Large-scale survey of a neglected agent of sparganosis Spirometra erinaceieuropaei (Cestoda: Diphyllobothriidae) in wild frogs in China

Background In China, frogs play an understudied role in the spread of human sparganosis (caused by the larval form of Spirometra). However, our knowledge about the prevalence of sparganum infection in frogs remains fragmented, and the taxonomic identification of the parasite is still controversial. Methodology/Principal findings The prevalence of sparganum infection in wild frogs was surveyed at 145 geographical locations from 28 of the 34 provinces/autonomous regions/municipalities in China for six years. The collected sparganum isolates from the different locations were subjected to molecular identification by a multiplex PCR assay and then were analysed with clustering analysis. In the survey, sparganum infection was found in 8 out of 13 of the collected frog species, and the most frequently infected species was Pelophylax nigromaculatus (the infection rate was up to 14.07%). Infected frogs were found in 80 of the 145 surveyed locations. The sparganum infection rates in the wild frogs in several regions of China were still high (above 10%), especially in South and Southwest China. A total of 72 spargana were selected for molecular identification, and the clustering analysis showed that sequences from the Chinese isolates were very similar to those identified as from Spirometra erinaceieuropaei. However, the taxonomy of the genus remains confused and further analysis is required. Conclusions Eating wild frogs is associated with considerable health risks in China. Several traditional Chinese folk remedies may increase the risk of infection. The sparganum isolates in China are most likely from S. erinaceieuropaei, but new studies, especially comprehensive morphological analyses, are needed in the future.


Introduction
Human sparganosis is a neglected food-borne parasitic disease caused by the larval forms (procercoid/plerocercoid) of the species in the genus Spirometra [1]. Despite its global distribution, most cases occur in Eastern and Southeastern Asia [2]. Frogs, as the second intermediate host in the life cycle of Spirometra, play an important role in the spread of the disease in China [3,4]. Humans can be infected through the consumption of raw or undercooked frog meat or by using raw frog flesh in traditional poultices [5]. In addition, eating raw frog meat is a traditional custom for many people in some areas of China [6][7][8]. Recently, more than 30 autochthonous human cases caused by the ingestion of live frog tadpoles have been recorded in central China [4]. The number of reported cases of human sparganosis has exceeded 1300 in China [3,9], but this number is believed to be an underestimate. The actual number of infections may be far higher than those estimated because many cases may not be recognized or reported. As a result, the investigation of infection by plerocercoid larva (sparganum) in frogs is therefore valuable for food safety and the prevention and control of human sparganosis in China [10].
The human sparganosis has been reported to be distributed in 26 of 34 provinces/autonomous regions/municipalities in China, with the majority of cases in Southern and Eastern China [3]. Sporadic investigations of the sparganum infection in frogs have been performed in several regions of China, such as in Henan, Hunan and Guangdong provinces [6,[11][12][13]; however, there is still a lack of data on the geographical distribution of sparganum infection in wild frogs in all regions of China where the parasite is endemic. In this study, we conducted the first large-scale survey of sparganum infection in wild frogs from 145 geographical locations that covered 88.9% of the endemic regions of human sparganosis.
Although a medically important genus, the identification and taxonomy of Spirometra species have been controversial for a long time [14]. The most recent review concluded that there were only 4 valid species (S. erinaceieuropaei, S. mansonoides, S. pretoriensis and S. theileri) in the genus Spirometra [15]; however, dozens of nominal species of Spirometra have been described in other publications [16][17][18]. In addition, several unidentified species have been reported in South America [19]. Using different genetic markers or the complete mitochondrial genome, Spirometra isolates from Australia, New Zealand, Indonesia, India, Japan and Korea [20], mainland China [6, 10, 12-14, 21, 22] and Hong Kong [23] have been assigned as S. erinaceieuropaei. In contrast, one isolate from Korea was identified as S. decipiens (KJ599679) through both morphological and genetic methods [24]. However, the morphological identification of the Korean S. decipiens did not have detailed evidence, such as the study of the type material of S. decipiens or collection of new material from the type locality, and two recent phylogenetic analyses have suggested that the Korean S. decipiens was likely conspecific with the Asian isolates of S. erinaceieuropaei [14,19], so we refer to the Korean S. decipiens as the "Korean S. decipiens genotype" here for careful consideration. Another Korean isolate, S. erinaceieuropaei (KJ599680) (which we call the "Korean S. erinaceieuropaei genotype"), was genetically more distinct from other Asian isolates of S. erinaceieuropaei [14,19]. In China, most reported cases of human sparganosis have been caused by S. erinaceieuropaei. In contrast, both the "Korean S. erinaceieuropaei genotype" and "Korean S. decipiens genotype" have been identified as infectious to humans [24]. In addition, several sparganum isolates collected from snakes (Dinodon rufozonatum and Agkistrodon saxatilis) in China were identified as the "Korean S. decipiens genotype" [17]. Therefore, the exact identification of Spirometra species requires further investigation.
To clarify the exact taxonomy of the collected sparganum isolates and to explore whether there were other species of Spirometra in China, we intended to identify the collected samples using the multiplex PCR assay developed by Jeon et al. [18]. The multiplex PCR method was useful for distinguishing the "Korean S. erinaceieuropaei genotype" and "Korean S. decipiens genotype" [18]. More specifically, the aims of this study were as follows: (1) to conduct a largescale survey of sparganum infections in wild frogs in mainland China; and (2) to identify the collected sparganum specimens using molecular identification.

Ethics statement
Our study was performed strictly based on the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Health Commission of China. The protocol was approved by the Life Science Ethics Committee of Zhengzhou University (Permission No. 2013-0137). All of the frog specimens were collected from paddy fields after obtaining the permission of farm owners. No specific permits were required by the authorities for the collection of frog samples.

Study site
According to the most recent review of human sparganosis in China [3], the endemic region for the disease encompassed 26 provinces/autonomous regions/municipalities (Fig 1A). In addition, 5 nonendemic regions were also investigated in our survey: Tianjin (TJ) municipality, the Shaanxi (SaX) and Shanxi (SX) provinces and the Ningxia (NX) and Inner Mongolia (IM) autonomous regions. In total, 145 geographical sampling sites in 28 of the 34 administrative regions in China were included ( Fig  1B). The precise locality (coordinates), origin and date of collection for each sampling site are presented in the Supplementary S1 Table.

Examination of frogs and sample collection
The presence of spargana in wild frogs was evaluated according to previously described methods [12]. Briefly, the frogs were caught in paddy fields or other wild environments, and then the collected frogs were euthanized using ethyl-ether anaesthesia and were weighed and skinned. The presence of spargana in the skeletal muscles was carefully observed visually. Once identified, the spargana were isolated using small scissors and forceps (Supplementary S1 Fig). The white spargana are highly active and able to change shape in physiological saline. The larvae show an enlarged head and many horizontal folds on the surface of the body when examined under a microscope. The number of identified spargana, parasitizing sites, infection intensity and length of each sparganum were counted and measured. The measurements were expressed as ranges, with the means followed by SDs. After that, all collected larvae were preserved in alcohol (99% ethanol) for molecular identification.

Multiplex PCR and sequencing analysis
The multiplex PCR method was used to identify the sparganum isolates collected from the different geographical locations. For each specific locality, one sparganum was selected for the multiplex PCR assay. In total, 72 spargana were used for the molecular identification, and specimens from 8 locations were excluded because of the close proximities of the sampling sites (Supplementary S2 Table). The multiplex PCR assay was performed using two sets of species-specific primers (Supplementary S3 Table) and the methods described in Jeon et al. [18]. In brief, two sets of species-specific primers, Se/Sd-1800F + Se-2018R plus Se/Sd-7955F + Se-8356R, and Se/Sd-1800F + Sd-2317R plus Se/Sd-7955F + Sd-8567R, were designed based on the mitochondrial nucleotide sequences of cytochrome b (cytb), NADH dehydrogenase subunit 4L (nad4L), NADH dehydrogenase subunit 4 (nad4), cytochrome c oxidase subunit 1 (cox1), and large subunit ribosomal DNA (lrDNA) from "S. erinaceieuropaei (KJ599680)" and "S. decipiens (KJ599679)". As described, using a combination of "Korean S. erinaceieuropaei genotype" specific primers (Se/Sd-1800F + Se-2018R; Se/Sd-7955F + Se-8356R) should generate two "S. erinaceieuropaei" specific bands of 239 bp and 401 bp, respectively. In contrast, the "Korean S. decipiens genotype" specific primers (Se/Sd-1800F + Sd-2317R; Se/Sd-7955F + Sd-8567R) should generate two "S. decipiens" specific bands of 540 bp and 644 bp, respectively. In addition, the amplified PCR products were sequenced to confirm the results of the electrophoretic analysis. The PCR products were purified using the EasyPure PCR Purification Kit (Transgen, China) and sequenced bidirectionally using an automated sequencer (ABI Prism 3730 XL DNA Analyzer; ABI Prism, Foster City, CA) at the Genwiz Company (Suzhou, China). All sequences were deposited in the GenBank database under accession numbers MN842190-MN842261 and MN861812-MN861883. Moreover, eight complete mitochondrial genomes of S. erinaceieuropaei (JQ267473, KY114886-114889, AP017668, AB374543 and KU852381), two mitochondrial genomes from the Korean genotypes, the "S. decipiens genotype" (KJ599679) and "S. erinaceieuropaei genotype" (KJ599680), from GenBank were included as reference sequences. One Schistocephalus species (Schistocephalus solidus), two Solenophoridae species (Duthiersia expansa and Scyphocephalus bisulcatus) and one Cephalochlamydidae species (Cephalochlamys namaquensis), were used as outgroups. These species were selected as the closest known groups to Spirometra based on the study of Waeschenbach et al. [25]. The sequences were aligned in MEGA v.6.06 [26] with the default settings. The variable sites, nucleotide compositions and pairwise distances were also estimated in MEGA. The clustering analysis was performed with two methods: the maximum likelihood (ML) method and Bayesian inference (BI). The substitution model of the dataset was selected by jModelTest v0.2 [27]. The ML analysis was conducted in MEGA, and the confidence levels in each node were assessed with the boot-strap method (1000 pseudoreplicates). The BI analysis was performed in MrBayes v.3.2 [28]. The analysis consisted of two runs, each with four MCMC chains running for 5,000,000 generations, which were sampled every 100th generation.

Molecular identification
As shown in Fig 2A, all sparganum isolates collected from the 72 different geographical locations revealed two bands (540 bp and 644 bp) after amplification with the "Korean S. decipiens genotype" specific primers; however, when using the "Korean S. erinaceieuropaei genotype" specific primers, no bands were detected. For further identification, the PCR products were sequenced for comparison with the 10 referenced mitochondrial genomes of the Spirometra tapeworms from GenBank. The first band (using the primer pair of Se/Sd-1800F + Sd-2317R) was sequenced and consisted of 622 bp after trimming. The corresponding tree topologies generated by the two methods (ML and BI) were identical. As shown in Fig 2B, the earliest divergence gave rise to the Cephalochlamydidae (C. namaquensis), followed by S. solidus, and then to the 2 samples of Solenophoridae (D. expansa and S. bisulcatus). The last divergence gave rise to the remaining Spirometra isolates; the isolates collected in this study and the 10 Spirometra mitochondrial genomes from GenBank made up a single, highly supported group. Within Spirometra, two main clades were revealed. One clade included only a single isolate of the "Korean S. erinaceieuropaei genotype" (KJ599680). The Chinese isolates, "Korean S. decipiens genotype" (KJ599679) and 8 other S. erinaceieuropaei mitochondrial genomes were clustered in the other clade. Using the sequences of the second band (515 bp, using the primer pair of Se/Sd-7955F + Sd-8567R), the clustering pattern of the Spirometra samples was consistent with that generated based on the sequences of the first band (622 bp) (Supplementary S2 Fig).
These phylogenetic patterns suggested that the taxonomic positions of the "Korean S. decipiens genotype" (KJ599679) and "Korean S. erinaceieuropaei genotype" (KJ599680) are controversial.

Discussion
After the first human case of sparganosis was reported in Xiamen, Fujian Province in 1882, more than 1300 cases were reported from 1949 to 2014 in China [3]. Eating raw or undercooked frog/snake meat, using raw frog/snake flesh in traditional poultices and ingesting live frog tadpoles are risk factors for infection [4]. Therefore, in addition to snakes, frogs also play an important role in the spread of human sparganosis in China. From 2013-2018, we conducted a large-scale survey of sparganum infection in wild frogs from 145 geographical locations to understand the prevalence of spirometrid tapeworms in wild frogs.  [6,12,29,30]. In our survey, sparganum infection was also found in P. nigromaculatus, F. limnocharis and H. chinensis. In addition, 5 other frog species: P. plancyi, S. latouchii, B. guentheri, Q. spinosa and O. margaretae, have been proven to be sensitive to sparganum infections. Among these frog species, the most frequently infected species was P. nigromaculatus, which indicates that the species is important in the prevention and control of human sparganosis in China. Although many specimens of R. chensinensis have been collected here, no sparganum-positive frogs were found. In addition, specimens of R. rugulosa and B. gargarizans were not collected in this survey.
Many human cases of sparganosis have been reported in Hunan province; accordingly, the sparganum infection in wild frogs in Hunan was the highest in our survey, but it was still slightly lower than that in a previous survey in this region [6]. High prevalence of human sparganosis has also been reported in Guangdong and Zhejiang [3]; however, the infection rates in these regions were moderate. For example, in the Guangdong province, which accounts for 10% of the reported human sparganosis cases in China [7], the infection rate was 7.73% in wild frogs. However, a previous survey of the province suggested that 35% of wild frogs had sparganum infections [8]. In addition, frog meat, as a type of "bushmeat", has an important role in the cuisine of Guangdong, and approximately 59.9% of the residents eat frog meat [8,11]. Thus, the risk of infection in Guangdong remains high. In Zhejiang province, the infection rate detected in this study was 10.86%; in contrast, a previous report found that the infection rate in Hangzhou, Zhejiang, reached 31.15% [31]. Compared with the other regions, we chose the most sampling sites and examined the highest number of frogs in Henan province in central China. Therefore, the data generated in this region were the most representative. The infection rate in Henan was lower than those in most of the southern and southwestern regions; however, the number of reported human cases was higher than those from the other regions. Before 2006, human sparganosis in Henan was rarely reported (only 3 imported cases from southern China). After 2006, twenty autochthonous cases caused by the ingestion of live tadpoles emerged because some rural villagers in this region believe that live tadpole consumption has a medicinal role in skin diseases [4,32], which suggests that several traditional Chinese folk remedies are problematic for the prevention of human sparganosis. This incident was not the only case of human sparganosis that was caused by folk remedies: many cases of human sparganosis in Fujian in Southeastern China were caused by using fresh frog flesh as a poultice for sore eyes [33]. Generally, the infection rates in most of the south and southwest regions (Guangxi, Sichuan, Hainan, Guizhou and Chongqing) were higher than those in the other regions. These regions contain many minorities who have different dietary customs, and frog meat is a delicacy for many people [13]. Although few human cases have been reported, the high sparganum infection rate suggests that people living in these regions are at an increased risk of infection. One surprising finding was that no infected frogs were found in Northeast China (Heilongjiang, Jilin and Liaoning) over the 6 year study period, although 7955F + Sd-8567R). Lanes 2,4,6,8,10,12,14,16,18,20,22,24,26,  Survey of neglected spirometrid tapeworms in wild frog in China many cases have been reported in these regions [34][35][36]. The possible reasons were that (1) the number of sampling sites and the sample sizes were small, especially in Liaoning and Jilin, as only 2 and 3 sampling sites were selected, respectively; (2) the main frog species collected in the northeastern regions was R. chensinensis, but no infection was found in R. chensinensis frogs in this study; or (3) the reported human cases in Northeast China were probably caused by the importation of infected frogs/snakes from other regions of China. Human sparganosis cases have also been reported in Beijing and Hebei in North China and Qinghai in Northwest China, but no sparganum-positive frogs were detected in our survey. It should come as no surprise that no infected frogs were found in the non-endemic regions of Shaanxi, Shanxi, Ningxia and Inner Mongolia in North and Northwest China. In summary, this survey indicated the following: (1) sparganum infection in wild frogs was detected in 16 of the 23 surveyed regions where human sparganosis is endemic in China; (2) eating wild frogs is associated with considerable health risks in China, especially in the southern and southwestern regions, and improper cooking methods may increase the risk of infection; and (3) several traditional Chinese folk remedies play an important role in the spread of human sparganosis; therefore, health education should be strengthened to prevent the transmission of this disease.
The accurate morphological identification of Spirometra tapeworm species must be based on the morphology of the adult worms; however, the specimens obtained in the field are usually larval forms. It is impossible to distinguish plerocercoids using only morphological characteristics. Therefore, an increasing number of researchers have utilized molecular methods to identify spirometrid tapeworms [19,20,37]. Recently, Jeon and colleagues developed a multiplex PCR assay to specifically distinguish the Korean "S. erinaceieuropaei" and "S. decipiens" genotypes [18]. Using the multiplex PCR system, isolates from 72 different locations revealed uniform electrophoretic bands. Furthermore, the clustering analysis of the sequenced PCR bands suggested that the taxonomic positions of the "Korean S. decipiens genotype" (KJ599679) and "Korean S. erinaceieuropaei genotype" (KJ599680) are controversial. In agreement with our results, using the cox1 gene, Almeida et al [19] suggested that "S. decipiens" (KJ599679) is likely conspecific with the Asian isolates of S. erinaceieuropaei and that "S. erinaceieuropaei" (KJ599680) is a different species of Spirometra. Therefore, the controversial results here probably indicate the following: (1) if "S. decipiens" (KJ599679) was correctly identified, all of the Chinese isolates and the 8 mitochondrial genomes in GenBank should be S. decipiens; otherwise, "S. decipiens" (KJ599679) should be a conspecific of S. erinaceieuropaei; and (2) "S. erinaceieuropaei" (KJ599680) might be a different species of Spirometra or a special genotype that differed from those of other isolates. Nevertheless, as described above, the accurate identification of a species must be based on its morphological characteristics, so new studies especially comprehensive morphological analyses, are needed to accurately identify Spirometra species.
Traditionally, the plerocercoids in the Far East were called Manson's tapeworm. The scientific name of Manson's tapeworm has been described in many text books as Spirometra mansoni, but as more detailed studies have been conducted, parasitologists have divided it into many "species" according to its morphological characteristics [16]. However, Iwata [16] proposed that only S. erinacei exists and that other species have been described only because the characteristics used for identification vary with the environmental conditions of the host and the developmental stage and especially the position of the proglottid in the strobila. S. erinacei was first reported by Rudoiphi in 1819 as a sparganum isolated from a European hedgehog (Erinaceus europaeus). Then, the name was changed to S. erinaceieuropaei in 1959 [38]. Currently, this species has been reported as the major aetiological agent of human sparganosis worldwide [3,5,39]. On the other hand, Mueller [40] concluded that S. mansonoides is also a valid species based on a uterine trait. Later, the validation of S. mansonoides was supported by the molecular data [41]. Recently, Jeon et al. [24] reported that the "species" S. decipiens (Gedoelst, 1911) in Korea is infectious to humans. They noted that the main characteristic for "S. erinaceieuropaei" and "S. decipiens" differentiation is the spiral coiled uterus; "S. erinaceieuropaei" has 5-7 complete coils, while "S. decipiens" has 4-4.5 coils. According to Iwata [16], the uterus generally has 3-6 coils but can have 7-8 coils in long, young proglottids or only 2 coils (rarely one) in the anterior proglottids of young worms or the posterior proglottids of overmature worms; thus, the number of uterine coils cannot be regarded as a specific characteristic. In addition, the forms of the uterus, testes, vitellaria, uterine pore, genital pouch, etc. are changed by the conditions of contraction and cannot be regarded as specific characteristics either. In the most recent reviews [15,42]

Conclusions
In this survey, 8 frog species were found to be sensitive to sparganum infection, and the most frequently infected species was P. nigromaculatus. The sparganum infection rates in wild frogs in several regions of China were still high, especially in South and Southwest China. Eating wild frogs especially with improper cooking methods, is associated with considerable health risks in China. Several traditional Chinese folk remedies may increase the risk of infection. The molecular identification suggested that the taxonomic positions of "S. decipiens" (KJ599679) and "S. erinaceieuropaei" (KJ599680) are controversial. The sparganum isolates collected here were more likely the species of S. erinaceieuropaei, but new studies, especially comprehensive morphological analyses, are needed in the future.
Supporting information S1 Table. The information of