Introduction

Genetic disorders are particularly prevalent in purebred dogs, largely due to the combined effects of artificial selection, reduced effective population size, and increased inbreeding coefficients [1–4]. Indeed, selective breeding has often prioritized morphological traits over health, leading to the fixation of deleterious alleles and a consequent rise in breed-specific hereditary conditions [2–6]. Among these, genodermatoses, monogenic hereditary skin disorders, are of particular concern. Dogs, second only to humans, present the highest number of identified inherited diseases, many of which affect the skin, such as Ichthyosis [7–9]. Ichthyosis is a rare congenital or hereditary skin disease caused by primary defects in the formation of the stratum corneum [7]. It is classified into epidermolytic and non-epidermolytic forms, depending on the presence or absence of keratinocyte vacuolization and lysis in the spinous and granular layers [10,11]. Epidermolytic ichthyosis has been reported in breeds such as the Norfolk Terrier due to mutations in keratin genes [12], while non- epidermolytic ichthyosis (NEI) is more common and has been described in Golden Retrievers, Jack Russell Terriers, American Bulldogs, Shar-Pey and German shepherd [8,13,14].

In Golden Retrievers, NEI is most commonly associated with an insertion–deletion variant in exon 8 of patatin-like phospholipase domain containing 1 (PNPLA1) gene, located on chromosome 11 (NC_051816.1: 5,725,170–5,771,521) [15,16]. More recently, a second form of ichthyosis, referred to as Golden Retriever ichthyosis type 2, has been described and associated with a 14 bp deletion in the ABHD5 gene (c.1006_1019del), although its occurrence appears to be much rarer within the breed [17]. The PNPLA1 gene encodes a patatin-like phospholipase, while ABHD5 encodes the α/β-hydrolase domain-containing protein 5; both proteins are involved in epidermal lipid metabolism, particularly in the biosynthesis of ω-O-acylceramides, which are essential for skin barrier integrity [18–23]. Loss-of-function variants in PNPLA1 or ABHD5 have been linked to autosomal recessive congenital ichthyosis in both dogs and humans [16,19].

In both mutations, affected Golden Retrievers usually develop clinical signs early in life, characterized by generalized scaling, hyperpigmented and rough skin in ventral glabrous areas, and laminated orthokeratosis on histopathology [10,13]. The condition is generally non-pruritic, although secondary infections may cause inflammation and discomfort [13].

Given the hereditary nature and the high prevalence of NEI, genetic testing is a key preventive tool [24,25]. It enables the identification of carriers and affected dogs guiding breeding strategies to reduce mutation prevalence while maintaining genetic diversity.

The aim of this study was to assess the prevalence of the two mutations associated with non-epidermolytic ichthyosis (NEI) in Golden Retrievers in Italy.

Materials and methods

Statistical analyses

Statistical analyses were conducted using R software (R Core Team). Differences in genotype frequencies among groups were first assessed using the Chi-square test with the “chisq.test” function in R applied to global contingency tables, in order to evaluate overall associations between grouping variables and genotype frequencies. A p-value < 0.05 was considered statistically significant. When a significant overall difference were detected, pairwise post-hoc comparisons were carried out using Fisher’s exact test between each pair of groups, to account for small sample sizes and sparse contingency tables. Allele frequencies for the wild-type (wt) and mutant (mut) variants were calculated according to the following formulas:

(1)(2)

Results

Among the 463 Golden Retrievers tested for PNPLA1 mutation between 2017 and September 2025, 42% (n = 192) were classified as clears (wt/wt), 37% (n = 173) as carriers (wt/mut), and 21% (n = 98) as affected (mut/mut). The calculated allele frequencies were 60% for the wild-type allele and 40% for the mutant allele, indicating an almost balanced distribution of the two variants in the analyzed population.

In addition, 45 dogs were tested for the ABHD5 gene deletion (type 2 NEI), including two in 2024 and 43 in 2025. All individuals were clears, resulting in a wild-type allele frequency of 100%.

Of the three dogs tested for both genes, one was affected, one a carrier, and one clear for the PNPLA1 mutation, while all three were wild-type for ABHD5.

Temporal distribution of PNPLA1 genotypes

The number of samples tested for PNPLA1 varied over the years, showing an overall increasing trend despite some fluctuations (Table 1). The number of analyzed dogs progressively rose from 2 in 2017 to 126 in 2025, reaching the highest value even though the dataset includes data only up to September. Except for 2017 (n = 2), which is too small to be representative of the genotypic frequencies, the early years of sampling showed relatively high proportions of affected dogs, peaking at 44% in 2020. In the following years, this percentage gradually declined, reaching 14% in 2025. The proportion of carriers remained stable, with minor annual fluctuation. Overall, these results indicate both a progressive reduction in the frequency of affected individuals and an increase in the proportion of clear dogs, alongside a steady rise in the total number of animals tested over time. Statistical analysis confirmed a significant difference in genotype distribution across years (p = 0.01). Post-hoc comparisons revealed that this significance was mainly driven by differences between 2020 vs. 2025 (p = 0.00003), 2020 vs. 2024 (p = 0.0008), and 2020 vs. 2023 (p = 0.003).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Annual distribution and frequencies of PNPLA1 genotypes.

https://doi.org/10.1371/journal.pone.0345595.t001

When considering the year of birth of the tested dogs instead of the year of sampling, a more accurate representation of the population trend emerges (Table 2). When genotypes were analyzed according to year of birth, a significant difference in genotype distribution was detected (p = 0.04). Dogs born in earlier years showed higher proportions of affected individuals, whereas more recent birth cohorts (2021–2025) exhibited lower frequencies of affected and higher proportions of clears, culminating in 2025-born dogs with only 10% affected and 60% clear individuals (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. PNPLA1 genotype frequencies by year of birth.

https://doi.org/10.1371/journal.pone.0345595.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Temporal trend of PNPLA1 genotypic frequencies by year of birth in tested Golden Retrievers from 2015 to 2025. The graph shows the frequency of clear (wt/wt) in green, carriers (wt/mut) in orange, and affected (mut/mut) in red.

https://doi.org/10.1371/journal.pone.0345595.g001

Post-hoc pairwise comparisons indicated that the significance was driven by differences between 2020 vs. 2022 (p = 0.003), 2020 vs. 2021 (p = 0.005), 2020 vs. 2023 (p = 0.005), 2020 vs. 2024 (p = 0.04), and 2018 vs. 2023 (p = 0.04).

Discussion

Non-epidermolytic ichthyosis (NEI) in Golden Retrievers is a hereditary skin disorder caused by a loss-of-function insertion–deletion variant in the PNPLA1 gene, which plays a key role in epidermal lipid metabolism and skin barrier formation [16,18–20]. More recently, a second form, linked to a 14 bp deletion in the ABHD5 gene, has been identified, although its occurrence appears to be much rarer within the breed [17]. Both variants result in clinically similar non-epidermolytic ichthyosis phenotypes.

Given the much higher number of samples analyzed for PNPLA1, this study primarily focuses on this mutation, which remains the major genetic cause of NEI in Golden Retrievers and the main target of screening programs.

In our study, 463 Golden Retrievers were tested in Italy between 2017 and September 2025 for the PNPLA1 mutation. The results showed a predominance of the wild-type genotype compared to previous reports, with 42% clears, 37% carriers, and 21% affected. The calculated allele frequencies were 60% for the wild-type allele and 40% for the mutant allele, indicating a moderate reduction in the prevalence of the mutation compared with previous reports.

For instance, Owczarek-Lipska et al. [25] in Switzerland found 20% clear, 49% carriers, and 32% affected, with a mutant allele frequency of 56.1%, slightly higher than our estimate. Similarly, Roething et al. [26] in Germany reported mutant allele frequencies of 54.5% in puppies and 58.8% in dams. By contrast, a multi-country study [27] highlighted marked geographical differences. The mutant allele frequencies were found lower in Australia (31.5%) and in the USA (38.5%), the latter closely aligning with our estimate (40%). In contrast European values were consistently higher, ranging from 48.5% in French samples to 64.5% in Swiss (Guaguère et al., 2013).

The consistently high prevalence of the PNPLA1 mutation reported across different countries suggests that its distribution within the Golden Retriever breed may be compatible with a founder effect, potentially amplified by the widespread use of popular sires and the restricted genetic diversity typical of purebred dog populations [1–4]. However, dedicated population genetic or pedigree-based analyses would be required to formally test this hypothesis.

A comparison with another Italian study based on 48 samples conducted in 2018 [24] reveals a higher prevalence of the mutant allele. In that work, carriers represented the largest group (48%) and the proportion of affected dogs was also higher (31%) than in our cohort (21%), resulting in a mutant allele frequency of 55.2%, markedly above the 40% observed in our dataset. Both studies also examined genotype distribution by sex; however, the strong imbalance reported by Graziano et al. [24], with males more frequently clears (30%) and females predominantly affected (46%), was not confirmed in our population (S3 Table).

When considering only the dogs born before 2018, our results align closely with those reported by Graziano et al. in 2018 [24], showing a mutant allele frequency of 50%, with 30% of affected individuals. These higher percentages are also consistent with those described in previous European studies, reinforcing the idea that the differences between our findings and earlier reports are mainly attributable to the time period covered.

Indeed, our data show a statistically significant temporal shift in genotype distribution, both when analyzed by year of sampling and by year of birth. As illustrated in Fig 1, since 2020 there has been a marked increase in the proportion of clear dogs and a concurrent decrease in affected individuals, while carriers have remained relatively stable despite minor annual fluctuations. This pattern is consistent with the increasing adoption of genetic screening and selective breeding strategies aimed at reducing the prevalence of the PNPLA1 mutation. However, the dataset is based on voluntary genetic testing requested mainly for breeding or diagnostic purposes and therefore does not represent a random sample of the Italian Golden Retriever population. Consequently, the observed temporal decline in mutant allele frequency may partly reflect changes in testing practices and selection strategies rather than being solely attributable to a reduction in the frequency of the PNPLA1 mutation.

It is also worth noting that the majority of the dogs were tested at a young age, with 64% of samples collected from individuals aged 0–2 years. This further supports the increasing focus on early genetic screening, reflecting breeders’ attention to preventive selection practices within the population.

However, despite the evident decreasing trend, the incidence of the mutation remains high, with an overall mutant allele frequency of 40%. This value is very high compared with other hereditary mutations investigated in Italian dog populations, which usually range between 6% and 10% [28–31]. When considering other hereditary conditions reported in Italian Golden Retrievers, such as progressive retinal atrophy, the prevalence was estimated at 4.2% for PRA1 and 6.5% for PRA2 [29], values markedly lower than those observed for ichthyosis. These results confirm that PNPLA1-related ichthyosis is among the most prevalent and clinically relevant inherited disorders in the breed, consistent with Donner et al. [1], who ranked it among the 20 most common hereditary diseases in both purebred and mixed-breed dogs.

Conversely, for the ABHD5 gene deletion (type 2 NEI), all 45 dogs tested were clears, resulting in a wild-type allele frequency of 100%. These dogs, evenly distributed between males (n = 20) and females (n = 25) and born between 2020 and 2024, represent the first data available for this recently identified mutation, described only in 2022 by Kiener et al. [17]. Given the recent introduction of ABHD5 testing and the limited sample size, these findings apply exclusively to the tested population and do not allow conclusions about the overall prevalence of this variant in the breed. To date, ABHD5-related ichthyosis has been reported only in a limited number of cases, and additional data will be required to better define its distribution within Golden Retrievers. Nevertheless, continued genetic testing for ABHD5 is advisable, as early monitoring of this variant could help prevent its spread in the population should it emerge in the future.

Moreover, as observed for PNPLA1, ABHD5 testing also showed increasing attention, with the number of samples rising from only two in 2024–43 in 2025.

Conclusions

This study provides an updated overview of non-epidermolytic ichthyosis in Golden Retrievers in Italy, focusing on both the PNPLA1 and ABHD5 variants. The results confirm that PNPLA1-related ichthyosis remains widespread, with a mutant allele frequency of 40%, although lower than that reported in previous Italian and European studies. The progressive decrease in affected and carrier dogs observed over recent years, together with the marked rise in the number of genetic tests performed, reflects growing awareness and the effectiveness of preventive breeding strategies, although changes in testing behavior and selective screening may also contribute to the observed reduction in the mutation frequency.

In contrast, all dogs tested for the ABHD5 deletion were clears, indicating that this recently described mutation was not detected in the tested population. However, given the limited and recent dataset, conclusions regarding the overall prevalence of ABHD5-related ichthyosis in the breed cannot be drawn. Continued surveillance and early genetic testing for both variants are essential to prevent the spread of deleterious alleles and to promote the long-term genetic health of the breed.

Supporting information

Acknowledgments

The authors are grateful to Vetogene Laboratory for access to the test data and the information that led to the compilation of Tables.