125
DIAGN MICROBIOLINFECTDIS 1991;14:125-130
VIROLOGY
Optimizing Recovery of Cytomegalovirus in the Shell Vial Culture Procedure Max Arens, Jody Owen, Colleen M. Hagerty, Charles A. Reed, and Gregory A. Storch
We have investigated three factors that may be related to the recovery of cytomegalovirus (CMV) using the shell vial culture procedure. First, we compared fluorescent-antibody staining of shell vial cultures using a monoclonal antibody to a CMV immediate early antigen at 16 vs 40 hr after inoculation. Of 332 routinely submitted specimens cultured in duplicate and stained at the different times, 25 (7.5%) were positive at 16 hr and 32 (9.6%) were positive at 40 hr. The increased yield was 28%. Second, we analyzed the effect of using duplicate shell vials (both stained at 40 hr) for all routinely submitted CMV cultures. During a 6-month period, 272 (12.5%) of the 2157 cultures processed with duplicate shell vials were positive, including 222 positive in both vials
and 50 positive in only one. Assuming that a single-vial setup would have detected 50% of those positive in only one of the two vials, the increased yield attributable to the duplicate vial was estimated at 10% (25/(222 + 25)). Third, we investigated the effects of seeding density and culture age on the shell vial assay. Cell age of greater than 1 day was associated with a decrease in sensitivity both in cultures that were confluent and in those that were subconfluent at the time of inoculation. Incorporating these findings in the routine shell vial culture procedure used in our Clinical Virology Laboratory has resulted in a greater overall detection of CMV in shell vial cultures than in conventional 6-week tube cultures.
INTRODUCTION
1989; Paya et al., 1988; Thiele et al., 1987). We report here an investigation of several factors that affect sensitivity, including the time of staining, number of shell vials per specimen, and age of the cells in the shell vial. Our studies were carried out in the setting of a clinical virology laboratory, thus allowing us to demonstrate the impact of optimizing the procedure on the detection of CMV in clinical sampies.
Cytomegalovirus (CMV) infections are being diagnosed more frequently in immunosuppressed patients. With the development of effective drug therapy (Bach et al., 1985; Balfour et al., 1989) their rapid recognition has assumed great clinical importance. The speed and sensitivity of diagnosis of CMV infections have been significantly improved with the use of the centrifugation shell vial procedure and detection of infected nuclei by fluorescent antibodies (Gleaves et al., 1985). We and others have found that the sensitivity of this procedure is markedly affected by a number of parameters (Ashley et al., From the Department of Pediatrics, WashingtonUniversity Medical School, St. Louis, MO 63110, USA. Address reprint requests to Dr. M. Arens, Department of Pediatrics, WashingtonUniversityMedicalSchool, St. Louis, MO 63110, USA. Received and accepted 1 August 1990. © 1991 ElsevierScience Publishing Co., Inc. 655 Avenue of the Americas, New York, NY 10010 0732-8893/91/$3.50
MATERIALS A N D M E T H O D S Cells and Virus All MRC-5 cells were obtained commercially (Viromed, Minneapolis, MN) either in shell vials on 12ram glass coverslips, as monolayers in plastic tissue culture flasks, or in roller tubes. CMV (Towne, ATCC no. VR-977) obtained from the American Type Culture Collection (Rockville, MD) was used as a positive control for routine clinical assays and was also used in all of the cell age studies.
126
M. Arens et al.
,T, u
-m
120
100
"£-"--a.
Z ~-
o u
80 \
~m
~ CL 80
\ KO
\
100
~, 6o C3~ ~ 0 a-
40
~
20
k
z
AD~69 L
40,
\. \\
"LG,", ""N~N'o- ........ o
~" 20 r.T©~ NE ~_ LU a-
2
°o
4
6
8
10
$
o AGE OF MRC-5 CELLS (days)
12
AGE OF MRC-5 CELLS (days)
FIGURE 1 Effect of the age of fibroblast cells on their sensitivity to infection by CMV. Shell vials were seeded with 100,000 cells of the same passage of MRC-5 cells on the days indicated prior to infection. All shell vials were infected at the same time and with the same dilution of virus. Day-1 positive nuclei were K.O., 26; AD169, 94; and Towne, 341. Each point represents the mean of four shell vials.
FIGURE 2 Sensitivity of MRC-5 shell vial cultures to CMV (Towne). Effect of age and seeding density. Shell vials were seeded with 10,000 cells (squares), 25,000 cells (solid circles), 75,000 cells (triangles), or 150,000 cells (open circles) on the day indicated prior to infection. Confluencies of cell monolayers at the time of infection were squares: day 1, 5%; day 2, 15%; day 3, 20%; and 4, 30%; solid circles: day 1, 20%; day 2, 40%; day 3, 50%, and day 4, 60%; triangles: day 1, 50%; day 2, 80%; day 3, 90%; and day 4, 95%; and open circles: day 1, 80%; day 2, 100%; day 3, 100%; and day 4, 100%.
Shell Vial Procedure After removal of media, shell vials were inoculated by adding 0.1 ml of CMV (Towne strain) or 0.25 ml of clinical specimen, centrifuging at 700 g for 45 rain and feeding with 1 ml of media (Earle's minimal essential media with 10% fetal calf serum and 200 units/ml penicillin, 200 i~g/ml streptomycin and 1.25 ~g/ml amphotericin). At the end of the incubation period, coverslips were washed with phosphatebuffered saline (PBS), fixed with acetone at -20°C, and stained with purified mouse monoclonal antibody against a CMV immediate early nuclear protein (DuPont, Doraville, GA; cat. no. NEA9221) and F1TCconjugated goat anti-mouse IgG antibody, heavy and light chain (Organon Technika, Cappel Division, Durham, NC; cat. no. 1211-0081).
Cell Age Studies For in-house preparation of shell vials used in cell age studies, we obtained a 150-cm 2 culture flask of low-passage (for instance, passage 21) MRC-5 cells from a commercial source (Viromed). This flask was split into 22 small culture flasks, and one of these was used to prepare the shell vials for each day of the time-course study. This ensured that all shell vials used for a particular experiment were from the same cell passage. Flasks and shell vials were fed with fresh media every 7 days. Shell vials used in the experiment shown in Figure 1 were seeded at 100,000 cells/vial. Shell vials used in the experiment of Figure 2 were seeded at the densities indicated in the legend. All shell vials in a particular experiment were infected at the same time and with the same dilution of stock virus. All coverslips were stained
at 40 hr after inoculation as described above. Quadruplicate shell vials were used at each point in all experiments.
Clinical Specimens Clinical specimens used were routinely submitted to the Clinical Virology Laboratory at St. Louis Children's Hospital. All urine specimens were neutralized by the addition of I N NaOH using 0.3% phenol red as the pH indicator. Lymphocytes from specimens of heparinized blood were purified by centrifugation on Mono-Poly Resolving Medium (Flow Labs, McLean, VA) 700 g for 30 rain at 22°C, and washed once with PBS before centrifugation onto the MRC-5 cells in shell vials as described above. Respiratory specimens included throat swabs, nasopharyngeal swabs or aspirates, sputums, bronchoalveolar lavages, bronchial washes, and aspirates. All other specimens received were classified as miscellaneous. These included tissues, cerebrospinal fluids, semen, drainages, and pleural fluids. All tube cultures were held for 6 weeks during which time they were fed with media at 7-day intervals and were observed for viral cytopathic effect everyday for the first 7 days and then weekly thereafter. Tubes were kept on roller drums for the initial 7 days and then on stationary racks until day 42 after inoculation.
Definitions The sensitivity of the shell vial culture in comparison to tube culture was defined as the percentage of all
Optimizing Recovery of CMV in Shell Vials
TABLE 1
127
Results of Performing FluorescentAntibody Staining on Duplicate Shell Vial Cultures at 16 or 40 hr After Inoculation
TABLE 2
Results of First/Second Shell Vial Cultures Pos/Pos/Pos/Total Neg/Tox/Tox/Total Total Pos Neg Tox Pos Neg Neg Tox Neg Done
No. Positive at Specimen
No. Inoculated
16 hr
40 hr
Urine Buffycoat Respiratory Other
140 124 47 21
19 3 3 0
24 4 4 0
Total
332
25a
32a
Results of Duplicate Shell Vial Cultures on the Recovery of CMV
Urine Buffycoat Respiratory Other
127 8 21 156 395 78 37 13 3 53 479 51 46 1 3 50 506 11 12 1 0 13 185 13
81 27 43 16
473 530 517 198
710 610 610 227
Total
222 23 27 272 1565 153 167 1718 2157
Pos, positive; Neg, negative; and Tox, toxicity.
ap = 0.10, McNemar's test. tube culture positives that were also positive by shell vial culture. The specificity of the shell vial culture in comparison to tube culture was defined as the percentage of all tube culture negatives that were also negative by shell vial culture. Specimens whose shell vial culture could not be read because of toxicity to the cell monolayer were included in the calculation of sensitivity but were excluded from the calculation of specificity. Specimens in which viruses other than CMV were detected in the tube culture were excluded from all analyses (93 specimens during the time period of April 1987-March 1988 (2% of the total specimens), and 80 specimens during the time period of April 1988-March 1989 (1.8% of total).
RESULTS First, we investigated the effect of performing fluorescent-antibody staining at two different times after inoculation of the culture. As shown in Table 1, of 322 specimens routinely submitted for CMV culture that were stained at both 16 and 40 hr after inoculation, 25 were positive at 16 hr compared to 32 at 40 hr, an increase of 28% (p = 0.10; McNemar's test). After the completion of the first experiment, all specimens routinely submitted for CMV culture were set up on duplicate shell vials that were both stained at 40 h after inoculation. As shown in Table 2, of 2157 specimens, 272 were positive in one or both shell vial cultures, including 222 in which both shell vial cultures were positive, 23 in which one shell vial culture was positive and the other was negative, and 27 in which one shell vial culture was positive and the other was unreadable because of toxicity. If it is assumed that 50% of the specimens in which only one shell vial culture was positive would have been detected if only one shell vial culture had been set up per specimen, then the additional yield attributable to the use of two shell vial cultures per spec-
imen was 25 cultures (50% of 23 + 27), or an increase of 10% compared to the number of cultures that would have been positive if only one shell vial had been set up per specimen [25/(222 + 25)]. We also evaluated the effect of the age of the fibroblast cells on their sensitivity to infection by CMV as detected by the shell vial procedure. As shown in Figure 1, the results depended on the test strain used. With each of the two lab strains (Towne and AD169), the sensitivity was maximal at day 1 after seeding, and declined sharply thereafter, with day-12 ceils having only 5%-25% as many positive nuclei as day-1 cells. Experiments with a patient isolate (K.O.) yielded a curve that was initially relatively fiat, with day-8 cells having 77% as many positive nudei as day-1 cells. By day 12, however, there was only 40% as many positive nuclei as in day-1 cells. We performed an additional experiment to investigate the relative effects of cell age and cell density on the decreasing sensitivity of cells with increasing age of the culture. The Towne strain was used in this experiment. The decreasing sensitivity with age was apparent in cultures seeded at four different densities yielding a range of confluencies, suggesting that the effect was independent of cell density or confluency of the monolayer (Figure 2). In the spring of 1988, we made changes in our routine laboratory procedures to incorporate staining of shell vials at 40 hr instead of 16 hr after inoculation, inoculation of all specimens onto duplicate shell vials, and age monitoring of shell vials so that shell vials older than 12 days from the time of seeding were not used. A comparison of 12-month periods before and after these changes is shown in Table 3. With approximately the same number of specimens in each time period and with no major changes in the personnel who read the coverslips, the sensitivity of the shell vial procedure compared
128
TABLE 3
M. Arens et al.
Results of Shell Vial and Tube Cultures for CMV During Two Different Time Periods
Shell Vial Culture Result for Specimens With Tube Culture Pos Time Period Specimen
Tube Culture Neg
Pos Neg Toxic Pos Neg Toxic Sensitivity Specificity % Toxic
April 1987-March 1988a Urine 133 Buffy coat 38 Respiratory 33 Other 8 Total
Shell Vial Test Parameters (Compared to Tube Culture Results)
86 80 31 5
39 9 1 0
33 15 1 5
924 353 1173 97 979 61 434 30
52 30 51 62
97 99 >99 99
25 8 6 6
212 202
49
54 3510 541
46
98
13
April 1988-March 1989b Urine 222 Buffy coat 76 Respiratory 77 Other 27
13 46 14 3
13 2 7 0
119 23 28 5
926 140 1071 32 1015 79 464 31
90 61 79 90
89 98 97 99
11 3 7 6
Total
76
22
175 3475 282
80
95
7
402
aDuring the time period of April 1987-March 1988, one shell vial culture was inoculated per specimen, staining was at 16 hr after inoculation, and the age of ceils in the shell vial cultures was not monitored. bDuring the time period of April 1988-March 1989, two shell vials were inoculated per specimen, staining was at 40 hr after inoculation, and shell vial cultures 13 days or older prior to inoculation were not used.
to tube culture increased from 46% to 80%, a n d the percentage of specimens for which a shell vial result could not be r e p o r t e d because of toxicity decreased from 13% to 7%. To c o m p a r e the sensitivity of b o t h the shell vial and tube cultures, w e u s e d the data from both time periods to calculate the sensitivity and specificity of both the shell vial culture a n d tube culture in reference to the total n u m b e r of cultures f o u n d to be positive for CMV by either m e t h o d . It is likely that specimens in which the shell vial culture was positive and the conventional culture was negative truly contained CMV, because m o s t of these specimens were from transplant patients w h o h a d other specimens positive for CMV by tube culture shortly before or after the specimen that yielded discrepant results. The results are s h o w n in Table 4. In this analysis, the sensitivities of the shell vial and tube culture w e r e 51% and 90%, respectively, in the first time period and 86% and 74% in the second time period, indicating that d u r i n g the second period, shell vial cultures w e r e more sensitive than w e r e tube cultures for the detection of CMV. The total percentage of cultures positive for CMV increased from 11.3% in the first time period, to 15.2% in the second time period, a n d was attributable largely to
the increased detection in shell vial cultures in the second time period.
DISCUSSION The shell vial culture fluorescent-antibody proced u r e has h a d a decisive impact o n the rapid diagnosis of CMV infections. H o w e v e r , it is clear that optimal sensitivity of the p r o c e d u r e can be attained only by attention to a n u m b e r of specific aspects of the procedure. The optimal time for cell fixing a n d staining in shell vials is controversial. In the initial descriptions of the procedure, fluorescent-antibody staining was p e r f o r m e d at 16 hr after inoculation, with excellent sensitivity (Gleaves et al., 1985; Shuster et al., 1985). However, other investigators have reported that only 61% of the ultimately shell vial positive specimens were positive at 16 hr (Thiele et al., 1987). A recent report 0 e s p e r s e n et al., 1989) described a multisite s t u d y using a different commercial m o n o c l o n a l antibody (Syva), and d e m o n s t r a t e d an a p p a r e n t effect of cell age on the optimal staining time. In this study, staining at 16-24 hr was optimal on cells p r e p a r e d in the site laboratory, but 40-48 hr was optimal for
Optimizing Recovery of CMV in Shell Vials
TABLE 4
129
Sensitivity of Shell Vial and Tube Cultures Calculated in Relation to Total CMV Isolation No. (%) of Total Cultures Pos for CMV
Time Period April 1987-March 1988 April 1988-March 1989
Total Pos Cultures for CMVa Shell Vial Tube Performed n (%)b Culture Culture 4568 4433
Sensitivity (% of Total Cultures Detected) of ShellVial Tube Culture Culture
517 (11.3) 266 (5.8) 463 (10.1) 675 (15.2) 577 (13.0) 500 (11.3)
51 86
90 74
qncludes specimens positive in either shell vial culture, tube culture, or both. bpercent of cultures performed.
commercial cells. The results of the present study, which used commercial cells, are consistent in that staining at 40 hr resulted in a 28% increase in positive cultures compared to staining at 16 hr. Paya et al (1988) have investigated the effects of inoculating multiple shell vials for detection of CMV using different types of specimens. Our data support their conclusion that multiple shell vials should be used. The use of duplicate shell vials on all of 2157 specimens yielded 50 specimens that were positive on only one shell vial. Presumably, about onehalf of these would have been missed had they been inoculated on only one shell vial. The increased yield attributable to use of duplicate shell vials was 10%. The results of Thiele at al. (1987) indicated that the sensitivity of MRC-5 cells to CMV declined significantly by about 12 days after seeding. Using lowpassage MRC-5 cells in shell vials (as opposed to 24well plates used by Thiele) and the Towne or AD169 strains, we found that sensitivity declined as early as 2 days after seeding. However, we also found that this phenomenon was dependent on the viral strain used. Experiments using a patient isolate showed a less precipitous decline in sensitivity with increasing cell age than when the lab strains were used. Both Thiele's results and our own experience prompted us to monitor the age of cells used in shell
vial cultures in the clinical laboratory, and not to use any that were older than 12 days. It was not possible to quantitate the contribution of this policy to the overall increase in CMV detection observed in the Clinical Virology Laboratory. It may be that preparation of shell vial cultures in the laboratory rather than reliance on commercial cells would permit use of even younger cells and increase overall detection further. However, it does not seem at all reasonable to advocate making shell vials every day so that day1 cells would always be available. The net impact of staining at 40 hr, inoculating duplicate shell vials and monitoring the age of cells was to increase the sensitivity of the shell vial culture from 46% to 80% in comparison to tube culture. It appears that staining at 40 hr rather than at 16 hr after inoculation may have been the single most important factor. However, the use of duplicate shell vials for each specimen resulted in a further increase in detection. Finally, the use of young cells is important for maximum sensitivity. We believe that attention to each of these aspects of the shell vial procedure is required to optimize detection of CMV. Our data indicate that when conditions are optimized, the shell vial culture can be more sensitive than tube cultures incubated for 6 weeks for the detection of CMV in patient specimens.
REFERENCES Ashley R, Peterson E, Abbo H, Gold D, Corey L (1989) Comparison of monoclonal antibodies for rapid detection of cytomegalovirus in spin-amplified plate cultures. J Clin Microb 27:2858-2860. Bach MC, Bagwell SP, Knapp NP, Davis KM, Hedstrom PS (1985) A-(1,3-dihydroxy-2-propoxymethyl)guanine for cytomegalovirus infections in patients with acquired immunodeficiency syndrome. Ann Intern Med 103:381382.
Balfour HD, Chace BA, Stapleton JT, Simmons RL, Fryd DS (1989) A randomized, placebo-controlled trial of oral acyclovir for the prevention of cytomegalovirus disease in recipients of renal allografts. N Engl J Med 320:13811387. Gleaves CA, Smith T, Shuster EA, Pearson GR (1985) Comparison of standard tube and shell vial cell culture techniques for the detection of cytomegalovirus in clinical specimens. J Clin Microbiol 21:217-221.
130
Jespersen DJ, Drew WL, Gleaves CA, Meyers JD, Warford AL, Smith TF (1989) Multisite evaluation of a monoclonal reagent (Syva) for rapid diagnosis of cytomegalovirus in the shell vial assay. J Clin Microbio127:15021505. Paya CV, Wold AD, Ilstruk DM, Smith TF (1988) Evaluation of number of shell vial cell cultures per clinical specimen for rapid diagnosis of cytomegalovirus infection. J Clin Microbiol 26:198-200. Thiele GM, Bicak MS, Young A, Kinsey J, White RJ, Purtilo
M. Arens et al.
DT (1987) Rapid detection of cytomegalovirus by tissue culture, centrifugation, and immunofluorescence with a monoclonal antibody to an early nuclear antigen. J Virol Methods 16:327-338. Shuster EA, Beneke JS, Tegtmeyer GE, Pearson GR, Gleaves CA, Wold AD, Smith TF (1985) Monoclonal antibody for rapid laboratory detection of cytomegalovirus infections: characterization and diagnostic application. Mayo Clinic Proc 60:577-585.
DIAGN MICROBIOLINFECTDIS 1991;14:125-130
VIROLOGY
Optimizing Recovery of Cytomegalovirus in the Shell Vial Culture Procedure Max Arens, Jody Owen, Colleen M. Hagerty, Charles A. Reed, and Gregory A. Storch
We have investigated three factors that may be related to the recovery of cytomegalovirus (CMV) using the shell vial culture procedure. First, we compared fluorescent-antibody staining of shell vial cultures using a monoclonal antibody to a CMV immediate early antigen at 16 vs 40 hr after inoculation. Of 332 routinely submitted specimens cultured in duplicate and stained at the different times, 25 (7.5%) were positive at 16 hr and 32 (9.6%) were positive at 40 hr. The increased yield was 28%. Second, we analyzed the effect of using duplicate shell vials (both stained at 40 hr) for all routinely submitted CMV cultures. During a 6-month period, 272 (12.5%) of the 2157 cultures processed with duplicate shell vials were positive, including 222 positive in both vials
and 50 positive in only one. Assuming that a single-vial setup would have detected 50% of those positive in only one of the two vials, the increased yield attributable to the duplicate vial was estimated at 10% (25/(222 + 25)). Third, we investigated the effects of seeding density and culture age on the shell vial assay. Cell age of greater than 1 day was associated with a decrease in sensitivity both in cultures that were confluent and in those that were subconfluent at the time of inoculation. Incorporating these findings in the routine shell vial culture procedure used in our Clinical Virology Laboratory has resulted in a greater overall detection of CMV in shell vial cultures than in conventional 6-week tube cultures.
INTRODUCTION
1989; Paya et al., 1988; Thiele et al., 1987). We report here an investigation of several factors that affect sensitivity, including the time of staining, number of shell vials per specimen, and age of the cells in the shell vial. Our studies were carried out in the setting of a clinical virology laboratory, thus allowing us to demonstrate the impact of optimizing the procedure on the detection of CMV in clinical sampies.
Cytomegalovirus (CMV) infections are being diagnosed more frequently in immunosuppressed patients. With the development of effective drug therapy (Bach et al., 1985; Balfour et al., 1989) their rapid recognition has assumed great clinical importance. The speed and sensitivity of diagnosis of CMV infections have been significantly improved with the use of the centrifugation shell vial procedure and detection of infected nuclei by fluorescent antibodies (Gleaves et al., 1985). We and others have found that the sensitivity of this procedure is markedly affected by a number of parameters (Ashley et al., From the Department of Pediatrics, WashingtonUniversity Medical School, St. Louis, MO 63110, USA. Address reprint requests to Dr. M. Arens, Department of Pediatrics, WashingtonUniversityMedicalSchool, St. Louis, MO 63110, USA. Received and accepted 1 August 1990. © 1991 ElsevierScience Publishing Co., Inc. 655 Avenue of the Americas, New York, NY 10010 0732-8893/91/$3.50
MATERIALS A N D M E T H O D S Cells and Virus All MRC-5 cells were obtained commercially (Viromed, Minneapolis, MN) either in shell vials on 12ram glass coverslips, as monolayers in plastic tissue culture flasks, or in roller tubes. CMV (Towne, ATCC no. VR-977) obtained from the American Type Culture Collection (Rockville, MD) was used as a positive control for routine clinical assays and was also used in all of the cell age studies.
126
M. Arens et al.
,T, u
-m
120
100
"£-"--a.
Z ~-
o u
80 \
~m
~ CL 80
\ KO
\
100
~, 6o C3~ ~ 0 a-
40
~
20
k
z
AD~69 L
40,
\. \\
"LG,", ""N~N'o- ........ o
~" 20 r.T©~ NE ~_ LU a-
2
°o
4
6
8
10
$
o AGE OF MRC-5 CELLS (days)
12
AGE OF MRC-5 CELLS (days)
FIGURE 1 Effect of the age of fibroblast cells on their sensitivity to infection by CMV. Shell vials were seeded with 100,000 cells of the same passage of MRC-5 cells on the days indicated prior to infection. All shell vials were infected at the same time and with the same dilution of virus. Day-1 positive nuclei were K.O., 26; AD169, 94; and Towne, 341. Each point represents the mean of four shell vials.
FIGURE 2 Sensitivity of MRC-5 shell vial cultures to CMV (Towne). Effect of age and seeding density. Shell vials were seeded with 10,000 cells (squares), 25,000 cells (solid circles), 75,000 cells (triangles), or 150,000 cells (open circles) on the day indicated prior to infection. Confluencies of cell monolayers at the time of infection were squares: day 1, 5%; day 2, 15%; day 3, 20%; and 4, 30%; solid circles: day 1, 20%; day 2, 40%; day 3, 50%, and day 4, 60%; triangles: day 1, 50%; day 2, 80%; day 3, 90%; and day 4, 95%; and open circles: day 1, 80%; day 2, 100%; day 3, 100%; and day 4, 100%.
Shell Vial Procedure After removal of media, shell vials were inoculated by adding 0.1 ml of CMV (Towne strain) or 0.25 ml of clinical specimen, centrifuging at 700 g for 45 rain and feeding with 1 ml of media (Earle's minimal essential media with 10% fetal calf serum and 200 units/ml penicillin, 200 i~g/ml streptomycin and 1.25 ~g/ml amphotericin). At the end of the incubation period, coverslips were washed with phosphatebuffered saline (PBS), fixed with acetone at -20°C, and stained with purified mouse monoclonal antibody against a CMV immediate early nuclear protein (DuPont, Doraville, GA; cat. no. NEA9221) and F1TCconjugated goat anti-mouse IgG antibody, heavy and light chain (Organon Technika, Cappel Division, Durham, NC; cat. no. 1211-0081).
Cell Age Studies For in-house preparation of shell vials used in cell age studies, we obtained a 150-cm 2 culture flask of low-passage (for instance, passage 21) MRC-5 cells from a commercial source (Viromed). This flask was split into 22 small culture flasks, and one of these was used to prepare the shell vials for each day of the time-course study. This ensured that all shell vials used for a particular experiment were from the same cell passage. Flasks and shell vials were fed with fresh media every 7 days. Shell vials used in the experiment shown in Figure 1 were seeded at 100,000 cells/vial. Shell vials used in the experiment of Figure 2 were seeded at the densities indicated in the legend. All shell vials in a particular experiment were infected at the same time and with the same dilution of stock virus. All coverslips were stained
at 40 hr after inoculation as described above. Quadruplicate shell vials were used at each point in all experiments.
Clinical Specimens Clinical specimens used were routinely submitted to the Clinical Virology Laboratory at St. Louis Children's Hospital. All urine specimens were neutralized by the addition of I N NaOH using 0.3% phenol red as the pH indicator. Lymphocytes from specimens of heparinized blood were purified by centrifugation on Mono-Poly Resolving Medium (Flow Labs, McLean, VA) 700 g for 30 rain at 22°C, and washed once with PBS before centrifugation onto the MRC-5 cells in shell vials as described above. Respiratory specimens included throat swabs, nasopharyngeal swabs or aspirates, sputums, bronchoalveolar lavages, bronchial washes, and aspirates. All other specimens received were classified as miscellaneous. These included tissues, cerebrospinal fluids, semen, drainages, and pleural fluids. All tube cultures were held for 6 weeks during which time they were fed with media at 7-day intervals and were observed for viral cytopathic effect everyday for the first 7 days and then weekly thereafter. Tubes were kept on roller drums for the initial 7 days and then on stationary racks until day 42 after inoculation.
Definitions The sensitivity of the shell vial culture in comparison to tube culture was defined as the percentage of all
Optimizing Recovery of CMV in Shell Vials
TABLE 1
127
Results of Performing FluorescentAntibody Staining on Duplicate Shell Vial Cultures at 16 or 40 hr After Inoculation
TABLE 2
Results of First/Second Shell Vial Cultures Pos/Pos/Pos/Total Neg/Tox/Tox/Total Total Pos Neg Tox Pos Neg Neg Tox Neg Done
No. Positive at Specimen
No. Inoculated
16 hr
40 hr
Urine Buffycoat Respiratory Other
140 124 47 21
19 3 3 0
24 4 4 0
Total
332
25a
32a
Results of Duplicate Shell Vial Cultures on the Recovery of CMV
Urine Buffycoat Respiratory Other
127 8 21 156 395 78 37 13 3 53 479 51 46 1 3 50 506 11 12 1 0 13 185 13
81 27 43 16
473 530 517 198
710 610 610 227
Total
222 23 27 272 1565 153 167 1718 2157
Pos, positive; Neg, negative; and Tox, toxicity.
ap = 0.10, McNemar's test. tube culture positives that were also positive by shell vial culture. The specificity of the shell vial culture in comparison to tube culture was defined as the percentage of all tube culture negatives that were also negative by shell vial culture. Specimens whose shell vial culture could not be read because of toxicity to the cell monolayer were included in the calculation of sensitivity but were excluded from the calculation of specificity. Specimens in which viruses other than CMV were detected in the tube culture were excluded from all analyses (93 specimens during the time period of April 1987-March 1988 (2% of the total specimens), and 80 specimens during the time period of April 1988-March 1989 (1.8% of total).
RESULTS First, we investigated the effect of performing fluorescent-antibody staining at two different times after inoculation of the culture. As shown in Table 1, of 322 specimens routinely submitted for CMV culture that were stained at both 16 and 40 hr after inoculation, 25 were positive at 16 hr compared to 32 at 40 hr, an increase of 28% (p = 0.10; McNemar's test). After the completion of the first experiment, all specimens routinely submitted for CMV culture were set up on duplicate shell vials that were both stained at 40 h after inoculation. As shown in Table 2, of 2157 specimens, 272 were positive in one or both shell vial cultures, including 222 in which both shell vial cultures were positive, 23 in which one shell vial culture was positive and the other was negative, and 27 in which one shell vial culture was positive and the other was unreadable because of toxicity. If it is assumed that 50% of the specimens in which only one shell vial culture was positive would have been detected if only one shell vial culture had been set up per specimen, then the additional yield attributable to the use of two shell vial cultures per spec-
imen was 25 cultures (50% of 23 + 27), or an increase of 10% compared to the number of cultures that would have been positive if only one shell vial had been set up per specimen [25/(222 + 25)]. We also evaluated the effect of the age of the fibroblast cells on their sensitivity to infection by CMV as detected by the shell vial procedure. As shown in Figure 1, the results depended on the test strain used. With each of the two lab strains (Towne and AD169), the sensitivity was maximal at day 1 after seeding, and declined sharply thereafter, with day-12 ceils having only 5%-25% as many positive nuclei as day-1 cells. Experiments with a patient isolate (K.O.) yielded a curve that was initially relatively fiat, with day-8 cells having 77% as many positive nudei as day-1 cells. By day 12, however, there was only 40% as many positive nuclei as in day-1 cells. We performed an additional experiment to investigate the relative effects of cell age and cell density on the decreasing sensitivity of cells with increasing age of the culture. The Towne strain was used in this experiment. The decreasing sensitivity with age was apparent in cultures seeded at four different densities yielding a range of confluencies, suggesting that the effect was independent of cell density or confluency of the monolayer (Figure 2). In the spring of 1988, we made changes in our routine laboratory procedures to incorporate staining of shell vials at 40 hr instead of 16 hr after inoculation, inoculation of all specimens onto duplicate shell vials, and age monitoring of shell vials so that shell vials older than 12 days from the time of seeding were not used. A comparison of 12-month periods before and after these changes is shown in Table 3. With approximately the same number of specimens in each time period and with no major changes in the personnel who read the coverslips, the sensitivity of the shell vial procedure compared
128
TABLE 3
M. Arens et al.
Results of Shell Vial and Tube Cultures for CMV During Two Different Time Periods
Shell Vial Culture Result for Specimens With Tube Culture Pos Time Period Specimen
Tube Culture Neg
Pos Neg Toxic Pos Neg Toxic Sensitivity Specificity % Toxic
April 1987-March 1988a Urine 133 Buffy coat 38 Respiratory 33 Other 8 Total
Shell Vial Test Parameters (Compared to Tube Culture Results)
86 80 31 5
39 9 1 0
33 15 1 5
924 353 1173 97 979 61 434 30
52 30 51 62
97 99 >99 99
25 8 6 6
212 202
49
54 3510 541
46
98
13
April 1988-March 1989b Urine 222 Buffy coat 76 Respiratory 77 Other 27
13 46 14 3
13 2 7 0
119 23 28 5
926 140 1071 32 1015 79 464 31
90 61 79 90
89 98 97 99
11 3 7 6
Total
76
22
175 3475 282
80
95
7
402
aDuring the time period of April 1987-March 1988, one shell vial culture was inoculated per specimen, staining was at 16 hr after inoculation, and the age of ceils in the shell vial cultures was not monitored. bDuring the time period of April 1988-March 1989, two shell vials were inoculated per specimen, staining was at 40 hr after inoculation, and shell vial cultures 13 days or older prior to inoculation were not used.
to tube culture increased from 46% to 80%, a n d the percentage of specimens for which a shell vial result could not be r e p o r t e d because of toxicity decreased from 13% to 7%. To c o m p a r e the sensitivity of b o t h the shell vial and tube cultures, w e u s e d the data from both time periods to calculate the sensitivity and specificity of both the shell vial culture a n d tube culture in reference to the total n u m b e r of cultures f o u n d to be positive for CMV by either m e t h o d . It is likely that specimens in which the shell vial culture was positive and the conventional culture was negative truly contained CMV, because m o s t of these specimens were from transplant patients w h o h a d other specimens positive for CMV by tube culture shortly before or after the specimen that yielded discrepant results. The results are s h o w n in Table 4. In this analysis, the sensitivities of the shell vial and tube culture w e r e 51% and 90%, respectively, in the first time period and 86% and 74% in the second time period, indicating that d u r i n g the second period, shell vial cultures w e r e more sensitive than w e r e tube cultures for the detection of CMV. The total percentage of cultures positive for CMV increased from 11.3% in the first time period, to 15.2% in the second time period, a n d was attributable largely to
the increased detection in shell vial cultures in the second time period.
DISCUSSION The shell vial culture fluorescent-antibody proced u r e has h a d a decisive impact o n the rapid diagnosis of CMV infections. H o w e v e r , it is clear that optimal sensitivity of the p r o c e d u r e can be attained only by attention to a n u m b e r of specific aspects of the procedure. The optimal time for cell fixing a n d staining in shell vials is controversial. In the initial descriptions of the procedure, fluorescent-antibody staining was p e r f o r m e d at 16 hr after inoculation, with excellent sensitivity (Gleaves et al., 1985; Shuster et al., 1985). However, other investigators have reported that only 61% of the ultimately shell vial positive specimens were positive at 16 hr (Thiele et al., 1987). A recent report 0 e s p e r s e n et al., 1989) described a multisite s t u d y using a different commercial m o n o c l o n a l antibody (Syva), and d e m o n s t r a t e d an a p p a r e n t effect of cell age on the optimal staining time. In this study, staining at 16-24 hr was optimal on cells p r e p a r e d in the site laboratory, but 40-48 hr was optimal for
Optimizing Recovery of CMV in Shell Vials
TABLE 4
129
Sensitivity of Shell Vial and Tube Cultures Calculated in Relation to Total CMV Isolation No. (%) of Total Cultures Pos for CMV
Time Period April 1987-March 1988 April 1988-March 1989
Total Pos Cultures for CMVa Shell Vial Tube Performed n (%)b Culture Culture 4568 4433
Sensitivity (% of Total Cultures Detected) of ShellVial Tube Culture Culture
517 (11.3) 266 (5.8) 463 (10.1) 675 (15.2) 577 (13.0) 500 (11.3)
51 86
90 74
qncludes specimens positive in either shell vial culture, tube culture, or both. bpercent of cultures performed.
commercial cells. The results of the present study, which used commercial cells, are consistent in that staining at 40 hr resulted in a 28% increase in positive cultures compared to staining at 16 hr. Paya et al (1988) have investigated the effects of inoculating multiple shell vials for detection of CMV using different types of specimens. Our data support their conclusion that multiple shell vials should be used. The use of duplicate shell vials on all of 2157 specimens yielded 50 specimens that were positive on only one shell vial. Presumably, about onehalf of these would have been missed had they been inoculated on only one shell vial. The increased yield attributable to use of duplicate shell vials was 10%. The results of Thiele at al. (1987) indicated that the sensitivity of MRC-5 cells to CMV declined significantly by about 12 days after seeding. Using lowpassage MRC-5 cells in shell vials (as opposed to 24well plates used by Thiele) and the Towne or AD169 strains, we found that sensitivity declined as early as 2 days after seeding. However, we also found that this phenomenon was dependent on the viral strain used. Experiments using a patient isolate showed a less precipitous decline in sensitivity with increasing cell age than when the lab strains were used. Both Thiele's results and our own experience prompted us to monitor the age of cells used in shell
vial cultures in the clinical laboratory, and not to use any that were older than 12 days. It was not possible to quantitate the contribution of this policy to the overall increase in CMV detection observed in the Clinical Virology Laboratory. It may be that preparation of shell vial cultures in the laboratory rather than reliance on commercial cells would permit use of even younger cells and increase overall detection further. However, it does not seem at all reasonable to advocate making shell vials every day so that day1 cells would always be available. The net impact of staining at 40 hr, inoculating duplicate shell vials and monitoring the age of cells was to increase the sensitivity of the shell vial culture from 46% to 80% in comparison to tube culture. It appears that staining at 40 hr rather than at 16 hr after inoculation may have been the single most important factor. However, the use of duplicate shell vials for each specimen resulted in a further increase in detection. Finally, the use of young cells is important for maximum sensitivity. We believe that attention to each of these aspects of the shell vial procedure is required to optimize detection of CMV. Our data indicate that when conditions are optimized, the shell vial culture can be more sensitive than tube cultures incubated for 6 weeks for the detection of CMV in patient specimens.
REFERENCES Ashley R, Peterson E, Abbo H, Gold D, Corey L (1989) Comparison of monoclonal antibodies for rapid detection of cytomegalovirus in spin-amplified plate cultures. J Clin Microb 27:2858-2860. Bach MC, Bagwell SP, Knapp NP, Davis KM, Hedstrom PS (1985) A-(1,3-dihydroxy-2-propoxymethyl)guanine for cytomegalovirus infections in patients with acquired immunodeficiency syndrome. Ann Intern Med 103:381382.
Balfour HD, Chace BA, Stapleton JT, Simmons RL, Fryd DS (1989) A randomized, placebo-controlled trial of oral acyclovir for the prevention of cytomegalovirus disease in recipients of renal allografts. N Engl J Med 320:13811387. Gleaves CA, Smith T, Shuster EA, Pearson GR (1985) Comparison of standard tube and shell vial cell culture techniques for the detection of cytomegalovirus in clinical specimens. J Clin Microbiol 21:217-221.
130
Jespersen DJ, Drew WL, Gleaves CA, Meyers JD, Warford AL, Smith TF (1989) Multisite evaluation of a monoclonal reagent (Syva) for rapid diagnosis of cytomegalovirus in the shell vial assay. J Clin Microbio127:15021505. Paya CV, Wold AD, Ilstruk DM, Smith TF (1988) Evaluation of number of shell vial cell cultures per clinical specimen for rapid diagnosis of cytomegalovirus infection. J Clin Microbiol 26:198-200. Thiele GM, Bicak MS, Young A, Kinsey J, White RJ, Purtilo
M. Arens et al.
DT (1987) Rapid detection of cytomegalovirus by tissue culture, centrifugation, and immunofluorescence with a monoclonal antibody to an early nuclear antigen. J Virol Methods 16:327-338. Shuster EA, Beneke JS, Tegtmeyer GE, Pearson GR, Gleaves CA, Wold AD, Smith TF (1985) Monoclonal antibody for rapid laboratory detection of cytomegalovirus infections: characterization and diagnostic application. Mayo Clinic Proc 60:577-585.
