ARTICLE
Detection of localized, clinically atypical cutaneous infections with
varicella zoster virus (VZV) has proven difficult, as serum antibody tests
are sometimes not suitable for that purpose. Direct identification of
the virus can be done with cell culture, which is rather time-consuming,
electron microscopy, or antigen detection, quite often based on immunofluorescence
techniques [1-7]. Reported sensitivities of different techniques vary
widely. In addition, amplification of DNA from skin swabs by polymerase
chain reaction (PCR) has also been established [6-10]. However, due to
its labor-intensity, PCR is still a comparatively time-consuming and therefore
expensive method, even if established in a routine manner. The purpose
of this study was to compare detection of VZV from skin swabs by PCR versus
immunofluorescence in zoster patients diagnosed clinically, with respect
to sensitivity and applicability in routine diagnostics.
Materials and methods
Patients
Cutaneous swabs were taken from lesions of patients visiting the Department
of Dermatology, University of Ulm, Germany, for suspicion of varicella
zoster or herpes simplex virus infections. The lesions the swabs were
taken from consisted of erythema, blisters, pustules or crusts.
Immunofluorescence (direct immunofluorescent
antibody test)
Thorough swabs were taken from the cutaneous lesions, fixed on a glass
slide with acetone for 5 min and washed 5 min with phosphate buffered
saline solution pH 7.2 (PBS). Then the specimen was overlaid with an anti-VZV
mouse antibody (ade labware, Munich, Germany) diluted 1:40 with PBS and
incubated in a moist chamber at room temperature for 20 min. Cross-reactivity
of the anti-VZV antibody with herpes simplex virus types 1 and 2, cytomegalovirus
and Epstein-Barr virus had been excluded by the manufacturer. After washing
with PBS for 5 min the specimen was overlaid with a FITC-labeled anti-mouse
antibody from rabbit diluted 1:30 with PBS/10% rabbit serum (ade labware)
and incubated at room temperature for 20 min in a moist chamber. After
washing with PBS for 5 min the moist specimen was investigated under a
fluorescence microscope.
Criteria for investigation
A specimen was considered positive if fluorescence was detectable in
three or more cells. Fluorescence was detectable either on the cell margin
or throughout the whole cell surface (Fig.
1).
Specimens for PCR
Dry skin swabs were taken from the same location as for immunofluorescence
and at the same time. They were thoroughly rinsed with 400 mul of sterile
distilled water, and from 200 mul of this solution DNA was extracted.
DNA-extraction and PCR
DNA was extracted from 200 mul of the swab solution using the Quiagen
blood kit (Quiagen, Hilden, Germany) following the manufacturer's protocol.
PCR was performed as described [11] in a GeneAmp 9700 thermocycler (Applied
Biosystems, Weiterstadt, Germany). The final PCR-mixture contained MgCl2
2.5 mM, TRIS 10 mM, KCl 50 mM, 0.5 muM of each primer, 200 muM each of
dATP, dTTP, dGTP and dCTP, and Taq polymerase (Boehringer Mannheim, Penzberg,
Germany) 25 mU/mul. The downstream-primer was 5'-labeled with digoxigenin.
Each PCR-tube contained 500 molecules of internal controls (= mimics)
per 40 mul, consisting of the identical primer sequences as the target,
but of a different intermediate sequence. PCR consisted of a first heating
step (95° C for 5 min), 42 amplification cycles (95° C for 15
seconds, 60° C for 30 seconds, 72° C for 30 seconds) and one
final extension step (72° C for 7.7 min). Primer sequences were ATGTCCGTACAACATCAACT
(upstream) and CGATTTTCCAAGAGAGACGC (downstream) for VZV [12], and GAAGAGCCAAGGACAGGTAC
(upstream) and CAACTTCATCCACGTTCACC (downstream) for beta-globin [13].
Fragment lengths were 267 bp for VZV and 546 bp for the VZV-mimic, and
268 bp for beta-globin and 546 bp for the beta-globin-mimic.
Construction and stabilization of internal controls
(mimics)
DNA-mimics were constructed using the PCR MIMIC Construction Kit (Clontech,
Palo Alto, CA, USA) following the manufacturer's protocol. These mimic
fragments were stabilized by cloning them into a plasmid using the T/A
Cloning Kit from Invitrogen (San Diego, CA, USA) following the manufacturer's
protocol. Plasmid DNA was extracted using the High Pure Plasmid Isolation
Kit (Boehringer Mannheim, Penzberg, Germany) and quantified by reading
the optical density (OD) at 260 nm. The molar concentration was calculated,
and the mimics were appropriately diluted.
PCR ELISA
This assay was performed using the principle of a commercially available
system (PCR ELISA, Boehringer) following the manufacturer's protocol.
In brief, two parts of 10 mul each of the respective PCR product were
denatured with 10 mul alkaline solution ("denaturation solution") for
10 min at room temperature and hybridized with 0.75 pmol of a 5'-biotinylated
oligonucleotide specific for the target (VZV or beta-globin) or the mimic,
respectively, in a total volume of 120 mul. Of this mixture 100 mul were
incubated in a streptavidin-coated microtiterplate at 55° C for one
hour, washed five times (washing buffer included in the kit), incubated
with 100 mul of a peroxidase-conjugated anti-digoxigenin antibody at 37°
C for 30 min, washed five times and incubated with 100 mul of the substrate
for peroxidase, 3,3'-5,5'-tetramethylbenzidine (TMB), at room temperature
for 15 min. Then the reaction was stopped by adding 100 mul of 2 M hydrochloric
acid, and the optical density was measured at a wavelength of 450 nm.
Sequences of the 5'-biotinylated oligoprobes were GGTGGAGACGACTTCAATAGC
(VZV), ACACAACTGTGTTCACTAGC (beta-globin), and CAAGTTTCGTGAGCTGATTG (mimic).
Agarose gel electrophoresis
If detection by ELISA was positive the respective products of VZV-PCR
were separated on a 2% agarose gel containing 0.5 mug/ml ethidium bromide
and visualized under UV-light.
Criteria for PCR analysis
A sample was considered positive if color development was visible in
the target ELISA (optical density > 0.1).
A sample was considered negative if no color developed in the target
ELISA and if the mimic ELISA and beta-globin-ELISA were positive.
Optical densities of the negative control always were below 0.1.
Results
A total of 84 consecutive patients presenting with cutaneous lesions
suggesting viral infection were routinely investigated for the presence
of VZV both by immunofluorescence and PCR. An example for immunofluorescence
is given in Figure 1.
Detection of PCR-products was done primarily by ELISA, which is much more
sensitive than agarose gel electrophoresis. Only ELISA-positive PCR-products
were further separated on an agarose gel. Of these 84 samples three samples
were positive by immunofluorescence and negative by PCR. However, in these
samples PCR for herpes simplex virus (HSV) was performed additionally
(data not shown), and all of them were positive for HSV. Clinical diagnoses
were bullous drug reaction, herpes simplex recidivans in loco and
acute postoperative infection with HSV, infection with VZV had not been
suspected at all. For these reasons the most likely explanation for these
discrepant results is a mistake in the laboratory, probably during immunofluorescence.
These samples were excluded from the analysis. Four additional patients
had herpes simplex virus infections clinically. In these patients immunofluorescence
was positive for HSV and negative for VZV (data not shown). Therefore
false-positive results due to unspecific cell fluorescence is highly unlikely.
In an additional five samples neither VZV-DNA nor beta-globin-DNA could
be detected. Thus correctness of DNA extraction can not be proven, and
these samples were also excluded.
Of the remaining 76 samples, 51 were positive by PCR; 43 were PCR-positive
both by ELISA and agarose gel (84%), and 8 were PCR-positive only by ELISA
(16%). Using immunofluorescence only 34 positive samples were found (67%).
Every sample positive by immunofluorescence was also PCR-positive, and
no specimen was positive by immunofluorescence and false-negative by PCR.
One inherent problem of immunofluorescence was the absence of cells
or a low number of cells present on the slide. In routine diagnostics
these specimens were counted as negative with the remark "no cells present"
or "few cells present". The correct interpretation would have been "no
analysis possible due to lack of material". When excluding these specimens
from the analysis 60 samples remain. Of these 44 were PCR-positive by
ELISA (44 = 100%). Of these ELISA-positive specimens 37 (84%) were also
positive by agarose gel, and 34 (77%) were positive by immunofluorescence.
Of the 37 samples PCR-positive both by ELISA and agarose gel 5 were negative
for beta- globin (14%) whereas of the 7 samples PCR-positive only by ELISA,
4 were beta-globin-negative (57%). Results are summarized in Table
I.
However, when only the results of the easy-to-perform agarose gel electrophoresis
were considered, 29 samples were positive both by PCR and immunofluorescence,
8 samples only by PCR, and 5 samples only by immunofluorescence. Two samples
PCR-positive by ELISA were missed by agarose gel and immunofluorescence.
Discussion
In the present study immunofluorescence was compared with PCR (detection
of PCR products by ELISA and, if positive, by agarose gel in parallel)
for detection of cutaneous VZV-infections. Immunofluorescence is a fast
and comparatively inexpensive method and is, among others, used for detection
of cutaneous VZV infections. Its reported sensitivity varies between 80%
and 100% and is far better than cell culture or antibody assays [1-4,
6]. Up to now immunofluorescence and PCR have only been compared for laboratory
diagnosis of zoster using vesicle specimens [6]. Also PCR products can
be detected by different methods. An easy-to-perform and fast, but not
very sensitive method is agarose gel electrophoresis. Sensitivity of detection
can be improved by hybridization with radioactive probes or by using an
ELISA method, with the sensitivity reaching a few or even single pre-PCR
copies. This high sensitivity is especially important in patients presenting
with initial or clinically abortive manifestations of VZV infection, such
as segmental erythema without blisters, as the viral amount on such lesions
has to be considered very low. However, this high sensitivity is associated
with the danger of detecting clinically irrelevant colonization. Also
criteria for correct interpretation of PCR results have to be observed.
Apart from correct negative and positive controls the presence of amplifiable
DNA should be proven. This can easily be done by detecting a human house-keeping
gene, for example beta-globin. However, in the absence of beta-globin-DNA
amplification of VZV from skin swabs is possible, as 9 of the 44 samples
positive for VZV were negative for beta-globin. This discrepancy most
likely is due to the absence or small amount of human DNA in the skin
swabs, as the cells of the stratum corneum do not contain DNA any more.
Also the correct amplification procedure should be proven for every individual
PCR tube as PCR inhibition may occur [14, 15]. This can easily be accomplished
with internal controls. Internal controls, or mimics, contain the same
primer sequences as the target, but a different intermediate sequence
and can thus be differentiated from the target [16]. Due to the identical
primer sequences, mimic and target are competitively amplified [16]. Therefore
mimics can be used as individual positive controls for every PCR-tube.
This PCR system based on competitive amplification of internal controls
and detection by a hybridization-based method complies with the Recommendations
for Employing Molecular Methods in Diagnostic Routine Microbiology Laboratories
and Measures for Internal Quality Assurance [17] of the German Society
for Hygiene and Microbiology on technical PCR performance. Using this
control system a positive sample has to show a positive signal in the
detection system, and a negative sample has to have a negative signal
with the internal controls and beta-globin being positive.
In 16 of 76 samples there were no or only a few cells present for immunofluorescence
analysis, so no clear statement can be made, although these samples had
been termed negative in the routine laboratory. Of the remaining 60 specimens
44 were PCR-positive by ELISA and 37 by agarose gel, and 34 by immunofluorescence.
No sample was positive by immunofluorescence and negative by PCR. As it
is impossible to determine which clinical specimen is really positive
for VZV the total number of samples VZV-positive by PCR or immunofluorescence
(44 specimens) was set as 100% [2]. With this reference, sensitivity of
PCR with ELISA detection was 100% (44 of 44), of PCR with agarose gel
detection 84% (37 of 44), and of immunofluorescence 77% (34 of 44). Of
the 37 specimens PCR-positive by ELISA and agarose gel, 5 were negative
for beta-globin (14%), whereas of the 7 samples PCR-positive only by ELISA,
4 were negative for beta-globin (57%). All these 7 patients had been diagnosed
clinically as zoster. Most likely this high percentage of beta-globin-negative
samples associated with low VZV detection is due to the low amount of
starting material, especially in zoster patients presenting with erythema,
but without blisters, erosions or crusts. Of the cases positive by PCR
and negative by immunofluorescence one had been clinically diagnosed as
varicella, and all others as zoster.
These results demonstrate that immunofluorescence is a fast, comparatively
inexpensive and quite sensitive method for confirmation of cutaneous VZV
infections. Results can be obtained after 1 hr. If care is taken that
enough cells are present for analysis a sensitivity of 77% can be achieved.
Therefore immunofluorescence can be used as a fast screening method. If
immunofluorescence is negative, but zoster is clinically assumed, diagnostic
accuracy can be augmented with PCR which takes longer to perform rendering
it quite expensive. Also conventional agarose gel electrophoresis will
not detect all VZV-positive samples, and detection of PCR products with
ELISA is a very time-consuming procedure taking at least 8 hrs to perform.
CONCLUSION
In conclusion, immunofluorescence is a sufficiently sensitive method
for detection of cutaneous VZV infections. Sensitivity can be improved
by screening specimens negative by immunofluorescence with the more time-consuming
and cost-intensive PCR.
REFERENCES
1. Dahl H, Marcoccia J, Linde A. Antigen detection: the method
of choice in comparison with virus isolation and serology for laboratory
diagnosis of herpes zoster in human immunodeficiency virus-infected patients.
J Clin Microbiol 1997; 35: 347-9.
2. Coffin SE, Hodinka RL. Utility of direct immunofluorescence
and virus culture for detection of varicella-zoster virus in skin lesions.
J Clin Microbiol 1995; 33: 2792-5.
3. Zirn JR, Tompkins SD, Huie C, Shea CR. Rapid detection and
distinction of cutaneous herpesvirus infections by direct immunofluorescence.
J Am Acad Dermatol 1995; 33: 724-8.
4. Brumback BG, Farthing PG, Castellino SN. Simultaneous detection
of and differentiation between herpes simplex and varicella-zoster viruses
with two fluorescent probes in the same test system. J Clin Microbiol
1993; 31: 3260-3.
5. Perez JL, Garcia A, Niubo J, Salva J, Podzamczer D, Martin
R. Comparison of techniques and evaluation of three commercial monoclonal
antibodies for laboratory diagnosis of varicella-zoster virus in mucocutaneous
specimens. J Clin Microbiol 1994; 32: 1610-3.
6. Sauerbrei A, Eichhorn U, Schacke M, Wutzler P. Laboratory
diagnosis of herpes zoster. J Clin Virol 1999; 14: 31-6.
7. Beards G, Graham C, Pillay D. Investigation of vesicular rashes
for HSV and VZV by PCR. J Med Virol 1998; 54: 155-7.
8. Rübben A, Baron JM, Grussendorf-Conen EI. Routine detection
of herpes simplex virus and varicella zoster virus by polymerase chain
reaction reveals that initial herpes zoster is frequently misdiagnosed
as herpes simplex. Br J Dermatol 1997; 137: 259-61.
9. Jacobs JJ, Folkers E, Vreeswijk J. Detection of varicella-zoster
virus and herpes simplex virus by the polymerase chain reaction with degenerate
primers. J Virol Methods 1999; 83: 155-67.
10. Nahass GT, Mandel MJ, Cook S, Fan W, Leonardi CL. Detection
of herpes simplex and varicella-zoster infection from cutaneous lesions
in different clinical stages with the polymerase chain reaction. J
Am Acad Dermatol 1995; 32: 730-3.
11. Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn
GT, Mullis KB, Erlich HA. Primer directed enzymatic amplification of DNA
with a thermostable DNA polymerase. Science 1988; 239: 487-91.
12. Puchhammer-Stöckl E, Popow-Kraupp T, Heinz FX, Mandl
CW, Kunz C. Detection of varicella-zoster virus DNA by polymerase chain
reaction in the cerebrospinal fluid of patients suffering from neurological
complications associated with chicken pox or herpes zoster. J Clin
Microbiol 1991; 29: 1513-6.
13. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT, Erlich
HA, Arnheim N. Enzymatic amplification of beta-globin genomic sequences
and restriction site analysis for diagnosis of sickle cell anemia. Science
1985; 230: 1350-4.
14. Cone RW, Hobson AC, Huang MLW. Coamplified positive control
detects inhibition of polymerase chain reaction. J Clin Microbiol
1992; 30: 3185-9.
15. Verkooyen RP, Luijendijk A, Huisman WM, Goessens WHF, Kluytmans
JAJW, van Rijsoort-Vos JH, Verbrugh HA. Detection of PCR inhibitors in
cervical specimens by using the AMPLICOR chlamydia trachomatis assay.
J Clin Microbiol 1996; 34: 3072-4.
16. Siebert PD, Larrick JW. Competitive PCR. Nature 1992;
359: 557-8.
17. Roth A, Mauch H, Göbel U. Quality standards for microbiological
diagnostic techniques for infectious disesases. Nucleic acid amplification
techniques. Gustav Fischer Stuttgart, Jena, Lübeck, Ulm, 1997.
|
Printable version
