© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Transplant Infectious Disease, ISSN 1398-2273
Cytomegalovirus antiviral resistance: characterization of results from clinical specimens S. Kleiboeker, J. Nutt, B. Schindel, J. Dannehl, J. Hester. Cytomegalovirus antiviral resistance: characterization of results from clinical specimens. Transpl Infect Dis 2014: 16: 561–567. All rights reserved Abstract: Background. Cytomegalovirus (CMV) infections are a major cause of disease among immunocompromised patients. Prolonged antiviral therapy is often necessary to prevent or treat CMV disease, and may lead to development of antiviral resistance (AVR). Timely identification of viral mutations conferring resistance is essential for effective patient management. Methods. Amplification by polymerase chain reaction followed by bi-directional nucleotide sequencing was performed for relevant regions of the CMV UL97 and UL54 genes. Results from 570 samples submitted to a commercial reference laboratory for testing were reviewed and characterized with respect to the frequency of mutations detected, association with viral load (VL), and consistency of results in a subset of patients with multiple samples. Only AVR mutations confirmed by marker transfer experiments were included in the analysis. Results. AVR mutations were identified in 176 (30.9%) of the 570 samples evaluated. A total of 17 different UL97 mutations and 29 different UL54 mutations were detected. A single mutation per sample was most commonly observed, although 61 samples (10.7%) had >1 mutation, with 40 samples (7.0%) having mutations in both UL97 and UL54 genes. The VL of samples with AVR mutations ranged from 2.03–7.15 log10 copies/mL, and the VL did not differ significantly from samples without AVR mutations. A subset of patients (N = 85) had >1 sample tested, and 48.2% of these patients returned the same result for each sample analyzed, while the remainder had a different results. Conclusions. Genomic mutations conferring resistance in CMV to antiviral drugs were commonly identified in samples submitted from clinical patients to a reference laboratory for AVR testing. Mutations were identified over the full range of VLs and no correlation was identified between VL and the presence of AVR mutations. In patients with multiple samples submitted for analysis, approximately half of the patients had samples with variable results when the initial result was compared to subsequent results.
Cytomegalovirus (CMV) infection, a common complication affecting solid organ and hematopoietic stem cell transplant recipients, is frequently treated with antiviral therapy in immunocompromised patients (1). Genomic mutations in the CMV UL97 phosphotransferase gene confer resistance to ganciclovir, while mutations in the viral UL54 DNA polymerase gene confer resistance to
S. Kleiboeker, J. Nutt, B. Schindel, J. Dannehl, J. Hester Viracor-IBT Laboratories, Lee’s Summit, Missouri, USA
Key words: cytomegalovirus; antiviral; resistance; clinical; UL97; UL54 Correspondence to: Steven Kleiboeker, Viracor-IBT Laboratories, 1001 NW Technology Dr, Lee’s Summit, MO 64086, USA Tel: 816 554 5141 Fax: 816 251 6645 E-mail: [email protected]
Received 10 December 2013, revised 27 January 2014, 18 February 2014, accepted for publication 1 March 2014 DOI: 10.1111/tid.12241 Transpl Infect Dis 2014: 16: 561–567
ganciclovir, foscarnet, and cidofovir (2, 3). Risk factors for drug antiviral resistance (AVR) include prolonged (e.g., several months) antiviral therapy, intermittent low-level viral replication in the presence of drug caused by profound immunosuppression or suboptimal drug levels, and lack of prior immunity to CMV (4). AVR should be suspected after 2 or more weeks of full
561
Kleiboeker et al: CMV antiviral resistance in clinical specimens
ganciclovir dosing with increasing or unchanged viral loads (VLs) (5, 6). With greater adoption of extended prophylactic and empiric treatment strategies for prevention and treatment of disease caused by CMV, awareness has increased regarding the importance of timely and comprehensive CMV AVR assessment (7), particularly because of frequent multidrug-resistant and multiple resistant strains in transplant recipients (8).
Methods As recommended for the clinical care of immunocompromised patients suspected of having a drug-resistant CMV infection (9), laboratory testing was performed at the request of treating physicians to determine if drugresistance mutations were present in the viral genome
of the infecting CMV strain. Nucleotide sequence analysis for CMV AVR mutations was performed for relevant regions of the CMV UL97 and UL54 genes. A total of 28 UL97 and 43 UL54 AVR mutations previously reported in the literature (see www.viracoribt.com/ Test-Catalog/Detail/CMV-Antiviral-Resistance-5600 for a full list of citations for the AVR mutations shown in Table 1) were included in the viral genomic regions analyzed (Table 1). Only AVR mutations confirmed by marker transfer experiments were included in the analysis. To perform this testing, conventional polymerase chain reaction (PCR) followed by bi-directional Sanger (dideoxy) nucleotide sequencing was performed using standard methods. Briefly, total nucleic acid extraction using a Qiagen QIAcube (Qiagen, Valencia, California, USA) general
Cytomegalovirus antiviral resistance genomic mutations assessed by nucleotide sequence analysis and drug-resistance profiles conferred by the mutations UL54 UL97 (Ganciclovir)
Mutation
Drug(s)
Mutation
Drug(s)
D301N
C, G
Q578H
C, F, G
M460I
L595S
M460T
L595W
N408D
C, G
D588E
F
M460V
del595
N408K
C, G
D588N
F, G
V466G
del595-603
N410K
C, G
T700A
F
H520Q
E596G
F412C
C, G
V715M
F
del590-593
G598S
F412L
C, G
E756D
F
del591-594
K599T
F412S
C, G
E756K
F
del591-607
del600
F412V
C, G
E756Q
F
C592G
del601-603
D413A
C, G
L776M
F, G
A594E
C603R
D413E
C, G
V781I
F, G
A594P
C603S
N495K
F
V787L
F, G
A594T
C603W
L501I
C, G
L802M
F, G
A594V
C607F
T503I
C, G
K805Q
C
L595F
C607Y
K513N
C, G
A809V
F, G
K513E
C, G
V812L
C, F, G
L516R
C, G
T813S
C, F, G
I521T
C, G
T821I
F, G
P522A
C, G
A834P
C, F, G
P522S
C, G
T838A
F
L545S
C, G
G841A
C, F, G
L545W
C, G
del981-982
C, F, G
A987G
C, G
Drugs: C, cidofovir; F, foscarnet; G, ganciclovir.
Table 1
562
Transplant Infectious Disease 2014: 16: 561–567
Kleiboeker et al: CMV antiviral resistance in clinical specimens
protocol for plasma was performed, followed by PCR amplification using proof-reading FailSafe Taq polymerase with optimized pre-mixes (Epicentre; Madison, Wisconsin, USA, cat. no. FS99250). A single UL97 gene fragment and 4 UL54 gene fragments were amplified. Figure 1 shows a diagrammatic representation of the amplified fragments and genomic regions sequenced. The following oligonucleotide primers were used to amplify genomic fragments: UL97 forward 50 -ACGTTG GCCGACGCTATCA-30 , UL97 reverse 50 -GAGCATTC GTGGTAGAAGCGG-30 , UL54 fragment A forward 50 -G CTGGTGCTCCGTGAATCG-30 761–779, UL54 fragment A reverse 50 -CGTAAAGAGATGAAGACCCGAG TG-30 , UL54 fragment B forward 50 -GCTACGAGACGG AGGGAAAC-30 , UL54 fragment B reverse 50 -GCCGAC GCTGCTACTACTGTTACT-30 , UL54 fragment C forward 50 -GGCGGCTATGTTTCAGATGTC-30 , UL54 fragment C reverse 50 -CGTTCCCACTACATAGTCTTCCT GAT-30 , UL54 fragment D forward 50 -GCGCGGTTCAT CAAAGACAA-30 , UL54 fragment D reverse 50 -TCGTCC GTGCCATCAATC-30 . The following oligonucleotide primers were used for sequencing of the amplified fragments: UL97 forward 50 -ACGTTGGCCGACGCTATCA-30 , UL97 reverse 50 -GC GACACGAGGACATCTTGG-30 , UL54 fragment A forward 50 -GCTGGTGCTCCGTGAATCG-30 , UL54 fragment A reverse 50 -CGTAAAGAGATGAAGACCCGAG TG-30 , UL54 fragment B forward 50 -GCACTCGGGTCT TCATCTCTTTAC-30 , UL54 fragment B reverse 50 -GGT AGAGATGATAGCGGCGTTAG-30 , UL54 fragment C forward 50 -CAGTCAGGACGGCGTTTCA-30 , UL54 frag-
ment C reverse 50 -GCAAAAAACACGGCTCTGAAA-30 , UL54 fragment D forward 50 -TTTCAGAGCCGTGTTT TTTGC-30 , UL54 fragment D reverse 50 -TGCGGCAGGT TAGATTGACG-30 . Agarose gel electrophoresis of the PCR products demonstrated that amplicons of the correct size and acceptable purity were generated. For samples to be sequenced, both negative and positive control PCR reactions must have generated the expected results. Following amplicon purification by Qiagen QIAquick PCR purification (Qiagen; cat. no. 28106), BigDyeTerminator v3.1â cycle sequencing (Life Technologies, Carlsbad, California, USA, cat. no. 4337456), BigDye XTerminatorâ Purification (Life Technologies; cat. no. 4376487), and capillary sequencing were performed. Sequencing reactions with custom oligonucleotide primers were run on an ABI 3130XL or ABI 3730 capillary sequencer with ABI POP-7 Polymer (Life Technologies; cat. no. 4363929). Acceptance criteria were trace scores ≥30; median PUP scores ≥8 and control sequence matching 100% with the reference sequence in the reportable range. Sequences were analyzed with ABI’s Variant Reporter v1.0 software. Genomic analysis for AVR was performed for 570 CMV-positive plasma samples collected from US clinical patients between 28 October 2011 and 26 December 2012 and submitted to a commercial reference laboratory with a physician’s request for AVR testing. All samples analyzed were submitted as part of clinical patient care and the individual results were provided to the ordering physician. Results were analyzed for this
UL97 142,751→143409
UL54
78,181→79,043
78,963→79,686
79,587→80,605
80,490→80,827
Fig. 1. Diagrammatic representation of polymerase chain reaction (PCR) amplification products and sequencing primers for cytomegalovirus UL 97 (upper) and UL54 (lower) genes. The relative location of each PCR amplification product (amplicon) is shown by vertical lines with the sequence range below or above the amplicon. The portion of PCR amplicons sequenced for antiviral resistance mutation analysis is shown in shaded boxes with the 50 (forward) or 30 (reverse) sequence primer position listed diagonally. All sequence numbers shown correspond to the UL 97 or UL54 nucleotide numbers from GenBank accession no. FJ527563 (human herpesvirus 5 strain AD169, complete genome).
Transplant Infectious Disease 2014: 16: 561–567
563
Kleiboeker et al: CMV antiviral resistance in clinical specimens
study using fully de-identified data. Mann–Whitney and Kruskal–Wallis tests (1-way analysis of variance) with Dunn’s multiple comparison post-test, Chi-square analysis, and Wilcoxon matched-pairs signed rank test were performed using GraphPad Prism (v. 5.04) with 2-tailed P-values and alpha 1 mutation
16 (2.8)
4.35 (1.02)
2.76–6.81
NS2
19 (3.3)
3.47 (0.72)
2.13–4.69
NS2
5 (0.9)
3.32 (0.80)
2.29–4.29
NS2
40 (7.0)
3.95 (0.93)
2.72–6.13
NS2
UL97 mutation(s) only
UL54 mutation(s) only Single mutation >1 mutation Mutations in both UL97 and UL54 1
Mann–Whitney test, compared to CMV viral load (VL) values from samples with no mutations detected. Kruskal–Wallis test (1-way ANOVA) with Dunn’s multiple comparison post-test, compared to VL values from samples with no mutations detected and other groups with single or multiple mutations in either UL97, UL54, or UL97 and UL54. N, number; SD, standard deviation; NS, not significant. 2
Table 3
80 No mutation detected
70
No. of Samples
AVR mutation(s) detected
60 50 40 30 20 10 0
0. 5 1. 0 1. 5 2. 0 2. 5 3. 0 3. 5 4. 0 4. 5 5. 0 5. 5 6. 0 6. 5 7. 0 7. 5 8. 0
ranged from 0.9%, for specimens with multiple AVR mutations in UL54, to 16.8% for specimens with a single AVR mutation in UL97. The mean VL for specimens with no AVR mutations was 3.92 log10 copies/mL and the mean VLs for specimen groups with AVR mutations ranged from 3.32 log10 copies/mL, for samples with >1 UL54 AVR mutation, to 4.35 log10 copies/mL for samples with >1 UL97 AVR mutation. Statistically relevant associations were not observed when VLs were compared between the groups shown in Table 3. The association of VL and frequency of AVR mutation detection was further assessed following stratification of samples by VL in a frequency distribution histogram (Fig. 2). Observed frequencies of AVR mutations across the range of VLs were not statistically different from expected values based on the overall prevalence of AVR mutations in UL97 and/or UL54 (chi-square test, P = 0.9764). For a subset of 85 patients, multiple specimens were submitted for analysis. The mean number of multiple samples analyzed from this subset of patients was 2.36 (range 2–5). The mean time between sample collections was 86.7 days (range 2–324 days); for patients with >2 samples analyzed, the mean and range values apply to the first and last sample collected. A total of 48.2% of patients with multiple samples either had no mutation detected at each sampling or the same mutation detected (34.1% and 14.1%, respectively) in all samples (Table 4). Among the remaining patients, 23.5% had samples that initially contained no AVR mutations, while subsequent samples had 1 or more
CMV Viral Load (log10 copies/mL) Fig. 2. Frequency distribution histogram of viral load in clinical plasma specimens with and without cytomegalovirus (CMV) antiviral resistance (AVR) mutations.
mutations identified. For 3.5% of patients, AVR mutations were initially identified but, upon analysis of a subsequent sample, no mutations were detected, while 17.6% of patients had AVR mutation(s) initially detected but then had different mutation(s) upon analysis of subsequent samples. More complex patterns were noted in a total of 6 (7.1%) patients, for whom 3 or more samples were analyzed. When paired VLs for 28 patients who had specimens both positive and negative for AVR mutations were compared, no significant difference was identified (Wilcoxon matched-pairs signed rank test, P = 0.3550).
Transplant Infectious Disease 2014: 16: 561–567
565
Kleiboeker et al: CMV antiviral resistance in clinical specimens
Summary of results for patients with multiple samples analyzed for cytomegalovirus antiviral resistance mutations
Result for first sample analyzed
Subsequent result(s)
No. of patients (%) (N = 85)
No mutations detected
No mutations detected
29 (34.1)
No mutations detected
Mutation(s) detected
20 (23.5)
Mutation(s) detected
No mutations detected
Mutation(s) detected
Same mutation(s) detected
12 (14.1)
Mutation(s) detected
Different mutation(s) detected
15 (17.6)
No mutations detected
Different mutations detected in separate subsequent samples1
3 (3.5)
Mutation detected
No mutations detected, followed by detection of the same mutation identified in the initial sample1
2 (2.4)
No mutations detected
Mutations detected followed by no mutations detected1
1 (1.2)
1
3 (3.5)
Three or more samples analyzed per patient for each of the indicated categories.
Table 4
Discussion In patients with CMV infection, resistance to commonly used antiviral drugs may develop because of treatment without elimination of VL; these patients may have increasing VLs and/or progression of CMV-related disease (1, 2, 6). Drug-resistant CMV infections have been documented in 5–10% of the D+/R subset of solid organ recipients (9), with even higher values reported for kidney (10) and lung (2) recipients. In the present analysis, 30.9% of specimens analyzed were found to have CMV genomic mutations conferring AVR to 1 or more of the approved drugs. The increased prevalence compared to previous studies is likely a result of selection bias, as samples were submitted from patients with suspected AVR by the treating physician. VL has been considered an indicator of potential AVR in patients with CMV disease (11). In the present study, the association of VL and the presence of AVR mutations was examined in 3 ways: 1) analysis of the VLs for samples with no identified mutations, and single or multiple mutations in UL97, UL54, or both; 2) analysis by a frequency distribution histogram of specimens with and without AVR mutations; and 3) analysis of paired samples from individual patients in whom the VL of samples with no AVR mutations identified were compared to samples with AVR mutations. Statistically significant associations were not observed in any approach. Even in samples with the lowest VLs (
Cytomegalovirus antiviral resistance: characterization of results from clinical specimens S. Kleiboeker, J. Nutt, B. Schindel, J. Dannehl, J. Hester. Cytomegalovirus antiviral resistance: characterization of results from clinical specimens. Transpl Infect Dis 2014: 16: 561–567. All rights reserved Abstract: Background. Cytomegalovirus (CMV) infections are a major cause of disease among immunocompromised patients. Prolonged antiviral therapy is often necessary to prevent or treat CMV disease, and may lead to development of antiviral resistance (AVR). Timely identification of viral mutations conferring resistance is essential for effective patient management. Methods. Amplification by polymerase chain reaction followed by bi-directional nucleotide sequencing was performed for relevant regions of the CMV UL97 and UL54 genes. Results from 570 samples submitted to a commercial reference laboratory for testing were reviewed and characterized with respect to the frequency of mutations detected, association with viral load (VL), and consistency of results in a subset of patients with multiple samples. Only AVR mutations confirmed by marker transfer experiments were included in the analysis. Results. AVR mutations were identified in 176 (30.9%) of the 570 samples evaluated. A total of 17 different UL97 mutations and 29 different UL54 mutations were detected. A single mutation per sample was most commonly observed, although 61 samples (10.7%) had >1 mutation, with 40 samples (7.0%) having mutations in both UL97 and UL54 genes. The VL of samples with AVR mutations ranged from 2.03–7.15 log10 copies/mL, and the VL did not differ significantly from samples without AVR mutations. A subset of patients (N = 85) had >1 sample tested, and 48.2% of these patients returned the same result for each sample analyzed, while the remainder had a different results. Conclusions. Genomic mutations conferring resistance in CMV to antiviral drugs were commonly identified in samples submitted from clinical patients to a reference laboratory for AVR testing. Mutations were identified over the full range of VLs and no correlation was identified between VL and the presence of AVR mutations. In patients with multiple samples submitted for analysis, approximately half of the patients had samples with variable results when the initial result was compared to subsequent results.
Cytomegalovirus (CMV) infection, a common complication affecting solid organ and hematopoietic stem cell transplant recipients, is frequently treated with antiviral therapy in immunocompromised patients (1). Genomic mutations in the CMV UL97 phosphotransferase gene confer resistance to ganciclovir, while mutations in the viral UL54 DNA polymerase gene confer resistance to
S. Kleiboeker, J. Nutt, B. Schindel, J. Dannehl, J. Hester Viracor-IBT Laboratories, Lee’s Summit, Missouri, USA
Key words: cytomegalovirus; antiviral; resistance; clinical; UL97; UL54 Correspondence to: Steven Kleiboeker, Viracor-IBT Laboratories, 1001 NW Technology Dr, Lee’s Summit, MO 64086, USA Tel: 816 554 5141 Fax: 816 251 6645 E-mail: [email protected]
Received 10 December 2013, revised 27 January 2014, 18 February 2014, accepted for publication 1 March 2014 DOI: 10.1111/tid.12241 Transpl Infect Dis 2014: 16: 561–567
ganciclovir, foscarnet, and cidofovir (2, 3). Risk factors for drug antiviral resistance (AVR) include prolonged (e.g., several months) antiviral therapy, intermittent low-level viral replication in the presence of drug caused by profound immunosuppression or suboptimal drug levels, and lack of prior immunity to CMV (4). AVR should be suspected after 2 or more weeks of full
561
Kleiboeker et al: CMV antiviral resistance in clinical specimens
ganciclovir dosing with increasing or unchanged viral loads (VLs) (5, 6). With greater adoption of extended prophylactic and empiric treatment strategies for prevention and treatment of disease caused by CMV, awareness has increased regarding the importance of timely and comprehensive CMV AVR assessment (7), particularly because of frequent multidrug-resistant and multiple resistant strains in transplant recipients (8).
Methods As recommended for the clinical care of immunocompromised patients suspected of having a drug-resistant CMV infection (9), laboratory testing was performed at the request of treating physicians to determine if drugresistance mutations were present in the viral genome
of the infecting CMV strain. Nucleotide sequence analysis for CMV AVR mutations was performed for relevant regions of the CMV UL97 and UL54 genes. A total of 28 UL97 and 43 UL54 AVR mutations previously reported in the literature (see www.viracoribt.com/ Test-Catalog/Detail/CMV-Antiviral-Resistance-5600 for a full list of citations for the AVR mutations shown in Table 1) were included in the viral genomic regions analyzed (Table 1). Only AVR mutations confirmed by marker transfer experiments were included in the analysis. To perform this testing, conventional polymerase chain reaction (PCR) followed by bi-directional Sanger (dideoxy) nucleotide sequencing was performed using standard methods. Briefly, total nucleic acid extraction using a Qiagen QIAcube (Qiagen, Valencia, California, USA) general
Cytomegalovirus antiviral resistance genomic mutations assessed by nucleotide sequence analysis and drug-resistance profiles conferred by the mutations UL54 UL97 (Ganciclovir)
Mutation
Drug(s)
Mutation
Drug(s)
D301N
C, G
Q578H
C, F, G
M460I
L595S
M460T
L595W
N408D
C, G
D588E
F
M460V
del595
N408K
C, G
D588N
F, G
V466G
del595-603
N410K
C, G
T700A
F
H520Q
E596G
F412C
C, G
V715M
F
del590-593
G598S
F412L
C, G
E756D
F
del591-594
K599T
F412S
C, G
E756K
F
del591-607
del600
F412V
C, G
E756Q
F
C592G
del601-603
D413A
C, G
L776M
F, G
A594E
C603R
D413E
C, G
V781I
F, G
A594P
C603S
N495K
F
V787L
F, G
A594T
C603W
L501I
C, G
L802M
F, G
A594V
C607F
T503I
C, G
K805Q
C
L595F
C607Y
K513N
C, G
A809V
F, G
K513E
C, G
V812L
C, F, G
L516R
C, G
T813S
C, F, G
I521T
C, G
T821I
F, G
P522A
C, G
A834P
C, F, G
P522S
C, G
T838A
F
L545S
C, G
G841A
C, F, G
L545W
C, G
del981-982
C, F, G
A987G
C, G
Drugs: C, cidofovir; F, foscarnet; G, ganciclovir.
Table 1
562
Transplant Infectious Disease 2014: 16: 561–567
Kleiboeker et al: CMV antiviral resistance in clinical specimens
protocol for plasma was performed, followed by PCR amplification using proof-reading FailSafe Taq polymerase with optimized pre-mixes (Epicentre; Madison, Wisconsin, USA, cat. no. FS99250). A single UL97 gene fragment and 4 UL54 gene fragments were amplified. Figure 1 shows a diagrammatic representation of the amplified fragments and genomic regions sequenced. The following oligonucleotide primers were used to amplify genomic fragments: UL97 forward 50 -ACGTTG GCCGACGCTATCA-30 , UL97 reverse 50 -GAGCATTC GTGGTAGAAGCGG-30 , UL54 fragment A forward 50 -G CTGGTGCTCCGTGAATCG-30 761–779, UL54 fragment A reverse 50 -CGTAAAGAGATGAAGACCCGAG TG-30 , UL54 fragment B forward 50 -GCTACGAGACGG AGGGAAAC-30 , UL54 fragment B reverse 50 -GCCGAC GCTGCTACTACTGTTACT-30 , UL54 fragment C forward 50 -GGCGGCTATGTTTCAGATGTC-30 , UL54 fragment C reverse 50 -CGTTCCCACTACATAGTCTTCCT GAT-30 , UL54 fragment D forward 50 -GCGCGGTTCAT CAAAGACAA-30 , UL54 fragment D reverse 50 -TCGTCC GTGCCATCAATC-30 . The following oligonucleotide primers were used for sequencing of the amplified fragments: UL97 forward 50 -ACGTTGGCCGACGCTATCA-30 , UL97 reverse 50 -GC GACACGAGGACATCTTGG-30 , UL54 fragment A forward 50 -GCTGGTGCTCCGTGAATCG-30 , UL54 fragment A reverse 50 -CGTAAAGAGATGAAGACCCGAG TG-30 , UL54 fragment B forward 50 -GCACTCGGGTCT TCATCTCTTTAC-30 , UL54 fragment B reverse 50 -GGT AGAGATGATAGCGGCGTTAG-30 , UL54 fragment C forward 50 -CAGTCAGGACGGCGTTTCA-30 , UL54 frag-
ment C reverse 50 -GCAAAAAACACGGCTCTGAAA-30 , UL54 fragment D forward 50 -TTTCAGAGCCGTGTTT TTTGC-30 , UL54 fragment D reverse 50 -TGCGGCAGGT TAGATTGACG-30 . Agarose gel electrophoresis of the PCR products demonstrated that amplicons of the correct size and acceptable purity were generated. For samples to be sequenced, both negative and positive control PCR reactions must have generated the expected results. Following amplicon purification by Qiagen QIAquick PCR purification (Qiagen; cat. no. 28106), BigDyeTerminator v3.1â cycle sequencing (Life Technologies, Carlsbad, California, USA, cat. no. 4337456), BigDye XTerminatorâ Purification (Life Technologies; cat. no. 4376487), and capillary sequencing were performed. Sequencing reactions with custom oligonucleotide primers were run on an ABI 3130XL or ABI 3730 capillary sequencer with ABI POP-7 Polymer (Life Technologies; cat. no. 4363929). Acceptance criteria were trace scores ≥30; median PUP scores ≥8 and control sequence matching 100% with the reference sequence in the reportable range. Sequences were analyzed with ABI’s Variant Reporter v1.0 software. Genomic analysis for AVR was performed for 570 CMV-positive plasma samples collected from US clinical patients between 28 October 2011 and 26 December 2012 and submitted to a commercial reference laboratory with a physician’s request for AVR testing. All samples analyzed were submitted as part of clinical patient care and the individual results were provided to the ordering physician. Results were analyzed for this
UL97 142,751→143409
UL54
78,181→79,043
78,963→79,686
79,587→80,605
80,490→80,827
Fig. 1. Diagrammatic representation of polymerase chain reaction (PCR) amplification products and sequencing primers for cytomegalovirus UL 97 (upper) and UL54 (lower) genes. The relative location of each PCR amplification product (amplicon) is shown by vertical lines with the sequence range below or above the amplicon. The portion of PCR amplicons sequenced for antiviral resistance mutation analysis is shown in shaded boxes with the 50 (forward) or 30 (reverse) sequence primer position listed diagonally. All sequence numbers shown correspond to the UL 97 or UL54 nucleotide numbers from GenBank accession no. FJ527563 (human herpesvirus 5 strain AD169, complete genome).
Transplant Infectious Disease 2014: 16: 561–567
563
Kleiboeker et al: CMV antiviral resistance in clinical specimens
study using fully de-identified data. Mann–Whitney and Kruskal–Wallis tests (1-way analysis of variance) with Dunn’s multiple comparison post-test, Chi-square analysis, and Wilcoxon matched-pairs signed rank test were performed using GraphPad Prism (v. 5.04) with 2-tailed P-values and alpha 1 mutation
16 (2.8)
4.35 (1.02)
2.76–6.81
NS2
19 (3.3)
3.47 (0.72)
2.13–4.69
NS2
5 (0.9)
3.32 (0.80)
2.29–4.29
NS2
40 (7.0)
3.95 (0.93)
2.72–6.13
NS2
UL97 mutation(s) only
UL54 mutation(s) only Single mutation >1 mutation Mutations in both UL97 and UL54 1
Mann–Whitney test, compared to CMV viral load (VL) values from samples with no mutations detected. Kruskal–Wallis test (1-way ANOVA) with Dunn’s multiple comparison post-test, compared to VL values from samples with no mutations detected and other groups with single or multiple mutations in either UL97, UL54, or UL97 and UL54. N, number; SD, standard deviation; NS, not significant. 2
Table 3
80 No mutation detected
70
No. of Samples
AVR mutation(s) detected
60 50 40 30 20 10 0
0. 5 1. 0 1. 5 2. 0 2. 5 3. 0 3. 5 4. 0 4. 5 5. 0 5. 5 6. 0 6. 5 7. 0 7. 5 8. 0
ranged from 0.9%, for specimens with multiple AVR mutations in UL54, to 16.8% for specimens with a single AVR mutation in UL97. The mean VL for specimens with no AVR mutations was 3.92 log10 copies/mL and the mean VLs for specimen groups with AVR mutations ranged from 3.32 log10 copies/mL, for samples with >1 UL54 AVR mutation, to 4.35 log10 copies/mL for samples with >1 UL97 AVR mutation. Statistically relevant associations were not observed when VLs were compared between the groups shown in Table 3. The association of VL and frequency of AVR mutation detection was further assessed following stratification of samples by VL in a frequency distribution histogram (Fig. 2). Observed frequencies of AVR mutations across the range of VLs were not statistically different from expected values based on the overall prevalence of AVR mutations in UL97 and/or UL54 (chi-square test, P = 0.9764). For a subset of 85 patients, multiple specimens were submitted for analysis. The mean number of multiple samples analyzed from this subset of patients was 2.36 (range 2–5). The mean time between sample collections was 86.7 days (range 2–324 days); for patients with >2 samples analyzed, the mean and range values apply to the first and last sample collected. A total of 48.2% of patients with multiple samples either had no mutation detected at each sampling or the same mutation detected (34.1% and 14.1%, respectively) in all samples (Table 4). Among the remaining patients, 23.5% had samples that initially contained no AVR mutations, while subsequent samples had 1 or more
CMV Viral Load (log10 copies/mL) Fig. 2. Frequency distribution histogram of viral load in clinical plasma specimens with and without cytomegalovirus (CMV) antiviral resistance (AVR) mutations.
mutations identified. For 3.5% of patients, AVR mutations were initially identified but, upon analysis of a subsequent sample, no mutations were detected, while 17.6% of patients had AVR mutation(s) initially detected but then had different mutation(s) upon analysis of subsequent samples. More complex patterns were noted in a total of 6 (7.1%) patients, for whom 3 or more samples were analyzed. When paired VLs for 28 patients who had specimens both positive and negative for AVR mutations were compared, no significant difference was identified (Wilcoxon matched-pairs signed rank test, P = 0.3550).
Transplant Infectious Disease 2014: 16: 561–567
565
Kleiboeker et al: CMV antiviral resistance in clinical specimens
Summary of results for patients with multiple samples analyzed for cytomegalovirus antiviral resistance mutations
Result for first sample analyzed
Subsequent result(s)
No. of patients (%) (N = 85)
No mutations detected
No mutations detected
29 (34.1)
No mutations detected
Mutation(s) detected
20 (23.5)
Mutation(s) detected
No mutations detected
Mutation(s) detected
Same mutation(s) detected
12 (14.1)
Mutation(s) detected
Different mutation(s) detected
15 (17.6)
No mutations detected
Different mutations detected in separate subsequent samples1
3 (3.5)
Mutation detected
No mutations detected, followed by detection of the same mutation identified in the initial sample1
2 (2.4)
No mutations detected
Mutations detected followed by no mutations detected1
1 (1.2)
1
3 (3.5)
Three or more samples analyzed per patient for each of the indicated categories.
Table 4
Discussion In patients with CMV infection, resistance to commonly used antiviral drugs may develop because of treatment without elimination of VL; these patients may have increasing VLs and/or progression of CMV-related disease (1, 2, 6). Drug-resistant CMV infections have been documented in 5–10% of the D+/R subset of solid organ recipients (9), with even higher values reported for kidney (10) and lung (2) recipients. In the present analysis, 30.9% of specimens analyzed were found to have CMV genomic mutations conferring AVR to 1 or more of the approved drugs. The increased prevalence compared to previous studies is likely a result of selection bias, as samples were submitted from patients with suspected AVR by the treating physician. VL has been considered an indicator of potential AVR in patients with CMV disease (11). In the present study, the association of VL and the presence of AVR mutations was examined in 3 ways: 1) analysis of the VLs for samples with no identified mutations, and single or multiple mutations in UL97, UL54, or both; 2) analysis by a frequency distribution histogram of specimens with and without AVR mutations; and 3) analysis of paired samples from individual patients in whom the VL of samples with no AVR mutations identified were compared to samples with AVR mutations. Statistically significant associations were not observed in any approach. Even in samples with the lowest VLs (
