Genetic Link Across Species: SIX6, a Major Human Glaucoma Gene, Confers Susceptibility to Glaucoma in Shiba-Inu Dogs

Abstract

Purpose

The SIX6 gene is a major susceptibility gene for human glaucoma, and its variants have been implicated in structural and functional alterations of the retina, such as nerve fiber layer thinning and retinal ganglion cell degeneration, which contribute to glaucoma development. This study investigated whether SIX6 polymorphisms are associated with glaucoma in dogs, focusing on the Shiba-Inu and Shih-Tzu breeds.

Methods

We genotyped 19 single nucleotide polymorphisms within the SIX6 gene region in 109 Shiba-Inus (49 cases and 60 controls) and 57 Shih-Tzus (18 cases and 39 controls), followed by association analyses. Comparative sequence analysis of canine and human SIX6 was performed to assess evolutionary conservation.

Results

In Shiba-Inus, rs851962234, located in the 3′-untranslated region of SIX6, was significantly associated with glaucoma (P = 0.0047; Pc = 0.038), with the minor A allele showing a frequency of 17.3% in cases and 4.2% in controls, and conferring an increased risk (odds ratio, 3.56). In contrast, rs851962234 showed no association in Shih-Tzus, likely owing to the rarity of the A allele in this breed. No other single nucleotide polymorphisms were associated with glaucoma in either breed. A comparative analysis showed a 98.4% amino acid identity between canine and human SIX6 across the coding regions.

Conclusions

This study identifies rs851962234 in canine SIX6 as being significantly associated with glaucoma in Shiba-Inus, suggesting its potential role in modulating SIX6 expression and retinal ganglion cell vulnerability. These findings offer new insight into the genetic basis of canine glaucoma and highlight parallels and distinctions with human disease.

Glaucoma is a degenerative optic neuropathy that encompasses a spectrum of eye disorders characterized by visual field defects, progressive loss of retinal ganglion cells (RGCs), and degeneration of optic nerve axons, and is often accompanied by an elevated IOP.^1,2 Clinically, glaucoma is classified into three major types: POAG, primary angle-closure glaucoma (PACG), and primary congenital glaucoma.^3–5 Among these, POAG, which includes normal tension glaucoma (NTG) as a subtype, is the most common type of glaucoma in humans.⁶

Canine glaucoma, like its human counterpart, ultimately results in irreversible vision loss owing to optic nerve damage and is broadly classified as either primary or secondary.^7–10 Primary glaucoma typically results from inherited anatomical abnormalities that impair aqueous humor drainage, independent of other ocular diseases, and often affects both eyes, although onset may occur at different times.^7,10 Secondary glaucoma develops as a consequence of other ocular disorders—such as lens displacement, uveitis, hyphema, intraocular neoplasia, or complications after cataract surgery or ocular trauma—which obstruct the aqueous humor outflow pathways.^9,10 Shiba-Inus have the highest incidence of glaucoma in Japan, which is attributed to their high rate of abnormal or narrow iridocorneal angles (ICAs), observed in both glaucomatous and nonglaucomatous eyes compared with other breeds.^11,12 Similarly, Shih-Tzus, which have the second highest incidence of glaucoma in Japan, are also prone to such anatomical abnormalities.¹² Although abnormal ICAs increase the risk of glaucoma, there are many Shiba-Inus and Shih-Tzus with such abnormalities who do not develop the disease. This finding suggests that additional genetic or environmental factors are involved in the risk of glaucoma in these breeds.

Glaucoma is a multifactorial disease, and both environmental and genetic factors are thought to contribute to the development of glaucoma.¹³ In the past several years, genome-wide association studies (GWASs) have identified many susceptibility loci for glaucoma in humans. As of July 11, 2025, the NHGRI-EBI GWAS Catalog (https://www.ebi.ac.uk/gwas/),¹⁴ a publicly available database of published GWASs, had reported more than 400 candidate loci with genome-wide significance (P < 5.0 × 10⁻⁸) for human glaucoma, including POAG, PACG, and NTG. In dogs as well, GWASs have identified several candidate genes for glaucoma across breeds. For POAG, ADAMTS10 and ADAMTS17 have been reported as candidate genes in Beagles and Petit Basset Griffon Vendéen dogs, respectively.^15,16 For PACG, loci on chromosomes 14, 24, and 37 have been identified in Basset Hounds, with COL1A2 and RAB22A among the proposed candidate genes.^17,18 In Dandie Dinmont Terriers and American Cocker Spaniels, glaucoma with PACG-like anatomy has been linked to specific loci on chromosomes 8 and 10.^19,20 In Border Collies, OLFML3 has been suggested as a candidate gene for severe goniodysgenesis and secondary glaucoma.²¹ Despite these advances, the genetic architecture of canine glaucoma remains incompletely understood, and further studies are needed to elucidate the underlying mechanisms across breeds.

Interestingly, SRBD1, originally identified as a susceptibility gene for NTG in humans through GWAS,²² has also been implicated in canine glaucoma. In our previous study, we showed that SRBD1 polymorphisms are associated with glaucoma risk in dogs as well, based on a genetic analysis of Shiba-Inus and Shih-Tzus.²³ These findings suggest that humans and dogs may share common genetic pathways underlying glaucoma susceptibility.

Among the numerous susceptibility genes identified for human glaucoma, SIX6 has emerged as one of the most robust and consistently associated genes for POAG across ethnically diverse populations,^24–32 and has also been implicated in NTG, particularly in East Asian populations.^30,33–35 SIX6 belongs to the sine oculis homeobox (SIX) gene family and encodes transcription factors involved in ocular development and morphogenesis.^36–39 Accumulating evidence suggests that glaucoma-associated SIX6 risk variants disrupt RGC development, promote their degeneration, and induce premature cellular senescence, thereby contributing to the structural and cellular basis of glaucomatous optic neuropathy.^27,40–44 Consequently, SIX6 has become a focal point in the investigation of glaucoma pathogenesis at the genetic level.

Building on this background, we conducted a comprehensive association analysis between glaucoma and genetic polymorphisms within the SIX6 gene in dogs, using the Shiba-Inu and Shih-Tzu breeds, to assess whether canine SIX6 contributes to glaucoma susceptibility.

Methods

Participants

A total of 109 Japanese Shiba-Inu dogs and 57 Shih-Tzu dogs were recruited from the Veterinary Teaching Hospital of Azabu University. Each dog underwent a comprehensive ophthalmologic evaluation, which included examinations using a hand-held slit-lamp biomicroscope (SL-14; Kowa, Tokyo, Japan), indirect ophthalmoscopy, and IOP measurement. Topical anesthesia (oxybuprocaine hydrochloride, Santen, Osaka, Japan) was applied prior to tonometry, which was performed using the Tono-Pen XL (Mentor O&O, Norwell, MA, USA). The diagnosis of glaucoma required elevated IOP (>25 mm Hg in at least one eye) together with at least one characteristic clinical sign, such as corneal edema, conjunctival hyperemia, visual impairment or vision loss, cupping or atrophy of the optic nerve head (ONH), or globe enlargement (buphthalmos). All diagnoses were confirmed by Diplomates of the Japanese College of Veterinary Ophthalmologists with expertise in canine glaucoma. Of the enrolled dogs, 49 Shiba-Inus and 18 Shih-Tzus were diagnosed with glaucoma, while the remaining 60 Shiba-Inus and 39 Shih-Tzus were classified as normal (<25 mm Hg IOP and absence of the above glaucomatous signs) (Table 1). Because glaucoma typically develops later in life, only dogs aged 4 years or older were included in the control group to minimize the likelihood of enrolling potential glaucomatous dogs.

Table 1.

Demographic Characteristics of Glaucoma Cases and Controls in Shiba-Inus and Shih-Tzus

	Shiba-Inus			Shih-Tzus
Characteristic	Glaucoma Cases	Controls	P Value	Glaucoma Cases	Controls	P Value
No. of subjects	49	60		18	39
Mean age, years	8.7 ± 3.0	9.3 ± 3.3	0.29	8.8 ± 2.1	10.7 ± 2.6	0.0098
Male/female, %	32.7/67.3	36.7/63.3	0.66	50.0/50.0	69.2/30.8	0.16

Differences in age were evaluated using the Mann–Whitney U test, and differences in sex distribution were assessed using the χ² test.

This study was conducted as part of a research project approved by the Ethical Committee of Azabu University (Permit Number: 110408-2). Written informed consent was obtained from the owner of each dog. All procedures were carried out in accordance with the Guide for the Care and Use of Laboratory Animals of Azabu University.

Single Nucleotide Polymorphism (SNP) Genotyping

Genomic DNA was extracted from peripheral blood samples of glaucomatous and normal dogs using a DNA Whole Blood Spin Kit (Fuji Film, Tokyo, Japan). DNA purity and concentration were assessed with GeneQuant Pro spectrophotometer (GE Healthcare, Cambridge, UK).

To comprehensively investigate the association between canine glaucoma and SIX6, we analyzed all 19 SNPs located within the SIX6 gene region as listed in the Ensembl database (CanFam3.1 assembly) (Table 2). SNP genotyping was performed using the TaqMan 5′ exonuclease assay and primer–probe sets supplied by Thermo Fisher Scientific (Foster City, CA, USA). The primer and probe sequences of the TaqMan SNP Genotyping Assays are proprietary to Thermo Fisher Scientific and are not publicly available. PCRs for each SNP were carried out in 10 µL reaction mixture containing 1 × TaqMan GTXpress Master Mix (Thermo Fisher Scientific), 1 × TaqMan SNP Genotyping Assay primer/probe mix, and 3 ng of genomic DNA. The PCR conditions were as follows: initial denaturation at 95°C for 20 seconds, followed by 40 cycles of denaturation at 95°C for 3 seconds and annealing/extension at 60°C for 20 seconds. Fluorescent probe signals were detected using the StepOnePlus Real-Time PCR System (Thermo Fisher Scientific), according to the manufacturer's instructions.

Table 2.

List of 19 SNPs in the SIX6 Gene Region Analyzed in This Study

					HWE in Controls
SNP	Chr	Position (CanFam3.1)	Gene Location	Alleles	Shiba-Inus	Shih-Tzus
rs24517935	8	35,561,567	5′-upstream	A/G	1.000	0.669
rs8988460	8	35,561,752	5′-upstream	C/T	0.201	0.478
rs24517937	8	35,562,417	5′-upstream	A/G	0.201	0.478
rs24488773	8	35,563,129	5′-upstream	C/T	0.201	0.478
rs8816203	8	35,565,492	5′-upstream	G/T	0.087	0.724
rs851613897	8	35,565,744	5′-upstream	A/G	1.000	0.669
rs852129966	8	35,566,092	Exon (synonymous)	C/T	0.00018	0.669
rs8816199	8	35,567,179	Intron	A/G	0.201	0.574
rs851839004	8	35,567,346	Intron	C/T	0.201	0.574
rs851859199	8	35,567,499	Intron	A/G	0.201	0.574
rs850948063	8	35,567,616	Intron	G/T	0.397	0.106
rs851228774	8	35,567,739	Exon (synonymous)	A/G	1.000	1.000
rs851962234	8	35,568,078	3′-UTR	A/G	0.736	0.935
rs24543712	8	35,568,765	3′-downstream	C/G	0.172	0.574
rs24543710	8	35,568,906	3′-downstream	C/T	0.172	0.574
rs24543694	8	35,569,842	3′-downstream	A/G	0.172	0.574
rs24543691	8	35,570,513	3′-downstream	C/T	0.727	0.106
rs853179248	8	35,572,116	3′-downstream	C/T	0.389	0.538
rs24542512	8	35,572,911	3′-downstream	C/T	0.693	0.785

Statistical Analysis

Hardy–Weinberg equilibrium (HWE) testing, association analysis, and linkage disequilibrium (LD) estimation were performed using SNP & Variation Suite software (version 8.8.3, Golden Helix, Bozeman, MT, USA). Association analyses were carried out under additive models with age and sex as covariates. Meta-analyses combining the Shiba-Inu and Shih-Tzu cohorts were conducted using the inverse-variance method under a random-effects model with the META software (https://mathgen.stats.ox.ac.uk/genetics_software/meta/meta.html), which accounts for heterogeneity in effect sizes (odds ratios [OR]) between cohorts. The obtained P values were corrected for multiple testing using the Bonferroni method, based on the number of SNPs remaining after LD pruning (r² > 0.80), HWE filtering (P < 0.05), and minor allele frequency filtering (<0.01) among the 19 analyzed SNPs (n = 8 for both Shiba-Inus and Shih-Tzus). LD pruning was conducted using the Tagger program in Haploview 4.2 software.⁴⁵ A corrected P (Pc) value of less than 0.05 was considered significant.

Homology Analysis

The nucleotide sequences of the SIX6 gene in dogs and humans were retrieved from the Ensembl Genome Browser (https://www.ensembl.org). The dog SIX6 gene was identified by Ensembl ID ENSCAFG00845022479, and the human gene by ENSG00000184302. Each gene was divided into regions corresponding to exons, introns, and the 3′-untranslated region (UTR). Because the annotated human 3′-UTR was approximately 1000 bp longer than that of the dog, only the dog-defined 3′-UTR segment was used for interspecies comparison (see Table 4 footnote). This approach ensured that the sequences compared corresponded to orthologous regions, minimizing bias owing to species-specific UTR elongation. Pairwise alignments of each corresponding region between dog and human were performed using the “Blast 2 Sequences” tool⁴⁶ of the NCBI BLASTN program (https://blast.ncbi.nlm.nih.gov), with the following parameters: E-value threshold = 0.05, word size = 16, match/mismatch scores = 2/–3, and gap costs = existence 5, extension 2. Percentage identity was recorded as the primary measure of sequence conservation.

Table 4.

Nucleotide and Amino Acid Sequence Identity of the SIX6 Gene Between Dog and Human

	Dog (CanFam3.1)		Human (GRCh38)		Identity (%)
SIX6 Region	Position on Chr. 8	Length (bp)	Position on Chr. 14	Length (bp)	Nucleotide	Amino Acid
Exon 1	35,566,018–35,566,589	572	60,509,399–60,509,970	572	93.5	100.0
Intron 1	35,566,590–35,567,669	1,080	60,509,971–60,511,083	1,113	69.2	–
Exon 2	35,567,670–35,567,838	169	60,511,084–60,511,252	169	90.5	92.7
3′-UTR	35,567,839–35,568,401	563	60,511,254–60,512,850 (60,511,254–60,511,816)*	1,598	82.3	–

^*For interspecies comparison, only the portion of the human 3′-UTR corresponding with the dog-defined segment (60,511,254–60,511,816) was used, although the full length of the human 3′-UTR extends to position 60,512,850 in GRCh38. Coordinates in parentheses indicate the segment used in the analysis.

Comparative Expression Analysis of SIX6 in Canine Ocular Tissues

Publicly available whole-genome RNA-seq data from canine ocular tissues were obtained from the NCBI Gene Expression Omnibus under accession number GSE227619.⁴⁷ This dataset includes a total of 22 samples derived from 7 dogs—2 Pit Bull mixes, 1 Mastiff mix, 1 Boxer mix, and 3 Beagle mixes. Among these, three samples were excluded from analysis because they represented mixed tissues containing both the retina and the RPE–choroid complex, which were unsuitable for evaluating tissue-specific expression patterns. Consequently, 19 samples from the same seven dogs were retained for further analysis, including central retina, peripheral retina, central RPE–choroid, and peripheral RPE–choroid tissues. In dogs, a true foveomacular structure is absent; however, in the original GSE227619 dataset, the term “macular” was used to describe the central fundus encompassing the area centralis, a cone-enriched region analogous to the human macula. In this study, we refer to this region simply as “central” for clarity.

We compared the expression levels of the SIX6 gene among the four tissue types. Additionally, the percentile rank of SIX6 among all expressed genes was calculated to evaluate its relative expression abundance within each tissue. Furthermore, the expression levels of SIX6 were compared with those of established glaucoma-related genes, including ADAMTS10, ADAMTS17, COL1A2, RAB22A, and OLFML3 for canine glaucoma; MYOC, EFEMP1, OPTN, TBK1, ATOH7, and TXNRD2 for human glaucoma; and SRBD1 for both canine and human glaucoma. To compare gene expression levels between retinal tissues (central and peripheral retina) and RPE–choroid tissues (central and peripheral RPE–choroid), the Mann–Whitney U test was performed using RNA sequencing (RNA-seq) read counts. P values were corrected for multiple testing using the Benjamini–Hochberg method, and a Pc value of less than 0.05 was considered significant.

Results

Baseline Characteristics of Shiba-Inu and Shih-Tzu Dogs

The demographic characteristics of the glaucoma cases and controls are summarized in Table 1. In Shiba-Inus, the average ages of glaucoma cases and controls were 8.7 ± 3.0 and 9.3 ± 3.3 years, respectively (P = 0.29, Mann–Whitney U test). In Shih-Tzus, the corresponding values were 8.8 ± 2.1 and 10.7 ± 2.6 years, respectively, showing a significant difference (P = 0.0098). Regarding sex distribution, 32.7% of the Shiba-Inu cases and 36.7% of the controls were male (P = 0.66; χ² test). In Shih-Tzus, 50.0% of the cases and 69.2% of the controls were male (P = 0.16).

Association Analysis

In the Shiba-Inu controls, all 19 SNPs were in HWE except for rs852129966, which showed a significant deviation (P = 0.00018). In contrast, all SNPs, including rs852129966, were in HWE (P > 0.05) in the Shih-Tzu controls (Table 2). Table 3 shows the results of the association analysis for the 19 SNPs. rs851228774 was monomorphic in both Shiba-Inus and Shih-Tzus. In Shiba-Inus, rs851962234, located in the 3′-UTR of SIX6, was significantly associated with glaucoma (P = 0.0047; Pc = 0.038), with the minor A allele showing a frequency of 17.3% in cases and 4.2% in controls, and conferring an increased risk (OR, 3.56). Nine additional SNPs (rs8988460, rs24517937, rs24488773, rs8816199, rs851839004, rs851859199, rs24543712, rs24543710, and rs24543694) showed moderate LD with rs851962234 (r² = 0.53–0.55) (Fig. 1A), and their minor alleles were also associated with an increased risk of glaucoma (OR, 1.64–1.76), although the associations were not statistically significant. After conditioning on rs851962234, the minor alleles of these SNPs no longer showed increased risk (OR < 1.0), indicating that rs851962234 likely accounts for the signals observed at these SNPs. Two additional SNPs (rs850948063 and rs24543691) also showed increased risk (OR, 1.49 and 1.66, respectively), but without statistical significance. These two SNPs were in moderate LD with each other (r² = 0.58), but had very low LD with rs851962234 (r² < 0.10). Notably, they continued to show increased risk after conditioning on rs851962234 (OR, 1.65–1.85), indicating that the observed increased risk at these SNPs is independent of rs851962234.

Table 3.

Association Results of 19 SNPs in the SIX6 Gene Region With Glaucoma in Shiba-Inus and Shih-Tzus

		Shiba-Inus					Shih-Tzus
		Allele Frequency (%)					Allele Frequency (%)				Meta-Analysis
SNP	Test Allele	Cases (n = 49)	Controls (n = 60)	P Value	Pc	OR (95% CI)	Cases (n = 18)	Controls (n = 39)	P Value	OR (95% CI)	P Value	Pc	OR (95% CI)
rs24517935	A	1.0	0.0	0.22		–	5.6	6.4	0.76	0.75 (0.12–4.70)	–		–
rs8988460	T	23.5	14.2	0.12		1.76 (0.86–3.63)	63.9	71.8	0.76	0.86 (0.35–2.15)	0.45		1.31 (0.66–2.60)
rs24517937	G	23.5	14.2	0.12		1.76 (0.86–3.63)	63.9	71.8	0.76	0.86 (0.35–2.15)	0.45		1.31 (0.66–2.60)
rs24488773	T	23.5	14.2	0.12		1.76 (0.86–3.63)	63.9	71.8	0.76	0.86 (0.35–2.15)	0.45		1.31 (0.66–2.60)
rs8816203	T	15.3	18.3	0.61		0.84 (0.42–1.67)	36.1	20.5	0.19	1.92 (0.73–5.04)	0.69		1.18 (0.53–2.63)
rs851613897	A	1.0	0.0	0.22		–	5.6	6.4	0.76	0.75 (0.12–4.70)	–		–
rs852129966	T	2.0	3.3	0.63		0.70 (0.15–3.16)	5.6	6.4	0.76	0.75 (0.12–4.70)	0.58		0.72 (0.22–2.31)
rs8816199	A	23.5	14.2	0.12		1.76 (0.86–3.63)	55.6	60.3	0.93	0.96 (0.40–2.31)	0.29		1.37 (0.77–2.46)
rs851839004	T	23.5	14.2	0.12		1.76 (0.86–3.63)	55.6	60.3	0.93	0.96 (0.40–2.31)	0.29		1.37 (0.77–2.46)
rs851859199	A	23.5	14.2	0.12		1.76 (0.86–3.63)	55.6	60.3	0.93	0.96 (0.40–2.31)	0.29		1.37 (0.77–2.46)
rs850948063	T	25.5	18.3	0.20		1.49 (0.81–2.74)	30.6	14.1	0.12	2.21 (0.82–5.92)	0.056		1.66 (0.99–2.79)
rs851228774	A	0.0	0.0	–		–	0.0	0.0	–	–	–		–
rs851962234	A	17.3	4.2	0.0047	0.038	3.56 (1.31–9.65)	0.0	1.3	0.50	–	–		–
rs24543712	G	23.5	15.0	0.17		1.64 (0.80–3.37)	55.6	60.3	0.93	0.96 (0.40–2.31)	0.32		1.33 (0.76–2.31)
rs24543710	C	23.5	15.0	0.17		1.64 (0.80–3.37)	55.6	60.3	0.93	0.96 (0.40–2.31)	0.32		1.33 (0.76–2.31)
rs24543694	A	24.5	15.0	0.13		1.70 (0.85–3.43)	55.6	60.3	0.93	0.96 (0.40–2.31)	0.27		1.36 (0.79–2.36)
rs24543691	T	34.7	24.2	0.089		1.66 (0.92–3.01)	27.8	14.1	0.18	1.98 (0.73–5.32)	0.033	0.26	1.74 (1.05–2.89)
rs853179248	T	4.1	10.0	0.074		0.35 (0.10–1.18)	8.3	9.0	0.66	0.70 (0.14–3.50)	0.11		0.45 (0.17–1.19)
rs24542512	T	34.7	34.2	0.95		0.98 (0.55–1.76)	30.6	24.4	0.84	1.10 (0.43–2.80)	0.96		1.01 (0.62–1.66)

Figure 1.

Regional association plots and LD structures of SNPs within the SIX6 gene in Shiba-Inus (A) and Shih-Tzus (B). (Top) Regional association plots for each SNP. The lead SNPs (rs851962234 in Shiba-Inus and rs850948063 in Shih-Tzus) are shown as yellow circles. The color of all other SNPs indicates LD with the lead SNP (r²), with brighter red denoting higher LD. The y axis represents the –log₁₀ P values for association with glaucoma. The horizontal blue line indicates the nominal significance threshold of P = 0.05. (Bottom) LD plot of 18 SNPs of the SIX6 gene. The r² value for each SNP pair is expressed as a percentage within each square, with darker shading (closer to black) indicating higher LD. Squares corresponding with r² = 1 are shown in black without numerical values. rs851228774 was excluded because it was monomorphic in both breeds.

In Shih-Tzus, rs850948063 showed the lowest P value among the tested SNPs, with the minor T allele showing an allele frequency of 30.6% in cases and 14.1% in controls, and conferring an increased risk of glaucoma (OR, 2.21), although the association was not significant (P = 0.12). Two other SNPs in strong LD with rs850948063, rs8816203 (r² = 0.70) and rs24543691 (r² = 0.94) (Fig. 1B), also exhibited the increased risk (OR, 1.92 and 1.98, respectively), but none attained statistical significance. Interestingly, rs851962234, which was significantly associated with glaucoma in Shiba-Inus, showed no association in Shih-Tzus. The A allele of rs851962234 was extremely rare in this breed.

In the meta-analysis combining both breeds, rs24543691 was the only SNP that reached nominal significance (P = 0.033), with the minor T allele associated with increased glaucoma risk (OR, 1.74), although this significance did not survive correction. rs850948063, which is in moderate-to-strong LD with rs24543691 in both breeds, also showed a similar effect size (OR, 1.66) but only borderline significance (P = 0.056).

Comparison of Canine and Human SIX6 Gene Sequences

The SIX6 gene consists of two exons in both dog and human. Pairwise nucleotide sequence alignments revealed high conservation in the exon regions, with 93.5% identity for exon 1 and 90.5% for exon 2. At the amino acid level, exon 1 showed complete identity (100%), and exon 2 showed 92.7% identity, resulting in an overall amino acid identity of 98.4% across the coding region. In contrast, the intronic region showed lower sequence identity (69.2%), while the 3′-UTR exhibited 82.3% identity between the two species (Table 4; Supplementary Fig. S1).

Tissue-Specific Expression of SIX6 in Canine Eyes

The difference in SIX6 expression between retinal and RPE–choroid tissues was statistically significant (Pc = 0.0011), with approximately 150-fold higher expression in the retina (Supplementary Table S1; Supplementary Fig. S2), clearly demonstrating a strong retina-enriched expression pattern of SIX6. Consistently, the median percentile rank of SIX6 among all expressed genes was approximately 84% in retinal tissues and approximately 50% in RPE–choroid tissues, indicating that SIX6 is among the most abundantly expressed transcripts in the canine retina, corresponding with the top ∼20% of all genes in retinal tissues (Supplementary Table S2; Supplementary Fig. S3).

We also compared the expression of SIX6 with that of other glaucoma-related genes previously implicated in canine and human glaucoma (Fig. 2; Supplementary Table S1). Among the genes analyzed, SIX6 exhibited the most pronounced retina-dominant expression. ATOH7 also showed strong retina-dominant expression, with approximately 11-fold higher expression in the retina compared with the RPE–choroid (Pc = 0.0011). In addition, ADAMTS17, OPTN, and SRBD1 displayed significantly higher expression in retinal tissues than in RPE–choroid tissues (Pc < 0.05). Except for ADAMTS17, which is associated with extracellular matrix (ECM) organization, the other genes, SIX6, ATOH7, OPTN, and SRBD1, are known to be involved in RGC development, survival, or neurodegenerative processes related to glaucoma.

Figure 2.

Expression of SIX6 and glaucoma-related genes in retinal and RPE–choroid tissues of canine eyes. The y axis shows log₁₀-transformed RNA-seq read counts [log₁₀ (RNA-seq read count + 1)], allowing comparison across genes with different expression magnitudes. “+1” was added to permit logarithmic transformation of zero read counts. Each point (diamond) represents one sample. Boxplots indicate the median (center line) and interquartile range (box), with whiskers showing the data range. Asterisks denote significant expression differences between retina (central and peripheral) and RPE–choroid (central and peripheral) tissues. “Associated species” indicates whether each gene has been previously implicated in glaucoma in dogs, humans, or both. “Functional category” classifies genes according to their predominant biological roles related to glaucoma pathogenesis: “Neural” genes are involved in RGC development, neuroprotection, or stress response, whereas “ECM” genes relate to ECM organization and structural maintenance. RAB22A was categorized as “intermediate” because it participates in vesicular trafficking and autophagy, linking neural and cellular homeostatic processes. Numerical summaries of RNA-seq read counts and Pc values are provided in Supplementary Table S1.

Conversely, ADAMTS10, COL1A2, OLFML3, MYOC, and TXNRD2 showed significantly higher expression in RPE–choroid than in retinal tissues (Pc < 0.05). In particular, COL1A2 and MYOC demonstrated clear RPE-choroid–dominant expression. Except for TXNRD2, which participates in oxidative stress response and neuronal protection, the remaining genes are primarily involved in ECM maintenance and remodeling in ocular tissues.

Discussion

Glaucoma is more commonly observed in certain dog breeds, and Shiba-Inus and Shih-Tzus are among those with a relatively high prevalence. In Japan, a previous study reported that glaucoma developed in 29% of Shiba-Inus and 9.6% of Shih-Tzus.¹² These breeds are known to possess anatomically narrow ICAs, which may contribute to their higher susceptibility to glaucoma. According to the same study,¹² more than 80% of nonglaucomatous Shiba-Inus and more than 70% of nonglaucomatous Shih-Tzus had closed, narrow, or slightly narrow ICAs, indicating that such anatomical predispositions may contribute to an increased risk of glaucoma development in these breeds. In dogs, PACG is associated with developmental or hereditary abnormalities of the ICA, such as narrowing or closure owing to pectinate ligament dysplasia or goniodysgenesis.^7,48 These structural abnormalities gradually obstruct aqueous humor outflow through the trabecular meshwork, leading to increased IOP and the development of glaucoma. In canine POAG, disease progression involves structural collapse or degeneration of the trabecular meshwork, further exacerbating aqueous outflow resistance and elevating IOP.⁴⁹ Therefore, although PACG and POAG differ in their primary etiologies, canine PACG may share a similar pathophysiological mechanism with POAG, namely, progressive optic nerve damage secondary to impaired aqueous humor drainage. ICA abnormalities have also been reported in several other breeds, even in dogs without glaucoma, including Basset Hounds,⁵⁰ Flat-Coated Retrievers,^50,51 Welsh Springer Spaniels,⁵² Dandie Dinmont Terriers,⁵⁰ English Springer Spaniels,⁵³ Leonbergers,⁵⁴ and American Cocker Spaniels.¹² Notably, PACG cases have been documented in some of these breeds, such as Basset Hounds,^55–57 Welsh Springer Spaniels,⁵⁸ Dandie Dinmont Terriers,⁵⁹ and American Cocker Spaniels.⁶⁰ Given that PACG with abnormal ICA morphology occurs at a high rate in Shiba-Inus and Shih-Tzus in Japan, these breeds are considered to be predisposed to PACG owing to their anatomical characteristics.

SNPs within the human SIX6 gene region, located on chromosome 14, have been reported to be significantly associated with both POAG and NTG.^24–35 Since the first report in 2011, SIX6 has been recognized as one of the major genetic risk factors for glaucoma in humans. The aim of this study was to assess whether polymorphisms in the SIX6 gene region, located on canine chromosome 8, affect the development of canine glaucoma. To this end, we genotyped 19 SNPs in SIX6 in 2 breeds of dogs, Shiba-Inus and Shih-Tzus, with cases of glaucoma or without glaucoma as controls. This study is the first investigation of the SIX6 region for association with glaucoma in dogs. We found that rs851962234, located in the 3′-UTR of SIX6, was significantly associated with glaucoma in Shiba-Inus, whereas none of the tested SNPs, including rs851962234, showed any association in Shih-Tzus. These findings suggest that this SNP may contribute to the development of glaucoma in a breed-specific manner, particularly in Shiba-Inus.

SIX6 is involved in regulating key processes, such as optic nerve formation, retinal progenitor cell differentiation, and the maintenance of RGC populations.^36,39,61 Experimental and clinical studies have demonstrated that risk alleles of SIX6 SNPs, rs33912345, and rs10483727, associated with human glaucoma exert deleterious effects on retinal structure and function. rs33912345 is a missense variant that results in an Asn141His substitution. The risk allele (His141) has been shown in zebrafish models to cause abnormal optic nerve formation⁴⁰ and, in humans, to be associated with reduced retinal nerve fiber layer (RNFL) thickness.^40,41 Furthermore, the risk allele of rs10483727, located 93 kb downstream of SIX6 and in strong LD with rs33912345, has also been associated with significant RNFL thinning.⁴² In addition, the His141 risk allele of rs33912345 has been shown to promote RGC senescence, thereby contributing to accelerated RGC loss in experimental models of glaucoma.⁴³ Shiga et al.²⁷ reported that the rs33912345 His141 risk allele, while associated with reduced RNFL thickness, was also linked to increased ONH blood flow. A patient-derived induced pluripotent stem cell study has shown that the rs33912345 His141 risk allele impairs RGC development, leading to decreased differentiation efficiency, abnormal neurite outgrowth, and increased vulnerability to degeneration.⁴⁴ Collectively, these studies indicate that SIX6 risk alleles contribute to glaucomatous optic neuropathy by disrupting normal RGC development, enhancing their susceptibility to degeneration, and accelerating age-related cellular changes in the retina.

Given this functional importance of SIX6 in human glaucoma, we compared the canine and human SIX6 sequences to assess their degree of evolutionary conservation. Our analysis revealed a very high sequence identity in the coding exons between dogs and humans, with 100% amino acid identity in exon 1 and 92.7% in exon 2, resulting in an overall coding sequence identity of 98.4%. Such strong conservation in the protein-coding regions suggests that the functional roles of SIX6 are likely to be highly similar between the two species. Notably, the most prominent glaucoma-associated missense SNP in humans, rs33912345, is located in exon 1, whereas the SNP identified in Shiba-Inu dogs, rs851962234, resides in the 3′-UTR. In general, missense SNPs tend to alter protein function, whereas variants in the 3′-UTR are more often implicated in modulating gene expression levels. No variant has been reported to date at the canine orthologous position corresponding with rs33912345, or at the human orthologous position corresponding with rs851962234 (Supplementary Fig. S1). These findings suggest that, whereas in humans, glaucoma susceptibility related to SIX6 may arise from functional changes in the protein owing to rs33912345, in dogs, rs851962234 may modulate SIX6 expression, thereby increasing RGC vulnerability and ultimately contributing to glaucoma development. Such differences highlight the need for future functional analyses in dogs, particularly to determine whether rs851962234 affects SIX6 expression.

Our transcriptomic analysis of publicly available RNA-seq data demonstrated that SIX6 is strongly expressed in the canine retina, consistent with its established role as a transcription factor essential for RGC differentiation and optic nerve development. This retina-enriched expression supports the notion that SIX6 contributes to glaucoma susceptibility through neural, rather than structural, mechanisms. When compared with other glaucoma-related genes, a tissue-dependent expression pattern was evident. Genes primarily involved in neuronal differentiation and stress response tended to show higher expression in retinal tissues, whereas those related to ECM organization and structural maintenance were predominantly expressed in the RPE–choroid. This pattern suggests that glaucoma pathogenesis may involve two complementary pathways: one related to neural vulnerability in the retina and another associated with ECM remodeling and biomechanical stress in non-neuronal tissues.

Although associations of SIX6 variants with POAG and NTG have been well-documented in humans, no report has addressed their potential relationship with PACG. In this study, the glaucoma phenotype observed in Shiba-Inus is more consistent with PACG, because this breed is anatomically predisposed to narrow ICAs. Canine PACG may share a similar pathophysiology with POAG, characterized by progressive optic nerve damage secondary to impaired aqueous outflow. In this context, RGC vulnerability conferred by SIX6 variants, may further increase susceptibility to IOP-induced optic nerve damage. Therefore, the association between rs851962234 and glaucoma in Shiba-Inus may be explained by the interaction between anatomical risk factors and the RGC vulnerability it confers.

A limitation of this study is the relatively small sample size, particularly for Shih-Tzus. Limited sample sizes can sometimes lead to false results in case control studies. In this study, although rs851962234 was the only SNP that attained statistical significance in Shiba-Inus, several other SNPs showed increased risk with ORs exceeding 1.5. Notably, SNPs not in LD with rs851962234 also exhibited increased risk in both Shiba-Inus and Shih-Tzus, and in the combined meta-analysis of both breeds, rs24543691 showed an OR of 1.74 with a P value of less than 0.05. This finding raises the possibility that rs24543691, independent of rs851962234, may represent a shared genetic risk factor for glaucoma across both breeds. However, this association did not remain significant after correction for multiple comparisons, and the small sample size limits our ability to draw definitive conclusions. Future studies with larger cohorts are needed to validate these findings and clarify the contribution of these SNPs to glaucoma in both breeds. Another important limitation of this study is the lack of detailed phenotypic information on ONH damage in glaucoma-affected dogs. Because many of the dogs were elderly, fundus evaluation was often hindered by lens opacity owing to nuclear sclerosis or corneal edema, making accurate assessment of ONH cupping and atrophy difficult. A previous survey reported that ONH atrophy or cupping was observed in approximately 82% of evaluable glaucomatous Shiba-Inu eyes,⁶² suggesting that ONH damage is common in glaucomatous eyes of this breed. Nevertheless, we were unable to assess the association between ONH damage and SIX6 rs851962234. Future studies incorporating detailed ophthalmic phenotyping, including ONH evaluation, will be important to clarify how this variant contributes to glaucomatous optic neuropathy in dogs.

Conclusions

To our best knowledge, this study is the first to investigate the association between SIX6 gene variants and glaucoma in dogs. We identified rs851962234 as significantly associated with glaucoma in Shiba-Inus, potentially through modulation of SIX6 expression and increased RGC vulnerability. These findings provide new insights into the genetic basis of canine glaucoma and highlight parallels and distinctions between canine and human disease.