Tests for coxsackie B virus-specific 1gM

J. Hyg., Camb. (1983), 90, 327-332 327
Printed in Great Britain
EDITORIAL
Tests for coxsackie B virus-specific 1gM
During the last 10-15 years there has been a relative decline in virus isolation
procedures and an increasing emphasis on both direct antigen detection by
immunological methods and serological tests for the diagnosis of recent virus
infections. This relative decline has been due to a number of factors. There have
been increasing difficulties in maintaining a suitably wide range of cell culture
systems for routine diagnostic purposes. There has been increasing dissatisfaction
with the time taken to isolate and identify a virus in cell culture and, finally, cell
culture systems for the isolation of many of the more recently recognized human
pathogens have not yet been developed.
The change in emphasis to serology is even present for the diagnosis of infection
due to enteroviruses, a group for which virus isolation has been the cornerstone
of diagnosis for many years. Virus isolation has proved satisfactory in many
instances for the diagnosis of acute enteroviral infection including meningitis and
severe neonatal disease. Serological tests, notably neutralizing antibody studies on
the isolated virus, had their place. They could be used to add weight to the
proposed aetiological role of a virus, particularly when the virus was isolated from
a system such as the gastro-intestinal tract which was not involved in the main
clinical symptoms and signs of the patient. However, there were other syndromes
of disease (e.g. myocarditis and pericarditis) which came to be regarded as
relatively late manifestations of enteroviral infection. Thus patients would present
for investigation at a time when virus was no longer being shed and therefore diagnosis by virus isolation was impossible. Furthermore, diagnosis by demonstrating
rising titres was not possible since the first specimen of serum would be taken
at a time when antibody levels were already elevated. Thus a virus-disease
association was based upon finding elevated amounts of neutralizing antibody or
amounts which declined over the succeeding months. This was plainly unsatisfactory, so attention was turned to the possibility of detecting enterovirus-specific
IgM in order to diagnose recent infection. Work has centred round the group B
coxsackie viruses (CBVs) since there are only a limited number of serotypes and,
with the decline of poliovirus infection in developed countries, it is these
enteroviruses which have been associated with more serious diseases of man, e.g.
myocarditis and pericarditis. There also remains the interesting question of the
role of CBVs in juvenile-onset diabetes. In this issue of the Journal there is a further
contribution to the field (Morgan-Capner & MeSorley, 1983) which again illustrates
the difficulties of devising tests for specific 1gM for individual coxsackie B viruses
which are sensitive, specific and suitable for widespread use.
The first test for CBV-specific IgM to be extensively evaluated was based on an
immunodiffusion technique (Schmidt & Lennette, 1962; Schmidt, Magoffin &
Lennette, 1973). Using high-titre antigen preparations containing a mixture of
12 HYG 90
328EDTRA
intact virions and empty capsids, two distinct lines of precipitation were obtained,
one being IgM antibody precipitating intact virions and the other IgG antibody
reacting with empty capsids. Initial studies (Schmidt & Lennette, 1962), showed
that this technique could be used for the detection of specific IgM to all CBVs except
CBV-2. Specific IgM was present in the sera of 77 of 79 patients from whom CBV-1,
3, 4, 5 or 6 had been isolated. It was noted that not all specific IgM antibody was
homotypic for the virus isolated but that heterotypic reactions occurred in 22 %
of patients. There was no strict control group in this study but CBV-specific IgM
was found in 25 % of 86 patients with CNS disease from whom CBVs had not been
isolated. Subsequently, the test was used to detect the presence of CBV-specific
IgM in a group of 259 patients with pericarditis (148 cases), myocarditis (92 cases)
and pleurodynia (19 cases) and a group of 259 control patients with a diagnosis
of viral or mycoplasma pneumonia (Schmidt et al. 1973). Virus-specific lgM
antibody was detected in 27 % of patients with pericarditis, 25 % of patients with
myocarditis and 37 % of those with pleurodynia. Virus-specific IgM was also found
in 21 (8 %) of the 259 controls, but on closer scrutiny of the clinical records it was
noted that six of these had evidence of cardiac or central nervous system disease.
Specific IgM antibody was not detected in any specimen from a patient in the test
group that was taken more than 42 days after the onset of illness. Thus the test
appeared to have a sensitivity and specificity that would make it useful for the
diagnosis of recent coxsackie B virus infection. Again the authors noted that some
sera contained IgM antibody that reacted with more than one CBV. Of the 70
IgM-positive sera in the test group 56 (80 %) gave a positive result with only one
CBV; 11 (16 %) gave a positive result with two virus types; one serum gave a
positive result with three, another with four and another with five types. Even
allowing for this cross-reactivity, the requirement to produce high-titre antigens
and the fact that a different method had to be used for IgM antibody to coxsackie
B2 virus, it is surprising that the immunodiffusion test for coxsackie B virus
antibody was not more widely used.
Subsequently Minor et al. (1979) utilized the same basic principle of two distinct
precipitin lines in a counter-immunoelectrophoresis test with a view to devising
a more sensitive and more rapid technique. Again concentrated antigen preparations were required and it was found necessary to use recent coxsackie B virus
isolates rather than prototype strains which gave inconsistent results. The test
group consisted of 29 patients from whom CBVs had been isolated and who
exhibited an > 4-fold rise in neutralizing antibody titre to the isolated virus.
Twenty-two of these patients were infected with CBV 1, 3, 4 or 5, and 20 (91 %)
of these were positive for IgM antibody; 19 had IgM antibody to the infecting
serotype and one patient infected with coxsackie B5 virus had heterologous IgM
antibody only. Again problems of detecting IgM antibody to coxsackie B2 virus
were demonstrated; only three of seven patients infected with this serotype were
positive and one of the three exhibited a heterologous reaction only. Overall of the
23 patients with detectable IgM antibody 14 (61 %) were positive for IgM antibody
to a CBV other than the one with which they had recently been infected. The
control group investigated by Minor and colleagues consisted of 22 patients from
whom coxsackie A or echoviruses had been isolated and 50 patients from whom
a non-enterovirus had been isolated. Two (9 %) of the former (both echovirus type
328 EDITORIAL
Coxsackie B virus-specific IgM
6 infections) and one (2 %o) of the latter (an influenza A infection) had IgM antibody
to a single CBV serotype.
The two studies described above suggested that detection of CBV-specific IgM
was a useful diagnostic test. Over 90 % of patients with a recent CBV infection had
detectable CBV-specific IgM, although the occurrence of heterologous reactions
made it difficult to predict accurately the infecting serotype. By contrast, less than
10 % of sera from patients with no evidence of recent infection with CBV had
detectable specific IgM antibody.
The detection of coxsackie B virus IgM next enters the era of solid-phase
immunoassays. El-Hagrassy, Banatvala & Coltart (1980) described such an assay
based on the antibody-capture principle with the enzyme-labelled final indicator
antibody. The antigen used in the assay was a pool of coxsackie B 1-5 antigens
and each sample was also tested against a control antigen prepared from uninfected
vero cell cultures. The reading for each sample was taken as the difference between
the readings obtained with the test and control antigens. Three positive and three
negative control sera were included in each test and any test sample with an
absorbance of more than twice the mean negative control value was considered
positive. Sera from 12 patients in whom a recent CBV infection had been confirmed
by virus isolation and/or an > 4-fold rise in neutralizing antibody titre were tested
and 10 (83 %) of these patients had detectable CBV-specific IgM antibody. A
further 10 sera shown to contain CBV-specific IgM by neutralization tests on IgM
fractions obtained by sucrose density-gradient centrifugation were all positive by
the ELISA test. As a control group 46 patients with recent infections other than
CBV (M. pneumoniae 12, influenza A4, influenza B8, Chlamydia psittaci 5,
adenovirus 4, parainfluenza viruses 3, cytomegalovirus 3, poliovirus 2, herpes
simplex virus 3 and varicella-zoster virus 2) were tested. None of them gave a
reaction in the ELISA test for CBV-specific IgM. In addition to the above results
on positive and negative control groups the authors found positive results in
approximately 37 % of patients with cardiologically confirmed acute myocarditis
(64 patients) and acute pericarditis (38 patients) and in patients with a clinical
diagnosis of Bornholm disease (57) or aseptic meningitis (8). Lower rates of
positivity were found in acute ischaemic heart disease (13 % of 30 patients),
congestive cardiomyopathy (7 % of 28 patients) and chronic heart valve disease
(none of 39 patients). Although only four patients with juvenile-onset diabetes were
tested, two were positive. The authors elect to use a polyvalent antigen in order
to diagnose recent CBV infection rather than infection with a specific serotype.
Bearing this in mind their results are very much in accord with those of Schmidt,
Magoffin & Lennette (1973) and Minor et al. (1979) with respect to patients with
confirmed CBV infection and those with myocarditis, pericarditis and pleurodynia.
The test seems highly specific in that no positive result was found with sera from
patients with infections other than those due to CBV, although it may be
significant that only two of the 46 were due to enteroviruses (both poliovirus
infections). CBV-specific IgM was found to persist for 6-8 weeks after the onset
of symptoms in those 10 patients from whom sequential samples were available.
Nevertheless, the sensitivity appears only moderate since the screening dilution
was 1 in 100 and 70 % of positive sera had a titre of 400 or less.
An indirect solid-phase immunoassay has been described by Dorries & ter Meulen
12-2
329
(1980, 1983). Viral antigen is adsorbed to the wells of microtitre plates and reacted
with test serum dilutions. Binding of specific IgM is then determined by reaction
with a labelled anti-human IgM. In their initial description of the detection of
enterovirus-specific IgM using this technique D6rries & ter Meulen (1980) used a
radiolabelled indicator antibody but changed to an enzyme label when investigating
the specificity of the IgM antibodies in acute coxsackie B virus infections (Dorries
& ter Meulen, 1983). In the latter test dilutions of serum are reacted with viral
and control antigen and a net absorbance value calculated. A positive result is
indicated by a net absorbance value of > 0-25, this being the mean plus three
standard deviations (SDs) of a 1 in 100 dilution of 40 sera from healthy newborn
children. To investigate the specificity of the IgM response Dorries & ter Meulen
(1983) tested sera from patients with meningoencephalitis in whom a recent CBV
infection had been confirmed by virus isolation and a significant change in
neutralizing antibody titre. By testing these sera for IgM antibody to each of the
CBVs 1-5 the authors were able to define three patterns of response, homotypic,
type-predominant and heterotypic. In the homotypic pattern sera gave a positive
reaction with only one of the five viral antigens and this was of the same serotype
as the virus isolated from the patient. In the type-predominant pattern there was
a strongly positive reaction with antigen equivalent to the isolated virus with
weakly positive reactions to the other antigens. In the heterotypic pattern
moderate to strongly positive reactions were found with all five antigens and the
most strongly positive reaction was not necessarily with the antigen equivalent
to the virus isolated from the patient. The authors further investigated the
problem ofhomotypic and heterotypic reactions by analysing the virus polypeptides
present in their antigen preparations. The antigens were found to consist of intact
infectious virus and incomplete provirions on the basis of the presence of large
amounts of VP0, VP 1 and VP3 and small amounts of VP2 and VP4 after sucrose
density-gradient centrifugation and SDS-polyacrylamide gel electrophoresis.
Reaction of these separated polypeptides with sera exhibiting a heterotypic or a
homotypic response indicated that both the type- and group-specific antigenic
determinants were located on VP 1. These results are very much in agreement with
those of Katze & Crowell (1980a, b), who investigated the homologous and
heterologous antigenic reactivity of CBV using IgG antibody in an ELISA-based
immunosorbent assay. They found that a type-specific antigen/antibody reaction
was dependent upon the use of native, purified virions. Disruption by urea or even
absorption of virus on microtitre wells revealed antigen determinants with group
reactivity. This latter phenomenon necessitates the use of a sandwich immunoassay
to ensure type specificity of the reaction. Interestingly this group reactivity was
confined within the CBVs and did not extend to coxsackie A2, poliovirus type 2,
echovirus type 6 or human rhinovirus 2. By contrast, the heterotypic response
described by Dorries & ter Meulen (1983) was demonstrated with the VP 1 of all
five CBVs and the serum from a patient with an acute cosackie A-9 infection. This
together with the two CBV IgM-reactive sera from cases of echovirus 6 infection
described by Minor et al. (1979) are the only previously documented cross-reactions
in CBV IgM tests involving infections among the enterovirus genus as a whole.
However, in this issue Morgan-Capner & McSorley (1983) produce more extensive
evidence for this.
330 EDITORIAL
Coxsackie B virus-8pecific IgM 331
The basis of their test is again the antibody-capture principle with, in this
instance, a radiolabelled final indicator antibody. The authors (Morgan-Capner &
MeSorley, 1983) describe assays for CBV 4 and CBV 5 1gM antibody and set strict
criteria for a positive reaction. Only sera giving a reaction greater than the mean
plus three SDs of 100 adult sera are considered positive. This value is equivalent
to a T N ratio of 2-5 for the CBV 4 assay and 2-9 for the CBV 5 assay. In spite
of this high threshold value all sera taken during the early convalescent period from
patients from whom CBV 4 (4 cases) or CBV 5 (6 cases) was isolated were positive
in the homotypic assay. As expected, heterotypic responses were also found in early
convalescent sera from cases of CBV infection. Sera from 10 of 20 cases of CBV
1, 2, 3, 5 or 6 infection reacted in the CBV 4 1gM assay and sera from seven (39 %)
of 18 CBV 1, 2, 3, 4 and 6 infections reacted in the CBV 5 assay. Moreover, early
convalescent sera from 9 (69 00) of 13 cases from whom coxsackie A or echoviruses
had been isolated gave a positive reaction in both the CBV 4 and CBV 5 IgM assays.
However, positive reactions were not found with sera from 41 cases of infection due
to viruses other than enteroviruses or five cases of Mycopla8ma pneumoniae
infection.
Thus it is clear that further work is required to develop IgM antibody tests which
are specific for a single CBV or even the group as a whole. At present there is general
agreement that positive reactions in CBV IgM tests do not occur as a consequence
of virus infections other than those due to enteroviruses. Of the combined total
of 142 sera from such infections examined by Minor et al. (1979), El-Hagrassy et
al. (1980) and Morgan-Capner & McSorley (1983), only one (from a case of influenza
A infection, Minor et al. 1979) was found to be positive. Sera from six other cases
of influenza A were tested by the other workers and none was positive. The more
likely explanation of the case of Minor et al. (1979) is that the individual had a
CBV infection shortly before the confirmed influenza A infection rather than that
CBV-reactive IgM is produced as a consequence of some influenza A infections.
Conclusions about the occurrence of IgM antibody which reacts with CBVs after
infection with other enteroviruses are more difficult to reach on the basis of present
data. The experiments of D6rries & ter Meulen (1983) show that IgM antibody
which reacts with the VP 1 of CBV 1-5 may be present in sera from patients who
have had a recent confirmed coxsackie A infection. The data of Morgan-Capner
& MeSorley (1983) suggests that such antibody is found in nearly 70 % (9 of 13)
of patients with recent coxsackie A or echovirus infection. By contrast Minor et
al. (1979) find the frequency to by 9 % (2 of 22) in a similar group.
There are two possible explanations for this difference. In both the immunodiffusion and counterimmunoelectrophoresis tests two distinct lines of precipitate
are formed, implying a degree of antigen separation during passage through the
gels at least with respect to the point at which optimum concentrations occur for
precipitation with IgM or IgG antibody. IgM antibody reacts with intact virus
particles, and group-specific antigens are less likely to be present on such intact
particles than incomplete or disrupted particles. Thus a greater specificity of IgM
antibody reaction may be anticipated in precipitation-in-gel tests. However,
heterotypic reactions within the CBV group are found with these tests, and this
implies that group-reactive antigens are exposed. Moreover the work of Katze &
Crowell (1980b) on the reactivity of IgG antibody with enterovirus antigens
indicates that the group reactivity of urea-disrupted viruses does not extend to
coxsackie A, echo- or rhinoviruses. It appears therefore that IgM produced after
enterovirus infection in man has a broader cross-reactivity than the group-specific
reactions defined by Katze & Crowell (1980b). The alternative explanation of the
differences between the results of Minor et al. (1969) and Morgan-Capner &
McSorley (1983) may lie in the difference in sensitivity of the two tests. Minor et
al. (1969) do not define the sensitivity of their counterimmunoelectrophoresis test,
but Schmidt & Lennette (1962) state that specific IgM antibody can rarely be
detected by the immunodiffusion test in sera diluted beyond 1 in 2. The positive
control serum diluted in 1 in 20 in negative serum still gives a positive reaction
in the MACRIA of Morgan-Capner & McSorley (1983). Thus it is probable that
there is a tenfold difference in sensitivity between the tests of Minor et al. (1979)
and Morgan-Capner & McSorley (1983) and this may explain the higher frequency
of the cross-reactivity in the latter test.
It is important to resolve the problem of the extent of heterotypic reactions in
CBV IgM tests. If a test detects heterotypic reactions which are confined to the
CBV group then the test still remains a useful diagnostic and research tool.
However, if a positive result merely indicates recent infection with one of the 67
human enteroviruses then the usefulness of such a test is severely limited. To
pursue the current fashion for solid-phase immunoassays it seems essential to use
the highly purified virion antigens of Katze & Crowell (1980a) and an antibodycapture principle. If extensive heterotypic reactions are still found then this would
appear to be due to an inherent property of the human IgM response to enterovirus
infection, and this will be difficult to overcome.
J. R. PATTISON
Editor
REFERENCES
DORRIES, R. & TER MEULEN, V. (1980). Detection of enterovirus specific IgG and IgM antibodies
in humans by an indirect solid phase radioimmunoassay. Medical Microbiology and Immunology
168,159-171.
D6RRIES, R. & TER MEULEN, V. (1983). Specificity of IgM antibodies in acute human
Coxsackievirus B infections, analysed by indirect solid phase enzyme immunoassay and
immunoblot technique. Journal of General Virology 64, 159-167.
EL-HAGRASSY, M. M. O., BANATVALA, J. E. & COLTART, D. J. (1980). Coxsackie B-virus-specific
IgM responses in patients with cardiac and other diseases. Lancet ii, 1160-1162.
KATZE, M. G. & CROWELL, R. L. (1980a). Indirect enzyme-linked immunosorbent assay (ELISA)
for the detection of Coxsackie group B antibodies. Journal of General Virology 48, 225-229.
KATZE, M. G. & CROWELL, R. L. (1980b). Immunological studies of the group B Coxsackieviruses
by the sandwich enzyme-linked immunosorbent assay (ELISA) and immunoprecipitation.
Journal of General Virology 50, 357-367.
MINOR, T. E., HELSTROM, P. B., NELSON, D. B. & D'ALESSIO, D. J. (1979). Counterimmunoelectrophoresis test for immunoglobulin M antibodies to group B Coxsackievirus. Journal
of Clinical Microbiology 9, 503-506.
MORGAN-CAPNER, P. & MCSORLEY, C. (1983). Antibody capture radioimmunoassay (MACRIA)
for coxsackievirus B4 and B5-specific IgM. Journal of Hygeine 90, 333-349.
SCHMIDT, N. J. & LENNETTE, E. H. (1962). Gel double diffusion studies with group B and group
A, type 9 Coxsackieviruses. IL. Serological diagnosis of Coxsackievirus infections by gel double
diffusion technique. Journal of Immunology 89, 96-105.
SCHMIDT, N. J., MAGOFFIN, R. L. & LENNETTE, E. H. (1973). Association of group B Coxsackieviruses with cases of pericarditis, myocarditis, or pleurodynia by demonstration of immunoglobulin M antibody. Infection and Immunity 8, 341-348.
332 EDITORIAL