key: cord-0941541-5rvltf2l authors: Kidd, A. H.; Esrey, S. A.; Ujfalusi, Maria J. title: Shedding of coronavirus‐like particles by children in lesotho date: 2005-12-07 journal: J Med Virol DOI: 10.1002/jmv.1890270219 sha: 4d55122b20d2674aaa426ae90cc4ac0fa6fac52e doc_id: 941541 cord_uid: 5rvltf2l Stools from 266 children in four districts of Lesotho were examined for viruses by electron microscopy (EM) over a 5‐week period. Eighty one (30.5%) of the children shed coronavirus‐like particles (CVLPs). Shedding was not significantly associated with diarrhea. The proportion of children shedding these particles increased with increasing age. In one district, the prevalence of CVLP shedding (67.9%) was at least twice the prevalence in any of the other three districts. This was the only district to be sampled during the first week of the study. It was shown that the proportion of children shedding CVLPs declined during each of the 5 weeks of study. The stools of six children who shed CVLPs in the early summer of 1984–85 were negative by EM 5 months later. There was no association between the shedding of pathogenic bacteria or parasites and the presence or absence of CVLPs in the stool. Several animal species are known to be susceptible to enteric infection by coronaviruses IGarwes, 19821, usually resulting in severe diarrhea. Particles resembling coronaviruses in human stool have been described by several workers, notably in India [Mathan et al., 19751, Australia ISchnagl et al., 1978, 19791, and Gabon in West Africa ISitbon et al., 19851 . It has been difficult to attribute a causative role to these particles in diarrhea, as healthy individuals can also shed them [for a review, see MacNaughton and Davies, 19811 . Nevertheless, there have been several reports of gastroenteritis in humans apparently caused by coronavirus-like particles (CVLPs) [Caul et al., 1975; Gerna et al., 1985; Rettig and Altshuler, 1985; Vaucher et al., 19821 . There is still controversy over whether these particles are truly viruses [Dourmashkin et al., 19801, although there have been reports of successful propac 1989 Alan R. Liss. Inc. gation in organ cultures [Caul and Egglestone, 1977; Resta et al., 19851 , of antigenic differences [Schnagl et al., 19861, and of preliminary biochemical characterization IResta et al., 1985; Battaglia et al., 19871. Schnagl et al. I19871 recently reported widely different electrophoretic profiles for the proteins of CVLPs from different sources, which sheds doubts on their homogeneity as a group. CVLPs have recently been described in the stools of individuals with AIDS or AIDS related complex (ARC) [Kern et al., 19851 and homosexual men whose HIV antibody status was unknown IRiordan e t al., 19861. The data described in this paper were collected as part of a health impact evaluation of a rural water supply project I Esrey e t al., 1987 I. No difference in the shedding of CVLPs was found between children from villages with a n improved water supply compared with children from villages which relied on traditional contaminated water sources [Esrey et al., 19871 . The present paper reports on the electron microscopic appearance of the CVLPs found and the prevalence of shedding in relation to factors commonly associated with infection by enteric pathogens in general. Apart from the study of Sitbon in the Republic of Gabon, West Central Africa [Sitbon, 1985 1, there have been no other large-scale studies on the shedding of coronavirus-like particles in Africa. Data used in this paper were collected from 21 rural villages in the Kingdom of Lesotho, which is landlocked by South Africa. All villages were situated in the lowlands or foothills (Fig. l ) , in which the majority of the population resides, and were selected from four of the ten administrative districts. Villages ranged in elevation from 1,400 metres to 2,000 m. Within each village households were selected randomly and all chil-was applied to the grid and the excess removed as before. The grid was examined in a Jeol 1200 EX electron microscope. All specimens which were negative for CVLPS by this method were re-examined after ultracentrifugation of the clarified stool suspension for 1 hour at 37,000 rpm and resuspension of the pellet in 1-2 drops of 0.1% bacitracin. These preparations were applied to grids and stained as above. Standard bacteriological procedures were used to isolate Campylobacter spp., Escherichia coli, Salmonella spp., Shigella spp. and Yersinia spp. IEsrey et al., 1987 J. Laboratory personnel were blind as to the age and sex of the child, village type and location, and other study variables. Data were analysed using SYSTAT [ Wilkinson, 1987 I on a personal computer. Coronavirus-like particles were found in 81 of the 266 stool specimens (30.5%). Astrovirus was found in one specimen, and adenovirus in six others, one of which also contained CVLPs. Since CVLPs were the only virus-like agents detected with any frequency, only results related to CVLPs are given below. The CVLPs were pleomorphic, ranging from 80 to 740 nm in their greatest dimension. They had regular projections and were similar morphologically to the CVLPs described by other authors IMathan et al., 1975; Caul and Egglestone, 1977; Riordan et al., 1986; Schnagl et al., 19871. Not all of the fringes (coronas) of CVLPs in specimens from different children were exactly alike (Fig. 2) , and in two cases particles with distinctly dissimilar fringes were seen in the same specimen. The proportion of children shedding CVLPs increased with increasing age, from 21.3%. among children less than 13 months of age to 44.4% among children aged between 37 and 48 months (Table I) . However, the difference in the proportion of children of different ages shedding CVLPs was of doubtful significance ( P > 0.05). Among 4-year-old children, the proportion of positives dropped to 19.0%. There was no bias of age groups in the different districts ( P > 0.5). More males (35.5%) than females (26.2% ) shed CVLPs, but this was not significant ( P > 0.1). According to the mother's definition of diarrhea, 9.4%. of the children were reported to have diarrhea during the %-hour period prior to stool collection. No significant association between the diarrhea status of children and the presence or absence of CVLPs in the stool was found ( P > 0.9; Table 11 ). In three of the districts shown in Figure 1 , the proportion of children shedding CVLPs did not exceed 25% (Table 111 ). In a fourth district, Butha Buthe, the prevalence of CVLP shedding was at least twice the prev-Ftg 1 Map of Lesotho showing the main towns and villages after which the administrative districts are named dren under 5 years of age within the household were enumerated, Starting in July-August 1984, and continuing for approximately 6 months, an average of 28 children were studied in each village. Stool samples were taken from approximately 45% of these children ( n = 266) over a 5-week period in October-November 1984, at the start of the wet summer season. Stools were collected by mothers and placed in 60-ml plastic containers (Sterilin). The specimens were packed in ice within 4 hours, were flown to Johannesburg, and arrived on the same day as collection. After samples were taken by the bacteriology laboratory of the South African Institute for Medical Research, the specimens were kept a t 4°C and forwarded to the Virology Institute usually on the following day. Stools were examined for viruses by electron microscopy (EM) using standard procedures. Briefly, stools were prepared as 10-20% suspensions in distilled water and clarified at 3,000 rpm for 30 minutes. A drop of the supernate was applied to a carbon-Formvar-coated EM grid and the excess fluid was removed with torn filter paper. Potassium phosphotungstate 12% ) at pH 7 alence in any of the other districts (67.94 ). Butha Buthe was the first district to be sampled. The pattern of shedding was further broken down into week of shedding (Table IV) . Shedding was highest in week 1 (67.9%) and declined every week to 10% in week 5. This comparison is confounded, however, by district of sampling changing during each of the first 4 weeks of sampling. In each of two districts, Mafeteng and Leribe, data were collected from four villages during weeks 2 and 3, respectively. During week 5 , data from two additional villages in each of these districts were collected. Shedding was lower during week 5 than in week 2 in Mafeteng district ( P < 0.01) and in week 5 compared to week 3 in Leribe district ( P < 0.01). Only one stool sample was taken from each child during October-November 1984. In an attempt to ascertain whether prolonged shedding of CVLPs occurred, stools were collected again from six children in March 1985, 5 months after initial collection. These children had been shedding large numbers of CVLPs during the first collection, but all six second stool samples were negative by EM. The proportion of children shedding bacterial pathogens was as follows. For Campylobacter spp. it was 6.4%, for enteropathogenic E. coli 12.0%,, for enterotoxigenic E. coli 4.7%, and for enteroinvasive E . coli 9.4%. There was no significant association between the rates of shedding of these bacterial pathogens and the presence or absence of CVLPs in the stool. In addition, the protozoan pathogen Giardia lamblia was present in 23.7% of stools and the protozoan Entamoeba coli in 17.7% [Esrey et al., 19881 . The rates of shedding of other parasites (other Entamoeba spp., Hymenolepis spp., Endolimax spp., Chilomastir spp., and Taenia spp.) were below 3.0% There was no significant association between the shedding of any of these parasites, or the total number of parasites detected in stools, and the presence or absence of CVLPs. Similarly, CVLP shedding was not significantly related to any of the following variables, which have been defined in detail elsewhere [Esrey et al., 1987, 19881: quantity of water used per household per day, exclusive use of the improved (standpipe) water supply, use of a latrine, degree of crowding in the home, quantity of household possessions, age of mother and marital status (data not shown) In this paper we report the shedding of coronaviruslike particles by almost one-third of 266 children examined in Lesotho. The term coronavirus-like particle has been used by a number of authors, but not all of the particles described have been exactly alike. In one of the earliest reports on the presence of CVLPs in stool [Mathan et al., 19751 , three different forms of fringe were described. Biochemical characterization of CVLPs to date has involved immunoblotting [Resta et al., 1985; Battaglia et al., 19851 and electrophoresis of proteins using sensitive silver staining [Schnagl et al., 19871 , and much remains to be learned about the biochemical and antigenic make-up of CVLPs with different appearances. In this study, at least two of the children were shedding particles with different appearances in the same stool. The significance of this finding is now known, and neither child was reported to have diarrhea. It is not known whether different "strains" of CVLP exist, but there is some evidence by immune electron microscopy that there may be antigenic variants ISchnagl et al., 19861 . One other line of evidence for more than one group of CVLPs came from the study in Gabon [Sitbon 19851 . Different seasonal variations in the shedding of CVLPs amongst the diarrheal and control groups indicated the existence of two or more groups of CVLPs with different epidemiological characteristics. The shedding of CVLPs was not significantly associated with diarrhea. This is in agreement with Schnagl et al. 11978, 19791 , who detected CVLPs in the stools of Australian aboriginal children with similar frequency whether healthy or with acute diarrhea. Moreover, Sitbon 119851 found a higher proportion of healthy children in Gabon shedding CVLPs (65.2% ) than children with diarrhea (38.5%). Thus it is unlikely that CVLPs shed by children in developing populations are a major cause of diarrhea. Nevertheless, causation of diarrhea cannot be ruled out, as prolonged infection could occur, starting with diarrhea and resulting in reduced pathogenicity and hence tolerance. Assuming that CVLPs are enteric viruses, spread could be by the fecal-oral andlor respiratory route. If the spread is fecal-oral, then transmission is most likely to be related to conditions of hygiene rather than to the quality of drinking water, because no association between CVLP shedding and quality of water was found [Esrey et al., 19871. Twenty-five 5% of children shed a t least one bacterial pathogen in this study, which is a further indication that hygiene standards were low. However, CVLP shedding was apparently unrelated to the amount of water used per household, the presence of a latrine, or characteristics of the mother that might be related to hygiene. Spread by the respiratory route may be possible, but CVLP shedding in this study was not related to crowding. The present study was conducted in a temperate climate, and the rates of shedding were comparable to those reported for tropical countries I Mathan et al., 1975; Sitbon 19851. Assuming that CVLPs are enteric viruses, it seems likely that transmission of these agents could occur in any climate, given the correct predisposing factors. It is not known what these factors might be. Poor hygiene and malnourishment may play a role. It is also possible that susceptible children such as those found in our study are effectively immunocompromised as a result of repeated exposure to pathogens. The figures reported by other authors, revealing high rates of shedding in some third world populations [Mathan et al., 1975; Sitbon, 1985; Schnagl et al., 1978, 19791 may also reflect exposure to multiple pathogens. Thus high rates of shedding in some populations may be more a result of environmental circumstances than a n inherited susceptibility to such agents. The increasing prevalence of CVLP shedding with increasing age up to 3-4 years in this study is of questionable significance. It might suggest chronic infection with a cumulative increase in the number of children infected from birth. The apparent drop in the proportion of children shedding CVLPs beyond the age of 4 years does not fit this possibility. Although we could not demonstrate prolonged shedding, chronic infection with seasonal shedding cannot be discounted. It seems more likely that the progressive decrease in the proportion of children shedding CVLPs over the 5-week study period was due to seasonal shedding. The results were confounded by geographical differences in the populations sampled in different weeks, but it was evident from studies done in the same districts at different times that the proportion of children shedding particles was decreasing significantly. The reason for such a pronounced drop in shedding rates between spring and summer is not understood. The controversy over the viral nature of CVLPs [Dourmashkin et al., 19801 has been compounded by the recent study reported by Schnagl et al. I1987 I on the protein profiles of CVLPs by polyacrylamide gel electrophoresis. At least 38 polypeptides were detected in highly purified preparations of Central Australian CVLPs, which is more than the number known for poxviruses, the most complex DNA viruses yet described. Thus, the question of whether or not CVLPs are viruses has still not been satisfactorily answered. These authors put forward some evidence that they are unlikely to be mycoplasmas, but did not rule out the possibility that they may be fragments of parasites. However, in the present study there was no significant association between the shedding of CVLPs and infestation by either of the two most commonly found protozoa (Giardia lamblia and Entamoeba coli). Human enteric coronaviruses: Further characterization and immunoblotting of viral proteins Further studies on human enteric coronaviruses Coronavirus particles in faeces from patients with gastroenteritis Bird RG t1980): Are coronavirus-like particles seen in diarrhoea stools really viruses? Lancet The risk of infection from Giardra lonzhlia due to drinking water supply, use of water, and latrines among preschool children in rural Lesotho Latham MC 119871: Infection, diarrhoea and growth rates of young children following the installation of village water supplies in Lesotho Virus Infections of the Gastrointestinal Tract Human enteric coronaviruses: Antigenic relatedness to human coronavirus OC43 and possible etiologic role in gastroenteritis Detection of coronavirus-like particles in homosexual men with acquired immunodeficiency and related lymphadenopathy syndrome Human enteric coronaviruses Pleomorphic virus-like particles in human faeces Isolation and propagation of a human enteric coronavirus Fatal gastroenteritis associated with coronavirus-like particles Enteric coronavirus in symptomless homosexuals Characteristics of Australian human enteric coronavirus-like particles: Comparison with human respiratory coronavirus '229E and duodenal brush border vesicles Antibody to human enteric coronavirus-like particles and indications of antigenic differences between particles fromdifferentareas Coronaviruslike particles in aboriginals and non-aboriginals in Western Australia Rotavirus. coronavirus-like particles, bacteria and parasites in Central Australia Sitbon M 119851: Human-enteric-coronaviruslike particles ICVLPI with different epidemiological characteristics Pleomorphic, enveloped, virus-like particles associated with gastrointestinal illness in neonates SYSTAT: The System for Statistics We wish to thank Professor H.J. Koornof and Dr. Marianne D. Miliotis of the Department of Microbiology and Mr. James Collett of the Department of Tropical Diseases, School of Pathology, University of the Witwatersrand and South African Institute for Medical Research, Johannesburg, for facilitating the laboratory analyses and liaising between Lesotho and RSA. This research was funded in part by USAID grant number 632-0088-5-00-4012-00, and by the Medical Research Council of South Africa.