4
7	FNIP-4_F	CCATGATCATTTCAAGTGTTGTGG	UU_Cfam_GSD_1.0	15	56,292,798	56,292,821	24	58.8	246
8	FNIP-4_R	GCTCAGAGATATGGTTAACATCAC	UU_Cfam_GSD_1.0	15	56,293,020	56,293,043	24	57.2
9	FNIP-5_F	CCAGTGAAGAGCGTGCAGAGGAAG	UU_Cfam_GSD_1.0	15	56,296,138	56,296,161	24	65.4	286
10	FNIP-5_R	GCCTCAAGCAGGTACTCTGCCCTGC	UU_Cfam_GSD_1.0	15	56,296,399	56,296,423	25	69.2
11	FNIP-6_F	CTGTTGCACTGCACATCATGTTGC	UU_Cfam_GSD_1.0	15	56,298,863	56,298,886	24	63.9	421
12	FNIP-6_R	CTACTCTATCCAGAAACAGCCAC	UU_Cfam_GSD_1.0	15	56,299,261	56,299,283	23	58.3
13	FNIP-7_F	GTCTTCCTTCCTTCCCCTCTCCGG	UU_Cfam_GSD_1.0	15	56,300,987	56,301,010	24	65.8	441
14	FNIP-7_R	GTGGGTCACACTGTACTCAGCATTC	UU_Cfam_GSD_1.0	15	56,301,403	56,301,427	25	63.5
15	FNIP-8_F	CCAGTCATCAATAGTGCTTGACTCC	UU_Cfam_GSD_1.0	15	56,302,828	56,302,852	25	61.4	254
16	FNIP-8_R	CCACCTACTTCACCATTACAGATG	UU_Cfam_GSD_1.0	15	56,303,058	56,303,081	24	58.9
17	FNIP-9_F	GAGCCTGTGTAGGCACGTGAGGATTC	UU_Cfam_GSD_1.0	15	56,316,421	56,316,446	26	67.0	282
18	FNIP-9_R	CCATTGACAACAGGTGACTACCACCAG	UU_Cfam_GSD_1.0	15	56,316,676	56,316,702	27	65.3
19	FNIP-10_F	CCTCGATGTGATTTGTTGATGCATGG	UU_Cfam_GSD_1.0	15	56,323,138	56,323,163	26	63.1	746
20	FNIP-10_R	GTCTCTGTGTATCCCTAGATCCAAGC	UU_Cfam_GSD_1.0	15	56,323,858	56,323,883	26	62.2
21	FNIP-11_F	GCTTGAATCTCTTAAGGGCTAC	UU_Cfam_GSD_1.0	15	56,325,162	56,325,183	22	56.6	603
22	FNIP-11_R	CTCCATTCCATCCCACAGAGTCTCTG	UU_Cfam_GSD_1.0	15	56,325,739	56,325,764	26	64.3
23	FNIP-12_F	GCTGCCATATCCATGCCATGCTTG	UU_Cfam_GSD_1.0	15	56,331,820	56,331,843	24	64.9	1,496
24	FNIP-12_R	CTAACGTGCTCCTGAGTGCTAAGTCAGG	UU_Cfam_GSD_1.0	15	56,333,288	56,333,315	28	66.5
25	FNIPseq-12_F	GAAGCTCAGCCAGCAGGCTGTCAG	UU_Cfam_GSD_1.0	15	56,332,446	56,332,469	24	67.7	-
26	FNIPseq-12_R	CAGGTGTGAAGAAGCACCTAG	UU_Cfam_GSD_1.0	15	56,332,736	56,332,756	21	58.3	-
27	FNIP-13_F	GGTGAAACGGACCACCCTGAACCTC	UU_Cfam_GSD_1.0	15	56,334,228	56,334,252	25	67.2	316
28	FNIP-13_R	GTCAGCAGCCAGAGGCTCCAAGCTG	UU_Cfam_GSD_1.0	15	56,334,519	56,334,543	25	69.4
29	FNIP-14_F	CTTGAGTGCCACTTGGCAGTTC	UU_Cfam_GSD_1.0	15	56,348,037	56,348,058	22	62.2	281
30	FNIP-14_R	GGTGATATGAGAACAAGTAC	UU_Cfam_GSD_1.0	15	56,348,298	56,348,317	20	50.7
31	FNIP-15_F	GGTAGAGATTTGAATCCTGGCCCTC	UU_Cfam_GSD_1.0	15	56,351,703	56,351,727	25	62.7	213
32	FNIP-15_R	GGTCAGTCCTATCTCAGGGAGAAAC	UU_Cfam_GSD_1.0	15	56,351,891	56,351,915	25	61.9
33	FNIP-16_F	GATCTGGCAAGCTGAGTACCTAG	UU_Cfam_GSD_1.0	15	56,353,776	56,353,798	23	60.2	194
34	FNIP-16_R	CTTCCAAGGAAGAACTGTCCTGAGC	UU_Cfam_GSD_1.0	15	56,353,945	56,353,969	25	63.3
tRNA splicing endonuclease subunit 54 (TSEN54)
35	TSEN-1_F	CCAGAGCGGTGTCATTAGTGTTC	UU_Cfam_GSD_1.0	9	5,890,412	5,890,434	23	61.5	201
36	TSEN-1_R	CTGTGCCTACCTCTCTGTCTAAC	UU_Cfam_GSD_1.0	9	5,890,290	5,890,612	23	59.9
37	TSEN-2_F	GTAGGGTCACTGGTCTCGTACCCAC	UU_Cfam_GSD_1.0	9	5,891,071	5,891,095	25	65.8	570
38	TSEN-2_R	CACTGTCCTTGCAGCCATAGCCCTTC	UU_Cfam_GSD_1.0	9	5,891,615	5,891,640	26	67.5
39	TSEN-3_F	CTGTGGGGCTTTGGCTGGGTGAG	UU_Cfam_GSD_1.0	9	5,892,236	5,892,258	23	67.8	1,032
40	TSEN-3_R	GGTCTGTTGCCCACCCAATTCTC	UU_Cfam_GSD_1.0	9	5,893,245	5,893,267	23	63.8
41	TSENseq-3_F	CTATCTGAAGCCATGTTGGGGAAG	UU_Cfam_GSD_1.0	9	5,892,594	5,892,617	24	61.2	-
42	TSENseq-3_R	GATGCCCAGAGGAGAAACCTCAG	UU_Cfam_GSD_1.0	9	5,892,785	5,892,807	23	62.6	-
43	TSEN-4_F	GGTCCCAGGACTGGAACAAATGG	UU_Cfam_GSD_1.0	9	5,894,906	5,894,928	23	63.3	138
44	TSEN-4_R	CAAAGGTGGACCTCTGATCTTTAC	UU_Cfam_GSD_1.0	9	5,895,020	5,895,043	24	59.1
45	TSEN-5_F	CACCTGATTCAATGGATGGGTGGCTC	UU_Cfam_GSD_1.0	9	5,895,957	5,895,982	26	65.7	184
46	TSEN-5_R	GAACACAAGCTGAGTGGTTGGCCTC	UU_Cfam_GSD_1.0	9	5,896,116	5,896,140	25	66.1
47	TSEN-6_F	CTTTCAGGTCCTTGCTCCCAACCCTTC	UU_Cfam_GSD_1.0	9	5,896,359	5,896,385	27	67.2	726
48	TSEN-6_R	GATGGAGCCCGAGCTCGAG	UU_Cfam_GSD_1.0	9	5,897,066	5,897,084	19	62.2
Cytochrome b (CYTB)
49	CYTB_F	CCAACATTCGAAAAACCCACCCAC	Dog10K_Boxer_Tasha	MT	14,187	14,210	24	63.1	1,125
50	CYTB_R	GAAGGTTGTTTTCGATAACGCTAAC	Dog10K_Boxer_Tasha	MT	15,287	15,311	25	59.2
51	CYTBseq_F	CAGCTCTAGCAATAGTACACCTCC	Dog10K_Boxer_Tasha	MT	14,751	14,774	24	60.3	-
52	CYTBseq_R	GGTTTGCAGGGGTGTAGTTATCTG	Dog10K_Boxer_Tasha	MT	14,940	14,963	24	61.4	-
Gap junction protein gamma 2 (GJC2)
53	GJC2_F	GTCAGATCCACACCGTGGTCTTG	UU_Cfam_GSD_1.0	14	489,673	489,695	23	63.3	1,309
54	GJC2_R	CCTATGACCAACATGAGCTGGAG	UU_Cfam_GSD_1.0	14	490,959	490,981	23	60.7
55	GJC2seq_F	CGAAGCCGTACAGCAGGTAC	UU_Cfam_GSD_1.0	14	490,258	490,277	20	60.8	-
56	GJC2seq_R	CTGATGCGCGTGTACGTG	UU_Cfam_GSD_1.0	14	490,328	490,345	18	59.3	-

For the genetic analyses, dogs that did not exhibit the clinical neurological signs seen in our cases were defined as “normal dogs”. Genetic data regarding the GJC2 gene in Poodles as “normal dogs” were provided by the Japan Kennel Club (JKC) (Tokyo, Japan). To ensure that no individual from the same breeder was included more than once within a 1-year period, a maximum of one individual was selected from each breeder. A total of 684 dogs were included in the study, selected at random from data covering a 2-year period.

In addition, data for 10 dogs (6 males, 4 females; 2–12 years old) were obtained from surplus ethylenediaminetetraacetic acid-treated blood samples collected during blood tests performed in conjunction with the prescription of heartworm preventive medication at the veterinary hospital where the first author (YH) and co-authors (RO, MO, JO) work. Prior to using the blood samples, the aforementioned veterinarians explained the purpose of this study to the owners and obtained their understanding and consent before receiving the blood samples.

RESULTS

Clinical diagnosis

All six Toy Poodles exhibited severe tremors, characterized by intense vertical shaking from the head to the torso. Although all dogs were able to raise the tail during defecation, the presence of tremor localized to the tail was not able to be determined. These tremors were observed even during rest periods, such as when confined in cages, but became more pronounced during activity. Neurological examinations could not be performed due to the severity of the tremors, and nystagmus could not be confirmed. Table 1 lists the abnormalities identified from blood testing and blood biochemistry.

MRI examinations

Brain MRI revealed signal hyperintensity compared to the cortex throughout the subcortical white matter on T2WI and FLAIR images in all affected dogs (Fig. 1A–D). The corpus callosum and cerebellar peduncles appeared isointense with gray matter. On the midsagittal view, the contrast between gray matter and white matter in the cerebellum was indistinct, and cerebellar dendritic structures could not be identified (Fig. 1E). Results of pre- and post-contrast T1WI appeared normal. No findings of ventricular enlargement were observed.

Fig. 1.

Findings from magnetic resonance imaging (MRI) of the brain. A, B) Brain MRI reveals increased signal intensity compared to the cortex throughout the subcortical white matter on T2-weighted imaging (T2WI) (A) and T2-weighted fluid-attenuated inversion recovery (FLAIR) (B) images in all affected dogs. The corpus callosum shows signal isointensity with gray matter. Transverse-section images at the level of the interthalamic bridge (A–D). E) On the midsagittal view, contrast between gray and white matter in the cerebellum is indistinct, and cerebellar dendritic structures cannot be identified. Results of pre- and post-contrast T1WI appear normal (C and D). No findings of ventricular enlargement are observed.

CSF examination

Concentrations of protein, glucose, and electrolytes in CSF samples collected from Dogs 1 and 5 were all within normal ranges.

Histopathological examination

Histopathological examination revealed thinning of the white matter throughout the cerebrum in HE-stained specimens. In addition, vacuoles were scattered throughout the periventricular white matter. LFB staining revealed reduced myelin formation throughout the cerebral white matter (Fig. 2). However, immunohistochemical staining using anti-GFAP, anti-Olig2, and anti-Iba-1 antibodies showed no significant change in glial cells. Furthermore, all six cases showed a correlation between myelin formation deficiency in the white matter and neurological signs. The consistent pathological findings observed in all six cases were determined to be HLD, a hereditary disorder.

Fig. 2.

Histopathological findings. A) Reduced myelin formation in the cerebral white matter. Luxol fast blue (LFB)−hematoxylin and eosin (HE) stain, bar=250 μm. B) Higher magnification of cerebral white matter. LFB-HE stain, bar=50 μm. C) Vacuolar changes in cerebral white matter. HE stain, bar=50 μm.

Genetic analysis

No nucleotide substitutions were identified in the PLP1 gene. For the FNIP2 gene, three base substitutions were identified in the CDS (Table 3). However, since no common base substitutions were detected among the five affected individuals, no association between the FNIP2 gene and the affected individuals could be confirmed. For the TSEN54 gene, one base substitution was identified (Table 3). However, the TSEN54 gene substitution (XM_038546615.1: c.618C >T) was a silent variant not resulting in an amino acid substitution. For the CYTB gene, four substitutions and four haplotypes were identified (Table 3). However, no common substitutions were detected among the five affected individuals, so no association between the CYTB gene and the affected individuals could be confirmed. Sequencing and alignment of the GJC2 gene exons revealed a 14-bp deletion (XM_038556433.1:c.920_933del) common to all five affected individuals, present in a homozygous state in each case (Fig. 3). The deletion was predicted to result in a premature stop codon, XP_038412362.1:p. (Ala307GlyfsTer115). Using data from 684 Poodles provided by the JKC as a control group, we investigated the presence of the newly discovered GJC2 gene variant (XM_038556433.1:c.920_933del). This variant was not detected in 671 dogs (98.1%). The remaining 13 dogs (1.90%) were carriers, and no homozygous affected dogs. The variant allele frequency (VAF) in the 684 dogs was therefore 0.0096, and the disease prevalence was estimated to be 0.009% (approximately 1 in 11,000 dogs).

Table 3. Base substitutions detected in folliculin-interacting protein 2 (FNIP2), tRNA splicing endonuclease subunit 54 (TSEN54), cytochrome b (CYTB) gene.

No.	FNIP2			TSEN54	CYTB
	c.75	c.121	c.195	c.618	c.489	c.748	c.901	c.1032
1	T/T	A/A	A/A	C/T	G	T	C	A
2	C/T	A/A	A/A	T/T	G	T	T	G
3	C/T	A/A	A/A	T/T	G	T	T	G
4	C/T	A/A	A/A	C/T	G	T	T	A
5	C/C	A/A	A/A	T/T	A	C	T	A
Non-affected control dog*	C/C	A/A	A/G	T/T	G	T	T	G
Reference sequence	C	T	G	C	G	T	T	G

*For genetic analyses, dogs that do not exhibit neurological clinical signs, such as those in our cases, are defined as "non-affected control dog". Among non-affected control dogs, Table 3 shows results for one Toy Poodle (female, 11 years old) that donated blood at a veterinary clinic.

Fig. 3.

Details of the gap junction protein gamma 2 (GJC2) frameshift deletion. Sequencing and alignment of the GJC2 gene exons reveals a 14-bp deletion (c.920_933del) common to all five of the affected individuals tested, and present in a homozygous state. Each asterisk indicates a match between affected dogs and a non-affected control dog.

Using samples from 10 normal Poodles provided by the animal hospital, we also investigated the presence of the newly discovered GJC2 gene variant (XM_038556433.1:c.920_933del). This variant was not detected in any of the 10 normal Poodles.

Comparing the GJC2 gene using reference sequences from wolves, cats, horses, cattle, sheep, and goats, no 14-bp deletions were observed in the surrounding sequences in these species (Fig. 4).

Fig. 4.

Sequence comparison around the deletion site of the gap junction protein gamma 2 (GJC2) gene in multiple animal species. Comparisons of the GJC2 gene using reference sequences from dog (Line 2), wolf (Line 3), cat (Line 4), horse (Line 5), cow (Line 6), sheep (Line 7), and goat (Line 8) reveal that the 14-base pair deletion (XM_038556433.1:c.920_933del) observed in the present case (Line 1) is not present in sequences from the other species.

DISCUSSION

HLD is fundamentally due to genetic variants, although cases in which demyelination was caused by autoimmune factors or infections have also been confirmed [6]. In our six cases, MRI examinations first ruled out malformations such as hydrocephalus and revealed MRI findings characteristic of human HLD, including increased signal intensity throughout the subcortical white matter on T2WI and FLAIR images. Second, these dogs of the same breed presented with neurological signs in common immediately after birth, leading to the clinical assumption of congenital HLD. Histopathological examination of the brains obtained on necropsy revealed common myelin formation disorders in the white matter of all six dogs. However, no abnormalities were observed in the numbers of glial cells, and immunomediated inflammatory diseases such as meningoencephalitis of unknown origin were also ruled out. CSF examination showed no signs suggestive of inflammatory disease. Based on these findings, genetic analyses were performed on strong suspicion of HLD with a genetic background.

Approximately 90% of human congenital HLD cases are PMD, caused by genetic abnormalities in PLP1, a crucial protein in myelin, and are distinguished from other congenital HLDs [16]. PMD, positioned as the representative form of congenital HLD, is an X-linked recessive disorder that is therefore primarily recognized as a congenital disease affecting male children. Consequently, in humans, patients with the PMD phenotype but without variants in the PLP1 gene are considered to have PMLD1 [19].

PMLD1 is a typical slowly progressive HLD that manifests in the neonatal period or early infancy. Inheritance in humans is considered to show an autosomal recessive pattern, with genetic variants in GJC2 gene often identified as the causative factor [11]. GJC2 is expressed almost exclusively in oligodendrocytes within the brain, encodes gap junction protein gamma 2, which has previsouly also termed Cx47 and is implicated in myelin formation [9]. Tress et al. (2011) introduced a missense variant identified from human PMLD1 into the mouse genome to investigate the effects of a variant in GJC2 resulting in an alteration to the GJC2 protein. However, mice harboring the human missense mutation exhibited only mild myelin abnormalities. This was attributed to differences in the proportion of white matter between humans and mice, and the results support the notion that loss of GJC2 channel function causes PMLD1 [18]. These findings suggest that loss of connexin function and dysfunction of the GJC2 gap junction channel may cause hypomyelination and contribute to the PMLD1 phenotype [2, 18].

One method used to assess the pathogenicity of variants in humans is the guideline proposed by the American College of Medical Genetics (ACMG), the Association of Molecular Pathology (AMP), and the College of American Pathologists (hereinafter referred to as the ACMG/AMP guideline) [14]. Using the ACMG/AMP guideline, the pathogenicity of a variant is first evaluated using the “Criteria for Classifying Pathogenic Variants” to determine whether the case meets the criteria for “Very strong evidence of pathogenicity” (PVS: PVS1), “Strong evidence of pathogenicity” (PS: PS1–4), “Moderate evidence of pathogenicity” (PM: PM1–6), or “Supporting evidence of pathogenicity” (PP: PP1–5). Next, based on the combination of criteria met, the “Rules for Combining Criteria to Classify Sequence Variants” are applied to ultimately categorize the variant into one of five categories to determine pathogenicity: “Pathogenic”, “Likely pathogenic”, “Uncertain significance”, “Likely benign”, or “Benign”. Through these methods, the ACMG/AMP guidelines enable evaluations that balance objectivity and accuracy [14].

In humans, “p. Val298fs” (NM_020435.4: c.893_942del) and “p. Pro305fs” (NM_020435.4: c.914_947del) [15, 20] are registered in ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/) as causative mutations for PMLD1, and the pathogenicity, as assessed by the ACMG/AMP guidelines, is classified as “Likely pathogenic” for both. The variant in the canine GJC2 gene identified in this study (XM_038556433.1:c.920_933del) is located at a nucleotide position very close to the human GJC2 gene variants “p.Val298fs” and “p.Pro305fs,” and is considered to fall under the definition of a PVS1: “a null variant in a gene where loss of function is a known mechanism of the disease.” Furthermore, the VAF in the 684 dogs was 0.0096, and the disease prevalence was estimated as 0.009%. Consequently, this genetic variant falls under the PM2 category (not present in the control group, or present at an extremely low frequency). As a result, (XM_038556433.1:c.920_933del) is considered to meet the PVS1 and PM2 criteria. Applying the “Rules for Combining Criteria to Classify Sequence Variants” from the ACMG/AMP guideline to assess pathogenicity, the variant was classified as “likely pathogenic.”

Kiktev et al. (2018) reported that an increase in GC content leads to a higher mutation rate, and they proposed that this is likely due to the possibility of deletion/duplication mutations arising from slippage caused by error-prone DNA polymerases [5]. Furthermore, Ellegren et al. (2004) reported that replication slippage is more likely to occur in sequences with an abnormally high GC content, even though they are microsatellites [3]. The average GC content of the entire human genome has been reported as 40.89% [13]. Furthermore, Lindblad-Toh et al. reported in 2005 that the GC content in the canine genome is very similar to that in the human genome [8]. Therefore, the GC content in dogs is presumed to be comparable to that in humans. The GC content of the 14-bp deletion in the GJC2 gene detected in this study (XM_038556433.1:c.920_933del) and the 50-bp flanking regions (totaling 114 bp) was 89.47%. Mutations are therefore likely to occur in the 14-bp deletion in the GJC2 gene identified in this study and its surrounding region, and these mutations could plausibly result from slippage.

Except for the experimental report by Tress et al. (2011) on mice with missense variants in the human PMLD1 gene, to the best of our knowledge, no variants in the GJC2 gene causing PMLD1 have been reported in any mammal. Furthermore, variants in the human GJC2 gene are listed in ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/?gene=GJC2&disease=Hypomyelinating+ leuko- dystrophy +2&term=%22GJC2%22% 5BGENE%5D+ AND+%22Hypomyelinating+leukodystrophy +2%22%5BDIS%5D), with 20 classified as “pathogenic” and 32 as “highly pathogenic.”

Comparison of the GJC2 gene using reference sequences from wolves, cats, horses, cattle, sheep, and goats revealed no 14-basepair deletion in the flanking regions from these species (Fig. 4). This further supports the conclusion that this deletion is not present in normal GJC2 genes, including those of other animal species, and that the conditions of the juvenile Toy Poodles examined in this study were consistent with reports in humans and highly likely to be associated with PMLD1, a congenital HLD.

A key limitation of this study was that genetic analyses were performed only on five of the six affected individuals. Testing data from related individuals such as parents or littermates were unable to be obtained, making a pedigree impossible to construct.

The findings of this study will allow genetic analysis of HLD associated with GJC2 gene variants in dogs. Furthermore, although no recessive homozygotes (affected dogs) were identified among the GJC2 genetic data provided by the JKC from 684 dogs, the presence of 13 carriers is of great significance in terms of the spread of the variant gene and the onset of this disease. Going forward, breeders should incorporate testing for this disease into the genetic screening protocols for breeding animals and need to strictly avoid high-risk matings (between carriers).

This represents the first reported pathogenic GJC2 variant in domestic animals. Congenital HLD in dogs is a lethal hereditary disease, but case reports remain limited, and many aspects remain unclear. The deletion detected in the GJC2 gene (XM_038556433.1:c.920_933del) in this study represents one of the few reported as a causative gene variant in canine congenital HLD. Furthermore, the fact that this deletion was observed at a similar location in both humans and dogs suggests that this study provides important insights not only for dogs, but also for congenital HLD in general.

Conflicts of Interest

The authors declare that they have no conflict of interest.