key: cord-0786574-cmjnb0al
authors: Gerna, Giuseppe; Campanini, Giulia; Rovida, Francesca; Percivalle, Elena; Sarasini, Antonella; Marchi, Antonietta; Baldanti, Fausto
title: Genetic variability of human coronavirus OC43‐, 229E‐, and NL63‐like strains and their association with lower respiratory tract infections of hospitalized infants and immunocompromised patients
date: 2006-05-23
journal: J Med Virol
DOI: 10.1002/jmv.20645
sha: f4e792c44ffb34aba6bffc26b2cbb3b4562e7048
doc_id: 786574
cord_uid: cmjnb0al

In the winter–spring seasons 2003–2004 and 2004–2005, 47 (5.7%) patients with acute respiratory infection associated with human coronavirus (hCoV) 229E‐, NL63‐, and OC43‐like strains were identified among 823 (597 immunocompetent and 226 immunocompromised) patients admitted to hospital with acute respiratory syndromes. Viral infections were diagnosed by either immunological (monoclonal antibodies) or molecular (RT‐PCR) methods. Each of two sets of primer pairs developed for detection of all CoVs (panCoV) failed to detect 15 of the 53 (28.3%) hCoV strains identified. On the other hand, all hCoV strains could be detected by using type‐specific primers targeting genes 1ab and N. The HuH‐7 cell line was found to be susceptible to isolation and identification of OC43‐ and 229E‐like strains. Overall, hCoV infection was caused by OC43‐like, 229E‐like, and NL63‐like strains in 25 (53.2%), 10 (21.3%), and 9 (19.1%) patients, respectively. In addition, three patients (6.4%) were infected by untypeable hCoV strains. NL63‐like strains were not found to circulate in 2003–2004, and 229E‐like strains did not circulate in 2004–2005, while OC43‐like strains were detected in both seasons. The monthly distribution reached a peak during January through March. Lower predominated over upper respiratory tract infections in each age group. In addition, hCoV infections interested only immunocompetent infants and young children during the first year of life, while all adults were immunocompromised patients. Coinfections of hCoVs and other respiratory viruses (mostly interesting the first year of life) were observed in 14 of the 47 (29.8%) patients and were associated with severe respiratory syndromes more frequently than hCoV single infections (P = 0.002). In conclusion, the use of multiple primer sets targeting different genes is recommended for diagnosis of all types of hCoV infection. In addition, the detection of still untypeable hCoV strains suggests that the number of hCoVs involved in human pathology might further increase. Finally, hCoVs should be screened routinely for in both infants and immunocompromised patients with acute respiratory infection. J. Med. Virol. 78:938–949, 2006. © 2006 Wiley‐Liss, Inc.

In the winter-spring seasons 2003-2004 and 2004-2005, 47 (5.7%) patients with acute respiratory infection associated with human coronavirus (hCoV) 229E-, NL63-, and OC43-like strains were identified among 823 (597 immunocompetent and 226 immunocompromised) patients admitted to hospital with acute respiratory syndromes. Viral infections were diagnosed by either immunological (monoclonal antibodies) or molecular (RT-PCR) methods. Each of two sets of primer pairs developed for detection of all CoVs (panCoV) failed to detect 15 of the 53 (28.3%) hCoV strains identified. On the other hand, all hCoV strains could be detected by using typespecific primers targeting genes 1ab and N. The HuH-7 cell line was found to be susceptible to isolation and identification of OC43-and 229E-like strains. Overall, hCoV infection was caused by OC43-like, 229E-like, and NL63-like strains in 25 (53.2%), 10 (21.3%), and 9 (19.1%) patients, respectively. In addition, three patients (6.4%) were infected by untypeable hCoV strains. NL63-like strains were not found to circulate in [2003] [2004] , and 229E-like strains did not circulate in [2004] [2005] , while OC43-like strains were detected in both seasons. The monthly distribution reached a peak during January through March. Lower predominated over upper respiratory tract infections in each age group. In addition, hCoV infections interested only immunocompetent infants and young children during the first year of life, while all adults were immunocompromised patients. Coinfections of hCoVs and other respiratory viruses (mostly interesting the first year of life) were observed in 14 of the 47 (29.8%) patients and were associated with severe respiratory syndromes more frequently than hCoV single infections (P ¼ 0.002). In conclusion, the use of multiple primer sets targeting different genes is recommended for diagnosis of all types of hCoV infection. In addition, the detection of still untypeable hCoV strains suggests that the number of hCoVs involved in human pathology might further increase. Finally, hCoVs should be screened routinely for in both infants and immunocompromised patients with acute respiratory infection. J. Med. Virol. 78:938-949, 2006 . ß 2006 Wiley-Liss, Inc.

Following the first reports in the 1960s [Tyrrell and Bynoe, 1965; Hamre and Procknow, 1966; McIntosh et al., 1967] , human coronaviruses (hCoVs) were shown to be pathogenic in volunteer studies [Bradburne and Somerset, 1972] and widespread in the community in a number of seroepidemiological surveys of the two known strains, hCoV-229E and hCoV-OC43 [McIntosh et al., 1970; Kaye et al., 1971 Kaye et al., , 1972 Hamre and Beem, 1972; Monto and Lim, 1974; Gerna et al., 1978 Gerna et al., , 1980 MacNaughton, 1982] . Initially, several other hCoVs were described, such as B814 [Tyrrell and Bynoe, 1965] , hCoV-OC16, hCoV-OC37, and hCoV-OC48 [McIntosh et al., 1967] . However, these viruses could not be grown in tissue culture or animal models. Attempts to identify hCoVs in clinical specimens or cell cultures by using either polyclonal or monoclonal antibodies were not satisfactory [McIntosh et al., 1978; Sizun et al., 1998 ].

Animal CoVs were then investigated extensively in a variety of animal species, while the study of the antigenic relationship of human and animal CoVs was addressed [Kaye et al., 1975; Gerna et al., 1981] , and the unique strategy of replication of CoVs was described [Lai and Holmes, 2001] . This allowed the rapid identification of the etiologic agent of the severe acute respiratory syndrome (hCoV-SARS) in 2003 [Peiris et al., 2003] . In parallel, the potential pathogenic role of hCoVs, and, in particular, hCoV-OC43 in diseases of the gastrointestinal tract was investigated by the study of the antigenic relationship of hCoV-OC43 and neonatal calf diarrhea coronavirus or bovine CoV (BCoV), and hCoV-OC43 and human enteric CoVs [Chany et al., 1982; Gerna et al., 1985] .

In addition, molecular techniques permitted the identification of known group I (229E-like) and group II (OC43-like) hCoVs in respiratory secretions [Sizun et al., 1998; El-Sahly et al., 2000; Vabret et al., 2001 Vabret et al., , 2003 Falsey et al., 2002; Pene et al., 2003; Vallet et al., 2004] and the discovery of new hCoVs. Recently, two groups of researchers from The Netherlands [Fouchier et al., 2004; van der Hoek et al., 2004] and one group from USA [Esper et al., 2005] reported a new hCoV, referred to as NL63 and New Haven, respectively. This virus was related distantly to hCoV-229E, and described as one viral agent associated with acute respiratory disease in hospitalized children. Furthermore, another hCoV (referred to as HKU1), distantly related to hCoV-OC43, was detected in two Hong-Kong patients with pneumonia [Woo et al., 2005a] , and then reported elsewhere [Woo et al., 2005b; Sloots et al., 2006; Vabret et al., 2006] . A similar virus was also reported in Sweden by molecular screening of respiratory tract samples [Allander et al., 2005] . Finally, a few studies have addressed the circulation of hCoVs with respect to other respiratory viruses in different countries, as well as their involvement in respiratory syndromes other than the common cold, and their association with asthma and chronic obstructive bronchopneumopathies, including hCoV pneumonia in transplanted patients [Lina et al., 1996; Folz and Elkordy, 1999; El-Sahly et al., 2000; Falsey et al., 2002] .

At present, several issues remain to be investigated or defined: (i) the reliability of currently available immunological and molecular methodologies for diagnosing infections caused by different groups of hCoVs; (ii) the epidemiology of respiratory infections caused by hCoVs; (iii) the role of known hCoVs in causing admission to the hospital of patients with respiratory infections as well as their relationship to the severity of the relevant respiratory syndromes; (iv) the comparative pathogenicity of different hCoVs in different age groups. In the present study, these issues were investigated in a hospitalized patient population affected by 229E-, NL63-, and OC43-like infections during two consecutive winter-spring seasons. To date, HKU1 has not been detected in Italy.

From December 2003 through May 2005, nasopharyngeal aspirates were collected prospectively from 823 patients admitted to hospital with an episode of acute respiratory infection. Of these, 333 were less than 1 year old, 168 2-5 years old, 87 6-20 years old, and 233 >21 years old. Study patients were further subdivided based on immunocompetence. In the four age classes, immunocompetents included 333, 143, 44, and 77 patients, while immunosuppressed patients included 2, 25, 43, and 156 subjects, respectively. Of the 226 immunosuppressed patients, 70 included in the first three age groups were hematopoietic stem cell transplant recipients. In the >21-year group 108 were solid organ transplant recipients, 46 stem cell transplant recipients, and two AIDS patients. Specimens were handled and aliquoted as reported previously , and used for immunological (monoclonal antibodies) and molecular (RT-PCR) assays, as well as for short-term and long-term virus isolation in cell culture. Respiratory samples were examined routinely for influenzaviruses A and B, parainfluenzaviruses 1-4, human respiratory syncytial virus, and human adenoviruses by both direct fluorescent staining and culture. In addition, human metapneumoviruses were examined by both monoclonal antibodies and RT-PCR as recently reported Gerna et al., 2006] . Finally, hCoVs were tested by monoclonal antibodies and RT-PCR, as detailed below, and human rhinoviruses were searched for by RT-PCR [Steininger et al., 2001] . In this study, the term coinfection indicates the simultaneous detection of two or more respiratory viruses in the same sample taken from a patient with respiratory infection, while the term sequential infection refers to the sequential identification of two different respiratory viruses in two nasopharyngeal aspirates taken from the same patient within 30 days from each other.

Direct fluorescent antibody staining was applied to both slides containing smears of respiratory epithelial columnar cells from nasopharyngeal aspirates and cell cultures 48 hr after inoculation of respiratory secretions, as reported [Rovida et al., 2005] . In either case, as a first step, cells were stained with a pool (SimulFluor Respiratory Screen reagent, Chemicon International, Inc., Temecula, CA) of fluorescein-labeled monoclonal antibodies to conventional respiratory viruses (influenzavirus types A and B, human parainfluenzavirus types 1-4, human respiratory syncytial virus, and human adenovirus). Then, as a second step, positive samples were stained with individual monoclonal antibodies obtained from the same source (Chemicon). Group-and type-specific monoclonals to human metapneumovirus were developed in the laboratory as reported Gerna et al., 2006] . As for hCoVs, commercially available monoclonal antibody to OC43 (Chemicon) was found to perform satisfactorily, while monoclonal antibodies to 229E were developed in the laboratory and found to be highly specific. Finally, monoclonal antibodies to NL63 were not available.

Inoculated cell cultures included mixtures (MIX) of A549 and Mv1Lu (ratio 1:1) cells [Huang and Turchek, 2000] , as well as LLC-MK2 and MDCK cell lines. Cell cultures were processed for virus isolation and identification as reported previously [Rovida et al., 2005] . In addition, in the last part of the study, HuH-7 [ [Nakabayashi et al., 1982] provided by Maura Pizzuti, Institute for Biomedical Research, Rome, Italy], Vero and LLC-MK2 cell lines, as well as secondary African green monkey kidney cell cultures, were inoculated with six NPAs RT-PCR-positive for hCoVs (one positive for 229E, one positive for OC43, and four positive for NL63) to investigate the susceptibility of these cells to isolation and propagation of hCoVs.

RT-PCR assays for respiratory viruses were optimized to detect at least 10 input plasmid copies, as described previously ]. In addition, as reported in Table I , NPAs were tested for hCoVs by RT-PCR by using: (i) a primer set (referred to as PanCoV-03) developed for detection of all coronaviruses and relevant to gene 1ab [Poutanen et al., 2003] ; (ii) a second set of primers (referred to as PanCoV-05) also developed for detection of both human and animal coronaviruses and relevant to gene 1ab ; (iii) three sets of primers specific for OC43, 229E, and NL63 hCoVs, respectively, and relevant to gene N; and, finally, three sets of primers specific for OC43, 229E, and NL63, respectively, derived from PanCoV-05, thus, relevant to gene 1ab.

The following fragments of the 1ab gene were amplified (Table I) : nt 14321-14536 of 229E (amplicon size, 215 nt) with Pan CoV-03, nt 14098-14348 of 229E (size, 250 nt) with Pan CoV-05, nt 14922-15172 (size, 250 nt) with OC43-specific primers, nt 14098-14348 (size, 250 nt) with 229E-specific primers, and nt 14017-14267 (size, 250 nt) with NL-63-specific primers. In addition, the following fragments of the N gene were amplified: nt 715-1073 (size, 378 nt) with OC43-specific primers, nt 578-921 (size, 343 nt) with 229E-specific primers, and nt 563-876 (size, 313 nt) with NL63specific primers. The indicated fragments of genes 1ab and N of hCoVs OC43, 229E, and NL63, following amplification with specific primers, were sequenced with the ABI PRISM 3100 automatic sequencer (Applied Biosystems, Foster City, CA). Viral sequences of the amplified fragments of genes 1ab and N of different hCoV strains as well as reference strains were aligned with the Clustal W program version 1.7. Distances between pairs of nucleotide sequences were calculated by using the DNAdist modules (with Kimura's twoparameter method) in the Phylip package, version 3.572 (Felsenstein, Department of Genetics, University of Washington, Seattle, WA). The Phylip (njplot) program was used to construct phylogenetic trees with nucleotide sequences by means of the neighbor-joining method from the same distance matrices. Bootstrap support was determined by 100 resamplings of the sequences.

Comparison of the distribution frequencies was performed with the Pearson's Chi square test. 

On the whole, 53 hCoV strains were identified from 47 patients admitted to hospital as follows: 27 (50.9%) OC43-like strains, 14 (26.4%) 229E-like strains, 9 (17%) NL63-like strains, and 3 (5.7%) untypeable hCoV strains (Fig. 1A) . Six strains (four 229E-like and two OC43-like strains) were recovered twice from the same patients within 2 weeks after initial virus recovery. All hCoV-positive samples were tested by two sets of primers aimed at detecting all hCoV strains (Table I) : (i) panCoV-03 [Poutanen et al., 2003] ; and (ii) panCoV-05 . Of the 53 hCoV-positive samples, 23 (43.4%) were found positive by both assays, while 15 (28.3%) were detected by panCoV-03 only, and 15 (28.3%) by panCoV-05 only (Fig. 1B,C) . All of the nine NL63-like strains were detected by the panCoV-05 primer pair only. In conclusion, 38/53 (71.7%) samples tested by panCoV-03 were positive for hCoV: 21 (39.6%) for OC43-, 14 (26.4%) for 229E-like strains, and three (5.7%) for untypeable strains, while 15 strains (28.3%) were negative. Similarly, of the 53 samples tested by panCoV-05, 38 (71.7%) were positive: 22 (41.5%) for OC43-like, 7 (13.2%) for 229E-like, and 9 (17%) for NL63-like strains, while 15 strains (28.3%) were negative (no untypeable strains with this primer set).

All 53 hCoV strains were typed by sequencing and phylogenetic analysis using the amplification products provided by both or either panCoV. Subsequently, to confirm results, each of the 53 hCoV strains were amplified by using two primer pairs specific for genes 1ab and N of OC43, 229E, and NL63 prototypes, respectively (Table I) . PanCoV typing results were consistently confirmed by type-specific primer sets, with no cross-reactivity among primers.

Primer-specific amplification products of gene 1ab (Fig. 2) and N fragments were used for phylogenetic analysis. OC43-like strains clustered in a group distinct from reference strain OC43 and much more distant from the new hCoV HKU1, except for a single strain (I-PV 02/ 04-2645) distinct from both OC43 and HKU1. Similarly, 229E-strains clustered into a single group distinct from the reference strain 229E and distant from NL63. Finally, NL63-like strains were grouped together substantially, and distinct from the reference strain NL63 and very distant from 229E. A summary of the genetic variability (nt and aa changes) of different hCoV strains (within the limits of the amplified fragment of genes 1ab and N) with respect to reference strains reported in Gene Bank is given in Table II . Weakness of PanCoV-03 amplification of the three untypeable strains did not allow its sequencing. However, negative results following amplification of N gene with 229E-, OC43-, and NL63-specific primers did exclude their similarity with the reference strains.

Following identification of hCoV-positive nasopharyngeal aspirates by RT-PCR, retrospective attempts were made to grow short-term OC43-, 229E-, and NL63like strains in cell cultures from a small number of specimens. One OC43-and one 229E-like strain were inoculated and found to grow in the HuH-7 cell line. Viruses were identified (Fig. 3A-D) and 229E-like strains (10 6 to 10 7 TCID 50 /ml). No crossreactivity of these strains with other human respiratory viruses was detected either by using RT-PCR or monoclonal antibodies. On the other hand, four NL63like-positive samples showed initial growth in secondary African green monkey kidney cell cultures as well as Vero and LLC-MK2 cell lines, as indicated by RT-PCR using type-specific primers (Fig. 3E) . However, further propagation was unsuccessful. These preliminary results suggest the possible recovery and long-term propagation of 229E-like and OC43-like hCoV strains in cell culture, whereas a cell culture system suitable for long-term propagation of NL63-like strains remains to be identified.

Among 823 patients admitted to hospital with an episode of acute respiratory tract infection in the winter-spring seasons 2003-04 and 2004-05, 47 (5.7%) patients were found to be affected by hCoV infections. The list of respiratory viruses associated with acute respiratory infections in 447/823 (54.3%) patients is reported in Table III . The most frequently detected virus was hRSV (26.7%), followed by rhinoviruses (24.2%) and influenza viruses (13.9%), while hMPV and hCoVs were circulating at a comparable rate (8.1 and 7.4%, respectively).

A great variation in the circulation of different hCoVs was observed between the two seasons examined. In fact, 229E-like strains circulated only in 2003-04 (10 infected patients), and NL63-like strains only in 2004-05 (9 infected patients), while OC43-like strains were detected in both seasons, thereby infecting 25 patients (Fig. 4A,B) . The peak circulation of hCoVs was reached in January through March, while a few strains were observed from November to May.

Stratification by age showed that 20/333 (6.0%) patients were affected by hCoV infection in the first year of life, while 8 (4.8%) were affected in the 2-5 years old group, 13 (14.9%) in the 6-20 years old group, and only 6 (2.6%) were > 20 years old (Fig. 5A ). In 33/47 patients (70.2%), respiratory infections were attributed to a single hCoV strain, while in 14/47 patients (29.8%) hCoVs were associated with one or more other respiratory viruses. Analysis of the distribution of single infections and coinfections by age showed that they affected an equal number (n ¼ 10) of patients in the first year of life, while the incidence of coinfections (Table IV) greatly decreased in the other age groups (only four patients in the 2-20 years old groups).

Among the 47 patients affected by respiratory viral infections associated with hCoVs, 35 (74.5%) showed lower (bronchitis, bronchiolitis, pneumonia) and 12 (25.5%) upper respiratory tract (rhinitis, pharyngitis, laryngitis) infections. The predominance of lower versus upper infections was observed in patients of each age group with no significant difference in the incidence among different age groups (Fig. 5B) .

As shown in Figure 5C ,D, the mean percentage of immunocompetent patients with hCoV infections was 4.5% (27/597), whereas the mean percentage of immunocompromised patients infected by hCoVs was 8.8% (20/226) with a statistically significant difference (w 2 test, P ¼ 0.02). The relative risk of contracting hCoV infection in immunocompromised patients was twice that of immunocompetent patients. The total number of hCoV infections observed during the first year of life included 20/333 (6.0%) immunocompetent infants and young children, while the proportion of immunocompetent children with hCoV infections in the older age groups included 5/143 (3.5%) and 2/44 (4.5%) patients in the 2-5 and the 6-20 years old groups, respectively. No hCoV infection (0/77) was observed in immunocompetent patients older than 20 years (Fig. 5C ). Conversely, no immunocompromised patients with hCoV infections, were observed in the first year of life (0/2 patients). In older age groups, hCoV infections of immunocompromised patients increased progressively up to 12.5% (3/25 patients) in the 2-5 years old group, and to 25.6% (11/43 patients) in the 6-20 years old group, including all infected patients older than 20 years (3.8%, 6/156 patients) (Fig. 5D) .

In the same group of 47 patients examined with hCoV respiratory infections, 25 (53.2%) were infected by OC43-like strains (six in association with other viruses), 10 (21.3%) by 229E-like strains (three coinfections), and 9 (19.2%) by NL63-strains (two coinfections). In addition, three patients (6.3%) were infected by untypeable hCoVs (three coinfections). No significant difference in the circulation of different hCoVs among different age groups was observed (Fig. 4C) . The most common symptoms were rhinorrhea (52.2%), fever (47.8%), and cough (47.8%) in OC43-infected patients. Similarly, predominant symptoms were rhinorrhea (70%), fever (50.0%), and cough (50.0%) in 229E-infected patients, while rhinorrhea (57.1%) and cough (42.8%) were most commonly observed in NL63-infected individuals, in association with infrequent episodes of fever (14.2%).

Severe lower respiratory tract syndromes (bronchiolitis, pneumonia) were observed at presentation in 10/15 lower respiratory tract infections associated with hCoV in the first year of life: six (three single infections and three coinfections) were associated with OC43-like strains, two with hCoV untypeable strains (both coinfections), one to 229E-like strain (coinfection), and one to an NL63-like strain (coinfection). Thus, severe hCoV infections in the first year of life appeared to be mostly associated with OC43-like strains. Conversely, none of the five adult immunocompromised adult patients with lower respiratory tract infections showed association with bronchiolitis or pneumonia.

Severe respiratory syndromes were observed in 9/14 (64.3%) coinfected patients (Table IV) , and only in 3/33 patients (9.1%) infected by hCoV alone (P ¼ 0.002). Thus, hCoV coinfections appeared to be associated with a significantly higher number of severe syndromes involving the lower respiratory tract.

After the 1970s, clinical studies on hCoVs were mostly abandoned due to either the presumed low clinical impact of these viruses, or to the great difficulties encountered in their isolation and identification. Renewed interest was triggered recently by the identification of the coronavirus responsible for SARS in 2003. In addition, the development of viral genome amplification procedures has provided a major contribution to the identification of hCoV infections. A major variant of group I hCoVs, with the 229E virus as a prototype, was identified in the NL63 strain detected in the Netherlands [Fouchier et al., 2004; Van der Hoek et al., 2004] , and in the USA [Esper et al., 2005] . More recently, a major variant of group II hCoVs, with the OC43 virus as a prototype, was first reported in Hong-Kong as HKU1 [Woo et al., 2005a] and, then, in France, Australia, and other regions of China [Woo et al., 2005b; Sloots et al., 2006; Vabret et al., 2006] . A wide range of antigenic variability has been reported by researchers recovering a series of viruses in organ cultures of human embryonic trachea [McIntosh et al., 1967; McIntosh, 2005] . However, a great deal of controversy has been raised about the genetic stability [St-Jean et al., 2004] or variability of hCoV OC43-like strains.

The genetic variability of hCoVs is supported by the present study, showing that both sets of PanCoV primers actually did not detect about 30% of circulating strains. This failure of both PanCoV primer sets was confirmed by the use of group-specific primer pairs relevant to two different genes. A similar conclusion was recently reported by another study [Chiu et al., 2005] . As a result, in order to detect the maximal number of circulating hCoVs, specific primers should be used, and results confirmed by a second set of primers, preferably targeting a different gene. In the present study, no HKU1 strain was detected by the two PanCoV primer sets used, while, in the absence of a positive control, HKU1-specific primers were not used.

A complement to the molecular techniques available for detection of hCoVs is represented by the use of the HuH-7 cell line, which has been recently employed for respiratory virus isolation [Vabret et al., 2001; Pene et al., 2003; Freymuth et al., 2005] . This cell line, which was shown to be susceptible to infection by a murine coronavirus, the mouse hepatitis virus [Koetters et al., 1999] , has been found to be helpful in increasing the rate of respiratory virus recovery from DFA-negative samples . In addition, HuH-7 was reported to be permissive for growth of hCoVs, rhinoviruses, and enteroviruses. However, in that study, hCoVs were cultured in HuH7 for 4 days only, and then identified by RT-PCR.

In the present study, it was possible to isolate and extensively propagate two strains of 229E-like and OC43-like strains from respiratory secretions, thus obtaining high virus yields after a few passages. A major morphological characteristic of both 229E-like and OC43-like strains was the presence of syncytial formations in inoculated cell cultures, appearing early (24-48 hr) after inoculation and progressively enlarging until detaching from the cell surface. In this respect, it seems important to recall that SARS-hCoV was reported to produce syncytia in vivo [Ksiazek et al., 2003] . As for NL63-like strains, initial growth was confirmed by the positive RT-PCR results in the supernatant of all three types of cell cultures inoculated with four NL63-positive respiratory samples, that is, two cell lines (LLC-MK2 and Vero) and one secondary culture of African green monkey kidney. However, at subsequent passages, the presence of the viral RNA signal in the medium weakened progressively and finally disappeared. Since inoculated NL63-like-positive samples were freezethawed more than once prior to inoculation, new isolation attempts will be made as soon as fresh positive specimens become available.

The phylogenetic analysis based on sequencing of small fragments from genes 1ab and N showed that each of the three groups of hCoVs detected in our study belong to a single cluster, which is clearly distinct from the relevant reference strain. The only exception regarded an OC43-like strain (I-PV 02/04-2645), which was distinct from both Pavia strains and the OC43 reference strain and was recovered from a patient who had recently arrived from Albany. These data confirm the genetic similarity of the strains circulating in the same geographical area without excluding the simultaneous presence of other unrelated or poorly related strains.

The pathologic impact of hCoV infections was investigated by analyzing the role of these infections in the context of the patient population admitted to hospital due to an acute respiratory infection during two consecutive winter-spring seasons. On the whole, the overall rate of hCoV infections (including both single infections and coinfections) followed that of human respiratory syncytial virus, rhinovirus, and influenza virus infections, and preceded that of human metapneumoviruses, adenoviruses, and human parainfluenza viruses, with comparable rates of coinfections and sequential infections. Although different studies have reported variations, these data correlate grossly with previous surveys [Lina et al., 1996; Freymuth et al., 2005] . In addition, coinfections by different respiratory viruses have been repeatedly reported [El-Sahly et al., 2000; Rovida et al., 2005] .

From an epidemiological standpoint, in the great majority of studies, 229E, OC43, and NL63 van der Hoek et al., 2004 outbreaks have been found to occur in the winter-spring season with recurrences every 2-4 years [McIntosh et al., 1970; Kaye et al., 1971; Monto and Lim, 1974] . In our study, while the peak of hCoV circulation was reached in January through March, the three hCoVs were found to circulate differently during the two consecutive seasons studied. While OC43-like strains were found to circulate in both seasons, 229E-and NL63-like strains were detected in either one of the two seasons, as observed previously by others. On the other hand, all three hCoVs were represented in the four age groups in which the hospitalized patient population was divided, as reported previously [El-Sahly et al., 2000; Vabret et al., 2003] .

The study of the age distribution showed a large predominance of hCoV infections in the first year of life, when 15/20 (75.0%) patients were affected by lower respiratory tract infections and 10 (66.6%) suffered from severe lower respiratory tract syndromes, such as bronchiolitis or pneumonia. It is important to emphasize that all infected infants and young children were immunocompetent subjects. These results strengthen previous observations on the great pathologic potential of hCoVs in very young patients [McIntosh et al., 1974; Sizun et al., 1995; El-Sahly et al., 2000; Vabret et al., 2003; Chiu et al., 2005] . In addition to 229E-and OC43like strains, also the newly identified NL63-like strains were shown to be associated with lower respiratory tract infections, including bronchiolitis and pneumonia [Arden et al., 2005; Bastien et al., 2005; Ebihara et al., 2005; Esper et al., 2005; Vabret et al., 2005] . Therefore, hCoVs, and, as a result of this study, particularly OC43like strains, appear to behave like other respiratory viruses, such as human respiratory syncytial virus, human metapneumovirus, and human parainfluenza viruses, displaying a greater virulence in infancy and early childhood.

On the other hand, adult patients included in this study were all immunocompromised transplanted patients, who had never suffered from severe respiratory syndromes. However, case reports of hCoV-related pneumonia have been described [Pene et al., 2003] , and the increasing number of transplanted patients undergoing immunosuppressive therapy render the investigation of these viruses mandatory in lower respiratory tract secretions from transplant recipients with bronchiolitis and pneumonia.

Finally, the significantly greater association of hCoV coinfections with severe clinical syndromes, that emerged in the present study, shows a rate of severe lower respiratory tract infections greater than 60% in patients with coinfections compared to less than 10% in patients with a single infection. The major pathologic effect of hCoV coinfections was observed again in infants and young children. However, whether the coinfection by hCoV was a factor increasing the severity of the associated viral infection remains hypothetical.

In conclusion, the genetic variability of hCoVs might make the detection of all circulating strains difficult. Cultures might be important for new hCoV detection. The recently renewed interest in hCoVs has allowed the identification of these viruses as major pathogens of infancy and early childhood, while their pathologic role in immunocompromised patients remains to be defined. Coinfection by hCoVs and other respiratory viruses is often detected in severe lower respiratory tract infections of infants and young children.

Cloning of a human parvovirus by molecular screening of respiratory tract samples

New human coronavirus, HCoV-NL63, associated with severe lower respiratory tract disease in Australia

Human coronavirus NL63 infection in Canada

Coronavirus antibody titres in sera of healthy adults and experimentally infected volunteers

Association of coronavirus with neonatal necrotizing enterocolitis

Human coronavirus NL63 infection and other coronavirus infections in children hospitalised with acute respiratory disease in Hong Kong

Detection of human coronavirus NL63 in young children with bronchiolitis

Spectrum of clinical illness in hospitalised patients with ''common cold'' virus infection

Evidence of a novel human coronavirus that is associated with respiratory tract disease in infants and young children

Rhinovirus and coronavirus infection-associated hospitalizations among older adults

Coronavirus pneumonia following autologous bone marrow transplantation for breast cancer

A previously undescribed coronavirus associated with respiratory disease in humans

Replication of respiratory viruses, particularly influenza virus, rhinovirus, and coronavirus in HuH7 hepatocarcinoma cell line

Determination of coronavirus 229E antibody by an immune-adherence hemagglutination method

Human coronavirus OC43 serum inhibitor and neutralizing antibody by a new plaque-reduction assay

Antigenic and biological relationships between human coronavirus OC43 and neonatal calf diarrhoea coronavirus

Human enteric coronaviruses: Antigenic relatedness to human coronavirus OC43 and possible etiologic role in viral gastroenteritis

Simultaneous detection and typing of human metapneumovirus strains in nasopharyngeal secretions and cell cultures by monoclonal antibodies

Virologic studies of acute respiratory disease in young adults. V. coronavirus 229E infections during six years of surveillance

A new virus isolated from the human respiratory tract

Mink lung cells and mixed mink lung and A549 cells for rapid detection of influenza virus and other respiratory viruses

Seroepidemiologic survey of coronavirus (strain OC43) related infections in a children's population

Detection of coronavirus 229E antibody by indirect hemagglutination

Calf diarrhea coronavirus

Mouse hepatitis virus strain JMH infects a human hepatocellular carcinoma cell line

Newly discovered coronavirus as the primary cause of severe acute respiratory syndrome

Coronaviridae: The viruses and their replication

Surveillance of community-acquired viral infections due to respiratory viruses in Rhone-Alpes (France) during winter 1994 to 1995

Occurrence and frequency of coronavirus infections in humans as determined by enzyme-linked imunosorbent assay

Coronaviruses in the limelight

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

Seroepidemiologic studies of coronavirus infection in adults and children

Coronavirus infection in acute lower respiratory tract disease of infants

Diagnosis of human coronavirus infections by immunofluorescence: Method and application to respiratory disease in hospitalized children

A novel pancoronavirus RT-PCR assay: Frequent detection of human coronavirus NL63 in children hospitalized with respiratory tract infections in Belgium

The Tecumseh study of respiratory illness. VI. Frequency of and relationship between outbreaks of coronavirus infection

Growth of human hepatoma cell lines with differentiated functions in chemically defined medium

Coronavirus as a possible cause of severe acute respiratory syndrome

Coronavirus 229E-related pneumonia in immunocompromised patients

Rapid detection of human metapneumovirus strains in nasopharyngeal aspirates and shell vial cultures by monoclonal antibodies

Identification of severe acute respiratory syndrome in Canada

Monoclonal antibodies versus reverse transcription-PCR for detection of respiratory viruses in a patient population with respiratory tract infections admitted to hospital

Detection and pathogenicity of human metapneumovirus respiratory infections in pediatric Italian patients during a winter-spring season

Neonatal nosocomial respiratory infection with coronavirus: A prospective study in a neonatal intensive care unit

Comparison of immunofluorescence with monoclonal antibodies and RT-PCR for the detection of human coronaviruses 229E and OC43 in cell culture

Evidence of human coronavirus HKU1 and human bocavirus in Australian children

Human respiratory coronavirus OC43: Genetic stability and neuroinvasion

Early detection of acute rhinovirus infections by a rapid reverse transcription-PCR assay

Cultivation of a novel type of commoncold virus in organ cultures

Direct diagnosis of human respiratory coronaviruses 229E and OC43 by the polymerase chain reaction

An outbreak of coronavirus OC43 respiratory infection in Normandy

Human coronavirus NL63

Detection of the new human coronavirus HKU1: A report of 6 cases

Detection of human coronavirus 229E in nasal specimens in large scale studies using an RT-PCR hybridisation assay

Identification of a new human coronavirus

Circulation of genetically distinct contemporary human coronavirus OC43 strains

Characterization and complete genome sequence of a novel coronavirus, coronavirus HKU1, from patients with pneumonia

Clinical and molecular epidemiological features of coronavirus HKU1-associated community-acquired pneumonia

We thank Linda D'Arrigo for revision of the English. We are indebted to the technical staff of the Servizio di Virologia. This research was partially supported by Ministero della Salute, Ricerca Finalizzata 2004 (grants 89282 and 89288).