Surveillance of land molluscs infected by Angiostrongylus cantonensis (Nematoda) reveals risk areas for zoonotic eosinophilic meningitis in the State of Rio de Janeiro, Brazil

Angiostrongylus cantonensis (Chen, 1935) is a nematode parasite endemic to Southeast Asia and the adjacent Pacific Islands, and is one of the principal etiological agents of the zoonosis Eosinophilic Meningitis (EoM).^{(1, 2, 3, 4)} Over the past three decades, this zoonosis has spread to a number of other countries around the world, with cases reported in humans or infected snails and rodents in Egypt, Madagascar, the United States, various Caribbean islands, Spain, Ecuador, Colombia, and Brazil.^(5-15)

The biological cycle of A. cantonensis is complex, with the adult stage of the helminth being found in the arterial pulmonary system of rodents. In urban areas, the principal definitive hosts include Rattus norvegicus (Berkenhout, 1769)and Rattus rattus (Linnaeus, 1758), while the intermediate hosts include a number of different species of terrestrial snails, and the paratenic hosts are flatworms, crustaceans and fish, for example.^{(11, 16)} The infection of humans by A. cantonensis is considered to be accidental and occurs mainly through the ingestion of snails and slugs parasitised with infective third stage (L₃) larvae. However, the ingestion and the consumption of raw or poorly-cooked paratenic hosts, fruits or vegetables contaminated with larvae, is also a potential route of infection in humans.^{(10, 17)}

Up to now, approximately 40 cases of EoM have been confirmed in Brazil, based on records published in scientific journals, with more than 100 suspected cases in eight states ― Amapá, Espírito Santo, Pará, Paraná, Pernambuco, Rio de Janeiro, Rio Grande do Sul, and São Paulo.^{(3, 9, 14, 15, 18, 19)} Eosinophilic meningitis is considered to be an emerging zoonosis in Brazil⁽¹⁴⁾ and a number of studies have associated the dissemination of this zoonosis with the spread of the exotic giant African snail, Achatina fulica Bowdich, 1822, an invasive species that has now spread to all 26 Brazilian states, as well as the Federal District.^{(4, 20, 21)}

Given the wide distribution of this invasive snail and the risk of infection by A. cantonensis for human, wild, and domestic animals, the present study investigated the occurrence of the larvae of this nematode in specimens of terrestrial molluscs collected in half of the municipalities of the Brazilian State of Rio de Janeiro. The study used parasitological and genetic analyses to identify the focal sites of the transmission of A. cantonensis. The objective of this study was to identify these areas and to provide guidelines for the surveillance of the local populations of these animals and the prevention of the spread of EoM in the study area.

MATERIALS AND METHODS

The Brazilian State of Rio de Janeiro has a total of 92 municipalities, which are divided into six mesoregions (Metropolitana do Rio de Janeiro, Centro Fluminense, Baixadas Litorâneas, Noroeste Fluminense, Norte Fluminense, and Sul Fluminense), as defined by the Brazilian Institute for Geography and Statistics (IBGE/2010). Two of these mesoregions, the Metropolitan and the Central (Centro Fluminense), together contain exactly half (46) of the state’s municipalities (Fig. 1), which are subdivided into 177 administrative districts. The Metropolitan mesoregion has more than 12,500,000 inhabitants (IBGE/2010) and is the second largest metropolitan area in Brazil (Censo/2010), with a total of 30 municipalities, many of which have a thriving tourism industry, a feature that also contributed to the selection of the study area.

The molluscs were collected manually from natural environments, such as vacant lots, parks, public squares, and peripheral areas with natural vegetation, in all of the 177 administrative districts that make up the study area, between February 2015 and March 2019. Each collection site was georeferenced, and the molluscs were taken to the National Reference Laboratory for Schistosomiasis and Malacology of the Instituto Oswaldo Cruz at the Fundação Oswaldo Cruz in Rio de Janeiro. In the laboratory, the specimens were maintained in captivity for one week for the identification of the species, based on their conchology and morphology, and then up to a month for the parasitological examination and the extraction of the nematode larvae by artificial digestion,⁽²²⁾ with each specimen being processed individually. The larvae were separated under an optical microscope and identified initially to superfamily, based on their external morphology, with the specific objective of identifying the nematodes of the superfamily Metastrongyloidea.

In the case of the specimens identified as angiostrongylids, two subsamples were separated for morphological and molecular analyses. For the morphological procedure, 30 larvae per sample were fixed in heated AFA solution (2% glacial acetic acid, 3% formaldehyde, and 95% ethanol), to preserve their morphological structures, and stored in Eppendorf tubes. These specimens were clarified in lactophenol, mounted on slides, and examined under a bright field optical microscope for morphometry and the preparation of diagrams using a camera Lucida. For the molecular analysis, 10 larvae per sample were stored at -18ºC in phosphate-buffered saline (PBS) (1x) for the subsequent sequencing of the mitochondrial cytochrome oxidase I gene (COI), which has been used successfully as a DNA barcode for the diagnosis of species in a wide range of animal groups.⁽²³⁾ The total number of nematodes per mollusc specimen was not estimated.

For the genetic analyses, the DNA was extracted by thermal shock with liquid nitrogen for 1 min and transferred immediately to a hot plate at 95ºC for 20 min, repeated three times. A 5 µL aliquot of this extract was reserved for amplification by polymerase chain reaction (PCR), using the protocol described by Prosser et al.:⁽²³⁾ 1 x 94ºC for 5 min, 35 x (94ºC for 40 s, then 45ºC for 40 s, and 72ºC for 1 min), followed by 1 x 72ºC for 1 min, and then 1 x 17ºC for 3 min. The amplicon was then purified using the Illustra GFX kit for the purification of PCR DNA and gel banding (GE Healthcare, Little Chalfont, United Kingdom), following the manufacturer’s protocol, and then sequenced using the Sanger method in an ABI 3730xl automatic sequencer (Applied Biosystems) with 96 capillaries at the Oswaldo Cruz Institute Sequencing Platform in Rio de Janeiro.

To identify the species, the sequences obtained in the present study were compared with angiostrongylid sequences deposited in GenBank (www.ncbi.nlm.nih.gov/genbank/) using the Basic Local Alignment Search Tool (BLAST), available on the site of the National Centre for Biotechnology (NCBI). The BLAST search identified 22 A. cantonensis sequences of similar size, which were added to the matrix for the phylogenetic study. Two sequences representing other species of the genus Angiostrongylus were obtained from GenBank, and were added to the database as the outgroup.

The sequences were edited in Seqman 7.0 and aligned using Muscle,⁽²⁴⁾ which was run in MEGA 11.0.13.⁽²⁵⁾ The resulting sequence was trimmed to eliminate poorly-aligned extremities, and converted to the Nexus format using Mesquite 3.51.⁽²⁶⁾ The Bayesian Inference (BI) analyses were run in MrBayes 3.2.7,⁽²⁷⁾ with command blocks of the GTR + I nucleotide substitution model, which were calculated in MrModelTest and added to the matrix using Mesquite 3.7.⁽²⁶⁾ The MrBayes analysis was run on the CIPRES Science Gateway V. 3.3.⁽²⁸⁾ The BI node supports were calculated every 100 generations after the exclusion of a burn-in of 25%. To evaluate the quality of the sample, the Tracer 1.7.2 Software⁽²⁹⁾ was used to calculate the effective sample size (ESS) of the parameters, with an ESS of over 200 being considered to be adequate for analysis.

RESULTS

Overall, 14 species of terrestrial molluscs, including both native and exotic taxa, were collected during the present study, with a total of 2600 specimens being available for parasitological analysis. Angiostrongylid larvae were observed in 230 of these specimens (8.8% of the total), while a further 1233 specimens (47.4%) were parasitised by nematodes of taxa other than the superfamily Metastrongyloidea. The diagnostic traits of the angiostrongylid larvae are the presence of a filariform oesophagus, pointed tail with bevelling, and a genital primordium (Fig. 2). The morphological structures of the larvae analysed here presented the diagnosis traits proposed by Ash⁽³⁰⁾ and Thiengo et al.,⁽³⁾ which confirmed that they were angiostrongylids (Table I).

The molecular analysis confirmed the occurrence of A. cantonensis in 26 (56.5%) of the 46 study municipalities (Table II). This nematode was found parasitising six species: the snails A. fulica, Subulina octona (Brugüière, 1789), Bradybaena similaris (Férussac, 1821) and Leptinaria unilamellata (d’Orbigny, 1835), and the slugs Sarasinula linguaeformis (Semper, 1885) and Angustipes erinaceus (Colosi, 1922). We also collected the snails Bulimulus tenuissimus (Férussac, 1832), Drymaeus sp., Ovachlamys fulgens (Gude, 1900), Megalobulimus ovatus (Müller, 1774), Streptaxis contusus (Férussac, 1821), Tamayoa banghaasi (Thiele, 1927), Succinea sp., Beckianum beckianum (Pfeiffer, 1846), but they were represented by shells or only few specimens that were not parasitologically examined.

The exotic invasive snail A. fulica was found infected by A. cantonensis in 22 municipalities, in samples collected in all four years of the study period. It is important to note that A. fulica was present in 41 of the 46 municipalities surveyed. Considering the study area, the natural infection rates by Angiostrongylus in the different gastropod species were 12.5% in A. erinaceus, 9.7% in A. fulica, 6.8% in B. similaris, 6.3% in S. linguaeformis, 3.9% in L. unilamellata, and 4.6% in S. octona. Overall, A. cantonensis was found in snails and slugs collected from natural biotopes in more than half (56.5%) of the municipalities surveyed in the present study. There was a clear predominance of infection in the Metropolitan mesoregion of the state (Fig. 3).

Angiostrongylus cantonensis was found infecting terrestrial molluscs collected from 21 of 30 of the municipalities (70%) of the Metropolitan mesoregion, but only five of the 16 municipalities (31%) of the Central mesoregion. While a much larger sample of molluscs (n = 2167) was collected in the Metropolitan mesoregion, given its larger number of municipalities, infection rates, which are a proportional measure, were also higher than those recorded in the Central mesoregion (Table III). A similar difference was also found in the number of terrestrial molluscs species, with five species being parasitised by A. cantonensis in the Metropolitan mesoregion, but only three in the Central mesoregion. The invasive A. fulica was the most frequent and widely distributed snail overall, and the species with the largest number of specimens parasitised by A. cantonensis in both mesoregions.

The molecular analysis of the larvae obtained from the six gastropod species produced COI sequences with a mean length of 600 base pairs (OR536625–OR536637, OR804433-OR804448). These sequences were highly similar to the A. cantonensis sequences available in the GenBank. Ten COI sequences were selected as reference sequences for the phylogenetic analysis, based on their size and quality (Table IV).

The GTR + I nucleotide substitution model was selected by the Akaike Information Criterion (AIC), run in MrModelTest, as having the best adjustment for the COI matrix compiled here. The BI phylogenetic tree (Fig. 4) allocated all the samples to a single clade. As the ESS values were well above 200, the sampling was considered to be adequate.

DISCUSSION

Natural infection by A. cantonensis wasobservedin six of the total of 14 species collected in the present study in Rio de Janeiro State, which reconfirms the lack of specificity of this nematode in relation to its intermediate hosts, as reported by Caldeira et al.,⁽⁹⁾ Thiengo et al.⁽⁴⁾ and Valente et al.⁽³¹⁾ A similar pattern has also been observed in endemic regions. After screening the principal Hawaiian islands to determine which gastropods act as hosts of A. cantonensis, Kim et al.⁽³²⁾ recorded the presence of the parasite in 16 species (two native, 14 non-native) of a total of 37 species, with 70 of the 1271 specimens testing positive. In this study,⁽³²⁾ the small gastropod S. octona had the highest mean concentration of parasites, while the veronicellid slug Laevicaulis alte had the highest individual parasite concentration. Kim et al.⁽³²⁾ identified only seven positive specimens of A. fulica in a sample of 62 (11%), and added three gastropod families to the 33 already known to host A. cantonensis. Jaume-Ramis et al.⁽³³⁾ recorded a similar scenario in Mallorca, Spain, where they collected and identified 398 gastropods belonging to 17 species (14 snails and three slugs), of which 11% tested positive for A. cantonensis. The gastropods were collected at sites ranging from tourist settlements to farmland, with the least abundant species being the less likely to be infected. A. fulica was not recorded in this study. Santana-Teles et al.⁽³⁴⁾ and Prociv et al.⁽⁶⁾ also reported a lack of specificity in the A. cantonensis vertebrate hosts. All the gastropod species found infected with A. cantonensis in the present study, except L. unilamellata, had already been reported being infected with this nematode.

The larger numbers of infected molluscs found in the Metropolitan mesoregion highlight the risk of the occurrence of cases of EoM in the urban areas of these municipalities, which are densely populated, in general, and tend to have large concentrations of domestic animals. Two other Angiostrongylus species are known to both occur in Brazil and parasitise molluscs, Angiostrongylus costaricensis (Morera & Céspedes, 1971), a parasite of rodents which can infect humans accidentally, and causes abdominal angiostrongyliasis, and Angiostrongylus vasorum (Baillet, 1866), a cardio-pulmonary parasite of canids. While neither of these species was detected in the present study, their potential occurrence in the study area cannot be ruled out altogether, given that
Angiostrongylus nematodes tend to be relatively unspecific in terms of their intermediate and definitive hosts.

Turck et al.⁽³⁵⁾ reviewed the reports of A. cantonensis infections in both academic papers and the grey literature, and found that the number of human cases of infection had more than doubled worldwide over the past 10 years. Brazil is likely to have undergone a similar increase in infection levels, given that infected A. fulica have now been recorded in 14 of the country’s 26 states, and in its Federal District.^{(4, 36, 37, 38, 39, 40)} Human cases of A. cantonensis infection have now been reported from southern, southeastern, northeastern, and northern Brazil.^{(9, 14, 15, 20)}

In addition, the very large population of A. fulica observed in the present study, together with the A. cantonensis infection rates, indicate that this snail may represent the principal intermediate host of this parasite in Rio de Janeiro, considering its ample distribution and dense populations, in particular in urban areas. The proximity of this snail to humans, especially in urban environments, where its coexistence with synanthropic rodents establishes an important link in the epidemiology of A. cantonensis, increases the risk of human infection.^{(36, 37, 38, 39, 40, 41)}

In comparison with the Central mesoregion, A. fulica was much more common in the Metropolitan mesoregion, which has a densely-populated urban centre, with many social and environmental problems, including the proliferation of low-income housing in areas that lack public sanitation or adequate garbage disposal. This favours the establishment of invasive species that act as urban pests, and are associated with the transmission of diseases to humans and domestic animals. The Central mesoregion has a much lower population density, however, with more rural areas.⁽⁴²⁾

In a previous study in the Brazilian state of Rio de Janeiro, Oliveira et al.⁽³⁷⁾ reported A. fulica and B. similaris being parasitised by A. cantonensis, with a prevalence of infection by A. cantonensis of more than 50% in A. fulica throughout the study area. In another study, Kim et al.⁽³²⁾ concluded that the level of infection of A. fulica varies widely among the different Hawaiian Islands and that the dissemination of A. cantonensis is most likely determined by local habitats, rather than the presence of A. fulica. These authors related the observed differences in infection rates primarily to abiotic factors such as temperature and humidity, and secondarily to the distribution of infected rats, the rat species on the different islands and the variation in the level of interaction between the rats and the local gastropods.

The results of the present study highlight the epidemiological risk of transmission of EoM in the study area, based on the wide distribution of A. cantonensis and the high densities of its intermediate host molluscs. Knowledge of this epidemiological risk, combined with the existing sound diagnostic criteria,⁽⁴³⁾ should enable the faster and more reliable detection of EoM by the local health services.

In conclusions - The present study showed that A. cantonensis is now widespread in both the Metropolitan and the Central mesoregions of the State of Rio de Janeiro, Brazil, and highlights the potential epidemiological risk of human infection in this state. The transmission of A. cantonensis is associated with six species of terrestrial molluscs, with A. fulica having an important role in the dissemination of this nematode, considering that it is the most frequent and widely-distributed snail, overall, in both mesoregions.

Understanding the distribution of infected intermediate hosts in the different municipalities will be extremely useful for epidemiological surveillance and investigations by public health services and should help local health authorities to provide a faster and more accurate diagnosis whenever neuroangiostrongyliasis is suspected.

ACKNOWLEDGEMENTS

To the teams of the IOC Laboratory of Malacology and the Laboratory of Biology and Parasitology of Wild Reservoir Mammals for their assistance during fieldwork and the morpho-molecular analysis of the nematodes. We would also like to thank the technical teams of the Municipal Health Secretariats of Rio de Janeiro State for their logistic support during the collection of the mollusc specimens.

AUTHORS’ CONTRIBUTION