Antiviral susceptibility testing of cytomegalovirus: criteria for detecting resistance to antivirals

Clinical and Diagnostic Virology, 1 (1993) 179-185 © 1993 Elsevier SciencePublishers B.V. All rights reserved 0928-0197/93/$06.00

DIAVIR 00020

Clinical and Diagnostic Virology

Antiviral susceptibility testing of cytomegalovirus: criteria for detecting resistance to antivirals W. L a w r e n c e D r e w , R i c h a r d M i n e r a n d Ezz Saleh Mount Zion Medical Center of the University of California, San Francisco, San Francisco, CA, USA

(Received 14 December 1992;revisionreceived 15 April 1993;accepted 19 April 1993)

Abstract Testing cytomegaloviruses for antiviral susceptibility is increasing especially since the reports of 'resistance' to ganciclovir and foscarnet (Erice et al., 1989; Knox et al., 1991). There is however no standardized method for susceptibility testing nor are there criteria for designating an isolate as sensitive or resistant. In a previous paper we utilized a plaque reduction assay and suggested that a resistant strain be defined as one requiring > 12 #M ganciclovir for inhibition of 50% of viral plaques. (Drew et al., 1991) This concentration was chosen because it was at least four-fold greater than the mean concentration required to inhibit pretherapy isolates. In this paper we present the results of testing a large number of isolates prior to and during therapy with either ganciclovir or foscarnet. By analyzing the results of these assays we propose revised criteria for susceptibility of cytomegalovirus ~<5 #M for ganciclovir and < 400 pM for foscarnet. Key words: Cytomegalovirus;Resistance; Ganciclovir;Foscarnet

Materials and Methods Patients HIV infected patients with CMV disease were enrolled in a series of studies

and were followed prospectively with cultures at baseline (prior to the onset of antiviral therapy) and at monthly intervals thereafter for up to 15 months. Urine was the most common specimen tested but blood and/or semen were occasionally obtained. Viral cultures Specimens for culture were processed by standard methods in human foreskin diploid fibroblast (MRHF) (diploid fibroblast) cells. Specimens developing characteristic cytopathic effect (CPE) were stained with specific fluorescent antibody (SYVA Company, Palo Alto, CA) to confirm the recovery of CMV.

Correspondence to." W. L. Drew, Department of Infectious Diseases and Microbiology, Mount Zion Medical Center of UCSF, 1600 Divisadero Street, P.O. Box 7921, San Francisco, CA 94120-7921,USA Tel.: +1(415)885-7315,Fax: +1(415)885-7780.

180

Testing susceptibility to ganciclovir (GCV) and foscarnet Susceptibility to antivirals was determined as follows: The titer of virus was increased by passing positive cultures in MRHF cells approximately 2 6 times. Monolayers of MRHF cells were grown in 24-well plates (Corning Glass Works, Corning, NY). Virus inoculum diluted to provide approximately 50-100 plaque forming units (pfu) per ml was added in 1 ml aliquots to each well. Unadsorbed virus was removed aRer 3 h incubation at 37°C and monolayers were washed in Puck's Saline A. Cultures were then incubated for 7 10 days at 37°C in minimal essential medium (MEM) containing 0.5% high titered human CMV antibody to inhibit secondary plaques and GCV at concentrations of 0, l, 2.5, 5, I0, 20, and 50/~M or Foscarnet (Foscavir, phosphonoformic acid, PFA) at concentrations of 0, 25, 100, 250, 500, and 1000 #M. Gancictovir (Cytovene, DHPG) was supplied by Syntex, Palo Alto, CA and Foscarnet was supplied by Astra Pharmaceuticals. Three replicates were prepared for each drug concentration. Cultures were terminated by removing medium and washing monolayers with Puck's Saline A. Monolayers were then fixed with methyl alcohol, stained with Wright-Giemsa stain and plaques were counted with an inverted microscope at 70 x magnification. The IDso was that concentration of drug inhibiting 50% of the number of the plaques compared to control The ID9o was the concentration inhibiting 90% of plaques compared to control. Heterogeneity of viral population To determine whether CMV isolates are homo or heterogenous in their GCV susceptibility, one resistant isolate (//6304) initialiy isolated from lung tissue and an IDso of 23.5 uM was plaque purified. Individual plaques were then randomly picked, passed 3 5 times in diploid fibrobtast cells to increase viral titer and tested for antiviral susceptibility as described above.

Results

Ganciclovir 130 isolates from patients who had never been treated with ganciclovir were tested for ganciclovir susceptibility and the frequency distribution of !Dsos are shown in Fig. 1. All but 1 isolate from untreated patients had an IDso of ~<5 uM. 45

NUMBER OF PATIENTS

0 - 1.0

I.! - 2.0

2.1 - 3.0

3.1 - 4.0

4.1 - 5.0

>5,0

IDs0 ( u M ) F i g . 1. D i s t r i b u t i o n o f I D s o s o f g a n c i c l o v i r f o r C M V i s o l a t e s f r o m u n t r e a t e d p a t i e n t s .

181

The mean IDso of this group of isolates was 1.94 #M. The ID9o for these isolates is shown in Fig. 2. More than 95% of these isolates had an |D9o ~<15 #M and the mean ID9o was 7.5/zM. Eighty-nine CMV isolates from patients receiving gancMovir for up to 15 months were tested for susceptibility to the drug and the results are shown in Fig. 3. 57 of these isolates (64%) had IDsos>5.0 FtM and 23 (26%) had IDsos>10.0 #M. ID9o results for these 89 isolates are shown in Fig. 4. Foscarnet 58 isolates from patients who had never received foscarnet were tested and the IDso results are Shown in Fig. 5. The distribution of IDsos is a typical

NI MBER

t)V

PA~]ENTS

W 4

40-

36 35-

30-

25-

20-

mmmi

15-

10-

5-

:

0-2.5

2.6-5

5.1-7.5

7.6-10

10.1-12.5

~ "~:

12.6-15

2

2

15.I-17.5

2

17.6-20

20.1-22.5

22.6-25

25.1-27.5

ID~ (uM)

Fig. 2. Distribution

of ID9os of

ganciclovirfor CMV isolates from untreated patients.

Wm NUMBER OF PATIENTS 5 !~ # ~ ! ~

0- 5

5.1 - 10

10.1 - 15

15.1 - 20

1

20.1 - 25

2

25.1 - 30

30.1-40

I

40.i - 50

>50

IDs0(uM) Fig. 3. Distribution of IDsos of ganciclovir for CMV isolates from patients treated with ganciclovir.

182

NUMBER OF PATIENTS

0 ~ 10

10.1 - 20

20.1 - 30

30.1-40

40. l - 50

>50

iD90(uM) Fig. 4. Distribution of IDgos of ganciclovir for CMV isolates from patients treated with ganciciov~r:

NUMBER OF PATIENTS

0-50

51-100

101-150

151-200

201-250

251-30(

301

350

351-400

>401

IDs0(uM) Fig. 5. Foscarnet new untreated patients. population distribution with a mean and median between 201 and 250 #M. Only 5 of 58 isolates had an I D 5 o > 3 0 0 pM and 2 an I D 5 o > 4 0 0 uM,

Heterogeneity of ganciclovir reszstant virus Table I shows that a resistant strain of C M V is composed of numerous subpopulations exhibiting a wide variation of IDsos, from 0.6 12.6/~M.

Discussion Given the results of testing susceptibility for C M V isolates from patients who had never been treated with these 2 drugs, we would propose that isolates tested by the method described (or a closely similar method) be regarded as susceptiNe if their G C V IDso is ~<5 # M or their foscarnet IDso is <400 ,aM. Isolates with higher IDsos are outside the usual distribution and can be considered to exhibit varying degrees

183 TABLE I IDso OF GCV FOR DIFFERENT CMV ISOLATES(PLAQUEPURIFIED FROM ISOLATE#6304 [IDso= 23.5]) Isolate 4Cs 2A2 4Cz 4A3 4Dz 2C1 2Cs 4B5 4D3 2D5

IDso of GCV(pM) 0.6 2.8 3.8 4.0 5.2 6.4 8.0 10.0 10.8 12.6

of resistance. Since these studies were performed only with isolates from AIDS patients the proposed susceptibility criteria may not be relevant to isolates from other patient populations, e.g., congenitally infected babies or organ transplant recipients. AIDS patients may be simultaneously infected with more than 1 strain of CMV and their isolates may represent mixtures which respond differently to antivirals than single strain isolates from other patient populations. Further validation of these breakpoints is that patients shedding isolates with ganciclovir IDsos greater than 5 #M have failed ganciclovir treatment but been successfully treated with the Foscarnet (Jacobson et al., 1991). The ID9o determination might appear to be more useful for detecting a resistant sub population but valid statistical analysis is hampered by the small number of plaques that are counted at this level of inhibition. Thus we prefer to use IDso breakpoints to distinguish resistant isolates. In fully sensitive viral isolates (GCV IDso < 5 #M) the overwhelming percentage of the virus population is inhibited by this concentration. In isolates with GCV IDso > 5 #M, less susceptible variants are present in increased concentration and the IDs0 is increased. Isolates with IDsos between 6 and 10 #M might be considered in an intermediate or indeterminate category but with continued exposure to ganciclovir the more resistant variants would be increasingly selected and the IDso of the viral isolate would be expected to progressively increase. We have previously reported that patients receiving GCV can shed CMV isolates with progressively increasing IDsos as the duration of therapy is lengthened. (Stanat et al., 1991) For example, one of our patients shed CMV in urine with an IDso of 2.6 #M prior to therapy. At 3 months, his urine isolate had an IDso of 8.5 and at 8 months an ID~o of 31 #M. These isolates exhibited progressively less phosphorylation of ganciclovir which helped to elucidate the principal mechanism of CMV resistance to ganciclovir. (Stanat et al, 1991; Biron et al., 1986) The evidence that 'resistant' viral isolates are mixtures of both susceptible and resistant populations is shown in Table I. Although the IDso of the isolate #6304 was 23.5 #M, none of the randomly picked plaques had an ID5o>12.6 #M. This might seem paradoxical but reflects the fact that only 10 plaque variants were tested for ganciclovir susceptibility. Presumably, if all of the hundreds of possible plaques had been systematically tested, many would have exhibited IDsos of > 23.5 #M. The important conclusion from these single plaque susceptibility assays is the documenta-

184

tion of the heterogeneity of resistant isolates. In order to determine whether 'sensitive' strains also contain heterogeneous sub populations, including resistant variants, it will be necessary to perform the same type of plaque susceptibility assays asing 'sensitive' isolates. Whether these are separate viral strains or the same strain wi.th differing sub populations is not known although we and others have previously reported that immunocompromised patients may be infected by more than one different strain of CMV (Drew et at., 1984; Spector etal., t984). Unfortunately several problems attend antiviral susceptibility testing° Most importantly there is no accepted, standardized method and individual laboratories use very variaNe methodology. In particular there is no standardization of the ptaque reduction assay in terms of the tissue culture (cell lines, media, overlay), inoculum, incubation time or concentrations of drug used. Despite this seemingly appalling state of affairs, the results of ganciclovir susceptibility testing of GCV susceptibility strains of CMV by plaque reduction are reasonably comparable between mdividua! laboratories (Plotkin et aL 1985). However, it is not known whether "resistant' stratus would be detected equally by differing methods nor what the criteria for resistance should be. Obviously, it is imperative that a standard method be agreed upon and utilized. A further major problem with antiviral susceptibility testing of CMV is the excessive time required to complete the assay. In our laboratory an average of 4, 6 weeks is required to isolate the virus and then perform a susceptibility test. The plaque reduction assay itself requires only 7 10 days but the bulk of the time is utilized in initially isolating the virus and then passing it serially in tissue culture to develop a suffÉcient titer of virus for an appropriate inoculum. Numerous methods have been developed using radioactive or nonradioactively labeled DNA probes in order to accelerate the actual susceptibility assay IErice et al.. 1989 t. These procedures substitute for the plaque reduction assay and may decrease the time for the susceptibility test by 2 4 days but in general they do not reduce the need for repeated passage to develop sufficient ~iral inoculum. Very recently we have reported on a DNA based method which appears to also reduce the need for multiple virus passages. IShen 1992) Until more rapid methods are developed and validated, antiviral susceptibility tests with CMV are of limited clinical utility, especially when confronted with a patient who appears to have progressive CMV disease despite antiviral therapy. In such instances, it is probably best to empirically employ an alternative antiviral unless there is a contraindication for that agent. The necessity for multiple passages of virus to obtain a sufficient titer for testing also carries the potential for altering antiviral susceptibility by the passages. We have not rigorously studied this but isolates we have sent to many other !aboratories have consistently maintained their susceptibility or resistance throughout many passages. A final problem of testing CMV for antiviral susceptibility is that the isolate shed in urine or semen may differ from the isolate causing end organ disease in e.g. the retina. We and other have previously shown that immunocompromised patients may harbor multiple different strains of CMV concurrently (Drew et at.. 1984; Spector etal., 1984). It is therefore possible that an isolate from blood or vitreous fluid of the eye wilt differ in susceptibility from an isolate m secretions such as urine or semen. Since CMV retinitis is thought to usually result from disseminated viremic disease, an isolate from blood is more likely to reflect the characteristics of the s~rain infecting the retina than is an isolate from urine or semen: accordingly susceptibility tests should be performed on a blood isolate in preference to one from the genitouri-

185

nary tract. Similarly, an isolate from respiratory secretions would be preferred for patients with CMV pneumonia.

References Biron, K.K. et al. (1986) A human cytomegalovirus mutant resistant to the nucleoside analog 9-[2(hydroxymethyl) ethoxymethyl] guanine (BWB759U) induces required level of BWB759U triphosphate. Proc. Natl. Acad. Sci. USA, 83, 8769-8773. Drew, W.L. et al. (1984) Multiple infections with CMV in AIDS patients: Documentation by southern blot hybridization. J. Infect. Dis., 150, 952. Drew, W.L. et al. (1991) Prevalence of resistance in patients receiving ganciclovir for serious cytomegalovirus infection. J. Infect. Dis., 163, 716-719. Erice, A. et al. (1989) Progressive disease due to ganciclovir-resistant cytomegalovirus in immunocompromised patients. N. Eng. J. Med., 320, 293-297. Jacobson, M. et al. (1991) Foscarnet therapy for ganciclovir resistant patients with AIDS. J. Infect. Dis., 163, 1348-1351. Knox, K. et al. (1991) Cytomegalovirus isolate resistant to ganciclovir and foscarnet from a marrow transplant patient. Lancet, 337, 1292-1293. Plotkin, S.A. et al. (1985) Sensitivity of clinical isolates of human cytomegalovirus to 9-(1,3-dihydroxy-2propoxymethyl)guanine. J. Infect. Dis., 152, 833-834. Shen L-P (1992) Direct detection of HCMV in clinical specimens using bDNA assay. Abstracts of the 32nd Interscience Conference on Antimicrobial Agents and Chemotherapy, 1244, p. 318, Anaheim, CA, USA, Oct. 11-14, 1992. Spector, S. et al. (1984) Identification of multiple cytomegalovirus strains in homosexual men with acquired immunodeficiency syndrome. J. Infect. Dis., 150, 953-956. Stanat, S. (1991) Ganciclovir-resistant cytomegalovirus clinical isolates: mode of resistance to ganciclovir. Antimicrob. Agents Chemother., 35, 2191-2197.