Occurrence and genotypes of Cryptosporidium spp., Giardia duodenalis, and Blastocystis sp. in household, shelter, breeding, and pet market dogs in Guangzhou, southern China

Cryptosporidium spp., Giardia duodenalis, and Blastocystis sp. are common intestinal protozoans that infect humans and animals worldwide. A survey that assessed the prevalence, molecular characteristics, and zoonotic potential of these pathogens was conducted on a variety of dogs in Guangzhou, southern China. A total of 651 canine stool samples from household (n = 199), shelter (n = 149), breeding (n = 237), and pet market dogs (n = 66) were collected from eight districts in Guangzhou. Cryptosporidium spp., Giardia duodenalis, and Blastocystis sp. were detected by PCR amplification of the SSU rRNA gene. Giardia duodenalis-positive specimens were further assigned into assemblages using the glutamate dehydrogenase gene. Cryptosporidium spp., G. duodenalis, and Blastocystis sp. were found in 21 (3.2%), 20 (3.1%), and 35 (5.4%) samples, respectively. The overall prevalence of shelter dogs (40.28%, 60/149) was significantly higher than that of household (3.0%, 6/199), breeding (2.1%, 5/237), and pet market dogs (7.5%, 5/66) (χ2 = 154.72, df = 3, P < 0.001). Deworming in the past 12 months had a strong protective effect on the risk of contracting parasite infections (P < 0.001). No significant differences were detected between age or sex groups (P > 0.05). Dog-specific C. canis (n = 19) and zoonotic C. parvum (n = 2) were the only two Cryptosporidium species. Sequence analysis revealed the presence of three G. duodenalis assemblages: dog-specific assemblages D (n = 14) and C (n = 5), and cat-specific F (n = 1). Zoonotic Blastocystis ST3 (n = 28) was the dominant subtype, followed by ST1 (n = 6) and ST10 (n = 1). To our knowledge, this is the first large-scale investigation on the occurrence and molecular characteristics of Blastocystis sp. in dogs in China. Our results indicated that the dogs seemed to play a negligible role as reservoirs for Cryptosporidium spp. and G. duodenalis transmission to humans, but they are potential novel suitable hosts of Blastocystis sp. A strict sentinel surveillance system of dogs should be established to minimise the zoonotic risk of spreading blastocystosis among humans and dogs.

www.nature.com/scientificreports/ region 12,13,19,36 ; however, nothing is known about the occurrence and molecular characterisation of Blastocystis sp.. The purpose of our study was to estimate the overall occurrence of Cryptosporidium spp., G. duodenalis, and Blastocystis sp. in dogs living in areas of Guangzhou and assess the zoonotic potential between humans and dogs.

Methods
Study design. The study was conducted in Guangzhou. Guangzhou is one of the largest metropolitan cities in southern China (coordinates 3° 28′-25° 31′ N and 108° 13′-119° 59′ E); it covers an area of 7434 m 2 and has a population of about 140 million. The annual average temperature is 20-22 °C and the average relative humidity is 77% (Guangzhou Statistics Bureau; https ://tjj.gz.gov.cn/). A total of 651 fresh faecal samples were randomly collected from 199 household dogs (from four pet hospitals located in four different districts in urban Guangzhou: Tianhe, Baiyun, Huadu, and Panyu Districts), 149 shelter dogs (from two shelters located in suburban Luogang and Huangpu Districts), 237 breeding dogs (from two breeding centres located in suburban Conghua and Nansha Districts), and 66 pet market dogs (from one pet market in urban Tianhe District), with or without a history of illness, on a single occasion between January and December 2018 ( Fig. 1). Faecal samples from shelters, breeding centres, and pet markets were collected as soon as practicably possible after defaecation by our research staff, either directly from the floor of the cage or per rectum. Care was taken to avoid sampling faecal material that had contacted the ground at the time of sampling. All samples from household dogs were collected immediately after natural defecation and donated by the dog owners, who provided consent for the use of samples from their animals in the survey. All collected fresh samples had no apparent diarrhoeal symptoms at the time of sampling. The samples were placed into clean plastic bags marked with ID numbers corresponding to the date, origin, age, sex, and whether the dog was dewormed in the past 12 months. The plastic bags were sealed and immediately placed onto ice packs in an insulated container. Samples were transported to the laboratory, stored at 4 °C, and processed no later than 24 h after collection.
DNA extraction. A  PCR detection. Cryptosporidium spp. was detected by nested PCR amplification of an approximately 830bp fragment of the 18S rRNA gene as previously described 37 . G. duodenalis was detected by nested PCR amplifications of a 290-bp product of the 18S rRNA gene and a 520-bp polymorphic fragment of the glutamate dehydrogenase (GDH) gene as described 38,39 . To detect Blastocystis sp., an approximately 600-bp fragment of the 18S rRNA gene was amplified by single primer PCR as previously described (see Supplementary Table S1 online) 40   www.nature.com/scientificreports/ Co., Ltd., Otsu, Japan), and 1 μl of genomic DNA. Each specimen was analysed in duplicate using positive (cattle-derived DNA) and negative (sterile water) controls.
Sequence and phylogenetic analysis. The positive secondary PCR products were directly sequenced by GENEWIZ (Suzhou, China). Sequence accuracy was confirmed with two-directional sequencing. All raw sequencing data were viewed and aligned by eye in Chromas Pro 1.33 (Technelysium Pty. Ltd., Helensvale, Queensland, Australia). The identity of species/assemblages/subtypes was established by comparing the obtained sequences with reference sequences from the National Center for Biotechnology Information (https ://www.ncbi. nlm.nih.gov/) database using Clustal X 2.1 (https ://www.clust al.org). Phylogenetic analysis was performed by a neighbour-joining (NJ) analysis in MEGA 7.01 (https ://www.megas oftwa re.net/) based on the Kimura 2-parameter model using 1000 bootstrap replicates. The p-distance model was selected as the most suitable model.

Statistical analysis.
Differences between prevalence and the dog's origin (households, pet market, breeding centres, and shelters), age (≤ 6 months vs. > 6 months), sex, and deworming conditions (dewormed vs. nondewormed in the past 12 months) were compared using a χ 2 test in SPSS 22.0 for Windows (SPSS Inc., Chicago, IL, USA) with 95% confidence intervals. Differences at P < 0.01 were considered significant.

Statement of informed consent, ethics approval and guidelines.
Prior to fecal specimen collection, we showed an informed consent to the dogs owner. This informed consent provides some information, including the purpose of the study, research approval number, the benefits and risks that may bring to their animals participating in the study. The informed consent was obtained from the dog owners. Appropriate permission in written form was obtained from the animal owners. During specimen collection, all animal work strictly followed the guidelines relating to the recommendations from the Guide for the Care and Use of Labora

Results
Overall prevalence of Cryptosporidium spp., G. duodenalis, and Blastocystis sp.. Faecal samples from 651 dogs were tested by PCR for the presence of Cryptosporidium spp., G. duodenalis, and Blastocystis sp. The prevalence and 95% confidence intervals are summarised in Table 4. Cryptosporidium spp., G. duodenalis,  G. duodenalis assemblages. Sequence analysis revealed the presence of three different G. duodenalis assemblages: D (n = 14), C (n = 5), and F (n = 1). Three SSU rRNA nucleotide sequences of G. duodenalis assemblages D, C, and F were 100% identical to the GenBank reference sequences DQ385549, DQ385548, and JX275387, respectively. The genotypes identified by the GDH gene were fully consistent with those identified by  Table 5. Species/assemblages/subtypes of Cryptosporidium, G. duodenalis and Blastocystis in dogs.

Factor Category
No. tested

G. duodenalis Cryptosporidium Blastocystis
Site Shelters 149 D (11), C (2) C. canis (11) No mixed assemblage infections were identified (see Supplementary Table S2 online). Moreover, four GDH subassemblage D nucleotide sequences were identified in this study. One sub-assemblage D sequence was identical to GenBank EF507636 reference sequences (n = 5), and the remaining three sequences had minor differences from EF507636, including two single nucleotide polymorphism (SNPs) in four specimens (C to T substitution at position 162 and T to G substitution at position 324), three SNPs in two specimens (C to T substitution at position 162, A to G substitution at position 174, and A to T substitution at position 311), and three SNPs in three specimens (T to C substitution at position 109, A to G substitution at position 183, and A to T substitution at position 312). All GDH sub-assemblage C sequences were identical to each other and showed 99% sequence similarity to the EF507621 reference sequence, with one SNP difference at position 12 (T → C). The assemblage F nucleotide sequence was identical to the corresponding KF993737 sequence of from a cat in China (Fig. 2).
Blastocystis subtypes. DNA sequencing of the SSU rRNA PCR products from the 35 Blastocystis-positive samples and sequence analysis indicated the existence of subtypes ST3 (n = 27), ST1 (n = 6), ST10 (n = 1), and unknown ST (n = 1). Twenty-seven nucleotide sequences identified as ST3 were identical to each other and had 100% similarity to the MK782518 reference sequence from urticaria patients in Brazil. Six ST1 nucleotide sequences included two different nucleotide sequences with 100% similarity to the GenBank reference sequences MK782501 in four specimens and MK782521 in two specimens. The ST10 nucleotide sequence had 100% similarity to a cattle-derived sequence from Malaysia in GenBank (MK240480). The remaining nucleotide sequence was not assigned a subtype by the sequence typing database, and was 100% homologous with the published sequence MK511788 from Malaysia. Phylogenetic analysis using NJ analyses clustered Blastocystis subtypes obtained in the present study into three subtypes (ST1, ST3, and ST10), and the unknown subtype was grouped into subtype ST1 (Fig. 3).

Discussion
In this study, Cryptosporidium spp., G. duodenalis, and Blastocystis sp. were found at considerably low prevalences (3.2%, 3.1%, and 5.4%, respectively) in the surveyed canine populations. This finding is comparable to the prevalences in Poland (2.0% for G. duodenalis) 41 , India (3.0% for G. duodenalis) 42 , the United States (2.0% and 3.8% for Cryptosporidium) 43,44 , China (3.8% and 4.9% for Cryptosporidium) 10,14 , Italy (3.3% for Cryptosporidium) 9 , Australia (2.5% for Blastocystis) 29 , Brazil (2.6% for Blastocystis) 31 , and France (3.4% for Blastocystis) 20 . However, higher prevalences of 21% of Cryptosporidium from Japan, 36.5% of G. duodenalis from Spain, and 37.5% of Blastocystis from Colombia were also previously found in dogs 22,24,33 . Aside from geographical considerations, many factors can contribute to this difference in the prevalence, including a dog's age, origin, health status, and examination methods used. Importantly, the true prevalences may be underestimated because of the intermittent shedding of oocysts/cysts, low parasitic burdens, and invalid amplification 23 . Therefore, it is important to perform a well-designed longitudinal study that includes appropriate sampling methods and a combination of diagnostic tests (e.g. microscopic examination, antigen assay, and PCR assay) to estimate the real prevalence.
In risk factor analysis, the overall and pathogen-specific prevalences of shelter dogs were significantly higher than those of household, breeding, and pet market dogs. Poor care conditions may be a significant factor that contributed to the high prevalence in shelter dogs 23 . Additionally, these shelter dogs originally roamed free in the nearby streets before they were found by local citizens and sent to the shelters, which increased their exposure to a variety of pathogens. No significant age-and sex-associated differences were detected in the prevalence of these three pathogens, which is consistent with observations of previous studies in China, Japan, Colorado, France 11,20,21,44 , and Australia 10,12,45 . As expected, deworming had a significant negative effect on the risk of overall and single pathogen infections. This could be related to the fact that the anthelminthic ingredients used for dogs in China mainly include ivermectin, pyrantel, praziquantel, pyrantel pamoate, febantel, nitazoxanide, metronidazole, and fenbendazole, most of which are also effective against protozoan infections, such as Giardia spp., Cryptosporidium spp., and Blastocystis sp. [46][47][48] . Thus, pet hygiene management is suggested to be a major risk factor for contracting these pathogens in dogs.
The sequencing data revealed that dogs were predominantly infected by the expected host-specific species C. canis. Another interesting outcome was the identification of C. parvum in two of the canine isolates genotyped. Cryptosporidium parvum is the most frequent species known in cryptosporidium infections of humans and has resulted in several zoonotic outbreaks 49 . However, because of the omnivorous nature of dogs, C. parvum detected from dogs in this study may have been present because of accidental acquisition or mechanical carriage of C. parvum oocysts of anthroponotic origin via environmental contamination. Although the host-specific species C. canis also colonised individuals in hospitals, including children, HIV patients, and even immunocompetent individuals [50][51][52] , the infections in humans were likely transient 53 . Based on our data and that of other studies, we conclude that dogs do not seem to be suitable reservoirs for Cryptosporidium spp. transmission to humans, therefore, posed a limited risk to humans. This is consistent with the findings reported in dog populations in eastern Spain, where most of the genotypes identified seemed to be primarily transmitted within canine cycles 24 .
Regarding G. duodenalis, dogs were infected by assemblages D and C. Surprisingly, the supposedly catspecific assemblage F was also found in one household dog, which was consistent with a previous report in Beijing, China 14 . Considering there was little possibility of the specimen being contaminated with cat faeces, as individual faecal samples were freshly collected by pet dog owners who kept only a single companion animal, it is more likely that the dog was transiently infected by ingesting parasite cysts of cat origin. Assemblages D and C have strong host specificities and have been mainly detected in canines. Both are considered of limited zoonotic relevance, although sporadic cases of human infections have been frequently detected in travellers, children,  [53][54][55] . The dominant appearances of G. duodenalis assemblages D and C in dogs in our study suggested that zoonotic transmission of giardiasis rarely occurs between humans and dogs. A high diversity of Blastocystis subtypes has been identified in dogs worldwide and the subtype constitution was observed to differ among geographical regions, such as ST1 and ST4 in China; ST1, ST4, ST5, and ST6 in India; ST1 and ST10 in the USA; ST1, ST3, and ST4 in Australia; and ST1 and ST2 in Thailand 27,29,35,56,57 . This is the first large-scale survey on the prevalence and genetic characteristics of Blastocystis sp. in dogs of various origins in China, and we detected the presence of subtypes ST3, ST1, and ST10. Among the subtypes in this study, ST3 and ST1 are the two predominant subtypes. The subtype distribution of Blastocystis in dogs in our study consistent with most populations in humans around the world 15,57,58 . In China, in a cross-sectional survey in humans from four epidemiological settings, ST3 was the predominant type, accounting for 60.4% (n = 116) of 192 positive specimens, followed by subtype ST1, accounting for 24.5% (n = 47) 58 . A report from Argentina also showed that Blastocystis ST3 was the most prevalent subtype (48 cases) among 76 patients infected with Blastocystis, and other subtypes identified were ST1 (14.9%), ST6 (7.5%), and ST2 (5.9%) 57 . Moreover, ST3 and ST1 were found in dogs and their owners in Australia, the Philippines, and Turkey 15,35,59 . All of these studies suggested that ST3 and ST1 might be the important sources of human-to-human or animal-to-human transmission of Blastocystis, although more environmental factors and or other animal sources should be included. Blastocystis ST10 was found in one dog sample; this subtype is frequently identified in common livestock, including cattle, www.nature.com/scientificreports/ sheep, goats, and deer 3,16,60 , and occasionally found in wild animals, pigs, dogs, and cats 20,27 . Taken together, our results revealed that canines are a novel reservoir for Blastocystis. We conducted partial assessment of the zoonotic potential of our canine isolates using only molecular epidemiological data, because assessing the risk for zoonotic transmission of these pathogens from dogs to/from humans is difficult. The only way to properly determine zoonotic transmission is by conducting case-control studies that assess the genotypes/subtypes of these pathogens by using appropriate molecular typing tools in human and canine populations that maintain permanent close contact in the same spatial and temporal setting 23,34,61,62 . However, our epidemiological study still generated baseline information and determined the genetic diversity of these pathogens in the investigated region.

Conclusions
The prevalence and molecular characteristics of Cryptosporidium spp., G. duodenalis, and Blastocystis sp. were determined in dogs in Guangzhou. Our risk factor analysis showed that management of pet hygiene may be a major risk factor for contracting these pathogens in dogs. Our results suggest that dogs do not seem to be suitable reservoirs of human giardiasis or cryptosporidiosis in the investigated region, but may act as novel suitable hosts of human blastocystosis. Strict sentinel surveillance of dogs, especially stray dogs, should be established to minimise the risk of spreading blastocystosis among humans and dogs.