This is the first comprehensive study to evaluate anti-S IgG and IgM for four non-SARS HCoVs among the general population in China. We first time show that anti-S IgM is only present in children, indicating that for all viruses first infection takes place during childhood (age < 14 years).
The SARS epidemic that originated from southern China in 2003 sparked interest in all areas of coronavirus research [4–9]. Four endemic non-SARS-related HCoVs (HCoV-OC43,-229E,-NL63 and –HKU1) are major contributors to respiratory tract infections and other clinical manifestations [6, 11–17, 22]. However, specific and feasible serological surveys of these HCoVs, especially for HCoV-NL63 and –HKU1, have to date been reported among the general population only rarely in China. Spike is the major structural protein of HCoVs [1, 10]. It contains multiple conformational epitopes that are major inducers of antibody neutralisation, and it has the lowest sequence conservation among coronavirus proteins, rendering it a specific target for serodiagnosis [1, 10, 23–26]. We chose the HCoV S gene for recombinant expression on the basis of current knowledge of immunodominant SARS-CoV antigens, which belongs to the same virus family as HCoV. Previous studies of SARS-CoV serology have successfully used the S protein in enzyme immunoassays, immunoblots, and IFA [23–27]. In addition, comparison of the S protein sequences of four HCoVs revealed that these proteins share <35% similarity [1, 10]. We hypothesised that the difference in the amino acid sequence is sufficiently high to ensure the usefulness of S protein as a specific antigen for antibody detection. The native surface S protein expressed in 293 T cells can be recognised in post-infection population serum by IFA. Some reports indicated that this cell-based S protein expression system can differentiate false-positive ELISA results using the more cross-reactive nucleoprotein antigen[18,24]. Furthermore, Woo et al. reported that the rS-based IgM ELISA is more sensitive than the rN-based IgM ELISA for SARS-CoV pneumonia . Our results showed that the S-based IFA enabled specific detection of IgG or IgM to four individual HCoVs.
Using IFA, we investigated the natural seroprevalence of four non-SARS-related HCoVs in blood samples from a general population that comprised a variety of age groups. Anti-S IgG antibodies to these four non-SARS-related HCoVs were detected at high rates (>70%) among healthy adults. Both anti-S IgG and IgM antibodies were found in the child group, and their prevalence increased with age up to 6 years, at which point it almost plateaued. These data suggest that exposure to HCoV is common in childhood and first infections by all four non-SARS HCoVs takes place during children. Moreover, we found evidence of anti-S IgG against double and multiple HCoVs in various combinations, which indicates that large proportions of the general population in Beijing may experience infections with more than one HCoV.
The seroprevalence of HCoV antibodies varies widely among studies, which used different antigens and methodologies [6, 13–18, 26–33]. Previous studies of non-SARS HCoVs demonstrated that seroprevalence varies greatly depending on the age of the population [6, 13–18, 26–33]. Our IFA study showed that the prevalence of the four anti-S IgG antibodies among this general Chinese population were slightly higher than those in another study from Germany using an Escherichia coli BL21-expressed recombinant N-based line immunoassay , which reported that the seropositivities in 25 healthy blood donors were 48% for HCoV-HKU1, 52% for -OC43, 56% for -229E, and 60% for -NL63. On the other hand, the seropositivity rate of three individual HCoVs excluding HCoV-HKU1 in this study was lower than that in another study of a US metropolitan adult population using baculovirus-expressed recombinant N-based ELISA , which showed seropositivity rates of 90.8% for HCoV-OC43, 91.3% for -229E, and 91.8% for -NL63, while that for -HKU1 was 59.2%. It was also supposed that the significantly different seropositivity rates for the various HCoVs might result in individuals with different demographic factors (e.g. ethnicity, smoking status, and socioeconomic status) having different susceptibilities to individual HCoVs . A recent study of 105 older adult veterans with underlying chronic obstructive pulmonary disease at seven US sites showed that serum IgG to HCoV-229E, -NL63, and -OC43 was detected in at least 98% of subjects, while antibodies to -HKU1 were identified in 96 subjects (91%) . Thus EIA assays that use whole virus or N protein as the antigen, using which apparent cross-reactivity of HCoV antibodies has been demonstrated previously [18, 24], may detect group- rather than type-specific antibody.
Our results regarding anti-S IgG to four individual HCoVs in children correlate with those of previous seroprevalence studies [6, 13–18, 26–33], although the data on HCoV-NL63 and -HKU1 were limited. Shao et al. found that antibodies directed to HCoV-NL63 and -229E in children 1 year of age and older were frequently detected using part of the C-terminal region of the N protein as an antibody capture antigen in an ELISA . A study of children in the Netherlands aged 2.5–3.5 years indicated that 75% and 65% of serum samples were positive for antibodies to HCoV-NL63 and -229E, respectively . We observed that the prevalence of these four non-SARS-related HCoV-directed IgG antibodies among children >1 year of age were almost identical.
Information on the prevalence of anti-S IgM to non-SARS-related HCoV is to-date lacking. We first investigated the prevalence of anti-S IgM to four individual HCoVs among a general population in Beijing using IFA. Anti-S IgM to individual HCoVs was detected in a portion of the asymptomatic child population. The anti-S IgM seropositivity appeared to increase with age up to 6 years and decline sharply after 14 years of age. However, no anti-S IgM to individual HCoVs was detected in the healthy adult population. These results suggest that primary seroconversion to these viruses occurs mainly during childhood and youth; In addition, HCoV infection might result in seroconversion in children with asymptomatic or subclinical manifestations. These results are also in agreement with HCoV molecular epidemiological surveys [12, 34], which have indicated that primary exposure occurs mainly in childhood and youth.