Identification of bacteria associated with canine otitis externa based on 16S rDNA high-throughput sequencing

Abstract

Bacteria are regarded as predisposing and perpetuating factors causing otitis externa (OE), whereas auricular anatomy is a predisposing factor. This study aims to investigate bacterial populations in the external auditory canals of healthy dogs and dogs with OE. Four categories of ear swabs included healthy erect-ear dogs, erect-ear dogs with OE, healthy pendulous-ear dogs and pendulous-ear dogs with OE. After bacterial DNA extraction, 16S rDNAs were amplified using specific primers within a V3/V4 region. Following DNA library construction, high-throughput sequencing was performed on MiSeq (Illumina). CLC Microbial Genomics Module was used to determine the rarefaction curve, bacterial classification, relative abundance, richness and diversity index. The results demonstrated that healthy dogs had higher bacterial richness and diversity than the dogs with OE. Comparable with culture-dependent methods described previously, this study revealed predominant Corynebacterium spp., Pseudomonas spp., Staphylococcus spp., and Proteus spp. in OE cases. Furthermore, high-throughput sequencing might disclose some potential emerging pathogens including Tissierella spp., Acinetobacter spp., and Achromobacter spp., which have not been reported in previous canine OE cases. Nevertheless, larger sample sizes are further required for an extensive evidence-based investigation.

Supplementary information

The online version contains supplementary material available at 10.1007/s42770-023-01166-0.

Introduction

Otitis externa (OE) is a disease characterized by skin inflammation in the external auditory canal. This disorder is commonly found in companion animal practice with a prevalence ranging from 5% to 20% [1–3]. Based on physical examination with clinical signs, the diagnosis is uncomplicated. Nonetheless, the etiology of OE involves multifactorial causes, being classified into predisposing, primary, secondary and perpetuating factors [4]. Predisposing factors such as auricular anatomy, maceration, obstruction, environmental temperature and humidity do not cause otitis externa directly but compel the ear more vulnerable to disease development. However, the most common primary factors initiating otitis externa are ectoparasites, allergies, foreign bodies, autoimmune diseases and keratinization disorders. Moreover, bacteria and yeasts are considered perpetuating and predisposing factors as well. Microorganisms are responsible for progressive exacerbation, evading spontaneous alleviation [4–8].

A diverse and plentiful microbial population resides on the skin, yet a large number have been insufficiently investigated [9, 10]. Some believe that the bacterial population involves both pathological causes and prevention of the disease [10]. Moreover, it was clear that the alteration in the microbial population plays an essential role in both infectious and non-infectious diseases. Since diagnostic test methods were limited, a minimally diverse composition of microbial abundance was reported in most of the earlier studies on skin microbiota in dogs and cats [9, 10]. To date, both culture-dependent and culture-independent methods have been used to identify microbial biodiversity in the canine microbiota, while some recent studies can link the human-associated microbes between household members and their dogs [9, 11, 12].

Most of the studies showed that culture-dependent methods had been utilized to isolate bacterial species which can cause skin diseases for decades. However, culture-based methods are labor- and time-consuming techniques. On the other hand, the latest advances in culture-independent sequencing technology revealed that the diversity of the skin microbiome had been drastically underestimated in canine and feline species [10]. Moreover, it was comprehensible that several species that appeared in low frequency or were unable to propagate in culture conditions may have a strong association between their proportions and skin diseases. In this prospect, high-throughput sequencing is required to explore the microbial biodiversity of the skin in high resolution.

To elucidate the pathological mechanisms of many skin diseases as well as to develop new preventive and therapeutic ways by manipulating the microbial population, a better understanding of this complex microbial population is required. The disorganization of the microbial community in external auditory canals may lead to the progression of the disease and bring about clinical translation. Therefore, the aim of this study is the identification and comparison of the bacterial community mainly presenting in external auditory canals of dogs with and without otitis externa, determining the bacteria associated with OE in dogs.

Material and methods

Sample collection and DNA extraction

Sample collection was performed at a private small animal hospital located in Bangkok, Thailand. Ear swabs were collected and submitted by licensed veterinarians from the external auditory canals as part of routine diagnostic screening. All protocols complied with Ethical Principles in Animal Experimentation guidelines and were conducted with the approval of the Chulalongkorn University Animal Care and Use Committee (Animal Use Protocol No. 1775005). Ear swabs were obtained from 24 dogs (Table 1), including 12 dogs with clinical signs of OE (disease group) and 12 dogs without clinical evidence of OE (healthy control group). Each group was divided equally into 2 subgroups based on the anatomical shape of the ear pinnae, including the pendulous (n=6) and erect group (n=6). Moreover, some clinical criteria were excluded prior to enrolment as follows: OE due to fungal/ yeast infection; use of systematic or topical antibiotics within the past 30 days; use of systemic or topical immunosuppressant (e.g. corticosteroids) within the past 30 days; use of topical ear medications within past 14 days. The cotton swabs were rolled deeply into the external ear canals. The swabs were stored in phosphate buffer saline (PBS) and refrigerated at 4°C until further processing. Bacterial genomic DNAs were extracted by using the Genomic DNA Extraction Kit (RBC Bioscience, New Taipei City, Taiwan) according to the manufacturer’s protocol.

Table 1.

Demographic distribution and the number of sequences obtained from different groups of auricular dogs

Group	Code	Breed	Age	Sex	PF reads
Erect-ear dogs with OE (DE)	DE01	Pomeranian	6 yr	Female	96,275
	DE02	Mixed	17 yr	Male	48,473
	DE03	Pomeranian	6 mth	Female	88,719
	DE04	Bull Terrier	8 mth	Male	76,719
	DE05	Mixed	N/A	Male	51,611
	DE06	Chihuahua	4 mth	Male	83,097
Healthy erect-ear dogs (HE)	HE01	Mixed	N/A	Female	54,587
	HE02	Chihuahua	2 yr	Female	68,111
	HE03	Chihuahua	3 yr	Male	159,760
	HE04	Chihuahua	2 yr	Male	84,270
	HE05	Chihuahua	1 yr 3 mth	Female	72,369
	HE06	French Bulldog	2 yr 10 mth	Female	96,519
Pendulous-ear dogs with OE (DP)	DP01	Shih Tzu	3 yr 8 mth	Male	33,777
	DP02	Poodle	5 mth	Female	27,209
	DP03	Mixed	10 yr	Male	62,620
	DP04	Mixed	4 yr	Male	48,162
	DP05	Golden Retriever	4 yr	Male	43,093
	DP06	Poodle	12 yr	Male	84,656
Healthy pendulous-ear dogs (HP)	HP01	Mixed	1 yr 9 mth	Female	80,222
	HP02	Yorkshire Terrier	3 yr	Female	86,450
	HP03	Yorkshire Terrier	2 mth	Female	44,137
	HP04	Mixed	3 yr	Male	55,883
	HP05	Mixed	8 mth	Female	61,546
	HP06	Mixed	10 yr	Female	63,479

Passed filter (PF), years (yr), months (mth), not applicable (N/A), erect-ear dogs with OE (DE), healthy erect-ear dogs (HE), pendulous-ear dogs with OE (DP) and healthy pendulous-ear dogs (HP)

Amplification of V3/V4 regions of 16S rDNA

Targeted high-throughput sequencing of the 16S rDNA gene was used to identify the bacterial community. The DNAs were amplified by using specific primers within a V3/V4 region of 16S rDNA (forward primer: 5'-ACTCCTACGGRAGGCAGCAG-3' & reverse primer: 5'-TACNVGGGTATCTAATCC-3'). The Polymerase Chain Reaction (PCR) was performed in the condition; of denaturation at 94°C for 50 sec, annealing at 40°C for 30 sec and extension at 72°C for 60 sec. This PCR condition was repeated for 25 cycles and then followed by a final amplification step at 72°C for 5 min. The amplified PCR products (approximately 460 bp in length) were separated by 2% agarose gel electrophoresis and then purified by HiYield Gel/PCR DNA Fragments Extraction Kit (RBC Bioscience, New Taipei City, Taiwan). The concentration of purified DNAs was measured by Qubit^® fluorometer (Invitrogen, Waltham, MA, USA) with a Qubit^® dsDNA BR Assay kit (Invitrogen, Waltham, MA, USA).

DNA library preparation

Purified 16S rDNA PCR product (500 ng) of each specimen was used to construct DNA libraries with different indexes using the NEBNext^® Ultra^™ DNA Library Prep Kit for Illumina (New England Biolabs, Beverly, MA, USA). The concentration of DNA libraries was quantified by using a quantitative real-time PCR with a KAPA SYBR FAST qPCR Master Mix (Kapa Biosystems, Wilmington, MA, USA) following the manufacturer’s instructions. All DNA libraries were pooled together with equal concentration, and the pool was then adjusted to the final concentration of 4 nM. The libraries were then paired-end sequenced (2x250 bp) using MiSeq v2 reagent kit (500 cycles) and run on a MiSeq platform (Illumina, San Diego, CA, USA).

Data analysis and bacterial classification

The primary analysis of sequencing data (FASTQ) was performed using MiSeq Reporter Software (MSR) version 2.4. The sequencing reads with adaptors or low-quality (Q-score<30) sequences were trimmed by CLC Genomics Workbench version 10.1.1. The pass-filter reads with the high-quality score (Q-score≥30) were further analyzed for microbiome analysis. Bacterial classification based on 16S rDNA was performed compared to the Greengene database 13.5. CLC Microbial Genomics Module was used to determine the rarefaction curve, taxonomic classification, abundance, richness, alpha diversity and beta diversity. The amounts of bacterial phyla and genera were analyzed based on the number of reads and then represented as a percentage of relative abundance in each group. The reporting of bacterial classifications was used for comparative microbiome analysis between healthy controls and dogs with otitis externa in order to identify the bacteria associated with the disease. The beta diversity among groups was calculated using Qiime2 (https://qiime2.org/) and was presented by Principle Coordinate Analysis (PCoA) based on Bray-Curtis dissimilarity. The data were statistically analyzed by using GraphPad Prism version 8.1. Unpaired t-test and one-way ANOVA were used for differences between two groups, and multiple group comparisons, respectively. P values less than 0.05 (P < 0.05) were considered as statistically significant.

Results

High-throughput sequencing data and rarefaction analysis

A total of 3,437,193 reads were gained from the MiSeq platform, of which 1,671,744 reads passed the filter with a Q-score ≥ 30. The average number of passed filter reads was 69,656 reads/sample ranging from 27,209 to 159,760 reads depending on the sample (Table 1). Moreover, rarefaction analysis is used to evaluate whether the depth of coverage (reads per sample) is enough. The rarefaction curve demonstrated the number of reads obtained from each sample was enough for effective OTUs classification. Due to the high individual variability of the rarefaction curve observed among samples, the mean of rarefaction curves obtained from each group was shown in Fig. 1A. Assignment of consensus taxonomy yielded 197 unique operational taxonomic units (OTUs) in all samples, with an average of 23 OTUs per sample. Twenty-two phyla colonized in the 24 samples with different proportions, while 9 phyla had a relative sequence abundance greater than 1% among the samples (Table S1).

Fig. 1.

Rarefaction curve, richness and diversity index of bacterial community among groups. (A) Representative rarefaction curves were obtained from each group. (B) The beta diversity between healthy dogs and dogs with OE was presented by Principle Coordinate Analysis (PCoA) based on Bray-Curtis dissimilarity. (C) Chao1 richness. (D) Shannon diversity index. (E) Simpson’s diversity index. * indicates statistically significant (P < 0.05). Abbreviations: erect-ear dogs with OE (DE), healthy erect-ear dogs (HE), pendulous-ear dogs with OE (DP) and healthy pendulous-ear dogs (HP)

Diversity of the bacterial community

To compare the diversity between the disease and healthy groups, beta diversity was investigated using Bray-Curtis dissimilarity. The result showed that the microbial community in the disease group was slightly different from the healthy group (Fig. 1B). To analyze the diversity within each sample, alpha diversity including Chao1 richness, abundance and evenness was examined in this study. The Chao1 richness was evaluated using the number of observed OTUs (Fig. 1C). The samples from the healthy pendulous-ear group had the highest Chao1 richness (191.32 ± 55.10),whereas those collected from the pendulous-ear group with OE had the lowest Chao1 richness (148.40 ± 60.00). The results demonstrated that the healthy groups had higher Chao1 richness, Shannon and Simpson’s diversity indices of bacterial community than those found in the OE groups (Fig. 1C-E). In addition, the Shannon and Simpson’s diversity indices (Fig. 1D and 1E, respectively) were statistically significant differences between the erect-ear dogs with OE and the healthy erect-ear group (P = 0.03). The result revealed that healthy erect-ear dogs contain more richness and diversity than those found in erect-ear dogs with OE.

Bacterial community differences between erect-ear dogs and pendulous-ear dogs

According to the different anatomy between erect-ear dogs and pendulous-ear dogs, the exposure and accumulation of bacterial community in the ear canal were different (Fig. 2A). The results of bacterial classification at the phylum level revealed that Proteobacteria was the most abundant in healthy erect-ear dogs (75.70%) whereas Firmicutes was the most dominant phylum in pendulous-ear dogs (58.62%). In addition, the bacterial classification at the genus level (Fig. 2B) demonstrated that the most dominant genus in healthy pendulous-ear dogs was Staphylococcus (48.34%) which was remarkedly different (P=0.02) when compared with that found in healthy erect-ear dogs (0.07%). Although Novosphingobium was the most abundant genus in healthy erect-ear dogs (33.57%), it was not significantly different (P > 0.05) when compared to that found in healthy pendulous-ear dogs (8.98%). In addition, Pseudomonas was more prevalent in healthy erect-ear dogs (14.68%) than those found in healthy pendulous-ear dogs (1.34%), but not dramatically different (P > 0.05).

Fig. 2.

Relative abundance of each OTUs in different groups of auricular dogs. (A) OTUs abundance at the phylum level. (B) OTUs abundance at the genus level. Abbreviations: erect-ear dogs with OE (DE), healthy erect-ear dogs (HE), pendulous-ear dogs with OE (DP) and healthy pendulous-ear dogs (HP)

Bacteria associated with otitis externa in erect-ear dogs

Bacterial community comparison between erect-ear dogs with OE and healthy erect-ear dogs was investigated in order to explore the bacteria that might be associated with OE. The results demonstrated the different proportions of bacterial phyla Firmicutes and Actinobacteria were found to be more abundant in erect-ear dogs with OE than healthy group (Fig. 2A). Furthermore, the bacterial classification at genus level showed that Corynebacterium (17.90%), Proteus (15.33%), Tissierella (3.95%), Acinetobacter (2.57%) and Achromobacter (2.07%) were relatively higher in erect-ear dogs with OE, but not statistically different (P > 0.05). However, there was a significant difference (P = 0.03) in the relative percentage of Staphylococcus (17.65%) found in erect-ear dogs with OE, implying that this bacterial genus might be associated with OE in erect-ear dogs.

Bacteria associated with otitis externa in pendulous-ear dogs

Comparison of the bacterial community in the phylum level between pendulous-ear dogs with OE and healthy pendulous-ear dogs revealed that Proteobacteria and Actinobacteria were found to be more proportion in the OE group than those found in the healthy group (Fig. 2A). For genus level (Fig. 2B), Corynebacterium (29.57%), Pseudomonas (18.19%), Acinetobacter (4.07%), Cloacibacterium (5.23%) and Dechloromonas (3.46%) were relatively higher in pendulous-ear dogs with OE. This preliminary data showed no significant difference in the percentage of bacterial genera.

Common and unique bacteria associated with OE

Interestingly, Corynebacterium and Acinetobacter were commonly found in both erect-ear dogs and pendulous-ear dogs with OE implying that the proportion of these two genera in bacterial communities might be associated with OE. For unique bacteria, Proteus, Staphylococcus, Tissierella, and Achromobacter were obviously observed in erect-ear dogs with OE, while the Pseudomonas, Cloacibacterium and Dechloromonas were dominantly found in pendulous-ear dogs with OE.

Discussion

Recent studies reported that OE is generally presented, and probably the most frequent disorder found in small animal clinical practice [1, 3]. This disease makes affected dogs uncomfortable, which could lead to some serious consequences such as hearing impairment and neurological signs [13, 14]. The etiology of OE may involve multiple factors including bacteria and auricular shape [4]. Therefore, this study investigated the bacterial community in the different shape of ear pinnae which might influence disease progression. In this observation, we noticed that older and male dogs presented in the OE group rather than the healthy one.

The external ear canal harbors a rich and diverse bacterial community, which might play an essential role in some ear diseases including OE. Staphylococcus, Streptococcus, Micrococcus, Proteus and Corynebacterium could be isolated from the ears of healthy dogs [15–17]. Although predominantly colonizing in healthy ears, Staphylococcus species were the most isolated bacteria from dogs with OE [7, 15, 18, 19]. This finding also revealed that Staphylococcus spp. could be identified in both healthy and OE groups. However, the prevalence of OE associated with a highly relative abundance of Staphylococcus spp. was 17.65% and 20.42% in erect-ear dogs and pendulous-ear dogs, respectively. In addition, the result of this study is consistent with previous reports that Coryneform bacteria are regarded as a part of the normal flora found in the canine skin Angus [1] and Yoshida, Naito et al. [17]. Moreover, Corynebacterium spp. might opportunistically dominate in OE cases [20, 21]. According to the result of this study, the relative abundance of Corynebacterium spp. was higher in OE groups. This investigation was also correlated with previous studies that OE might be caused by Pseudomonas spp. and Proteus spp. [5, 7, 19]. Pseudomonas spp. was found not only in otitis cases but also in healthy dogs with a lower relative abundance. Interestingly, Proteus spp. was not found in healthy dogs, but was predominant in erect-ear dogs with OE (15.33%).

In most of the previous studies mentioned, culture-based techniques have been intensively used to identify bacterial pathogens associated with canine OE. Nevertheless, this conventional method consumes time and labor. In addition, biases may arise due to the selectivity of the cultural conditions and media [19, 22]. While a fraction of the bacterial populations are selectively favored, the diversity of low-frequent or unculturable bacteria is utterly underestimated. For instance, Almeida and colleagues [15] showed that bacteria were not culturable from some of the specimens with OE (4.7%) and without OE (20%). Moreover, multiple bacterial species were grown from both OE (77.3%) and healthy (63.3%) specimens. Instead of the obsolete culture methods, recent molecular techniques, particularly high-throughput sequencing, have been used to study microbiota at different body sites, such as skin [9, 10], nose [23], intestines [24–26], urinary tract [27], and oral cavity [28].

A previous study using high-throughput sequencing found that Staphylococcus spp., Pseudomonas spp., Parvimonas spp., Bacteroides spp. and Actinomyces spp. were over-represented in the OE groups [29]. Our observation also showed that Staphylococcus spp. and Pseudomonas spp. were largely found in the affected groups. On the other hand, Corynebacterium spp., and Proteus spp. were found instead of Parvimonas spp., Bacteroides spp. and Actinomyces spp. Despite no report of canine otitis, Tissierella spp., Achromobacter spp., and Acinetobacter spp. caused otitis cases in humans [30–34]. Interestingly, Tissierella spp. and Achromobacter spp. were found in erect-ear dogs with OE (3.95% and 2.07%, respectively) in this study. Furthermore, Acinetobacter spp. colonized dominantly in erect-ears and pendulous-ears dogs. According to previous studies [35–38], Porphyromonas spp. were predominantly found in the canine oral cavity. The present result gained from NGS revealed the presence of Porphyromonas spp. in both healthy and OE groups. Therefore, it might be presumed that microflorae from canine oral were probably transmitted to ears by licking behavior. Thus far, there have not been any reports regarding OE in dogs. However, otitis infections with Porphyromonas spp. were evident in humans [39] and a Parma wallaby [40].

It was observed that canine breeds with pendulous ears had a higher prevalence of OE when compared to breeds with erect ears [15, 41]. Environmental factors such as airflow, heat retention, and humidity might be different in each breed affecting susceptibility to the disease. However, Yoshida and colleagues [17] demonstrated that these factors were not significantly different between the healthy and the otitis dogs. Therefore, the current study demonstrated that the richness and diversity of the bacterial community in the OE groups seemed lower than those in the healthy controls, particularly the Shannon and Simpson’s diversity indices between healthy and OE erect-ear dogs. This study is consistent with the previous one using next-generation sequencing, reporting that dogs with OE had decreased alpha diversity when compared to healthy dogs [29].

Conclusions

The present study demonstrated bacterial residents in external auditory canals that might be associated with OE in dogs. Comparable with previous studies using culture-dependent methods, predominant Corynebacterium spp., Pseudomonas spp., Staphylococcus spp., and Proteus spp. in OE cases were also revealed in this study. Furthermore, high-throughput sequencing might disclose some potential emerging pathogens including Tissierella spp., Acinetobacter spp., and Achromobacter spp., which have not been reported in previous cases of canine OE. Despite the relatively small sample size examined, the feasibility and cost-effectiveness of utilizing high-throughput sequencing to study microbiome profiles of canine auditory canals were considered in this pilot study. However, a deeper and broader diversity of the microbiome in canine external auditory canals was demonstrated in this case-control study. Therefore, larger sample sizes are further required for an evidence-based investigation.

Supplementary information

Below is the link to the electronic supplementary material.

Funding

This research was funded by the Thailand Science Research and Innovation Fund (TSRI) (Grant Number CU_FRB65_hea (27)_034_30_15). Besides, Suthat Saengchoowong is grateful to his doctoral funding through Graduate School and Faculty of Medicine, Chulalongkorn University (the 100^th Anniversary Chulalongkorn University Fund for Doctoral Scholarship; the 90^th Anniversary of Chulalongkorn University Ratchadaphiseksomphot Endowment Fund; and the Overseas Research Experience Scholarship for Graduate Students of Chulalongkorn University).

Data availability

All data and materials are available upon request.

Declarations

Ethical approval

This research was done with the approval of the Chulalongkorn University Animal Care and Use Committee (Animal Use Protocol No. 1775005).

Competing interests

The authors declare no competing interests.