Abstract
Fecal specimens (271 samples) from wild deer, Cervus nippon centralis, were collected from nine different areas in Japan; these samples were subjected to a real-time reverse transcription PCR for Cryptosporidium-and Giardia-specific 18S ribosomal RNA to investigate the prevalence of Cryptosporidium and Giardia infection. The incidence of Cryptosporidium and Giardia in the nine areas ranged from 0% to 20.0% and 0% to 3.4%, respectively. The prevalence of Cryptosporidium among male and female deer was 8.1% and 3.9%, respectively, while that of Giardia was 0.7% and 0.8%. Sequence analysis identified the Cryptosporidium deer genotype, Cryptosporidium bovis, Cryptosporidium ryanae and Cryptosporidium meleagridis from the sequence of Cryptosporidium-specific partial 18S ribosomal RNA and Giardia intestinalis assemblage A from the partial sequence of Giardia-specific 18S rRNA. The variation in regional prevalence indicates that Cryptosporidium infection depends on environmental factors, and that bovine Cryptosporidium was detected more frequently than cervine Cryptosporidium. These data suggest that wild deer might be a healthy carrier of bovine Cryptosporidium.
Key words: : Cervus nippon centralis, Cryptosporidium, Giardia, Japan, pathogen, wild deer
1. Introduction
Cryptosporidium spp. and Giardia spp. are protozoan parasites that infect a wide range of vertebrates, including humans. They cause different types of severe diarrhea with shedding of a small number of oocysts1). Infection is caused by the ingestion of viable oocysts via food or water contaminated with the feces of the infected host. Both Cryptosporidium and Giardia are chlorine resistant. This makes water sanitation difficult and results in sporadic outbreaks of water-borne disease in both developing and developed countries. Thus far, many wild animals, including ruminants and domestic animals, have been reported to be infected with several species of Cryptosporidium1). Among these, cattle are known to be typically infected by five different species: Cryptosporidium parvum, C. andersoni, C. bovis, C. ryanae, and C. ubiquitum2). In contrast, among a large number of Giardia species and genotypes, only Giardia intestinalis assemblage A and assemblage E are known to infect cattle3,4). Recently, the population of wild deer has markedly increased in Japan, which has led to severe damage to farming industries. To cope with these problems, regional governments are trying to control deer populations, and simultaneously utilize their meat. Cases of food-borne cryptosporidiosis and giardiasis have increased in Japan and abroad5); however, there are few reports of epidemiological studies on pathogenic microbes in wild animals. Studies of Cryptosporidium have mostly focused on species that infect livestock such as cows and hogs. Thus, while there have been many epidemiological reports from foreign countries on Cryptosporidium and Giardia infections in wild animals6), the distribution of these parasites in wild animals in Japan remains largely unknown. In the present study, we attempted to determine the distribution of Cryptosporidium and Giardia species in wild deer in Japan in order to facilitate the establishment of preventive measures against Cryptosporidium and Giardia infection.
2. Materials and Methods
2-1. Collection of Samples
Between October 2012 and June 2016, the sampling of sika deer (Cervus nippon centralis) fecal samples was carried out during the hunting season in nine prefectures in Japan. A total of 271 (Season: January, n = 30; February, n = 13; May, n = 35; April, n = 11; May, n = 16; June, n = 19; July, n = 9; August, n = 4; September, n = 11; October, n = 38; November, n = 25; December, n = 60; Sex: male, n = 135; female, n = 129; undetermined at sample collection, n = 7) deer fecal samples were collected. Individual fecal samples were taken directly from the rectum of each deer during evisceration. The samples were transported to the laboratory in an insulated cooler and stored at 4°C until DNA extraction.
2-2. DNA Extraction
DNA was extracted from each sample within 24 hours after arrival and stored at -20°C until analyzed. Feces were diluted in deionized water and filtered through gauze into centrifuge tubes. Ethyl acetate was then added at a ratio of 1:4 (v/v). The tubes were vigorously shaken and centrifuged at 1000 × g for 5 min at room temperature and the supernatants were decanted. The pellet was suspended again in deionized water and then centrifuged at 1000 × g for 3 min at room temperature to recover the pellet. The recovered pellet was resuspended in the extraction buffer to a volume of 200 µl which was immediately thereafter subjected to lysing and DNA extraction. Briefly, the samples were lysed by 5 cycles of freezing at -196°C and thawing at 85°C, and then DNA was extracted using the Qiagen DNA Stool Mini Kit (QIAGEN, Hilden, Germany). DNA was eluted in water and then stored at -20°C before assay.
2-3. Real-time PCR
The molecular characterization and estimation of parasite density of Cryptosporidium were performed using a Cycleave RT-PCR Cryptosporidium (18S rRNA) detection kit (Takara Bio, Shiga, Japan) and a Cycleave RT-PCR Giardia (18S rRNA) detection kit (Takara Bio) according to the manufacturer’s protocol.
2-4. Statistical Analysis
Prevalence with respect to sampling area and sex were statistically analyzed using Fisher’s exact test. The percentage of the parasites in each deer group were compared by the nonparametric Kruskal-Wallis test and Mann-Whitney U-test.
2-5. DNA Sequencing
All PCR-positive samples were directly sequenced targeting the partial sequence of the Cryptosporidium and Giardia 18S rRNA genes to determine the species. For Cryptosporidium, the forward primer (5ʹ-aagctcgtagttggatttctg-3ʹ) corresponded to nucleotides 601 to 621, and the reverse primer (5ʹ-taaggtgctgaaggagtaagg-3ʹ) corresponded to nucleotides 1015 to 1035 (GenBank accession number L16996)7). For Giardia, the forward primer (5’-CATCCGGTCGATCCTGCC-3’) corresponded to nucleotides 1 to 18, and the reverse primer (5’-AGTCGAACCCTGATTCTCCGCCAGG-3’) corresponded to nucleotides 268 to 292 (GenBank accession number AF473852)8,9). The sequence data were examined using an ABI 3730xl Genetic Analyzer/DNA Sequencer (Life Technologies, Tokyo, Japan) with a Big Dye Terminator V.3.1 cycle sequencing kit (Applied Biosystems, Massachusetts, USA). The 18S rRNA of Giardia was also sequenced by Takara Bio Inc. using a Cycleave RT-PCR Giardia kit.
2-6. DNA Sequencing and The Phylogenetic Analysis
Sequence assemblies were analyzed using the ATGC (Ver. 7) Sequence Assembly Software (GENETYX, Tokyo, Japan) and Molecular Evolutionary Genetics Analysis (MEGA version 6: http://www.megasoftware.net/) software programs10). To identify Cryptosporidium and Giardia at the species level, each of the resulting sequences were compared to those of the Cryptosporidium and Giardia species genotypes submitted to GenBank (National Institutes of Health, Bethesda, MD, USA), using the BLAST software program (v.2.2.12; http://www.ncbi.nlm.nih.gov/BLAST/, National Center for Biotechnology Information, Bethesda, MD, USA), and the homology was determined. We used the neighbor joining method for the phylogenetic analysis of our sequencing results and compared the sequences to other genotypes deposited in the GenBank database.
3. Results
For Cryptosporidium, a total of 18 samples (6.6%) were confirmed to be Cryptosporidium-positive (Table 1). The percentage of Cryptosporidium positive samples by prefectures were as follows: Chiba (11.1%), Shizuoka (15.6%), Nagano (2.5%), Kyoto (13.8%), Miyazaki (20.0%), and Nagasaki (3.4%). The percentage by region were as follows: east (9.8%), west (4.4%), north (7.9%) and south (3.3%). The percentage by season were as follows: December to February (8.7%), March to May (3.2%), June to August (6.3%) and September to November (4.1%) (data not shown). The percentages by sex were 8.1% and 3.9%, respectively.
Table 1. The prevalence of Cryptosporidium and Giardia in Sika deer in Japan.
| Cryptosporidium | Giardia | Cryptosporidium / Giardia | ||||||
|---|---|---|---|---|---|---|---|---|
| Location / Sex |
No. of deer | No. PCR- positive for 18S-rRNA |
% |
P value (Fisher’s exact test*) |
No. PCR- positive for 18S-rRNA |
% |
P value (Fisher’s exact test*) |
P value (Fisher’s exact test*) |
| Prefecture | ||||||||
| Chiba | 27 | 3 | 11.1 | 0 | 0.0 | |||
| Shizuoka | 45 | 7 | 15.6 | 0 | 0.0 | |||
| Nagano | 40 | 1 | 2.5 | 0 | 0.0 | |||
| Shiga | 25 | 0 | 0.0 | 0 | 0.0 | |||
| Mie | 23 | 0 | 0.0 | 0 | 0.0 | |||
| Kyoto | 29 | 4 | 13.8 | 1 | 3.4 | |||
| Nagasaki | 58 | 2 | 3.4 | 1 | 1.7 | |||
| Kumamoto | 19 | 0 | 0.0 | 0 | 0.0 | |||
| Miyazaki | 5 | 1 | 20.0 | 0.055 | 0 | 0.0 | 0.817 | |
| Region | ||||||||
| East | 112 | 11 | 9.8 | 0 | 0.0 | |||
| West | 159 | 7 | 4.4 | 0.088 | 2 | 1.3 | 0.513 | |
| North | 189 | 15 | 7.9 | 1 | 0.5 | |||
| South | 82 | 3 | 3.7 | 0.288 | 1 | 1.2 | 0.514 | |
| Total | 271 | 18 | 6.6 | 2 | 0.7 | 0.0004 | ||
| Sex | ||||||||
| Male | 135 | 11 | 8.1 | 1 | 0.7 | |||
| Female | 129 | 5 | 3.9 | 1 | 0.8 | |||
| Total | 264† | 16 | 6.6 | 0.2 | 2 | 0.7 | 0.974 | |
The prevalence of the Cryptosporidium and Giardia deer genotype according to area and sex. Cryptosporidium and Giardia were identified by a reverse transcription polymerase chain reaction targeting the partial sequence of 18S ribosomal RNA (18S rRNA).
*Fisher's exact test was conducted to analyze the statistical analyses according to area (Prefecture: Chiba, Shizuoka, Nagano, Shiga, Mie, Kyoto, Nagasaki, Kumamoto and Miyazaki. Region: Chiba, Shizuoka and Nagano for east; Shiga, Mie, Kyoto, Nagasaki, Kumamoto and Miyazaki for west; Chiba, Shizuoka, Nagano, Shiga, Mie and Kyoto for north; Nagasaki, Kumamoto and Miyazaki for south) and sex (male and female).
†The sex was unknown for seven samples.
Two samples (0.7%) were confirmed to be Giardia-positive (Table 1). The percentage of positive samples of Giardia in the two prefectures were Kyoto (3.4%) and Nagasaki (1.7%) and by region, were west (1.3%), north (0.5%) and south (1.2%). The percent positive samples by season were as follows: December to February (1.0%) and September to November (1.4%) (data not shown); and that by sex were 0.7% and 0.8%, respectively.
There was a significant difference (P<0.01) between the total number of positive samples of Cryptosporidium and Giardia collected. There were no statistically significant differences among the sampling areas and seasons for either Cryptosporidium or Giardia.
The results of the partial sequence analysis of 18S ribosomal RNA of Cryptosporidium and Giardia are summarized in Table 2. Sample C1 and C2 from Chiba prefecture showed 100% identity with C. sp. deer genotype (Genbank accession number KR260681), C. ryanae (AY587166), and C. bovis (KU168249). With regard to the samples from Shizuoka, S1 showed 100% identity with C. ryanae (KT922235) and C. sp. genotype (KR260681); S2 showed 100% identity with C. ryanae (AY587166); S3, S4 and S5 showed 100% identity with C. ryanae (KT922235), C. bovis (KU168249) and C. sp. genotype (KR260681); and S6 and S7 were suggested to have 100% identity with C. ryanae (KT922235). In the samples from Kyoto, K1 showed 100% identity with C. ryanae (KT922235) and C. sp. genotype (KR260681); K2 showed 100% identity with C. sp. genotype (KT260681); and K3 and K4 showed 100% identity with C. ryanae (AY587166). M1 from Mie prefecture showed 100% identity with C. sp. genotype (KR260681) and N1 from Nagasaki showed 100% identity with C. meleagridis (AF112574).
Table 2. The identification of the identification of Cryptosporidium and Giardia species in Sika deer in Japan.
| Date of capture | Sampling area | Sample no. | Sex | Identity (>99% homology) |
|---|---|---|---|---|
| Cryptosporidium | ||||
| 2015 Jun.13 | Chiba | C1 | F | C. ryanae, C. bovis, C. sp. deer genotype |
| 2016 Jun. 26 | C2 | M | C. ryanae, C. bovis, C. sp. deer genotype | |
| 2012 Oct. 16 | Shizuoka | S1 | M | C. ryanae, C. sp. deer genotype |
| 2012 Oct. 16 | S2 | F | C. ryanae | |
| 2012 Oct. 18 | S3 | F | C. ryanae, C. sp. deer genotype, C. bovis | |
| 2012 Dec. 14 | S4 | M | C. ryanae, C. sp. deer genotype, C. bovis | |
| 2012 Dec. 14 | S5 | M | C. ryanae, C. sp. deer genotype, C. bovis | |
| 2013 May 28 | S6 | M | C. ryanae | |
| 2013 May 28 | S7 | M | C. ryanae | |
| 2015 Jun. 6 | Kyoto | K1 | M | C. ryanae, C. sp. deer genotype |
| 2013 Dec. 11 | K2 | F | C. sp. deer genotype | |
| 2013 Dec. 13 | K3 | M | C. ryanae | |
| 2014 Jan. 9 | K4 | M | C. ryanae | |
| 2014 Jan. 11 | Miyazaki | M1 | Unknown | C. sp. deer genotype |
| 2014 Dec. 18 | Nagasaki | F | ND | |
| 2016 Jan. 21 | N1 | M | C. meleagridis | |
| Giardia | ||||
| 2014 Jan. 16 | Kyoto | Kyoto 29 | F | G. intestinalis |
| 2015 Nov. 27 | Nagasaki | Nagasaki 28 | M | G. intestinalis |
The identification of Cryptosporidium and Giardia species and genotypes by the 18S rRNA sequence in wild deer samples
ND: not determined.
Both of the Giardia-positive samples from Kyoto and Mie were suggested to have 100% identity with G. intestinalis (Table 2).
The phylogenetic analysis placed all of the Cryptosporidium-positive samples, with the exception of the Nagasaki sample (C1-2, S1-7, K1-4, M1), in a single cluster with ruminant-derived Cryptosporidium species; N1 was clustered in bird-derived Cryptosporidium (Fig. 1). The phylogenetic analysis of Giardia revealed that the sequences of both samples from Kyoto and Nagasaki were clustered with Giardia intestinalis assemblage A (Fig. 2).
Fig. 1.
The phylogenetic relationships of the Cryptosporidium deer genotype obtained in this study to other available Cryptosporidium species and genotypes from the NCBI database as inferred by a neighbor-joining analysis, based on 18S ribosomal RNA (18S rRNA) sequences. A phylogenetic tree was constructed using the maximum-likelihood method with a maximum composite likelihood substitution model.
Samples C1-C2 from Chiba, S1-S4 from Shizuoka, K1-K4 from Kyoto, M1 from Mie, and N1 from Nagasaki correspond to Table 2.
Fig. 2.
Phylogenetic relationships of the Giardia deer genotype obtained in this study to other available Giardia species and genotypes from the NCBI database as inferred by neighbor-joining analysis, based on the 18S ribosomal RNA (18s rRNA) sequences. A phylogenetic tree was constructed using the maximum-likelihood method with a maximum composite likelihood substitution model.
Samples Contig Nagasaki 28 from Nagasaki and Contig Kyoto 29 from Kyoto correspond to Table 2.
4. Discussion
The results of this study demonstrate the difference in the prevalence of Cryptosporidium spp. and Giardia spp. in Cervus nippon centralis in Japan. This result suggests that Cryptosporidium had already spread widely throughout a large part of central Japan but that Giardia had not. The difference in the number of positive samples of Cryptosporidium and Giardia was statistically significant (P<0.01). The Cryptosporidium positive male deer was twice that of female deer (Table 1). This situation may be due to the males’ independent behavior and wide habitation range, which increases opportunities for exposure to pathogens. All of the deer examined in this study were asymptomatic. This may be due to the fact that the samples collected in this study were mostly obtained from adult deer of >2 years of age2). The estimated number of Cryptosporidium and Giardia oocysts in each positive sample calculated by the manufacturer’s protocol was ≤4 oocysts per 1 g of feces (data not shown), which may be another reason of asymptomatic presentation. In this study, most of the samples were clustered in ruminant-derived Cryptosporidium (Fig. 1). Wild deer have recently caused significant feeding damage to grazing land in Japan; this represents a major problem for the country’s cattle industry and suggests that some cases of Cryptosporidium infection in deer result from contact with feces of cattle.
Zoonotic outbreaks of enterohaemorrhagic Escherichia coli (EHEC), Salmonella, hepatitis E, and Trichinella from the consumption of game meat have been reported in Japan; most cases were caused by eating raw meat12). In the guidelines for processing game meat, information about pathogenic microbes refers only to those affecting farm animals, which are generally raised under relatively sanitary conditions. The detection of Cryptosporidium spp. and Giardia spp. in wild deer in the present study may indicate that the guidelines need to be updated to include preventive measures against Cryptosporidium and Giardia contamination that can be implemented during the slaughter of Sika deer.
The variable prevalence of these pathogens in different areas may suggest that there are some regional trends regarding frequency of contact between wild deer and cattle in Japan. A similar situation has been reported in China and Spain6,11). A possible reason for the observed diversity in regional prevalence of Cryptosporidium and Giardia in deer feces is that some of the study areas were located close to cattle farms so wild deer may have frequent contact with cattle. On the other hand, since other study areas were located deep in the mountains, frequency of contact between deer and cattle may be low. In this study, Shiga, Mie, and Kumamoto are forest areas, which may be why no Cryptosporidium or Giardia were detected from deer feces.
In some areas, two or three different species were detected with >99% sequencing homology. This does not necessarily indicate coinfection; however, there are some reports on both mixed13) and single infection11,14) with Cryptosporidium species in cattle. These species are highly prevalent in cattle, indicating that C. bovis and C. ryanae may infect both cattle and deer, reflecting the results of the phylogenetic analysis (Fig. 1). The deer that harbored Cryptosporidium spp. were asymptomatic, suggesting that these healthy carrier deer may spread Cryptosporidium species to cattle or other farm animals.
Human infection with Cryptosporidium derived from deer has been reported in several countries15), and the genotypes that caused sporadic cases in humans were also detected in Cervus nippon Temminck in China6). In the present study, C. meleagridis was detected from deer in Nagasaki (Table 2). C. meleagridis has been reported in humans, but the most common host of C. meleagridis is turkey or parrots16), so the current study may represent the first report from wild deer.
Two G. intestinalis assemblage A isolates were detected in the current study. This is consistent with the finding that among 12 species of Giardia, only G. intestinalis assemblage A has been reported to use wild animals as a host. Since G. intestinalis assemblage A is a human pathogenic species, it will be prudent to monitor the prevalence of G. intestinalis assemblage A to prevent human infection. In fact, the Giardia-positive deer in Nagasaki were caught just 10 km from an urban area.
In conclusion, we detected Cryptosporidium and Giardia in wild deer in Japan. These data could provide valuable information in relation to preparing Japanese wild game in a hygienic manner, and evidence to support the control of deer populations to protect water sources from contamination.
Acknowledgments
Our work was supported by a Grant-in-Aid for Young Scientists from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
Abbreviations: EHEC: enterohaemorrhagic Escherichia coli, PCR: polymerase chain reaction, rRNA: ribosomal RNA, BLAST: Basic Local Allignment Search Tool, PCR: polymerase chain reaction, RT-PCR: real-time PCR
Footnotes
Conflict of interest: The authors have no conflict of interest.
References
- 1.Xiao L, Sulaiman IM, Ryan UM, et al. Host adaptation and host–parasite co-evolution in Cryptosporidium: implications for taxonomy and public health. International Journal for Parasitology. 2002; 32: 1773–1785. 10.1016/S0020-7519(02)00197-2 [DOI] [PubMed] [Google Scholar]
- 2.Santín M, Trout JM, Fayer R. A longitudinal study of cryptosporidiosis in dairy cattle from birth to 2 years of age. Veterinary Parasitology. 2008; 155: 15–23. 10.1016/j.vetpar.2008.04.018 [DOI] [PubMed] [Google Scholar]
- 3.O’Handley RM, Olson ME. Giardiasis and cryptosporidiosis in ruminants. Veterinary Clinics of North America: Food Animal Practice. 2006; 22: 623–643. 10.1016/j.cvfa.2006.07.002 [DOI] [PubMed] [Google Scholar]
- 4.Thompson RCA, Palmer CS, O’Handley R. The public health and clinical significance of Giardia and Cryptosporidium in domestic animals. The Veterinary Journal. 2008; 177: 18–25. 10.1016/j.tvjl.2007.09.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Yoshida H, Matsuo M, Miyoshi T, et al. An outbreak of cryptosporidiosis suspected to be related to contaminated food, October 2006, Sakai City, Japan. Jpn J Infect Dis. 2007; 60: 405–407. [PubMed] [Google Scholar]
- 6.Wang R, Wang J, Sun M, et al. Molecular characterization of the Cryptosporidium cervine genotype from a sika deer (Cervus nippon Temminck) in Zhengzhou, China and literature review. Parasitology Research. 2008; 103: 865–869. 10.1007/s00436-008-1069-2 [DOI] [PubMed] [Google Scholar]
- 7.Johnson DW, Pieniazek NJ, Griffin DW, Misener L, Rose JB. Development of a PCR protocol for sensitive detection of Cryptosporidium oocysts in water samples. Appl Environ Microbiol. 1995; 61: 3849–3855. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hopkins RM, Meloni BP, Groth DM, Wetherall JD, Reynoldson JA, Thompson RCA. Ribosomal RNA sequencing reveals differences between the genotypes of Giardia isolates recovered from humans and dogs living in the same locality. The Journal of Parasitology. 1997; 83: 44–51. 10.2307/3284315 [DOI] [PubMed] [Google Scholar]
- 9.Nantavisai K, Mungthin M, Tan-ariya P, Rangsin R, Naaglor T, Leelayoova S. Evaluation of the sensitivities of DNA extraction and PCR methods for detection of Giardia duodenalis in stool specimens. Journal of Clinical Microbiology. 2007; 45: 581–583. 10.1128/JCM.01823-06 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: Molecular Evolutionary Genetics Analysis Version 6.0. Molecular Biology and Evolution. 2013; 30: 2725–2729. 10.1093/molbev/mst197 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.García-Presedo I, Pedraza-Díaz S, González-Warleta M, et al. The first report of Cryptosporidium bovis, C. ryanae and Giardia duodenalis sub-assemblage A-II in roe deer (Capreolus capreolus) in Spain. Veterinary Parasitology. 2013; 197: 658–664. 10.1016/j.vetpar.2013.07.002 [DOI] [PubMed] [Google Scholar]
- 12.Food Safety Commission of Japan Fact Sheets, Zoonosis through the Jibie [in Japanese]. http://www.fsc.go.jp/sonota/factsheets/140805_gibier.pdf. Published August 2014, Accessed February 28, 2018.
- 13.Amer S, Zidan S, Adamu H, et al. Prevalence and characterization of Cryptosporidium spp. in dairy cattle in Nile River delta provinces, Egypt. Experimental Parasitology. 2013; 135: 518–523. 10.1016/j.exppara.2013.09.002 [DOI] [PubMed] [Google Scholar]
- 14.Venu R, Latha BR, Basith SA, Raj GD, Sreekumar C, Raman M. Molecular prevalence of Cryptosporidium spp. in dairy calves in Southern states of India. Veterinary Parasitology. 2012; 188: 19–24. 10.1016/j.vetpar.2012.02.025 [DOI] [PubMed] [Google Scholar]
- 15.Ong CSL, Eisler DL, Alikhani A, et al. Novel cryptosporidium genotypes in sporadic cryptosporidiosis cases: first report of human infections with a cervine genotype. Emerging Infectious Diseases. 2002; 8: 263–268. 10.3201/eid0803.010194 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Fayer R. Cryptosporidium: a water-borne zoonotic parasite. Veterinary Parasitology. 2004; 126: 37–56. 10.1016/j.vetpar.2004.09.004 [DOI] [PubMed] [Google Scholar]