key: cord-263658-dlmzcl9g
authors: MCINTOSH, KENNETH; KAPIKIAN, ALBERT Z.; TURNER, HORACE C.; HARTLEY, JANET W.; PARROTT, ROBERT H.; CHANOCK, ROBERT M.
title: SEROEPIDEMIOLOGIC STUDIES OF CORONAVIRUS INFECTION IN ADULTS AND CHILDREN(1)
date: 1970-06-17
journal: Am J Epidemiol
DOI: 10.1093/oxfordjournals.aje.a121171
sha: 
doc_id: 263658
cord_uid: dlmzcl9g

McIntosh, K. A. Z. Kapikian, H. C Turner, J. W. Hartley, R. H. Parrott and R. M. Chanock. (Lab. of Infectious Diseases, NIAID, NIH, Bethesda, Md. 20014) Sero-epidemiologic studies of coronavirus infection in adults and children. Amer. J. Epid., 1970, 97: 585–592-A seroepidemiologic study of infection by coronavirus strains 229E, OC38, OC43, and mouse hepatitis virus (MHV) strain A-59, is described. In adults with upper respiratory disease, two “outbreaks” of coronavirus infection occurred, one during the winter of 1965–1966 associated with complement fixing (CF) antibody responses to OC38, OC43 and MHV, and the other during the following winter associated with CF antibody responses to 229E. In hospitalized children, infection with 229E was rare; infection with OC38, OC43, and MHV occurred less often in hospitalized children with lower respiratory tract disease (3.5%) than in a control group with non-respiratory tract disease (8.2%). The limitations of the CF test using available coronavirus antigens are discussed.

Several investigators have recently reported the recovery from adults with upper respiratory infection of ether-labile viruses morphologically indistinguishable from mouse hepatitis virus (MHV) and avian infectious bronchitis virus (EBV) (1) (2) (3) (4) (5) (6) . The distinctive appearance of these viruses by electron microscopy suggested that they be designated as a new group, the "coronaviruses" (7) . Members of the group are medium-sued, RNA-containing, and etherlabile, and bear characteristic club-shaped surface projections. Nine of the 23 strains recovered from the human respiratory tract were originally isolated in human embryonic tracheal organ culture (HETOC) (1, 3, 6, 8) . Three of the nine were subsequently adapted to growth in monolayer tissue cultures and two others (OC38 and OC43) were successfully adapted to growth in suckling mice and shown to be antigenically related to several strains of MHV by CF tests and fluorescence staining (6, 9, 10) . The 14 remaining human strains were originally recovered in tissue culture monolayers and appeared antigenically similar or identical to the prototype tissue culture strain, 229E (2, 5) .

Little is known of the epidemiology of human infection with coronaviruses. CF and neutralizing antibodies to several strains of 685 MHV were found in military recruits before any coronavirus strains had been recovered from man (11) . Epidemiologic analysis of these findings indicated that, although fourfold rises in antibody were frequent, there was no discernible association with disease. Both strain 229E and strain B814 have been shown to cause colds in adult volunteers (1, 12) . Preliminary seroepidemiologic surveys performed in this laboratory suggested that 229E, OC38, and OC43 infection occurred with sporadic frequency in adults with upper respiratory tract disease (5, 9) . This report describes a more extensive seroepidemiologic study using the available CF antigens from human strains and an antigen of MHV, strain A-59, shown by Hartley and others to detect CF antibody in human sera (11) .

Adults. Nasopharyngeal washings and acute phase sera were obtained from employees of the National Institutes of Health with upper respiratory tract disease on or before the fourth day of illness, and convalescent phase sera three weeks later. There were 466 patients admitted to the study, some of them on multiple occasions, so that 577 serum pairs were available for study. The periods covered were October, 1962 through May, 1964 , and September, 1965 through September 5, 1967 . Some aspects of the study have been published (5, 13) .

Infants and children. Throat or nasal swabs and acute phase sera were obtained from pediatric patients on admission to Children's Hospital, Washington, D.C., and convalescent phase sera about three weeks later. A diagnosis of lower respiratory tract disease (LRTD) indicated croup (laryngotracheobronchitis), bronchitis, pneumonia, or bronchiolitis. Control sera were obtained at similar intervals from hospitalized patients with non-respiratory tract disease. Individuals with incidental respiratory tract disease at the time of selection were excluded from the group. There was difficulty in finding sufficient control infants under 12 months of age (see table 1 ). Members of the control group were not matched with respect to their length of hospital stay. The sera chosen for study were obtained from October, 1965 through April, 1966, and from October, 1966 through April, 1967 (14) .

The preparation and maintenance of human embryonic intestine (HEI) diploid cell cultures, and their use in the recovery of coronaviruses have been described (5) . Human embryonic tracheal organ cultures were prepared and maintained as previously reported (3) . The methods used for recovery of coronaviruses in organ culture have also been described (3).

Dr. Hamre kindly supplied coronavirus strain 229E which had been purified by the terminal dilution technique. It was passaged several times in human diploid cell strain WI38 in this laboratory. CF antigen was prepared as a single lot of infected WI38 cells (5) . Eight units of antigen as measured with a standard convalescent human serum were used in all tests.

CF antigen of MHV, strain A-59, was prepared from infected mouse liver NCTC 1469 tissue culture as previously described (11) . A single lot of antigen was used for the study reported here. Four to eight units of antigen as measured with hyperimmune mouse serum were used.

The adaptation of organ culture grown coronavirus strains OC38 and OC43 to Swiss mice and the preparation of CF antigens from infected mouse brain have been described (9) . The antigens used in this study were from a single lot for each virus and were free of contamination with mycoplasma and cytopathic agents. Moreover, screening CF tests were performed for possible contamination with Theiler's agent, lymphocytic coriomeningitis virus, Sendai virus, mouse leukemia, reoviruses, rubella, mumps, measles, respiratory syncytial (ES) virus, parainfluenza virus types 1, 2, 3, and 4, and influenza virus types A, B, and C; all such tests were negative. The techniques of insuring that the antigens were also free of MHV strains have been previously described (9) . Eight to 16 units of antigen as measured with hyperimmune mouse serum were used.

Control uninfected WI38 cell, NCTC 1469 cell and mouse brain antigens were prepared in parallel to infected antigen lots and were included in each test. Sera reacting with control antigens were not included in • the data reported. These omissions account for the differences in the total number of serum pairs analyzed for each antigen (see table 2 ).

CF tests were performed by the microtiter technique using overnight fixation at 4 C and 1.7-1.8 units of complement as previously described (15) . Figure 1 is a diagram representing the number of and per cent antibody responses to the various coronavirus antigens tested in MHV, and because of their reported close antigenic relationships (9), in certain analyses these agents are grouped together. Table 1 shows the proportion of infants and children with and without LRTD showing fourfold or greater antibody responses to the three related coronavirus antigens OC38, OC43, and MHV, strain A-59. When all age groups were combined, a positive correlation between coronavirus infection and LRTD did not exist. However, the youngest age group (under one year) with LRTD tended to have more coronavirus infections than the control group, but the difference was not statistically significant (p > .05). In all age groups taken together there was a significant negative correlation with LRTD (x* = 7.4, p <.01). This negative correlation was particularly striking in the children over the age of one (x* = 10.6, p <.01).

In table 2 is shown the proportion of individuals tested who had measurable (1:4 or greater) CF antibody to coronavirus antigens. During the period 1965-1967 CF antibody to strain 229E was rare in children (0.6 per cent) and quite common in adults (41 per cent). CF antibody to OC38 and OC43 was increasingly common in children up to the age of three years, when it was measurable in approximately 50 per cent of those tested. In adults, 72 per cent of those tested had measurable CF antibody to one or both strains during 1962-1964, and 67 per cent during 1965-1967. Antibody to MHV, strain A-59, was of somewhat lower prevalence (approximately 10 per cent in children and 34 per cent in adults during the periods noted above).

Information obtained from serologic, tissue culture, and organ culture studies is summarized in figures 2 and 3. Among adults (figure 2) coronaviruses appeared to be rarely associated with upper respiratory tract disease during 1963 and 1964. The frozen nasopharyngeal washings of a small group of patients, who in preliminary studies had shown CF antibody responses to strain 229E, were examined in HEI tissue culture. Three of these yielded coronaviruses serologically similar to strain 229E. Sporadic frozen specimens were examined in tracheal organ In the second "outbreak" antibody responses to OC38, OC43, and MHV antigens were rare, and no coronaviruses were recovered in organ culture. In children, on the other hand, (figure 3) there appeared to be no meaningful temporal pattern of coronavirus CF antibody responses, and no strains were recovered in organ culture. 

The children under study were examined for evidence of infection by other respiratory viruses, and a certain number of dual infections with myxo-or paramyxoviruses and coronaviruses were found. There was no evidence that coronavirus infection potentiated the pathogenic effect of infection by either parainfluenza virus 3 or RS virus, the two viruses found most often in conjunction with coronaviruses (table 3) . DISCUSSION These studies were undertaken to define more clearly the seroepidemiology of coronavirus infection in adult and pediatric popu-lations. Serum samples and nasopharyngeal washings were studied from adults with colds occurring in two two-year periods. During the first period (1962-1964) detectable coronavirus infection was uncommon. However, in the winter of 1965-1966, members of the OC38-OC43-MHV group were prevalent and detectable both by serologic and organ culture techniques. One year later members of the 229E group were prevalent and again detectable by both serologic responses and virus isolation. It is of interest that among adults all the coronavirus recoveries and most of the coronavirus serologic responses occurred in the four winter months of December, January, February, and March. In several surveys of colds in adults, respira-tory disease was common during these winter months but rhinovirus infection, though frequent in the fall and spring, was rare (13, 16, 17) . This pattern of rhinovirus infection prevailed in this study during the period 1965-1967. Thus, during these winter months coronaviruses although still accounting for less than 50 per cent of colds, became the predominant identifiable organisms associated with adult upper respiratory disease.

It is evident from the studies of hospitalized children that infection with coronaviruses was not significantly associated in this survey with pediatric lower respiratory tract disease. Indeed, a negative correlation of coronavirus infection with LRTD was found. It may be that the presence of severe LRTD requiring hospitalization interfered in some way with infection of the respiratory epithelium by coronaviruses. On the other hand, it seems more likely that this negative correlation reflects the fact that it was not possible to select a completely comparable control group. Control children were drawn from different pediatric wards, and their lengths of stay in the hospital frequently differed from those of the test group. In any case, the data reported suggest that the coronaviruses tested were not an important cause of LRTD in infants and children during the period covered by this study.

Our studies did not provide information on the etiologic association of coronavirus infection with respiratory tract disease in adults. Two coronavirus strains, 229E and B814, were administered to adult volunteers and in both instances a significant number of common colds occurred (1, 12) . Data are still lacking, however, to show that members of the OC38-OC43 serologic type cause disease under natural conditions. It is of interest that in the seroepidemiologic studies of Hartley and others, where several strains of MHV were tested with sera from military recruits, no epidemiologic association of MHV or MHV-like virus infection with respiratory tract disease was made (11) .

The infrequency of detectable CF antibody to 229E in the pediatric population studied here was surprising. Although 229E virus infection may indeed be rare in children, several other explanations for this anomalous finding are possible. The 229E antigen, in contrast to the other coronavirus antigens, may have reacted with CF antibody to a narrow range of coronavirus serologic types, owing either to the type of tissue in which it was made, or to some characteristic of the virus itself. Possibly serum CF antibody (as opposed to serum neutralizing, or secretory antibody) did not develop in children in response to 229E virus infection. Studies of 229E neutralizing antibody in different age groups would contribute to the clarification of this finding.

Antigenic studies of human coronaviruses have shown that those members of the group which were originally recovered in tissue culture are all closely related to the prototype virus strain 229E (2, 5) . In contrast, the strains isolated in organ culture include at least two and probably several more anti-genie types (10) . One type, exemplified by strains OC38 and OC43, is clearly related to the MHV group. The other type or types have an indefinite relationship with MHV and with the OC38-OC43 group. In particular, patients yielding coronaviruses in organ culture varied markedly in their antibody responses to the CF antigens used in this study. One such patient showed in multiple tests no rise to any of the coronavirus antigens. Two others showed responses in some tests and not in others. Infection with strain B814 was likewise difficult to detect serologically, using OC38, OC43, MHV and 229E antigens (10) . The heterogeneity of antibody response in subjects with known or presumed coronavirus infection, and the probable insensitivity of presently available serologic tests in the detection of coronavirus infection indicate that the serologic studies reported here do not describe a complete picture of coronavirus infection. It is probable that undetected coronavirus infections occurred in these populations during the study periods. CF antigens from the other known coronavirus strains, and further efforts to isolate and characterize new strains will contribute to further definition of the epidemiology of coronavirus infection.

Cultivation of a novel type of common-cold virus in organ culture

A new virus isolated from the human respiratory tract

Recovery in tracheal organ cultures of novel viruses from patients with respiratory disease

The morphology of three previously uncharacterized human respiratory viruses that grow in organ culture

Isolation from man of "avian infectious bronchitis viruslike" viruses (coronaviruses) similar to 229E virus, and some epidemiologic observations

Sensitivity of L132 cells to some "new" respiratory viruses

Cultivation of difficult viruses from patients with common colds

Growth in suckling-mouse brain of "IBV-like" viruses from patients with upper respiratory tract disease

Antigenic relationships among the coronaviruses of man and between human and animal coronaviruses

Antibodies to mouse hepatitis virus in human sera

Effects of a "new" human respiratory virus in volunteers

Etiology of upper-respiratory-tract illness among civilian adults

Respiratory syncytial virus. II. Serologic studies over a 34-month period of children with bronchiolitis pneumonia, and minor respiratory diseases

Application of a microtechnique to viral serological investigations

Virologic studies of acute respiratory disease in young adults. IV. Virus isolations during four years of surveillance

Rhinovirus infections in an industrial population. I. The occurrence of illness