- Email: [email protected]
A sensitive enzyme-linked immunosorbent assay for the detection of herpes simplex virus antigens
Journal of Virological Methods, 17 (1987) 159-174 Elsevier
159
JVM 00628
A sensitive enzyme-linked immunosorbent assay for the detection of herpes simplex virus antigens J.M. Middeldorp’, A.M. Hooymans’, A.J.H.F. Kocken’, A.M. van Loon2, J.A. Emsbroek3 and R.A. Coutinho3 ‘Scientific Development Group. Organon International B.V., Oss, The Netherlands; ‘Department of Medical Microbiology, State University of Nijmegen, Nijmegen, The Netherlands; ‘Department of Infectious Diseases, Municipal Health Service, Amsterdam, The Netherlands (Accepted
27 March
1987)
Summary A sandwich ELBA for the detection of herpes simplex virus (HSV) antigens was developed using sheep anti-HSV F(ab’)2 fragments for capture and an indirect antibody system for detection. Current detection limits are 0.5 ng protein for HSVl and 1.5 ng protein for HSV2. This compares to a single HSVl-infected Vero-cell in a background of lo6 non-infected cells or 10 plaque forming units (PFU) of HSVl in culture supernatants as determined in separate experiments. Limiting dilution experiments show that one PFU of HSVl can be detected after overnight culture in both supernatant and cell extracts. The use of F(ab’)2 for capture completely eliminated binding of Staphylococcus aureus. No cross-reactivity was observed with other human herpes viruses. When evaluated with 245 random ‘left-overs’ of genital swab specimens in transport medium the test showed a sensitivity and specificity of 77.2 and 97.8%, respectively, with respect to virus isolation in culture. In a preliminary study on 16 direct ELISA swab-specimens extracted in 0.5 ml ELISA sample buffer both sensitivity and specificity were 100% with respect to culture. In both clinical series there was a proportional relationship between the ELISA value and the estimated amount of infectious virus in the specimen. Herpes
simplex
virus;
Antigen
detection;
ELISA
Correspondence to: J.M. Middeldorp. Organon International Room RE 2204, P.O. Box 20, 5340 BH Oss, The Netherlands. 0166-0934/87/$03.50
0
1987 Elsevier
Science
Publishers
B.V.
B.V.,
Scientific
(Biomedical
Development
Division)
Group.
160
Introduction Herpes simplex virus (HSV) infections are widely spread and form a still increasing medical problem. Infections with HSV may cause serious, life-threatening complications with long-term sequelae in newborns and immunocompromised patients. In addition, many non-compromised normal individuals suffer from herpes virus infections presenting either as oral/facial infection - mostly caused by HSV type 1 - or as genital/anal infection, mostly associated with HSV type 2 (Corey et al., 1983; Corey and Spear, 1986). The reported number of yearly new cases in the latter population is about 2~~ to 500~ in the U.S.A. alone (Guinnan, 1986). Fortunately an elegant and effective anti-viral therapy has recently become available to control active virus replication of both HSV subtypes (King and Galasso, 1982). However, success of treatment and relief of symptoms strongly depend on rapid initiation of treatment after onset of infection. This is of particular importance for newborns and immuno~ompromis~d patients who are at risk for the most severe forms of HSV infection (Corey and Spear, 1986). The need for rapid initiation of treatment calls for rapid diagnosis of ongoing HSV infection, which can be achieved by a variety of techniques, including serology, culture isolation and detection of viral antigens or DNA. At present isolation of the virus is most widely used for the diagnosis of active HSV infection, although culture techniques have obvious disadvantages. Immunological detection of viral antigens has been shown to be feasible by a variety of techniques including ELISA and immunofluorescence. The major merits of ELISA for diagnosis are that it is rapid and does not require complex sample pre-treatment. In addition, no precautions are needed to maintain virus infectivity or sample integrity as required for culture and fluorescence techniques, respectively. Therefore we recently started with the development of an ELISA for the detection of herpex simplex virus antigens in clinical specimens.
Materials and Methods Virus cultures and viral antigens
HSVl strains KOS and McIntyre and HSV2 strain MS and 333 were grown in Vero cell and human fetal fibi-oblast cultures and virions were purified as described elsewhere (Welling-Wester et al., 1985). Infectivity of viral stocks prepared from the supernatant of these cultures was determined by plaque titration under an agar overlay on human fetal fibroblasts grown in 24-well cluster trays (Costar, Badhoevedorp, The Netherlands) and expressed as plaque forming units (PFU) per ml. Limiting dilution experiments used for detection of HSV antigens were performed similarly without an agar overlay. HSV-positive ELISA control antigen was prepared by detergent extraction of HSVI (McIntyre) or HSV2 (333) infected Vero cells using 0.35% NP40 (Sigma, St. Louis, MO, U.S.A.) as described by Van Loon et al. (1981). ~ytomegalovirus
161
(CMV), varicella zoster virus (VZV) antigens were prepared identically from infected fibroblast cultures showing complete cytopathology. Epstein-Barr virus (EBV) antigens were prepared similarly from chemically-induced cultures of the P3HRl producer cell line. Staphylococcus aureus organisms were obtained from the Organon Microbiology Department at a protein concentration of 100 g/l. Protein concentrations were determined by the Lowry-method as modified by Dulley and Grieve (1975). Antisera
Hyperimmune anti-HSV antisera were produced in sheep and rabbits by infection or immunization or both using a variety of application routes. Routinely animals were infected on the eye with lo6 PFU or HSVl-KOS followed by a series of booster immunizations with purified HSV virions. Sheep anti-rabbit IgG was prepared by repeated immunization with purified rabbit IgG in complete and incomplete Freund’s adjuvants. IgG fractions were isolated by caprylic acid treatment (Steinbuch and Audran, 1969) and F(ab’)2 fragments were isolated by pepsin-digestion and chromatography over Sephacryl S-200 (Pharmacia, Sweden). Horseradish peroxidase (HRP)-conjugates were prepared using the heterobifunctional reagent SPDP as described by the manufacturer (Pharmacia). HSV-antigen
specimens
HSVl and HSV2 antigens from different tissue culture sources were isolated as described above. Clinical specimens were obtained from the Municipal Health Service, Amsterdam and consisted of random ‘left-overs’ of swab specimens in 2 ml transport medium sent in over a period of 3 mth for HSV-isolation by tissue culture and direct swab specimens specially collected for ELISA. Most samples were from genital/anal lesions with a male-to-female ratio of 1. Female samples were both from cervical and external sites. The direct swab specimens were taken from the same lesion that first served as source for the routine culture specimen. The culture ‘left-over’ specimens were sent to Organon biweekly in separate cohorts. All samples were stored at -20°C until use. Eleven swab specimens of culture proven varicella zoster virus lesions were obtained from Dr. Y. Perol, Hospital St. Louis, Paris, France. Sample preparation
Tissue culture antigens extracted in 0.35% NP40 were diluted serially in phosphate buffered saline (PBS) containing 1% Triton X-100. Intact cells and virions were treated with a specially developed extraction buffer containing preservatives and a mixture of detergents to release viral antigens and to inactivate virus infectivity. ‘Left-overs’ of specimens in transport medium received 10% (v/v) of a 10~ concentrated extraction buffer and were placed for 10 min at room temperature to ensure antigen solubilization before testing. Direct swab specimens
162
were extracted in 0.5 ml extraction buffer for 10 min at room temperature which the swab was squeezed out and discarded. HSV-antigen
after
ELBA
A number of different procedures have been used during test development resulting in the final procedure described below. Modifications of this procedure are mentioned in the text. Wells of polystyrene microtiter plates, coated with sheep anti-HSV F(ab’)2 for HSV capture or with F(ab’)2 fragments from a non-immune sheep for control, were filled with 100 ~1 of sample, pretreated or prediluted as mentioned above, and incubated for 1 h at 37°C. After three washes with PBS containing 0.05% Tween20 (PBS-T) 100 ~1 of a 1:600 dilution of rabbit anti-HSV serum was added to each well and again incubated for 1 h at 37°C followed by three washes with PBS-T. Subsequently, 100 ~1 of sheep anti-rabbit IgG HRP-conjugate was added, incubated for 1 h and 37°C and washed as above and finally the presence of HRP activity was detected using tetramethylbenzidine and ureum peroxide as substrate (Bos et al., 1981). After 30 min the reaction was stopped with 2 M H,SO, and the absorbance was measured at 450 nm using a Multiscan photometer. The cut-off value was defined as 0.150 OD units above the value of the control well incubation as applied for clinical specimens and as 0.150 above the (PBS-T) negative control for tissue cultured antigens for which no control incubation was needed.
Results Test development
A first series of experiments was initiated to develop a sensitive procedure for the detection of HSV antigens. Hyperimmune antisera were raised in different animal species by infection and immunization with HSV using a variety of application protocols. For capture and detection of HSV we used a combination of human, sheep and rabbit anti-HSV antisera either as serum, as purified IgG or as HRP-labeled IgG. Both direct and indirect HSV-capture sandwich ELISA procedures were evaluated as shown in Fig. 1. The antigen in all experiments consisted of a detergent-soluble fraction of HSVl (McIntyre)-infected Vero cells serially diluted in PBS-T. The most sensitive procedure consisted of sheep antiHSV for capture and rabbit anti-HSV plus sheep anti-rabbit HRP conjugate for detection, giving a detection limit of 5 ng HSVl protein. Subsequently this procedure was further optimized by improving the quality of the antisera and elimination of non-specific reactions resulting in the procedure as described in Materials and Methods. Sensitivity further increased to a level of 0.5 ng protein of HSVl and 1.5 ng protein of HSV2 as described below.
163
Test
I---+ h-a-HSV
->
r-a-HSV
r-a-HSV
ha>
Sensitivity
system
@
=
r-a-HSV=
s-a-WV
@
s-a-rIgG
@
s-a-rIgG
@
40
nglwell
,.
ng,we,,
5 ng,we,,
20
nglwell
40
nglwell
Fig. 1. Schematic representation of the different methods evaluated for detection of HSV-antigen. Sensitivity is expressed as the lowest amount of HSV-protein den&able per testwell (100 ~1 antigen solution) using a serially diluted detergent-soluble fraction of HSVl (McIntyre) infected Vero-cells. ha-HSV= human antibody against HSV; r-a-HSV=rabbit antibody against HSV; s-a-HSV= sheep antibody against HSV; HRP = horseradish peroxidase.
Specificity
The specificity of the test was evaluated with respect to S. aureus and other human herpesviruses CMV, VZV and EBV. Binding of S. aureuS was completely eliminated upto a level of at least 1 mg/ml bacterial protein by using F(ab’)2 fragments for capture without affecting the HSV-binding capacity. Antigens from CMV, VZV and EBV infected cells were harvested at maximal antigen expression using an identical procedure as for HSV. After adjusting the protein concentration in each extract to the level of the HSV standard, serial dilutions in PBS-T were tested in ELISA (Fig. 2). HSV antigen was detectable at 6.2 @ml (0.62 @test well), whereas CMV, VZV, and EBV antigens only were detectable at concentrations of 50 &ml or higher; i.e. about lo4 times higher than found for HSV.
164
24 \
~ _________
HSV CM”
----
EBV VI\’
20.
T
\
1.6 .:
51
‘2
\ ‘\
:
\ \ \ \
0 a-
\ \
\
\
\ ‘.
04.
‘1
\.
\.
--_
0
204.600
12.600
50
800 Protein
concentrotlon
30
0
(ngimli
Fig. 2. Comparison of cross-reactions with cultured antigens of human herpes viruses. Detergent-extracts prepared from human fibroblasts infected with herpes simplex virus (HSVI, McIntyre), varicella zoster virus (VZV) or cytomegalovirus (CMV) or from P3HRl-cells induced for Epstein-Barr virus capsid antigen expression were tested for reactivity in HSV-antigen ELISA. All extracts were prepared by identical procedures, adjusted to the same protein concentration and serially diluted in the PBS-T before testing.
/ 14.
12-
,o E
Ob-
0 10
0.4.
0
0
0
00 000
0
0 0
0 0
I
0.2
04
0.6
0.6
1.0 A450
1.2 HSV-test
I
1.4
1.6
1.6
2.0
2.2
2.4
well
Fig. 3. Comparison of absorbance values of 112 clinical ‘left-over’ samples in transport medium as tested on control and HSV-capture ELISA plates. Details are described in the text. Closed circles represent culture negative samples and open circles represent culture positive samples,
165 TABLE Detection
1 limits of current
HSV-antigen
ELISA
with respect
to cultured
HSV-antigens.
Type of specimen
Sensitivity
HSVl culture supernatant HSV2 culture supernatant Purified HSVl virions Purified HSV2 virions Detergent soluble HSVl-antigens Detergent soluble HSV2-antigens Intact HSVl-infected cells Overnight cultured HSVl
10 PFU 40 PFU 0.5 ng protein 1.5 ng protein 0.5 ng protein 2.5 ng protein 1 infected cell 1 PFU
This level is not considered to be physiologically relevant as confirmed by the negative reactions found in 11 clinical samples from vesicular fluid obtained from patients with culture confirmed varicella. Although no significant cross-reactions were observed in the above described experiments, individual samples from clinical lesions - especially from genital sites - may still produce non-specific binding, resulting in false-positive reactions when tested in a single-well capture experiment. For this reason a control incubation was used for all clinical specimens tested. Results from a study on 112 ‘left-over’ specimens in transport medium - mostly from genital lesions - demonstrate that non-specific binding to the control plate was not observed to a significant extent except for the strongly HSV-positive samples (Fig. 3). However, the reactivity of these samples on HSV-capture plate in all cases clearly was much stronger, resulting in a clear net-positive response. Using a cut-off level of 0.150 OD units above the control a specificity of 100% and a sensitivity of 76.3% with respect to culture were obtained for this particular panel. Sensitivity The detection limits of the HSV antigen ELBA were evaluated using a variety of HSV antigen preparations as shown in Table 1. First, supernatants from HSVl (McIntyre)-infected human fibroblast cultures were treated with 10% (v/v) of the concentrated extraction buffer and then assayed for HSV antigen reactivity. Fig. 4 demonstrates that up to 100 PFU/ml (i.e. 10 PFU per test well) were detectable at a significant level. Secondly, using gradient purified virions of HSVl (KOS) and HSV2 (333), treated similarly and tested in ELISA in serial dilutions, a detection limit of 0.5 ng protein for HSVl and 1.5 ng protein for HSV2 was found as tested independently for different isolations. Similar results were obtained with detergentextracted antigens from HSVl- and HSV2-infected Vero cells or human fibroblasts. Thirdly, we made serial dilutions of HSVl-infected Vero cells in non-infected Vero cells keeping the total number of cells constant at lo6 cells in 1 ml. Subsequently HSV antigens were solubilized by adding 10% (v/v) concentrated extraction buffer and 100 l.~lsamples were tested in ELISA. As shown in Fig. 5 a single HSV-infected cell was detectable above the level of control that consisted
166
24 I
I,,, 12800
3200
800
200
50
12.5
PFUlml
Fig. 4. Detection of HSV-antigen in culture medium. A dilution curve is shown of serial dilutions made in PBS-T of a culture supematant from HSVl (McIntyre)-infected human fibroblast cultures, pre-treated with 10% (v/v) 10 x concentrated extraction buffer. 100 l.~l samples were tested in ELISA. Infectivity of the original supernatant is expressed as plaque forming units (PFU) per ml.
of lo6 non-infected Vero cells. Also indicated is the effectiveness of the optimized extraction buffer in solubilizing HSV antigen as compared to extraction with PBST. The high background in this experiment is due to the residual anti-Vero reactivity of the animal anti-HSV sera used in the test, since these were raised by immunizing with HSV prepared from Vero cell cultures. Finally, serial dilutions of HSVl (McIntyre) were inoculated onto human fibroblast cultures (105 cells per well in 24 well cluster trays) and cultured overnight. At 1 h post inoculation and following overnight culture the supernatant was removed and HSV antigens were extracted from the cell monolayer using 0.1 ml extraction buffer and tested in ELISA. The results of such an experiment are shown in Fig. 6. Already at 1 h after inoculation (day 0) viral antigens are detectable in cell extracts using an inoculum of lo3 PFU and higher (no control incubation was used in this experiment). A very strong increase, however, is seen after overnight culture allowing the detection of single PFU of input virus. The amount of antigen extracted
161
-----
Optimized PBST
extraction
buffer
0.21
I 16 Dilution
6
4 of
HSV
2
1
infected
vero
0 cells
(cells/well)
Fig. 5. Detection of HSV-antigen in intact HSVl-infected Vero-cells. Intact HSVl (McIntyre)-infected Vero-cells were diluted in a total and constant number of lo6 non-infected Vero-cells in 1 ml volumes of PBS. The cell suspensions were then treated with 10% (v/v) of 10X concentrated extraction buffer or 10x concentrated PBS-T for 30 min at room temperature and 100 pl samples were subsequently tested in ELISA. Indicated on the horizontal axis is the total number of HSV-infected cells per testwell; that is in each case in the presence of lo5 non-infected cells.
at day 1 from the cell layer infected with 1 PFU is comparable with 16 HSVlinfected Vero cells. Similar results were obtained by testing cellular antigens sampled on a swab or by testing the culture supernatant. Although responses were somewhat lower, 1 PFU of input virus remained clearly detectable by ELISA in these cases. Cytopathic effects were not detectable at this time (i.e. 1624 h post inoculation) using an input of 1PFU but these appeared between the second and third day after inoculation, involving 5-10% of the cell layer. Clinical specimens
To evaluate the performance of HSV antigen ELISA in clinical specimens we performed a study on 245 random ‘left-overs’ from clinical samples in transport medium sent in for virus culture to the Amsterdam municipal health service. The
168
Fig. 6. Detection of HSV-antigen in infected cell extracts before and after overnight culture. Decreasing amounts of infectious HSVl (McIntyre) were inoculated onto 10s human fibroblasts cultured in 24well clustertrays and incubated for 1 h. Subsequently the supernatant was removed and replaced by fresh medium for overnight culture or the cell layer was washed twice with PBS and 100 ~1 extraction buffer was pipetted directly onto the monolayer (day 0). The overnight cultures were processed identically (day 1). Antigen-extraction was performed at room temperature for 30 min with repeated shaking and 100 ul samples were tested in ELISA.
TABLE 2 HSV-antigen detection in 245 random culture ‘left-over’ specimens. ELISA Culture
+
_
+ (n=lOl) - (n=144)
80 (77.2%) 3
21 144 (97.8%)
169 Culture
I OE-
.t
06-
l
positive
Negative
. . . i .
.
.
. 0.2 c”t_or~-_--l_~__l_~_‘__----- ;
___*__
,i@ 2
1 4
Days
6 in
l
8
10
*
--
culture
Fig. 7. ELBA reactivity of 245 clinical (‘left-over’) samples in transport medium. All samples were stored at -20°C before shipment to Organon and were treated with 10% (v/v) 10x concentrated extraction buffer before ELBA testing. 100 p,l samples were tested in both control and in HSV-capture wells. Other details are mentioned in the methods and results sections.
panel was sent to Organon in four cohorts representing all routine specimens sent in over a period of 2 mth and consisted of 144 samples with a negative HSV culture result and 101 positive ones, of which 84% were typed as HSV2. The results are shown in Table 2 and in more detail in Fig. 7. The overall sensitivity was 77.2% with respect to culture with a spread of 73-86% in the four cohorts. The specificity varied from 96-100% with a mean of 97.8% for the whole panel. Upon retesting one of the false positive ELISA results (AA,,,: 0.155) became negative, the other two remaining positive, possibly due.to the presence of non-infectious virus in the specimen. However, no confirmatory tests have been applied to these specimens yet. A positive relation exists between the ELISA value and the time needed to obtain a positive culture result as indicated in Fig. 7. The relative sensitivity of ELISA as compared to culture positivity at day 2 is 98%, at day 4, 87% and at day 6, 82%, respectively.
170
2 o-
18-
.
.
1.6-
. .
.
14-
1.25: B
l.O-
.
. :
o.a-
0.6-
. .
0.4 -
. 0.2-
____-___________-____J_______________?______cut_off
o-
1
1 2
0
’ 4
Days
’
’ 6
’
1 8
of positive
’
’
3
10
culture
0
12
’
’
14
result
Fig. 8. ELISA-reactivity of 16 clinical specimens specially collected for ELISA-testing. Swabs containing lesion material were replaced in the swab-container and stored dry at -20°C. For testing in ELISA 0.5 ml extraction buffer was added to the container and the swab was incubated and squeezed out. Subsequently a 100 ~1 sample was tested in both control and HSV-capture wells. For other details see Materials and Methods.
In a second and preliminary study we evaluated the value of direct-swab sampling, especially for ELISA, in a limited number of 16 ‘direct-swab’ samples and their corresponding ‘left-over’ culture specimens in transport medium. Parallel samples in this study were taken from the same lesion (the culture sample was always taken first), being 14 female, 2 male genital lesions, with one negative TABLE 3 HSV-antigen detection in 16 direct swab and corresponding
culture ‘left-over’ specimens.
ELISA ‘Left-over’
Direct swab
Culture
+
_
+
-
+ (n=15) - (n=l)
7 (46.7%) 0
8 1 (100%)
15 (100%) 0
0 1 (100%)
171
culture result and 15 cultures positive for HSV2. One specimen yielded both HSVl and HSV2. Direct ELISA swabs were stored dry at - 20°C and were extracted in 0.5 ml extraction buffer shortly before use. The results are indicated in Table 3 and Fig. 8. Although the ELISA results on ‘left-over’ samples were disappointingly low in this study (sensitivity 46.7%) the increase of sensitivity by using a swab sample directly for ELISA is dramatic (sensitivity 100%). Specificity could not be evaluated since only one culture-negative specimen was present in this panel; however this sample was also negative in ELISA. As indicated in Fig. 8, again a positive relation was found between the absorbance value in ELISA and the estimated quantity of infectious virus present in a sample being related to the development of cytopathology. No relationship was found between the absorbance value of direct swab specimens and ‘left-over’ specimens (data not shown).
Discussion Virus isolation still is the most widely used technique for the identification of active HSV infection in the human host, although it has obvious disadvantages. Alternative and more rapid techniques have been introduced in recent years, including immunofluorescence and ELISA techniques. Both techniques have been applied for direct antigen detection (Smidt et al., 1983; Adler-Storthz, 1983; Grillner and Landquist, 1983; Warford et al., 1985; Sewell and Horn, 1985), culture confirmation (Sutherland et al., 1986; Frame et al., 1984), or for rapid detection following a short pre-culture for antigen amplification (Nerurkar et al., 1984; Morgan and Smith, 1984; Gleaves et al., 1985; Warford et al., 1984). For rapid diagnosis direct detection techniques are preferred, especially when no extensive sample pretreatment is required and results can be obtained easily and objectively. ELISA techniques fulfill these requirements but their reported sensitivity remains disputable (Grillner and Landquist, 1983; Sewell and Horn, 1985; Warford et al., 1985). In the present study we have described our initial experiments in developing a sensitive ELISA for the detection of HSV antigens in clinical specimens. Special emphasis was placed on sensitivity by selecting specific procedures for the production of hyperimmune antibody reagents in different animal species and combining these in an indirect double antibody sandwich ELISA procedure. Specificity was checked against genetically and antigenically closely related members of the human herpes virus family. No significant cross-reactions were observed. In addition the binding of S. aurem species, a common contaminant in ulcerative lesions (Nerurkar et al., 1984), was completely eliminated by using F(ab’)2 fragments for HSV capture. In clinical studies involving more than 300 specimens only three ELISA-positive specimens were found that had been found negative by culture. However, the possible presence of non-infectious virus in these specimens has to be investigated further. The sensitivity of the present test is about 5-10 times greater with respect to cultured HSV antigens than described by other investigators (Ziegler et al., 1983; Clayton et al., 1985; Adler-Storthz et al., 1985; Bos et al., 1986), and allows the
172
detection of one single HSVl-infected cell, 0.5 ng HSVl protein or 10 PFU of infectious virus in cultured supernatants. The sensitivity towards HSV2 antigens was found to be about 2-5 times less. This most likely is due to the fact that only anti-HSVl reagents have been used in the experiments described in this paper. However, recent experiments have shown a clear increase of sensitivity towards HSV-2 antigens by using antisera from animals infected with HSVl and boosted by immunization with HSV2 (data not shown). Still the clinical sensitivity of the present procedure indicates that most HSV2-isolates (comprising 84% of all tested genital samples) can be identified. This may be further improved by using immunological reagents with increased anti-HSV2 activity. The results from the first clinical trial on 245 ‘left-overs’ of randomly collected (mostly) genital swab specimens in 2 ml transport medium show an overall sensitivity and specificity with respect to culture isolation that is comparable or even better than described by others (Grillner and Landquist. 1983; Sewell and Horn, 1985; Warford et al., 1984; Morgan and Smith, 1984). No differences in sensitivity were observed when the data were analysed separately for male and female specimens or for female cervical and external lesion specimens (data not shown). The results from this trial also suggest that the absorbance value found in ELISA may be used as an indication of the amount of infectious virus present in the sample as indicated in Figs. 7 and 8, assuming that the development of cytopathic effects in culture is directly related to the amount of input virus in the sample. A similar finding was described by Morgan and Smith (1984). A dramatic increase in sensitivity with respect to HSV antigen detection in culture ‘left-over’ specimens was found when swab samples were directly extracted with a specially developed extraction buffer. Although more data are clearly needed to substantiate this finding the results indicate that this method may be suitable as an alternative for culture isolation. In addition to direct detection the present assay procedure may also be used for rapid confirmation of HSV culture within 16-24 h after inoculation. As suggested by data shown in Fig. 6 this allows the detection of a single PFU of infectious virus, well before the development of cytopathological effects. However, these experiments were done using culture-adapted strains of HSV and need further confirmation with fresh culture isolates. These studies are presently in progress. As suggested by other investigations (Nerurkar et al., 1984; Morgan and Smith 1984; Warford et al., 1984) this culture-amplified ELISA-technique approaches the sensitivity of culture isolation but results can be obtained more rapidly. In addition, due to the increased amount of viral antigens available for testing, direct typing of the HSV subtype becomes feasible (Buckmaster et al., 1984; Sutherland et al., 1986). As to date, typing of HSV infections has been considered unnecessary by many physicians. However, with the advent of more antiviral drugs showing different effectiveness on the two viral subtypes the need for accurate typing of HSV in clinical specimens may increase.
173
Acknowledgements We thank Dr. S. Welling-Wester (Department of Medical Microbiology, State University Groningen, The Netherlands) for providing infectious HSV stocks and purified virion preparations and Ms. Marja de Ridder (Municipal Health Service, Amsterdam, The Netherlands) for providing information on the clinical specimens used in this study and Ms. Carla Persons for typing the manuscript.
References Adler-Storthz, K., Kendall, C., Kennedy, R.C., Henkel, R.D. and Dreesman G.R. (1983) Biotin-avidinamplified enzyme immunoassay for detection of herpes simplex virus antigen in clinical specimens. J. Clin. Microbial. 18, 1329-1334. Bos, E.S., Van der Doelen, A.A., Van Rooy, N. and Schuurs, A.H.W.M. (1981). 3,3’, 5,5’Tetramethylbenzidine as an Ames test negative chromogen for horse-radish peroxidase in enzyme immunoassay. J. Immunoassay 2, 187-204. Bos, C.A., Wilterdink, J.B. and Scheffer, A.J. (1986). Detection of viral antigens in cerebrospinal fluid of rabbits with experimental herpes simplex virus type 1 encephalitis. Arch. Virol. 91, 73-81. Buckmaster, E.A., Cranage, M.P., McLean, C.S., Coombs, R.R.A. and Minson A. (1984). The use of monoclonal antibodies to differentiate isolates of herpes simplex types 1 and 2 by neutralization and reverse passive haemagglutination tests. J. Med. Virol. 13, 19%202. Clayton, A., Beckford, U., Roberts, C., Sutherland, S., Druce, A., Best, J. and Shandler. S. (1985). Factors influencing the sensitivity of herpes simplex virus detection in clinical specimens in a simultaneous enzyme-linked immunosorbent assay using monoclonal antibodies. J. Med. Virol. 17, 275-282. Corey, L. and Spear, P.G. (1986). Infections with herpes simplex viruses (second of two parts). N. Engl. J. Med. 314, 749-756. Corey, L., Adams, H.G., Zane. A., Brown, M.D. and Holmes, K.K. (1983) Genital herpes simplex virus infections: clinical manifestations, course and complications. Ann. Intern. Med. 98, 958-972. Dulley, J.R. and Grieve, P.A. (1975). A simple technique for eliminating interference by detergents in the Lowry method of protein determination. Anal. Biochem. 64, 136141. Frame, B., Mahony, J.B. Balachandran, N., Rawls, W.E. and Chemeskey, M.A. (1984). Identification and typing of herpes simplex virus by enzyme immunoassay with monoclonal antibodies. J. Clin. Microbial. 20, 162-166. Gleaves, C.A.. Wilson, D.J., Wold, A.D. and Smith T.F. (1985). Detection and serotyping of herpes simplex virus in MRC-5 cells by use of centrifugation and monoclonal antibodies 16 h postinoculation. J. Clin. Microbial. 21, 29-32. Grillner, L. and Landquist, M. (1983). Enzyme-linked immunosorbent assay for detection and typing of herpes simplex virus. Eur. J. Clin. Microbial. 2, 39-42. Guinnan, M.E. (1986). Sexually transmitted diseases may reverse the revolution. J. Am. Med. Assoc. 255, 1665-1672. King, D.H. and Galasso, G. (1982). Proceedings of a Symposium on acyclovir. Am J. Med. 73, lA-393A. Morgan, M.A. and Smith, T.F. (1984). Evaluation of an enzyme-linked immunosorbent assay for the detection of herpes simplex virus antigen. J. Clin. Microbial. 19, 730-732. Nerurkar, L.S., Namba, M., Brashears, G., Jacob, A.J., Lee, Y.J. and Sever, J.L. (1984). Rapid detection of herpes simplex virus in clinical specimens by use of capture biotin-streptavidin enzymelinked immunosorbent assay. J. Clin. Microbiolf 20, 109-144. Sewell. D.L. and Horn, S.A. (1985). Evaluation of a commercial enzyme-linked immunosorbent assay for the detection of herpes simplex virus. J. Clin. Microbial. 21, 457-458. Smidt, N.J., Dennis, J., Devlin, V., Gallo, D. and Mills, J. (1983) Comparison of direct
174 immunofluorescence and direct immunoperoxidase procedures for detection of herpes simplex virus antigen in lesion specimens. J. Clin. Microbial. 18, 445-448. Steinbuch, M. and Audran, R. (1969). The isolation of IgG from mammalian sera with the aid of caprylic acid. Arch. Biochem. Biophys. 134, 279-284. Sutherland, S.. Morgan, B., Mindel, A. and Chan. W.L. (1986) Typing and subtyping of herpes simplex isolates by monoclonal fluorescence. J. Med. Virol. 18, 235-245. Van Loon, A.M., Heessen. F.W.A., Van der Logt, J.Th. M. and Van der Veen. J. (1981). Direct enzyme-linked immunosorbent assay that uses peroxidase-labeled antigen for determination of immunoglobulin M antibody to cytomegalovirus. J. Clin. Microbial. 13, 416-422. Warford. A.L., Levy, R.A. and Rekrut. K.A. (1984). Evaluation of a commercial enzyme-linked immunosorbent assay for detection of herpes simplex virus antigen. J. Clin. Microbial. 20,490-493. Warford, A.L., Levy, R.A., Strong, C.A. and Rekrut, K.A. (1985). Comparison of two commercial enzyme-linked immunosorbent assays for the detection of herpes simplex virus antigen. Am. J. Clin. Pathol. 85, 229-232. Welling-Wester, S., Popken-Boer. T., Wilterdink, J.B., Van Beeumen, J. and Welling. G.W. (198.5). Isolation by high-performance liquid chromatography and partial characterization of a 57OCQ-dalton herpes simplex virus type 1 polypeptide. J. Virol. 54, 265-270. Ziegler, T., Meurman. O.H., Arstila, P.P. and Halonen, P.E. (1983) Solid-phase enzyme-immunoassay for the detection of herpes simplex virus antigens in clinical specimens. J. Virol. Methods. 7, l-9.
159
JVM 00628
A sensitive enzyme-linked immunosorbent assay for the detection of herpes simplex virus antigens J.M. Middeldorp’, A.M. Hooymans’, A.J.H.F. Kocken’, A.M. van Loon2, J.A. Emsbroek3 and R.A. Coutinho3 ‘Scientific Development Group. Organon International B.V., Oss, The Netherlands; ‘Department of Medical Microbiology, State University of Nijmegen, Nijmegen, The Netherlands; ‘Department of Infectious Diseases, Municipal Health Service, Amsterdam, The Netherlands (Accepted
27 March
1987)
Summary A sandwich ELBA for the detection of herpes simplex virus (HSV) antigens was developed using sheep anti-HSV F(ab’)2 fragments for capture and an indirect antibody system for detection. Current detection limits are 0.5 ng protein for HSVl and 1.5 ng protein for HSV2. This compares to a single HSVl-infected Vero-cell in a background of lo6 non-infected cells or 10 plaque forming units (PFU) of HSVl in culture supernatants as determined in separate experiments. Limiting dilution experiments show that one PFU of HSVl can be detected after overnight culture in both supernatant and cell extracts. The use of F(ab’)2 for capture completely eliminated binding of Staphylococcus aureus. No cross-reactivity was observed with other human herpes viruses. When evaluated with 245 random ‘left-overs’ of genital swab specimens in transport medium the test showed a sensitivity and specificity of 77.2 and 97.8%, respectively, with respect to virus isolation in culture. In a preliminary study on 16 direct ELISA swab-specimens extracted in 0.5 ml ELISA sample buffer both sensitivity and specificity were 100% with respect to culture. In both clinical series there was a proportional relationship between the ELISA value and the estimated amount of infectious virus in the specimen. Herpes
simplex
virus;
Antigen
detection;
ELISA
Correspondence to: J.M. Middeldorp. Organon International Room RE 2204, P.O. Box 20, 5340 BH Oss, The Netherlands. 0166-0934/87/$03.50
0
1987 Elsevier
Science
Publishers
B.V.
B.V.,
Scientific
(Biomedical
Development
Division)
Group.
160
Introduction Herpes simplex virus (HSV) infections are widely spread and form a still increasing medical problem. Infections with HSV may cause serious, life-threatening complications with long-term sequelae in newborns and immunocompromised patients. In addition, many non-compromised normal individuals suffer from herpes virus infections presenting either as oral/facial infection - mostly caused by HSV type 1 - or as genital/anal infection, mostly associated with HSV type 2 (Corey et al., 1983; Corey and Spear, 1986). The reported number of yearly new cases in the latter population is about 2~~ to 500~ in the U.S.A. alone (Guinnan, 1986). Fortunately an elegant and effective anti-viral therapy has recently become available to control active virus replication of both HSV subtypes (King and Galasso, 1982). However, success of treatment and relief of symptoms strongly depend on rapid initiation of treatment after onset of infection. This is of particular importance for newborns and immuno~ompromis~d patients who are at risk for the most severe forms of HSV infection (Corey and Spear, 1986). The need for rapid initiation of treatment calls for rapid diagnosis of ongoing HSV infection, which can be achieved by a variety of techniques, including serology, culture isolation and detection of viral antigens or DNA. At present isolation of the virus is most widely used for the diagnosis of active HSV infection, although culture techniques have obvious disadvantages. Immunological detection of viral antigens has been shown to be feasible by a variety of techniques including ELISA and immunofluorescence. The major merits of ELISA for diagnosis are that it is rapid and does not require complex sample pre-treatment. In addition, no precautions are needed to maintain virus infectivity or sample integrity as required for culture and fluorescence techniques, respectively. Therefore we recently started with the development of an ELISA for the detection of herpex simplex virus antigens in clinical specimens.
Materials and Methods Virus cultures and viral antigens
HSVl strains KOS and McIntyre and HSV2 strain MS and 333 were grown in Vero cell and human fetal fibi-oblast cultures and virions were purified as described elsewhere (Welling-Wester et al., 1985). Infectivity of viral stocks prepared from the supernatant of these cultures was determined by plaque titration under an agar overlay on human fetal fibroblasts grown in 24-well cluster trays (Costar, Badhoevedorp, The Netherlands) and expressed as plaque forming units (PFU) per ml. Limiting dilution experiments used for detection of HSV antigens were performed similarly without an agar overlay. HSV-positive ELISA control antigen was prepared by detergent extraction of HSVI (McIntyre) or HSV2 (333) infected Vero cells using 0.35% NP40 (Sigma, St. Louis, MO, U.S.A.) as described by Van Loon et al. (1981). ~ytomegalovirus
161
(CMV), varicella zoster virus (VZV) antigens were prepared identically from infected fibroblast cultures showing complete cytopathology. Epstein-Barr virus (EBV) antigens were prepared similarly from chemically-induced cultures of the P3HRl producer cell line. Staphylococcus aureus organisms were obtained from the Organon Microbiology Department at a protein concentration of 100 g/l. Protein concentrations were determined by the Lowry-method as modified by Dulley and Grieve (1975). Antisera
Hyperimmune anti-HSV antisera were produced in sheep and rabbits by infection or immunization or both using a variety of application routes. Routinely animals were infected on the eye with lo6 PFU or HSVl-KOS followed by a series of booster immunizations with purified HSV virions. Sheep anti-rabbit IgG was prepared by repeated immunization with purified rabbit IgG in complete and incomplete Freund’s adjuvants. IgG fractions were isolated by caprylic acid treatment (Steinbuch and Audran, 1969) and F(ab’)2 fragments were isolated by pepsin-digestion and chromatography over Sephacryl S-200 (Pharmacia, Sweden). Horseradish peroxidase (HRP)-conjugates were prepared using the heterobifunctional reagent SPDP as described by the manufacturer (Pharmacia). HSV-antigen
specimens
HSVl and HSV2 antigens from different tissue culture sources were isolated as described above. Clinical specimens were obtained from the Municipal Health Service, Amsterdam and consisted of random ‘left-overs’ of swab specimens in 2 ml transport medium sent in over a period of 3 mth for HSV-isolation by tissue culture and direct swab specimens specially collected for ELISA. Most samples were from genital/anal lesions with a male-to-female ratio of 1. Female samples were both from cervical and external sites. The direct swab specimens were taken from the same lesion that first served as source for the routine culture specimen. The culture ‘left-over’ specimens were sent to Organon biweekly in separate cohorts. All samples were stored at -20°C until use. Eleven swab specimens of culture proven varicella zoster virus lesions were obtained from Dr. Y. Perol, Hospital St. Louis, Paris, France. Sample preparation
Tissue culture antigens extracted in 0.35% NP40 were diluted serially in phosphate buffered saline (PBS) containing 1% Triton X-100. Intact cells and virions were treated with a specially developed extraction buffer containing preservatives and a mixture of detergents to release viral antigens and to inactivate virus infectivity. ‘Left-overs’ of specimens in transport medium received 10% (v/v) of a 10~ concentrated extraction buffer and were placed for 10 min at room temperature to ensure antigen solubilization before testing. Direct swab specimens
162
were extracted in 0.5 ml extraction buffer for 10 min at room temperature which the swab was squeezed out and discarded. HSV-antigen
after
ELBA
A number of different procedures have been used during test development resulting in the final procedure described below. Modifications of this procedure are mentioned in the text. Wells of polystyrene microtiter plates, coated with sheep anti-HSV F(ab’)2 for HSV capture or with F(ab’)2 fragments from a non-immune sheep for control, were filled with 100 ~1 of sample, pretreated or prediluted as mentioned above, and incubated for 1 h at 37°C. After three washes with PBS containing 0.05% Tween20 (PBS-T) 100 ~1 of a 1:600 dilution of rabbit anti-HSV serum was added to each well and again incubated for 1 h at 37°C followed by three washes with PBS-T. Subsequently, 100 ~1 of sheep anti-rabbit IgG HRP-conjugate was added, incubated for 1 h and 37°C and washed as above and finally the presence of HRP activity was detected using tetramethylbenzidine and ureum peroxide as substrate (Bos et al., 1981). After 30 min the reaction was stopped with 2 M H,SO, and the absorbance was measured at 450 nm using a Multiscan photometer. The cut-off value was defined as 0.150 OD units above the value of the control well incubation as applied for clinical specimens and as 0.150 above the (PBS-T) negative control for tissue cultured antigens for which no control incubation was needed.
Results Test development
A first series of experiments was initiated to develop a sensitive procedure for the detection of HSV antigens. Hyperimmune antisera were raised in different animal species by infection and immunization with HSV using a variety of application protocols. For capture and detection of HSV we used a combination of human, sheep and rabbit anti-HSV antisera either as serum, as purified IgG or as HRP-labeled IgG. Both direct and indirect HSV-capture sandwich ELISA procedures were evaluated as shown in Fig. 1. The antigen in all experiments consisted of a detergent-soluble fraction of HSVl (McIntyre)-infected Vero cells serially diluted in PBS-T. The most sensitive procedure consisted of sheep antiHSV for capture and rabbit anti-HSV plus sheep anti-rabbit HRP conjugate for detection, giving a detection limit of 5 ng HSVl protein. Subsequently this procedure was further optimized by improving the quality of the antisera and elimination of non-specific reactions resulting in the procedure as described in Materials and Methods. Sensitivity further increased to a level of 0.5 ng protein of HSVl and 1.5 ng protein of HSV2 as described below.
163
Test
I---+ h-a-HSV
->
r-a-HSV
r-a-HSV
ha>
Sensitivity
system
@
=
r-a-HSV=
s-a-WV
@
s-a-rIgG
@
s-a-rIgG
@
40
nglwell
,.
ng,we,,
5 ng,we,,
20
nglwell
40
nglwell
Fig. 1. Schematic representation of the different methods evaluated for detection of HSV-antigen. Sensitivity is expressed as the lowest amount of HSV-protein den&able per testwell (100 ~1 antigen solution) using a serially diluted detergent-soluble fraction of HSVl (McIntyre) infected Vero-cells. ha-HSV= human antibody against HSV; r-a-HSV=rabbit antibody against HSV; s-a-HSV= sheep antibody against HSV; HRP = horseradish peroxidase.
Specificity
The specificity of the test was evaluated with respect to S. aureus and other human herpesviruses CMV, VZV and EBV. Binding of S. aureuS was completely eliminated upto a level of at least 1 mg/ml bacterial protein by using F(ab’)2 fragments for capture without affecting the HSV-binding capacity. Antigens from CMV, VZV and EBV infected cells were harvested at maximal antigen expression using an identical procedure as for HSV. After adjusting the protein concentration in each extract to the level of the HSV standard, serial dilutions in PBS-T were tested in ELISA (Fig. 2). HSV antigen was detectable at 6.2 @ml (0.62 @test well), whereas CMV, VZV, and EBV antigens only were detectable at concentrations of 50 &ml or higher; i.e. about lo4 times higher than found for HSV.
164
24 \
~ _________
HSV CM”
----
EBV VI\’
20.
T
\
1.6 .:
51
‘2
\ ‘\
:
\ \ \ \
0 a-
\ \
\
\
\ ‘.
04.
‘1
\.
\.
--_
0
204.600
12.600
50
800 Protein
concentrotlon
30
0
(ngimli
Fig. 2. Comparison of cross-reactions with cultured antigens of human herpes viruses. Detergent-extracts prepared from human fibroblasts infected with herpes simplex virus (HSVI, McIntyre), varicella zoster virus (VZV) or cytomegalovirus (CMV) or from P3HRl-cells induced for Epstein-Barr virus capsid antigen expression were tested for reactivity in HSV-antigen ELISA. All extracts were prepared by identical procedures, adjusted to the same protein concentration and serially diluted in the PBS-T before testing.
/ 14.
12-
,o E
Ob-
0 10
0.4.
0
0
0
00 000
0
0 0
0 0
I
0.2
04
0.6
0.6
1.0 A450
1.2 HSV-test
I
1.4
1.6
1.6
2.0
2.2
2.4
well
Fig. 3. Comparison of absorbance values of 112 clinical ‘left-over’ samples in transport medium as tested on control and HSV-capture ELISA plates. Details are described in the text. Closed circles represent culture negative samples and open circles represent culture positive samples,
165 TABLE Detection
1 limits of current
HSV-antigen
ELISA
with respect
to cultured
HSV-antigens.
Type of specimen
Sensitivity
HSVl culture supernatant HSV2 culture supernatant Purified HSVl virions Purified HSV2 virions Detergent soluble HSVl-antigens Detergent soluble HSV2-antigens Intact HSVl-infected cells Overnight cultured HSVl
10 PFU 40 PFU 0.5 ng protein 1.5 ng protein 0.5 ng protein 2.5 ng protein 1 infected cell 1 PFU
This level is not considered to be physiologically relevant as confirmed by the negative reactions found in 11 clinical samples from vesicular fluid obtained from patients with culture confirmed varicella. Although no significant cross-reactions were observed in the above described experiments, individual samples from clinical lesions - especially from genital sites - may still produce non-specific binding, resulting in false-positive reactions when tested in a single-well capture experiment. For this reason a control incubation was used for all clinical specimens tested. Results from a study on 112 ‘left-over’ specimens in transport medium - mostly from genital lesions - demonstrate that non-specific binding to the control plate was not observed to a significant extent except for the strongly HSV-positive samples (Fig. 3). However, the reactivity of these samples on HSV-capture plate in all cases clearly was much stronger, resulting in a clear net-positive response. Using a cut-off level of 0.150 OD units above the control a specificity of 100% and a sensitivity of 76.3% with respect to culture were obtained for this particular panel. Sensitivity The detection limits of the HSV antigen ELBA were evaluated using a variety of HSV antigen preparations as shown in Table 1. First, supernatants from HSVl (McIntyre)-infected human fibroblast cultures were treated with 10% (v/v) of the concentrated extraction buffer and then assayed for HSV antigen reactivity. Fig. 4 demonstrates that up to 100 PFU/ml (i.e. 10 PFU per test well) were detectable at a significant level. Secondly, using gradient purified virions of HSVl (KOS) and HSV2 (333), treated similarly and tested in ELISA in serial dilutions, a detection limit of 0.5 ng protein for HSVl and 1.5 ng protein for HSV2 was found as tested independently for different isolations. Similar results were obtained with detergentextracted antigens from HSVl- and HSV2-infected Vero cells or human fibroblasts. Thirdly, we made serial dilutions of HSVl-infected Vero cells in non-infected Vero cells keeping the total number of cells constant at lo6 cells in 1 ml. Subsequently HSV antigens were solubilized by adding 10% (v/v) concentrated extraction buffer and 100 l.~lsamples were tested in ELISA. As shown in Fig. 5 a single HSV-infected cell was detectable above the level of control that consisted
166
24 I
I,,, 12800
3200
800
200
50
12.5
PFUlml
Fig. 4. Detection of HSV-antigen in culture medium. A dilution curve is shown of serial dilutions made in PBS-T of a culture supematant from HSVl (McIntyre)-infected human fibroblast cultures, pre-treated with 10% (v/v) 10 x concentrated extraction buffer. 100 l.~l samples were tested in ELISA. Infectivity of the original supernatant is expressed as plaque forming units (PFU) per ml.
of lo6 non-infected Vero cells. Also indicated is the effectiveness of the optimized extraction buffer in solubilizing HSV antigen as compared to extraction with PBST. The high background in this experiment is due to the residual anti-Vero reactivity of the animal anti-HSV sera used in the test, since these were raised by immunizing with HSV prepared from Vero cell cultures. Finally, serial dilutions of HSVl (McIntyre) were inoculated onto human fibroblast cultures (105 cells per well in 24 well cluster trays) and cultured overnight. At 1 h post inoculation and following overnight culture the supernatant was removed and HSV antigens were extracted from the cell monolayer using 0.1 ml extraction buffer and tested in ELISA. The results of such an experiment are shown in Fig. 6. Already at 1 h after inoculation (day 0) viral antigens are detectable in cell extracts using an inoculum of lo3 PFU and higher (no control incubation was used in this experiment). A very strong increase, however, is seen after overnight culture allowing the detection of single PFU of input virus. The amount of antigen extracted
161
-----
Optimized PBST
extraction
buffer
0.21
I 16 Dilution
6
4 of
HSV
2
1
infected
vero
0 cells
(cells/well)
Fig. 5. Detection of HSV-antigen in intact HSVl-infected Vero-cells. Intact HSVl (McIntyre)-infected Vero-cells were diluted in a total and constant number of lo6 non-infected Vero-cells in 1 ml volumes of PBS. The cell suspensions were then treated with 10% (v/v) of 10X concentrated extraction buffer or 10x concentrated PBS-T for 30 min at room temperature and 100 pl samples were subsequently tested in ELISA. Indicated on the horizontal axis is the total number of HSV-infected cells per testwell; that is in each case in the presence of lo5 non-infected cells.
at day 1 from the cell layer infected with 1 PFU is comparable with 16 HSVlinfected Vero cells. Similar results were obtained by testing cellular antigens sampled on a swab or by testing the culture supernatant. Although responses were somewhat lower, 1 PFU of input virus remained clearly detectable by ELISA in these cases. Cytopathic effects were not detectable at this time (i.e. 1624 h post inoculation) using an input of 1PFU but these appeared between the second and third day after inoculation, involving 5-10% of the cell layer. Clinical specimens
To evaluate the performance of HSV antigen ELISA in clinical specimens we performed a study on 245 random ‘left-overs’ from clinical samples in transport medium sent in for virus culture to the Amsterdam municipal health service. The
168
Fig. 6. Detection of HSV-antigen in infected cell extracts before and after overnight culture. Decreasing amounts of infectious HSVl (McIntyre) were inoculated onto 10s human fibroblasts cultured in 24well clustertrays and incubated for 1 h. Subsequently the supernatant was removed and replaced by fresh medium for overnight culture or the cell layer was washed twice with PBS and 100 ~1 extraction buffer was pipetted directly onto the monolayer (day 0). The overnight cultures were processed identically (day 1). Antigen-extraction was performed at room temperature for 30 min with repeated shaking and 100 ul samples were tested in ELISA.
TABLE 2 HSV-antigen detection in 245 random culture ‘left-over’ specimens. ELISA Culture
+
_
+ (n=lOl) - (n=144)
80 (77.2%) 3
21 144 (97.8%)
169 Culture
I OE-
.t
06-
l
positive
Negative
. . . i .
.
.
. 0.2 c”t_or~-_--l_~__l_~_‘__----- ;
___*__
,i@ 2
1 4
Days
6 in
l
8
10
*
--
culture
Fig. 7. ELBA reactivity of 245 clinical (‘left-over’) samples in transport medium. All samples were stored at -20°C before shipment to Organon and were treated with 10% (v/v) 10x concentrated extraction buffer before ELBA testing. 100 p,l samples were tested in both control and in HSV-capture wells. Other details are mentioned in the methods and results sections.
panel was sent to Organon in four cohorts representing all routine specimens sent in over a period of 2 mth and consisted of 144 samples with a negative HSV culture result and 101 positive ones, of which 84% were typed as HSV2. The results are shown in Table 2 and in more detail in Fig. 7. The overall sensitivity was 77.2% with respect to culture with a spread of 73-86% in the four cohorts. The specificity varied from 96-100% with a mean of 97.8% for the whole panel. Upon retesting one of the false positive ELISA results (AA,,,: 0.155) became negative, the other two remaining positive, possibly due.to the presence of non-infectious virus in the specimen. However, no confirmatory tests have been applied to these specimens yet. A positive relation exists between the ELISA value and the time needed to obtain a positive culture result as indicated in Fig. 7. The relative sensitivity of ELISA as compared to culture positivity at day 2 is 98%, at day 4, 87% and at day 6, 82%, respectively.
170
2 o-
18-
.
.
1.6-
. .
.
14-
1.25: B
l.O-
.
. :
o.a-
0.6-
. .
0.4 -
. 0.2-
____-___________-____J_______________?______cut_off
o-
1
1 2
0
’ 4
Days
’
’ 6
’
1 8
of positive
’
’
3
10
culture
0
12
’
’
14
result
Fig. 8. ELISA-reactivity of 16 clinical specimens specially collected for ELISA-testing. Swabs containing lesion material were replaced in the swab-container and stored dry at -20°C. For testing in ELISA 0.5 ml extraction buffer was added to the container and the swab was incubated and squeezed out. Subsequently a 100 ~1 sample was tested in both control and HSV-capture wells. For other details see Materials and Methods.
In a second and preliminary study we evaluated the value of direct-swab sampling, especially for ELISA, in a limited number of 16 ‘direct-swab’ samples and their corresponding ‘left-over’ culture specimens in transport medium. Parallel samples in this study were taken from the same lesion (the culture sample was always taken first), being 14 female, 2 male genital lesions, with one negative TABLE 3 HSV-antigen detection in 16 direct swab and corresponding
culture ‘left-over’ specimens.
ELISA ‘Left-over’
Direct swab
Culture
+
_
+
-
+ (n=15) - (n=l)
7 (46.7%) 0
8 1 (100%)
15 (100%) 0
0 1 (100%)
171
culture result and 15 cultures positive for HSV2. One specimen yielded both HSVl and HSV2. Direct ELISA swabs were stored dry at - 20°C and were extracted in 0.5 ml extraction buffer shortly before use. The results are indicated in Table 3 and Fig. 8. Although the ELISA results on ‘left-over’ samples were disappointingly low in this study (sensitivity 46.7%) the increase of sensitivity by using a swab sample directly for ELISA is dramatic (sensitivity 100%). Specificity could not be evaluated since only one culture-negative specimen was present in this panel; however this sample was also negative in ELISA. As indicated in Fig. 8, again a positive relation was found between the absorbance value in ELISA and the estimated quantity of infectious virus present in a sample being related to the development of cytopathology. No relationship was found between the absorbance value of direct swab specimens and ‘left-over’ specimens (data not shown).
Discussion Virus isolation still is the most widely used technique for the identification of active HSV infection in the human host, although it has obvious disadvantages. Alternative and more rapid techniques have been introduced in recent years, including immunofluorescence and ELISA techniques. Both techniques have been applied for direct antigen detection (Smidt et al., 1983; Adler-Storthz, 1983; Grillner and Landquist, 1983; Warford et al., 1985; Sewell and Horn, 1985), culture confirmation (Sutherland et al., 1986; Frame et al., 1984), or for rapid detection following a short pre-culture for antigen amplification (Nerurkar et al., 1984; Morgan and Smith, 1984; Gleaves et al., 1985; Warford et al., 1984). For rapid diagnosis direct detection techniques are preferred, especially when no extensive sample pretreatment is required and results can be obtained easily and objectively. ELISA techniques fulfill these requirements but their reported sensitivity remains disputable (Grillner and Landquist, 1983; Sewell and Horn, 1985; Warford et al., 1985). In the present study we have described our initial experiments in developing a sensitive ELISA for the detection of HSV antigens in clinical specimens. Special emphasis was placed on sensitivity by selecting specific procedures for the production of hyperimmune antibody reagents in different animal species and combining these in an indirect double antibody sandwich ELISA procedure. Specificity was checked against genetically and antigenically closely related members of the human herpes virus family. No significant cross-reactions were observed. In addition the binding of S. aurem species, a common contaminant in ulcerative lesions (Nerurkar et al., 1984), was completely eliminated by using F(ab’)2 fragments for HSV capture. In clinical studies involving more than 300 specimens only three ELISA-positive specimens were found that had been found negative by culture. However, the possible presence of non-infectious virus in these specimens has to be investigated further. The sensitivity of the present test is about 5-10 times greater with respect to cultured HSV antigens than described by other investigators (Ziegler et al., 1983; Clayton et al., 1985; Adler-Storthz et al., 1985; Bos et al., 1986), and allows the
172
detection of one single HSVl-infected cell, 0.5 ng HSVl protein or 10 PFU of infectious virus in cultured supernatants. The sensitivity towards HSV2 antigens was found to be about 2-5 times less. This most likely is due to the fact that only anti-HSVl reagents have been used in the experiments described in this paper. However, recent experiments have shown a clear increase of sensitivity towards HSV-2 antigens by using antisera from animals infected with HSVl and boosted by immunization with HSV2 (data not shown). Still the clinical sensitivity of the present procedure indicates that most HSV2-isolates (comprising 84% of all tested genital samples) can be identified. This may be further improved by using immunological reagents with increased anti-HSV2 activity. The results from the first clinical trial on 245 ‘left-overs’ of randomly collected (mostly) genital swab specimens in 2 ml transport medium show an overall sensitivity and specificity with respect to culture isolation that is comparable or even better than described by others (Grillner and Landquist. 1983; Sewell and Horn, 1985; Warford et al., 1984; Morgan and Smith, 1984). No differences in sensitivity were observed when the data were analysed separately for male and female specimens or for female cervical and external lesion specimens (data not shown). The results from this trial also suggest that the absorbance value found in ELISA may be used as an indication of the amount of infectious virus present in the sample as indicated in Figs. 7 and 8, assuming that the development of cytopathic effects in culture is directly related to the amount of input virus in the sample. A similar finding was described by Morgan and Smith (1984). A dramatic increase in sensitivity with respect to HSV antigen detection in culture ‘left-over’ specimens was found when swab samples were directly extracted with a specially developed extraction buffer. Although more data are clearly needed to substantiate this finding the results indicate that this method may be suitable as an alternative for culture isolation. In addition to direct detection the present assay procedure may also be used for rapid confirmation of HSV culture within 16-24 h after inoculation. As suggested by data shown in Fig. 6 this allows the detection of a single PFU of infectious virus, well before the development of cytopathological effects. However, these experiments were done using culture-adapted strains of HSV and need further confirmation with fresh culture isolates. These studies are presently in progress. As suggested by other investigations (Nerurkar et al., 1984; Morgan and Smith 1984; Warford et al., 1984) this culture-amplified ELISA-technique approaches the sensitivity of culture isolation but results can be obtained more rapidly. In addition, due to the increased amount of viral antigens available for testing, direct typing of the HSV subtype becomes feasible (Buckmaster et al., 1984; Sutherland et al., 1986). As to date, typing of HSV infections has been considered unnecessary by many physicians. However, with the advent of more antiviral drugs showing different effectiveness on the two viral subtypes the need for accurate typing of HSV in clinical specimens may increase.
173
Acknowledgements We thank Dr. S. Welling-Wester (Department of Medical Microbiology, State University Groningen, The Netherlands) for providing infectious HSV stocks and purified virion preparations and Ms. Marja de Ridder (Municipal Health Service, Amsterdam, The Netherlands) for providing information on the clinical specimens used in this study and Ms. Carla Persons for typing the manuscript.
References Adler-Storthz, K., Kendall, C., Kennedy, R.C., Henkel, R.D. and Dreesman G.R. (1983) Biotin-avidinamplified enzyme immunoassay for detection of herpes simplex virus antigen in clinical specimens. J. Clin. Microbial. 18, 1329-1334. Bos, E.S., Van der Doelen, A.A., Van Rooy, N. and Schuurs, A.H.W.M. (1981). 3,3’, 5,5’Tetramethylbenzidine as an Ames test negative chromogen for horse-radish peroxidase in enzyme immunoassay. J. Immunoassay 2, 187-204. Bos, C.A., Wilterdink, J.B. and Scheffer, A.J. (1986). Detection of viral antigens in cerebrospinal fluid of rabbits with experimental herpes simplex virus type 1 encephalitis. Arch. Virol. 91, 73-81. Buckmaster, E.A., Cranage, M.P., McLean, C.S., Coombs, R.R.A. and Minson A. (1984). The use of monoclonal antibodies to differentiate isolates of herpes simplex types 1 and 2 by neutralization and reverse passive haemagglutination tests. J. Med. Virol. 13, 19%202. Clayton, A., Beckford, U., Roberts, C., Sutherland, S., Druce, A., Best, J. and Shandler. S. (1985). Factors influencing the sensitivity of herpes simplex virus detection in clinical specimens in a simultaneous enzyme-linked immunosorbent assay using monoclonal antibodies. J. Med. Virol. 17, 275-282. Corey, L. and Spear, P.G. (1986). Infections with herpes simplex viruses (second of two parts). N. Engl. J. Med. 314, 749-756. Corey, L., Adams, H.G., Zane. A., Brown, M.D. and Holmes, K.K. (1983) Genital herpes simplex virus infections: clinical manifestations, course and complications. Ann. Intern. Med. 98, 958-972. Dulley, J.R. and Grieve, P.A. (1975). A simple technique for eliminating interference by detergents in the Lowry method of protein determination. Anal. Biochem. 64, 136141. Frame, B., Mahony, J.B. Balachandran, N., Rawls, W.E. and Chemeskey, M.A. (1984). Identification and typing of herpes simplex virus by enzyme immunoassay with monoclonal antibodies. J. Clin. Microbial. 20, 162-166. Gleaves, C.A.. Wilson, D.J., Wold, A.D. and Smith T.F. (1985). Detection and serotyping of herpes simplex virus in MRC-5 cells by use of centrifugation and monoclonal antibodies 16 h postinoculation. J. Clin. Microbial. 21, 29-32. Grillner, L. and Landquist, M. (1983). Enzyme-linked immunosorbent assay for detection and typing of herpes simplex virus. Eur. J. Clin. Microbial. 2, 39-42. Guinnan, M.E. (1986). Sexually transmitted diseases may reverse the revolution. J. Am. Med. Assoc. 255, 1665-1672. King, D.H. and Galasso, G. (1982). Proceedings of a Symposium on acyclovir. Am J. Med. 73, lA-393A. Morgan, M.A. and Smith, T.F. (1984). Evaluation of an enzyme-linked immunosorbent assay for the detection of herpes simplex virus antigen. J. Clin. Microbial. 19, 730-732. Nerurkar, L.S., Namba, M., Brashears, G., Jacob, A.J., Lee, Y.J. and Sever, J.L. (1984). Rapid detection of herpes simplex virus in clinical specimens by use of capture biotin-streptavidin enzymelinked immunosorbent assay. J. Clin. Microbiolf 20, 109-144. Sewell. D.L. and Horn, S.A. (1985). Evaluation of a commercial enzyme-linked immunosorbent assay for the detection of herpes simplex virus. J. Clin. Microbial. 21, 457-458. Smidt, N.J., Dennis, J., Devlin, V., Gallo, D. and Mills, J. (1983) Comparison of direct
174 immunofluorescence and direct immunoperoxidase procedures for detection of herpes simplex virus antigen in lesion specimens. J. Clin. Microbial. 18, 445-448. Steinbuch, M. and Audran, R. (1969). The isolation of IgG from mammalian sera with the aid of caprylic acid. Arch. Biochem. Biophys. 134, 279-284. Sutherland, S.. Morgan, B., Mindel, A. and Chan. W.L. (1986) Typing and subtyping of herpes simplex isolates by monoclonal fluorescence. J. Med. Virol. 18, 235-245. Van Loon, A.M., Heessen. F.W.A., Van der Logt, J.Th. M. and Van der Veen. J. (1981). Direct enzyme-linked immunosorbent assay that uses peroxidase-labeled antigen for determination of immunoglobulin M antibody to cytomegalovirus. J. Clin. Microbial. 13, 416-422. Warford. A.L., Levy, R.A. and Rekrut. K.A. (1984). Evaluation of a commercial enzyme-linked immunosorbent assay for detection of herpes simplex virus antigen. J. Clin. Microbial. 20,490-493. Warford, A.L., Levy, R.A., Strong, C.A. and Rekrut, K.A. (1985). Comparison of two commercial enzyme-linked immunosorbent assays for the detection of herpes simplex virus antigen. Am. J. Clin. Pathol. 85, 229-232. Welling-Wester, S., Popken-Boer. T., Wilterdink, J.B., Van Beeumen, J. and Welling. G.W. (198.5). Isolation by high-performance liquid chromatography and partial characterization of a 57OCQ-dalton herpes simplex virus type 1 polypeptide. J. Virol. 54, 265-270. Ziegler, T., Meurman. O.H., Arstila, P.P. and Halonen, P.E. (1983) Solid-phase enzyme-immunoassay for the detection of herpes simplex virus antigens in clinical specimens. J. Virol. Methods. 7, l-9.