INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Enteroviruses are the most common etiological agents of human viral myocarditis and are associated with some cases of dilated cardiomyopathy (DCM) (5, 26, 38). DCM alone afflicts approximately 5 to 8 persons per 100,000 per year worldwide (25, 36). Results from recent studies have also demonstrated a possible etiological role of enterovirus infection in a particular form of heart muscle disease, selenium deficiency-related endemic cardiomyopathy (Keshan disease [KD]), seen in China (23, 24, 45). Enteroviruses also cause outbreaks in some subpopulations, such as young children within maternity units or nurseries (7). Such outbreaks, occurring in selenium-deficient areas, may affect the general population and are more severe, with high mortality. The incidence of outbreaks of enteroviral myocarditis in selenium-deficient regions of southwestern China has increased in the last decade (15 and our unpublished data). During an outbreak in 1991, 67 cases in patients 2 to 75 years old were reported, with 16 fatalities within 24 h of onset (15).

KD is endemic exclusively in selenium-deficient rural areas of China, from the northeast to the southwest, including 14 provinces and autonomous regions (13). Its clinical features are low body selenium content and acute or chronic episodes of heart disorder characterized by cardiogenic shock, arrythemia, and/or congestive heart failure, with an enlarged heart. Four types of the disease are seen: acute, subacute, chronic, and latent or compensated. Pathologically, it is characterized by multifocal myocardial necrosis and fibrous replacement throughout the myocardium, with various degrees of cellular infiltration and calcification, depending upon the type of disease. In some cases, the pathologic changes are similar to those of myocarditis or DCM (13, 22). Selenium deficiency has been considered a major cause since KD was first reported in 1935. Selenium supplementation has led to a decrease in the incidence of KD in areas of endemicity (11) but has not eliminated KD. In addition, a seasonal and annual fluctuation of KD incidence is seen. In the Chuxiong region of southwestern China, for instance, the prevalence of endemic KD is higher in summer and higher in some years than in others (13). This fact suggests an infectious etiology of KD in southwestern China.

Various enteroviruses have been isolated from patients with KD or during outbreaks of myocarditis in the selenium-deficient Chuxiong region of southwestern China. Most were coxsackievirus group B (CVB) strains, but some of them could not be identified antigenically using traditional serotyping methods (8, 15, and our unpublished data). Selenium deficiency increases the cardiovirulence of CVB3 in animal models by changing the viral genomic sequence (4). However, little is known about the biological properties of enteroviruses prevalent in selenium-deficient areas or areas in which KD is endemic, partly because of the inability to identify some isolates using conventional serotyping methods.

Sixty-six human enterovirus serotypes have been distinguished on the basis of a neutralization assay (28). Although this assay is generally reliable and is used routinely, it is labor-intensive and time-consuming and may fail to identify a clinical isolate because of antigenic variation, recombination, or the presence of multiple serotypes. Nucleotide sequencing of the 5' nontranslated region (5'NTR) and the capsid protein VP4-VP2 junction has been used as a diagnostic and epidemiologic tool for some enteroviruses. However, the sequence of these regions does not correspond to the serotype (2, 32), which is determined mainly by sequences encoding epitopes on viral capsid protein VP1. Recent studies have found that sequences coding for VP1, particularly the 3' half, correlate with results of the neutralization assay. This information has been confirmed for prototype strains and for clinical isolates of various serotypes (30, 31) and is likely to be useful for isolates that are difficult or impossible to type using standard immunological reagents.

In the present study, six enteroviruses isolated from patients with subacute KD or during outbreaks of viral myocarditis in selenium-deficient areas were characterized by nucleotide sequencing of the 5'NTR and the VP1 coding region and determination of their pathogenicity in adult and suckling mice.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Enteroviruses. Of six virus isolates in this study, two were isolated from cardiac blood obtained at autopsy from patients with subacute KD in 1985 and 1987 in the Chuxiong region of Yunnan Province, southwestern China. Selenium deficiency is evident there in the environment, food chain, and inhabitants. The remaining four isolates were obtained from blood or stool of four patients during outbreaks of acute myocarditis in the same region in the summers of 1991 and 1992 (Table 1), initially by use of primary cultures of monkey kidney cells. All these patients had a history of upper respiratory tract infection and a clinical diagnosis of myocarditis. Patients 3 and 4 had a fourfold increase in neutralizing antibody titer against CVB6 (from 1:80 to 1:320), while patients 5 and 6 had a twofold increase in neutralizing antibody titer against CVB4 and CVB6. These isolates were serotyped in two separate institutions, a virology laboratory in China and the Public Health Laboratory Service in the United Kingdom, using a standard neutralization assay described by Melnick et al. (28, 29). A cardiovirulent laboratory strain of CVB3 was used as a control (40).

Cell culture, plaque assay, and single-step growth. To adapt these isolates to established cell lines, confluent Vero cell (European Collection of Cell Cultures [ECACC] no. 88020401) or HEp-2 cell (ECACC no. 86030501) monolayers were inoculated with 0.1 PFU of virus per cell. After absorption at room temperature for 1 h, the cultures were washed twice in phosphate-buffered saline (PBS), maintained in Dulbecco's modified Eagle medium (DMEM) with 2% fetal bovine serum, and incubated until there was greater than a 50% cytopathic effect. After three cycles of freezing-thawing from 70°C to room temperature, cell debris was removed by centrifugation and the supernatants were used as virus stock for the following experiments.

RNA extraction, RT, and PCR. Total RNA was extracted from infected cells (25-cm² flask) using Tri-reagent (Sigma) and dissolved in 50 µl of diethyl pyrocarbonate-treated H₂O. The enterovirus-specific primers used for reverse transcription (RT)-PCR and sequencing of the 5'NTR and the VP1 coding region are shown in Table 2. RT was carried out as recommended by the manufacturer (Superscript II RNase H-free reverse transcriptase; GIBCO-BRL) and as described previously (1, 42). PCR was carried out with a 50-µl mixture containing 1 mM deoxynucleoside triphosphates, 0.5 mM forward primer (various), 0.5 mM reverse primer (various), 10 µl of reaction buffer (Promega), 1.0 to 2.5 mM MgCl₂, 2.5 U of Taq (Promega), and 5 µl of the RT reaction product. The PCR program was 1 cycle of 2 min at 95°C, 2 min at N°C, and 3 min at 72°C, followed by 30 cycles of 1 min at 95°C, 1 min at N°C, and 2 min at 72°C, where N is 5°C below the annealing temperature of the primer with the lower melting temperature. A negative control containing 45 µl of the reaction mixture plus 5 µl of UV-treated H₂O and a positive control containing 45 µl of the reaction mixture plus 5 µl of the RT reaction product from cells infected with the virulent laboratory strain of CVB3 were included in each set of PCRs.

Cycle sequencing and sequence analysis. The cycle sequencing reaction was carried out using 4 µl of DNA polymerase FS-terminator mix (PE-ABI), 3.2 mM primer, and 5 ng of PCR product per 100-bp length of template DNA, with the volume made up to 10 µl with H₂O. Reactions were run for 25 cycles with the following program: 96°C for 30 s, 50°C for 15 s, and 60°C for 4 min. DNA was precipitated, dissolved in loading buffer (16.7% formamide, 73.3% dextran), and separated in a 4.1% denaturing polyacrylamide gel on an ABI model 377 DNA sequencer (1, 42). Sequence alignment and analysis were carried out as previously described (24, 42).

Animal experiments. Five-week-old male MF-1 and SWR mice were obtained from HARLEN OLAC Ltd. or Charles River and housed in negative-pressure isolators. The animals were inoculated intraperitoneally with 10⁶ PFU of virus and sacrificed by cervical dislocation at 7 days postinfection. The heart and pancreas were removed and fixed in 10% formalin in PBS for histologic study. One- to 2-day-old MF-1 suckling mice were inoculated subcutaneously with 10³ PFU of virus and inspected daily for 10 days, and survival or death was recorded. Mice were inoculated similarly with a cardiovirulent laboratory strain of CVB3 or DMEM as positive and negative controls, respectively, for both adult and suckling mice.

Histology. Paraffin-embedded sections were stained with hematoxylin and eosin by standard procedures. Two slides with two or three sections each, cut 40 µm apart, were prepared from each sample, and histopathologic changes were examined under a light microscope by two investigators independently.

Nucleotide sequence accession numbers. The sequences reported here have been deposited in the GenBank sequence database under accession numbers AF225467 to AF225473.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Serotype and growth properties of enterovirus isolates. The serotypes of the isolates were determined by the standard neutralization assay as either CVB2 or CVB6 for four of six strains (Table 1). The remaining two isolates were identified in one laboratory as CVB3 but were reported as CVB4 by another laboratory; these two isolates are therefore regarded as nontypeable by the neutralization assay.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Single-step growth curves for enterovirus isolates in HEp-2 cells. Single-step growth of all six isolates (isolate 1 [], isolate 2 [], isolate 3 [], isolate 4 [], isolate 5 [], and isolate 6 []) was measured under the infection and culture conditions described in Materials and Methods. The curves were plotted based on duplicate results for each isolate.

Sequence analysis and molecular typing. Genomic sequences of the 5'NTR (700 bp) and the 3' half of or the complete VP1 coding region of all six isolates were determined in both orientations without ambiguity. Isolates 1 and 2 had identical sequences in the 5'NTR (700 of 700 bp) and more than 99% sequence identity in VP1 (359 of 360 bp). Isolates 3, 4, 5, and 6 also had more than 99% sequence identity in the VP1 region, but the 5'NTR sequences of isolates 3 and 4 were different from those of isolates 5 and 6 by 17.8%. The different sequences of the 5'NTR or VP1 among the six isolates indicated no cross contamination during virus propagation or RT-PCR and cycle sequencing procedures.

Pathogenicity of isolates in mice. Animal experiments confirmed that the virulent laboratory strain of CVB3 caused typical myocarditis and pancreatitis in adult MF-1 mice, similar to that seen in the characterized SWR mouse model (41). This result validates MF-1 mice as an alternative model for viral myocarditis which could be used to test the pathogenicity of enteroviruses. There were no obvious pathologic changes in the hearts of MF-1 mice 7 days after inoculation with any of the isolates (10⁶ PFU/mouse). Isolate 2 was also inoculated into SWR mice, with similar results. However, pancreatitis was evident in all mice inoculated with isolate 3, 4, 5, or 6, and enterovirus RNA was detected in myocardium by in situ hybridization (data not shown). To further characterize pathogenicity, groups of 10 suckling mice were inoculated with 10³ PFU of particular viruses. Isolate 3, 4, 5, or 6 caused death within 6 days after inoculation, comparable to that seen with the virulent laboratory strain of CVB3. No mortality was observed for mice inoculated with isolate 1 or 2 or in negative controls mock infected with culture medium (Fig. 2).

View larger version (21K): [in this window] [in a new window]   FIG. 2.   Survival curves for suckling mice after infection with enterovirus isolates. Groups of 10 suckling mice were inoculated with isolate 1 (), 5 (), or 6 () or mock infected with DMEM as negative controls (); a group of 11 mice was given isolate 2 () or 4 () or the cardiovirulent laboratory strain of CVB3 as positive controls (); and a group of 8 mice was given isolate 3 (). For details, see Materials and Methods.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

KD has been studied for over 60 years within China. One of the key issues has been the complex etiology of this endemic disease, with a focus on environmental-nutritional factors and infectious agents (6, 8, 11, 13). Selenium deficiency in the food chain has been recognized as a major but not exclusive environmental-nutritional factor, and increasing evidence supports an etiological role of enteroviruses in KD (6, 23, 24, 34, 44, 45). In addition, annual summer outbreaks of fulminant enteroviral myocarditis occurring in similar selenium-deficient areas as KD appear to be more severe, with a higher mortality (15 and our unpublished data). It is important to identify the prevalent virus types and strains in these areas and to investigate their biological properties. Rapid and correct serotyping will make it possible to use hyperimmune globulin or type-specific immune globulin for treatment or passive immunization before and during an outbreak (27, 39). Serotyping will also provide information on the spectrum of viruses involved and will be useful in designing specific vaccines (9, 14, 33, 43).

Various enteroviruses have been isolated from patients with KD or during outbreaks of myocarditis in the Chuxiong region. They have been serotyped by a neutralization assay, but some have failed to be typed by this approach. A recent study has shown that certain sequences within the region coding for capsid protein VP1 correlate with serotype. It is known that VP1 contains major neutralizing antigenic sites, the basis of serotyping by the neutralization assay (29), and an enterovirus can be typed genetically by comparison of partial VP1 sequences to a database of sequences of enterovirus prototype strains or other typed strains (31). These sequences are determined by PCR amplification of part of or the entire VP1 gene with generic RT-PCR primers, designed to react with all known human enterovirus serotypes, followed by nucleotide sequencing of the products. This molecular typing approach has been used to discriminate between prototype strains of all human enterovirus serotypes (30, 31), to identify isolates from clinical specimens, or and to identify isolates refractory to antigenic typing and so to identify potential new viruses.

In this study, we determined the sequences of the 5'NTR and the 3' half of the VP1 coding region of six enterovirus isolates from cases of KD or outbreak myocarditis. Not surprisingly, isolates 1 and 2, from subacute KD cases occurring in the same geographic area, show high nucleotide sequence similarity in both the 5'NTR and the VP1 region (>99%). Similarly, isolates 3 and 4 or isolates 5 and 6, from myocarditis cases occurring in the same geographic area in 1991 or 1992, respectively, have more than 99% nucleotide sequence identity. A partial VP1 sequence alignment correlates with serotype for isolates 1, 2, 3, and 4 and identifies isolates 5 and 6, nontypeable by a conventional neutralization test, as CVB6. However, the 5'NTR sequences of isolates 3 and 4 are dissimilar from those of isolates 5 and 6, indicating that CVB6 strains recovered during outbreaks in 1991 and 1992 are not the same.

CVB6 infection is uncommon in humans. Lau (21) analyzed the level of CVB-specific antibodies in 1,020 normal subjects 1 to 60 years old and from all 18 health districts of New Zealand. CVB6 was the least prevalent, and 970 (95.1%) of the subjects had no antibody to this serotype, compared with 30.4 to 71.3% with antibodies against CVB1 to CVB5. Similar results were obtained from a 5-year clinical and epidemiologic study in France (35), but some studies have suggested that CVB6 is associated with human disease, such as pancreatitis (19) and respiratory tract infection (12), and CVB6 has been isolated from aborted fetal heart, brain, liver, kidney, and spleen tissue (3). Few studies have reported CVB6-induced heart muscle disease, and so it is intriguing to find CVB6 in outbreaks of myocarditis in selenium-deficient areas.

Infection with the same CVB3 strain can cause severe myocarditis and pancreatitis in some strains of mice but can cause only pancreatitis or is avirulent in others because of their different genetic backgrounds (16). SWR mice have been successfully used as a model of CVB3 myocarditis (41, 43), but little is known of their susceptibility to other CVB strains. MF-1 mice have not been used previously as a model for CVB infection. Our study shows the typical pathologic changes of myocarditis and pancreatitis in MF-1 mice after infection with the laboratory strain of CVB3, similar to those seen in SWR mice, suggesting that MF-1 mice could be an alternative model for CVB3-induced myocarditis.

We assessed the pathogenicity of the six isolates in MF-1 mice. Isolate 1 or 2 (CVB2) did not cause pathologic changes in the heart or pancreatitis in adult MF-1 or SWR mice (isolate 2 only) or death in suckling mice (MF-1). A similar result was obtained with a mouse model and a prototype CVB2 strain from the Virus Reference Laboratory, Public Health Laboratory Service, Colindale, London, England, by other authors (20). It is possible that our isolates are not virulent in these mouse strains, although the presence of enterovirus RNA in mouse myocardial tissue suggests virus replication in the heart (17, 37). The remaining four isolates (serotype CVB6) caused pancreatitis in adult MF-1 mice, despite no apparent myocarditis, and caused death in suckling mice, similar to the laboratory strain of CVB3. These results indicate differences in pathogenicity among these isolates and imply that virulence in an experimental mouse model may not reflect pathogenicity in humans (10, 18). Investigation of more isolates from selenium-deficient areas and a search for virulence determinants in viral genomic RNA sequences (4) will provide further information on pathogenicity.