1Institute of Medical Virology and 2Department II of Internal Medicine, University Hospital of Tübingen, Tübingen, Germany
Received 20 May 2002; revised 25 September 2002; electronically published 13 December 2002.
A 9-month posttransplantation course of an allogeneic stem-cell transplant recipient (human cytomegalovirus [HCMV] serostatus, donor positive/recipient negative), in whom ganciclovir (GCV) resistance developed (UL97 mutations M460V, L595S, and C603W) on day 164 after transplantation and who developed HCMV retinitis and fatal HCMV encephalitis is presented. Virus strains isolated from secondary cultures were analyzed by UL97 restriction assays and sequencing and were compared with primary DNA extracts of the same specimens, which resulted in molecular proof of an initial HCMV strain-specific in vitro selection of the in vivo nondominant UL97 L595S–
C603 mutant strain from 3 viral variants present in vivo. In addition, compartmentalization of virus present in blood and cerebrospinal fluid was found. The influence of rapidly increasing plasma virus load (to >106 copies/mL) and oral administration of GCV on the emergence of GCV resistance is shown. These findings have strong implications for the diagnosis of HCMV drug resistance.
Reprints or correspondence: Dr. Klaus Hamprecht, Institute of Medical Virology, University Hospital of Tübingen, Elfriede-Aulhorn-Str. 6, 72076 Tübingen, Germany .
There are only a few reports on phenotypically and genotypically characterized drug-resistant human cytomegalovirus (HCMV) strains isolated from patients who have undergone allogeneic stem cell transplantation (SCT) [1–
5]. For diagnosis of HCMV resistance to ganciclovir (GCV), foscarnet (PFA), and cidofovir (CDV), it is necessary to identify the HCMV DNA polymerase (UL54) and phosphotransferase (UL97) gene mutations that induce resistance, as well as patterns of antiviral susceptibility of isolated strains to these drugs. However, the isolation procedure itself appeared to select the predominant virus strain that was in vivo. [6]. It has been suggested that determination of HCMV susceptibility to GCV directly from primary cultures can be performed without the drawback of possible in vitro selection by several culture passages [7].
We present molecular proof of an initial HCMV strain-specific in vitro selection of an in vivo nondominant UL97 mutant strain after only 1 culture passage. Potential risk factors for the development of GCV-resistant HCMV disease after allogeneic SCT are described. Our findings have strong implications for the diagnosis of HCMV drug resistance.
Case report. The patient was a 48-year-old HCMV-seronegative woman who was diagnosed with acute myeloblastic leukemia (AML) (FAB M3) in 1995. She underwent an allogeneic-unrelated HLA-A–
, HLA-B–
, and HLA-DRB1–
matched SCT from a HCMV-seropositive donor in November 1997 for treatment of the second relapse of AML. Graft-versus-host disease prophylaxis was performed with antithymocyte globuline, prednisolone, and mycophenolatemofetil. HCMV primary infection was detected on day 46 after transplantation by polymerase chain reaction (PCR) assay. According to our protocol [8], preemptive antiviral therapy with intravenous GCV was initiated (induction, 5 mg/kg twice daily for 2 weeks; maintenance, 5 mg/kg/day), during which viremia occurred. An intermediate switch to PFA induction (60 mg/kg 3 times daily) therapy on day 90 was ceased because of significant nephrotoxicity. Therefore, treatment was maintained by administration of oral GCV (1 g 3 times daily) from days 115 to 164. Because of an increasing plasma virus load, intravenous GCV was readministered. Despite an initial decrease in virus load during combined treatment with GCV and PFA, the HCMV copy number increased again, and the patient developed fatal HCMV encephalitis and retinitis and died on day 233 after transplantation, in July 1998.
HCMV monitoring. Virus isolation in tube cell culture was performed from leukocytes, throat washes, and urine specimens, using monolayers of human foreskin fibroblasts. DNA extraction and qualitative nested PCR were done as described elsewhere [9]. Quantitative PCR was performed retrospectively with the Cobas Amplicor CMV Monitor Test (Roche Diagnostics).
Phenotypic drug sensitivity. Primary virus isolates from tube cultures were propagated by a single in vitro passage before drug susceptibility assays were performed by use of a simplified plaque reduction assay with cell-associated virus [10]. The cutoff limits for drug resistance were defined as follows: GCV, >6 M; PFA, >400 M; and CDV, >4 M.
UL97 restriction fragment–
length polymorphism (RFLP) analysis. Both primary DNA extracts from different specimens and DNA extracts from corresponding virus isolates after 1 in vitro passage were used for nested PCR of the HCMV UL97 gene. First-round PCR (forward primer 460 and reverse primer 650) yielded a 776-bp product [3]. Comprehensive restriction enzyme analysis of the specific HCMV UL97 nested PCR products of 266 bp (forward primer 460 and reverse 3′
-mismatch primer 520, 5′
-GCATGGGTCGGAAAGCAaG), 189 bp [3] (forward primer 595 and reverse primer 595), and 121 bp (forward primer 595 and reverse 3′
-mismatch primer 595, 5′
-GCAGTGCGTGAGCTTGCCGTTCTt) was performed to identify mutations in the following codons (recognition site [reference]): M460V (NlaIII [11]), H520Q (AluI [12]), A591V (HaeIII [3]), C592G (FseI [3]), A594V (Hin6I [11]), L595S (TaqI [11]), L595F (MseI [11]), and C603W (AvaII [3]). The restriction fragments of TaqI and AvaII digests were separated in 8% (wt/vol) polyacrylamide gels and were visualized with silver staining. To determine the relative proportion of different coexisting HCMV variants, the silver stained gels were evaluated by densitometry using LabImage 2.6 (Kapelan). On the assumption that each UL97 mutation is located on a single HCMV strain, the relative proportion of the GCV-sensitive population was calculated as follows: GCVsensL595–
C603 (%) = 100 (%) - [GCVresL595–
C603W (%) + GCVresL595S–
C603 (%)] ( and ), where GCVsens is the sensitive strain (UL97 wild type), and GCVres is the resistant strain (UL97 mutant strain).
fig.ommitted
Table 1. Synopsis of genotypic human cytomegalovirus UL97 screening.
fig.ommitted
Figure 1. In vitro selection of the in vivo nondominant UL97 L595S–
C603 strain detected by comparison of polymerase chain reaction–
based restriction fragment–
length polymorphism patterns of primary DNA extracts from leukocytes (L) and urine (U), with corresponding DNA extracts after 1 culture passage (C) on days 208 and 215 after stem cell transplantation (SCT). *, Based on the assumption of 1 mutation (mut)/virus strain; relative proportion of wild-type (wt) strain calculated as follows: GCVsensL595–
C603 (%) = 100 (%) - [GCVresL595–
C603W (%) + GCVresL595S–
C603 (%)], where GCVsens is the sensitive strain (UL97 wt strain), and GCVres is the resistant strain (UL97 mut strain). Relative proportion of strains is given in . AD169, human cytomegalovirus laboratory strain representing wt UL97; GCV, ganciclovir; M 123, 123-bp ladder; P, plasma; TC, throat after a single in vitro culture passage. Boldface type indicates UL97 mutation characterizing a GCV-resistant human cytomegalovirus variant.
Sequencing of UL97 and UL54. The UL97 region, including codons 439 (nt 1314) to 696 (nt 2089), was amplified, purified (QuiaQuick PCR Purification Kit; Qiagen), and sequenced, using primers 460F and 650R [3] with a terminator cycle sequencing kit (BigDye; Perkin-Elmer Cetus) by means of an automated genetic analyzer (ABI-310; Perkin-Elmer Cetus). For sequencing of the UL54 region, 4 overlapping PCR products were generated and were used to sequence codons 365 (nt 1091) to 1084 (nt 3254).
Molecular heterogeneity of virus strains. To distinguish selected virus strains in vitro from in vivo strains of different origin, a 2-kb amplification product of UL10–
13 was digested with RsaI and Hin6I [3].
Results. Evidence of HCMV primary infection in an initially seronegative STC recipient (donor positive/recipient negative [D+/R-]) was found on day 46 after transplantation, as indicated by low-level viral plasma DNAemia . Viremia was observed after 3 weeks of intravenous GCV treatment on day 74. Six days later, the patient's HCMV virus load had increased 2-fold, and viremia persisted. However, genotypic UL97 screening did not detect a mutant virus population up to at least day 122, when GCV was administered orally. Immunophenotyping of leukocyte subpopulations on day 140 revealed a total lack of CD4+ T cells. After 49 days of oral GCV administration and 101 days of cumulative GCV therapy (intravenous and oral), the UL97 mutations L595S, C603W, and M460V were detectable on day 164 after transplantation. The manifestation of genotypic GCV resistance coincided with a drastic 30-fold increase in the virus load, which reached 1.4 × 106 copies/mL . In all 3 associated UL97 codons, genotypic analysis resulted in the detection of a mixture of mutant and wild-type HCMV populations in leukocytes. The following week, the patient developed HCMV encephalitis. In contrast to the presence of GCV-resistant L595S–
C603 and L595–
C603W strains in blood, the viral DNA from cerebrospinal fluid (CSF) on day 208 exhibited a wild-type sequence . Of interest, the virus population carrying the M460V mutation was no longer detectable, whereas the UL97 mutations C603W and L595S persisted. On day 208, 44 days after genotypic resistance originated, HCMV was isolated from leukocytes, and the phenotypic drug susceptibility assay confirmed GCV resistance with an ID50 of 20 M. Despite the high ID50 result, this isolate was sensitive to PFA (ID50, 131 M) and CDV (ID50, 2.1 M) and showed no UL54 mutations. The virus load decreased significantly during intravenous GCV administration and intensified antiviral therapy with GCV and PFA, from 1.2 × 106 copies/mL on day 171 to <1000 copies/mL on day 224 . However, despite the marked decrease in virus load, the patient died from HCMV encephalitis on day 233 after transplantation.
Retrospectively, all available virus isolates and primary DNA extracts from different sites were screened for UL97, L595S, and C603W mutations by RFLP analysis and sequencing. In all primary DNA isolates (from leukocytes, plasma, urine, and throat washes) except those from CSF, a persistent infection with mutant HCMV strains carrying L595S and C603W was demonstrated from days 164 to 224. Of interest, the comparison of genotypic analysis of primary DNA extracts from leukocytes (day 208) and urine (day 215) with DNA extracts of the corresponding virus isolates after only 1 in vitro passage revealed an in vitro selection of an in vivo nondominant mutant strain . The relative proportion of the UL97 mutant strains (GCV resistant: L595–
C603W and L595S–
C603) and the wild-type strain (GCV sensitive: L595–
C603) is given . Short-term in vitro culture selected positively for only the in vivo nondominant L595S–
C603 GCV-resistant strain in leukocytes and urine from a mixture of in vivo coexisting GCV-sensitive (L595–
C603) and GCV-resistant (L595–
C603W and L595S–
C603) viral variants . All RFLP results were reproducible and were confirmed by sequencing. UL54 mutations were not found in any sample. Genotypic analyses of virus strains from leukocytes and urine (primary DNA extract vs. virus isolate) did not illustrate any differences in the UL10–
13 restriction patterns.
Discussion. This is the first report to present molecular evidence of initial in vitro selection of an HCMV UL97 in vivo nondominant mutant strain in a patient who had undergone bone-marrow graft from an HLA-matched unrelated donor with fatal HCMV encephalitis. Our patient exhibited all known clinical and virologic risk factors during the emergence of drug resistance, such as D+/R- status, oral (49 days) and long-term GCV treatment (>3 months), and extremely high virus load (>106 copies/mL).
Ten years ago, primary cultures for GCV susceptibility determination were described [7]. It was suggested that the drawback of possible in vitro selection by cell passages of HCMV strains would be overcome [7]. Recently, Baldanti et al. [6] have provided evidence that the UL97 C607Y mutation was detectable only by a genotypic assay, whereas the corresponding virus isolate showed GCV susceptibility and a wild-type genotype. They concluded that the isolation procedure itself appeared to select the in vivo predominant virus strain, providing a misleading picture of the true composition of the entire in vivo HCMV population. In our case, on days 208 and 215, 3 different virus strains coexisted in vivo: besides the dominant GCV-sensitive wild-type strain (L595–
C603), 2 nondominant GCV-resistant strains (L595–
C603W and L595S–
C603) were present. In contrast to a former report [6], our short-term in vitro culture selected positively for the in vivo nondominant L595S–
C603 strain in isolates from leukocytes and urine and negatively for the second mutant virus strain (L595–
C603W) and the wild-type strain.
In a previous study that involved a pediatric patient who had undergone SCT with multidrug resistance, we demonstrated that, in vivo under GCV selection, some UL97 mutant virus strains (C603W and M460I) may only be temporarily detectable beside a preexisting and persisting mutant species (A591V) [4]. This implicates the potential instability or fitness loss of additional mutant strains occurring during prolonged treatment in vivo. We also have shown that the fitness of the L595 wild-type strain in vivo has overcome the L595S mutant strain for a restricted period in the absence of GCV and the presence of CDV [4]. In the present case, it appears contradictory that the in vivo predominant L595 wild-type virus was not able to repopulate in vitro. Although the primary DNA specimen from a throat culture was not available, the RFLP analysis of the corresponding virus isolate (10% wild type) showed inverse restriction patterns with regard to the relative proportion of wild-type and mutant strains compared with the digestion profile from leukocytes. Therefore, we assume that a growth advantage for a specific mutant strain, such as L595S–
C603, does not exist. Additional factors, such as the initial proportion of wild-type to mutant strains in vivo and special in vitro culture conditions, may have an influence on the in vitro repopulation of in vivo preexisting UL97 virus species.
Furthermore, we gathered evidence for a compartmentalization of virus strains. In contrast to blood, we did not detect the GCV-resistant HCMV variants L595–
C603W or L595S–
C603 in CSF. Similar findings that compared UL97 mutant strains from blood (A591V) and CSF (A591 and P521L), blood (C603W) and urine (C603W and A594V), and blood (L595) and throat (L595S) have been reported elsewhere [3, 4].
During the emergence of the L595S and C603W mutant strains between days 122 and 164, a 30-fold increase in viral plasma DNA load was observed. Additionally, our patient showed severe CD4+ lymphocytopenia on day 140. This may have enhanced the possibility for survival and immune evasion of the fittest mutant strains [13, 14]. Prolonged oral GCV therapy also may have contributed to the emergence of mutant virus strains because of the lastingly reduced efficacy of orally administered GCV [13]. In a recent study, we found evidence that GCV-resistant HCMV infections were emerging early (days 44–
95) during immune recovery in children who had undergone SCT [15]. In contrast, our adult patient developed GCV resistance late after having undergone SCT (day 164) in the context of a severe CD4+ lymphocytopenia on day 140.
In conclusion, the present report demonstrates the specific and early in vitro selection of the fittest but in vivo nondominant mutant HCMV strain. In this context, phenotypic drug resistance screening of virus isolates may produce a major bias away from rapid and reliable diagnosis of GCV resistance. Therefore, additional longitudinal genotypic analyses of drug-resistant HCMV strains, such as RFLP or sequencing, may help detect the proportion of in vivo UL97 mutant or wild-type strains early enough to make adequate decisions on the continuation, alteration, or termination of GCV therapy.
References
1.
Erice A, Borrell N, Li WY, Miller WJ, Balfour HH. Ganciclovir susceptibilities and analysis of UL97 region in cytomegalovirus (CMV) isolates from bone marrow recipients with CMV disease after antiviral prophylaxis. J Infect Dis 1998; 178:531–
4.
2.
Wolf DG, Yaniv I, Honigman A, Kassis I, Schonfeld T, Ashkenazi S. Early emergence of ganciclovir-resistant human cytomegalovirus strains in children with primary combined immunodeficiency. J Infect Dis 1998; 178:535–
8.
3.
Prix L, Hamprecht K, Holzhüter B, Handgretinger R, Klingebiel T, Jahn G. Comprehensive restriction analysis of the UL97 region allows early detection of ganciclovir-resistant human cytomegalovirus in an immunocompromised child. J Infect Dis 1999; 180:491–
5.
4.
Eckle T, Prix L, Jahn G, et al. Drug-resistant human cytomegalovirus infection in children after allogeneic stem cell transplantation may have different clinical outcomes. Blood 2000; 96:3286–
9.
5.
Seo SK, Regan A, Cihlar T, et al. Cytomegalovirus ventriculoencephalitis in a bone marrow transplant recipient receiving antiviral maintenance: clinical and molecular evidence of drug resistance. Clin Infect Dis 2001; 33:e105–
8.
6.
Baldanti F, Simoncini L, Sarasini A, et al. Ganciclovir resistance as a result of oral ganciclovir in a heart transplant recipient with multiple human cytomegalovirus strains in blood. Transplantation 1998; 66:324–
9.
7.
Pepin J-M, Simon F, Dussault A, Collin G, Dazza M-C, Brun-Vezinet F. Rapid determination of human cytomegalovirus susceptibility to ganciclovir directly from clinical specimen primocultures. J Clin Microbiol 1992; 30:2917–
20.
8.
Einsele H, Ehninger G, Hebart H, et al. Polymerase chain reaction monitoring reduces the incidence of cytomegalovirus disease and the duration and side effects of antiviral therapy after bone marrow transplantation. Blood 1995; 86:2815–
20.
9.
Hamprecht K, Mikeler E, Jahn G. Semi-quantitative detection of cytomegalovirus DNA from native serum and plasma by nested PCR: influence of DNA extraction procedures. J Virol Methods 1997; 69:125–
35.
10.
Prix L, Maierl J, Jahn G, Hamprecht K. A simplified assay for screening of drug resistance of cell-associated cytomegalovirus strains. J Clin Virol 1998; 11:29–
37.
11.
Chou S, Erive A, Jordan MC, et al. Analysis of the UL97 phosphotransferase coding sequences in clinical cytomegalovirus isolates and identification of mutations conferring ganciclovir resistance. J Infect Dis 1995; 171:576–
83.
12.
Hanson MN, Preheim LC, Chou S, Talarico CL, Biron KK, Erice A. Novel mutation in the UL97 gene of a clinical cytomegalovirus strain conferring resistance to ganciclovir. Antimicrob Agents Chemother 1995; 39:1204–
5.
13.
Emery VC. Progress in understanding cytomegalovirus drug resistance. J Clin Virol 2001; 21:223–
8.
14.
Einsele H, Ehninger G, Steidle M, et al. Lymphocytopenia as an unfavorable prognostic factor in patients with cytomegalovirus infection after bone marrow transplantation. Blood 1993; 82:1672–
8.
15.
Eckle T, Lang P, Prix L, et al. Rapid development of ganciclovir-resistant cytomegalovirus infection in children after allogeneic stem cell transplantation in the early phase of immune cell recovery. Bone Marrow Transplant 2002; 30:433–
9.
(Klaus Hamprecht Tobias Eckle Lothar Prix Christoph Faul Hermann Einsele and Gerhard Jahn)
|
