- Email: [email protected]
High-level resistance to zidovudine but not to zalcitabine or didanosine in human immunodeficiency virus from children receiving antiretroviral therapy
High-level resistance to zidovudine but not to zalcitabine or didanosine in human immunodeficiency virus from children receiving antiretroviral therapy Robert N. Husson, MD,* T a k u m a Shirasaka, MD, PhD, Karina M. Butler, MB,BCh, Philip A. Pizzo, MD, a n d Hiroaki Mitsuya, MD, PhD From the Experimental Retrovirology Section, Medicine Branch, and the Infectious Diseases Section, Pediatric Branch, National Cancer Institute, Bethesda, Maryland Human i m m u n o d e f i c i e n c y virus type I (HIV-t) isolates from children receiving long-term therapy with an alternating regimen of zidovudine and zalcitabine, or with didanosine monotherapy, were e v a l u a t e d for resistance to zidovudine, zaIcitabine, and didanosine, and for mutations known to be associated with zidovudine or didanosine resistance. HIV-t from four of six patients receiving zid o v u d i n e with zalcitabine d e v e l o p e d high-level resistance to zidovudine. A mutation in the HIV-t reverse transcriptase that is highly associated with zidovudine resistance was identified in all four zidovudine-resistant posttherapy isolates. In contrast, none of the HIV-t isolates from the seven patients receiving didanosine d e v e l o p e d high-level resistance to this agent, despite the identification of a didanosine-associated mutation in six of these posttherapy isolates, although small decreases in sensitivity to didanosine were observed. These results indicate that nucleoside a n a l o g - a s s o c i a t e d mutations in HIV-t occur frequently in children receiving long-term antiretroviral therapy and that alternating c o m b i n a t i o n therapy does not prevent the d e v e l o p m e n t of resistance to zidovudine. They also suggest that there may be differences in the d e g r e e of resistance conferred by mutations that result from therapy with different nucleoside analogs. These findings underscore the need for studies to define the c l i n i c a l importance of these mutations, and for treatment strategies to overc o m e the e m e r g e n c e of viral resistance in vivo. (J PEDIATR1993;123:9-16)
The development of nucleoside analog for the treatment of infection with human immunodeficiencyvirus has resulted in improved survival and decreased morbidity rates for individuals infected with this virus. 1-3The use of these agents, however, is limited by their toxicity and by their lack of long-term efficacy in preventing progression of HIV disease in many patients. 25 The development of viral resistance to Submitted for publication Dec. 30, 1992; accepted March 3, 1993. Reprint requests: Robert N. Husson, MD, Division of Infectious Diseases, Children's Hospital, 300 Longwood Ave., Boston, MA 02115. *Now at Children's Hospital, Harvard Medical School, Boston, Massachusetts. 9/25/46919
these agents, initially described for zidovudine6 and subsequently for didanosine,7 in clinical isolates of HIV- 1 may be a significant factor in the incomplete efficacy of these agents in vivo. Indeed, the emergence of HIV-1 isolates with decreased susceptibility to AZT has been linked to deteriSee related article, p. 1. oration in clinical status in children being treated with thisagent. ~ The occurrence of viral resistance to nucleoside analogs has led to changes in approaches to the development and clinical use of antiretroviral agents. It is now essential for new antiretroviral agents to be assessed early in their devel-
10
H u s s o n et al.
AZT CN5o ddC dd[ HIV- 1 PBMC PCR TCID5o
3' -Azido-2',3'-dideoxythymidine(zidovudine) Drug concentration required to give negative results (no p24 gag protein) in 50% of culture wells 2 ' ,3 ' -Dideoxycytidine (zalcitabine) 2 ',3 '-Oideoxyinosine (didanosine) Human immunodeficiency virus type 1 Peripheral blood mononuclear cell Polymerase chain reaction Median (50%) tissue culture infective dose
opment for their ability to induce viral resistance in vitro and in initial clinical studies. Treatment approaches increasingly focus on combination therapies in an attempt to enhance antiviral activity and prevent or slow the development of resistance. The identification of mutations in HIV-I reverse transcriptase associated with the development of the resistant phenotype, 7, 9, m together with new information on the three-dimensional structure of this enzyme, Ll provides insight into the molecular basis of this resistance and have led to the development of m o l e c u l a r techniques for the rapid identification of resistance mutations in clinical isolates of HIV-1. 7' 12-t4 Although A Z T resistance occurs frequently in adults receiving long-term therapy with this agent, its incidence in children, the factors that contribute to its occurrence, and its clinical implications are less well understood. Resistance to ddI is less well characterized, and its frequency in both adults and children receiving long-term therapy, as well as its clinical significance, is unknown. In this study, we evaluated the occurrence of resistance to A Z T in HIV- 1 isolates from children receiving long-term therapy with a sequential combination of A Z T and zalcitabine, and the occurrence of resistance to ddI in HIV-1 isolates from children receiving long-term therapy with this drug as a single agent. In addition, we optimized a selective polymerase chain reaction technique to identify in these isolates a mutation known to be highly associated with A Z T resistance, and the one previously described ddI-associated mutation. METHODS
Nucleosides. The A Z T was purchased from Sigma Chemical Co. (St. Louis, Mo.); ddC and ddI were provided by the Developmental Therapeutics Program, Division of Cancer Treatment, National Cancer Institute (Bethesda, Md.). Isolation and titration of pretherapy and posttherapy clinical isolates of HIV-I. Patients were enrolled in antiretroviral therapy protocols incorporating monthly blood sampling in the Pediatric Branch, National Cancer Institute.I5, l6 Informed consent was obtained from the patient's parent or guardian before enrollment. The AZT-ddC treatment regimen consisted of 8 weeks of ddC followed by 4
The Journal of Pediatrics July 1993
weeks without therapy, after which an alternating schedule of 3 weeks of AZT followed by 1 week of ddC continued throughout the study period. In the alternating regimen, the AZT dose was 180 m g / m 2 per dose administered orally every 6 hours, and the ddC dose was 0.02 to 0.04 mg/kg per dose administered orally every 6 hours. None of the patients enrolled in this study had been treated previously with antiretroviral agents. The ddI treatment regimen consisted of ddI administered orally as monotherapy throughout the study period, at doses ranging from 40 to 120 m g / m 2 per dose administered every 8 hours. Two patients (Nos. I- 1 and 1-4) whose HIV-1 isolates were evaluated in this study had been treated previously with AZT-containing regimens, including one (No. 1-4) who had been treated with the alternating AZT-ddC regimen. All pretherapy HIV-1 isolates were obtained from previously frozen nonfractionated peripheral blood mononuclear cells remaining after flow cytometric analysis of T-cell subsets. Posttherapy isolates were obtained from frozen or fresh PBMCs. Patients' cells were cultured in the presence of HIV-1 seronegative donor phytohemagglutinin-stimulated PBMCs as previously describedJ 7 When the p24 gag protein concentration in the culture supernatant was >2 ng/ml by radioimmunoassay (Du Pont Co., Wilmington, Del.), the supernatant was collected and stored in aliquots at - 7 0 ~ C. These primary virus isolates were titered in quadruplicate by culturing fivefold serial dilutions of virus with 106 phytohemagglutinin-stimulated PBMCs. The supernatant was tested on day 7 for p24 gag protein by radioimmunoassay, and the median tissue culture infective dose was determined by the Karber method3 s Each pair of pretherapy and posttherapy HIV-1 isolates was titered in the same experiment with the same batch of PBMCs to minimize the variation inherent in the use of PBMCs as target cells. Sensitivity testing of HIV-1 clinical isolates. Titered primary virus isolates of pretherapy and posttherapy HIV-I isolates were diluted in medium to a concentration of approximately I TCID50/~I, and 20 TCIDs0 was then added to 106 phytohemagglutinin-stimulated PBMCs in a 24-well plate in the presence of serial dilutions of AZT, ddC, or ddI. Isolates were cultured in the continuous presence of drug, and a 50% medium change with drug-containing medium was performed on day 3 or 4. On day 7, the culture supernatant was tested for p24 gag protein by radioimmunoassay (Du Pont). The sensitivity of each strain was defined as the drug concentration required for 50% of the wells to contain no p24 gag protein, as previously described) 7 The drug concentrations used in these assays were, for AZT: 25, 5.0, 1.0, 0.2, and 0.04 ~mol/L; for ddC: 5.0, 1.0, 0.2, 0.04, and 0.008/~mol/L; and for ddl: 25, 12.5, 6.3, 3.1, 1.6, and 0.8 ~mol/L. Each assay was performed in
The Journal o f Pediatrics Volume 123, Number 1
quadruplicate, and each pair of pretherapy and posttherapy HIV-1 isolates was tested in the same experiment by using the same batch of PBMCs. In addition, in each sensitivity assay, a limited titration was performed to verify the accuracy of the initial titration.
Selective amplification of wild-type and mutant proviral DNA. Each pretherapy and posttherapy HIV-1 primary isolate was cultured in the absence of drug with phytohemagglutinin-stimulated target PBMCs. Once the culture supernatant contained p24 gag protein, the cells were harvested, washed with phosphate-buffered saline solution, and lysed by incubation of cells in lysis buffer (KC1, 50 mmol/L; Tris-HCl [pH 8.3], 10 retool/L; MgC12, 2 retool/L; a nonionic detergent [Nonidet P-40; Sigma Chemical Co,, St. Louis, Mo.], 0.045% [vol/vol]; and proteinase K, 100 #g/ ml) at 55 ~ C for 60 minutes, followed by incubation for 12 minutes at 95 ~ C to inactivate proteinase K. Cell lysates were subjected to two-step PCR assay. The first amplification step used previously described primer pairs and reaction conditions to amplify a 1.7-kilobase fragment containing the HIV-1 reverse transcriptase. 9 The product of this amplification was then diluted 10-fold, and 1 to 10 #1 of this dilution was then subjected to selective PCR assay to identify wild-type or mutant sequences in the reverse transcriptase coding region. For codon 215, AZTrelated mutations in two bases result in a change from threonine to tyrosine or phenylalanine; for codon 74, the ddl-related mutation in a single base results in a change from leucine to valine. Previously published primer pairs 12 were used with modified conditions to enhance the specificity of these reactions. For detection of the codon 215 wildtype or mutant sequence, reaction components in a final volume of 100 #1 were as follows: KC1, 25 retool/L; Tris-HC1 (pH 8.3), 10 mmol/L; MgCIz, 1.5 mmol/L; deoxyadenosine triphosphate, deoxyguanosine triphosphate, deoxycytidine triphosphate, and deoxythymidine triphosphate, 0.25 mmol/L each; sonicated salmon sperm DNA, 10 ~g/ml; and 50 pmol of each primer. The amplification was performed in a thermal cycler (Perkin-Elmer/ Cetus, Norwalk, Conn.), at temperatures of 94 ~ C for 1 minute, 55 ~ C for 30 seconds, and 72 ~ C for 30 seconds, for 30 cycles. For detection of the codon 74 wild-type or mutant sequence, the conditions were identical except that the annealing temperature was 47 ~ C. The sensitivity and specificity of the selective amplification step under these conditions were determined by using plasmid ERS200E1, a plasmid containing the reverse transcriptase gene from a clinical isolate that had been sequenced and found to be a wild type at both the 74 and 215 codons, and plasmid ERS203L2-8, a plasmid containing the reverse transcriptase gene from a clinical isolate that had
H u s s o n et al.
11
been sequenced and found to contain both the codon 215 AZT-associated mutations and the codon 74 ddI-associated mutation. ~7 Copy numbers of the reverse transcriptase template were calculated from the absorbance at 260 nm of the purified plasmid DNA. Amplification products were electrophoresed through 1.2% agarose, stained with ethidium bromide, and visualized with ultraviolet light. Each pretherapy and posttherapy lysate was amplified with primers to detect the wild type or mutant base in parallel reactions in the same experiment, and negative controls were included in each experiment. Selective amplification results were categorized as wild type if there was a clearly visible band of the correct size resulting from amplification with the wild-type primers and no clearly visible band resulting from amplification with the mutant primers, m i x e d if bands of the appropriate size were clearly visible after amplification with wild-type and mutant primers, and m u t a n t if there was a clear band of the correct size resulting from amplification with the mutant primers and no clearly visible band resulting from amplification with the wild-type primers. RESULTS Pretherapy and posttherapy HIV- 1 strains were obtained from six patients treated with an alternating regimen of A Z T and ddC, and from seven patients treated with ddI as a single agent. All patients had symptomatic HIV infection (Centers for Disease Control and Prevention class P-2). The median CD4 cell count at the time of study entry of the patients treated with AZT plus ddC was 744 cells/ram 3 (range, 145 to 1944); of the patients treated with ddI, it was 265 cells/mm 3 (range, 0 to 1544). At the time the posttherapy isolate was obtained, the patients receiving A Z T with ddC had been treated on this regimen for a median of 97 weeks (range, 81 to 128), and the patients receiving ddI had been treated with this agent for a median of 66 weeks (range, 49 to 92). The results of sensitivity testing for AZT, ddC, and ddI for each pair of isolates are shown in Table I. Substantial increases in CN50 for AZT were seen in four of six posttherapy isolates from patients receiving AZT plus ddC, indicating the development of decreased sensitivity of these isolates to AZT. In contrast, with the possible exceptions of patients AC-1 and AC-3, appreciable changes in sensitivity to ddC or ddI in isolates from patients on this regimen were not observed. Among the posttherapy isolates from patients receiving ddI monotherapy, no substantial changes in CNs0 for AZT, ddC, or ddI were observed (Table II). Despite the lack of marked increase in CNso after ddI therapy, five of seven posttherapy isolates did show small increases in CNs0 of 1.4 to 2.9 times the pretherapy value. Although the magnitude
12
Husson et al.
The Journal of Pediatrics July 1993
Table I. Changes in sensitivity of HIV-I isolates from patients receiving AZT with ddC Patient No. AC-1 AC-2 AC-3 AC-4 AC-5 AC-6
Age at entry (yr) 5.5 1.3 0,7 10.7 12.2 3.8
AZT-ddC close* 180/0.02-0.04 180/0,03 180/0.04 180/0.01-0.03 180/0.02-0.03 180/0.02-0.03
Weeks on study
AZT CNso (pmol/L) Pre/post
ddC CN50 (pmol/L) Pre/post
ddl C N 5 0 (pmol/L) Pre/post
Codon 245 genotype Pre/post
92 101 80 82 103 128
0.13/2.24 0.04/0,13 0.04/25.0 0.09/0.67 0.07/5.0 0.09/0.09
0.67/2.24 0.45/0.45 0.06/0.45 0.09/0.09 0,30/0.30 0,47/0.30
5.3/8,9 3.8/7,5 5.3/15 8.9/8.9 3.1 / 3.1 2.1/2.2
WT/MIX WT/WT WT/MUT WT/MIX WT/MUT WT/WT
Pre, Pretherapy;post, posttherapy; WT, wild-typesequenee;MIX, mixedsequence;MUT, mutantsequence.
*Thedoseof AZT is presentedas milligramsper squaremeterper dose administeredevery6 hours;the doseof ddC is presentedas milligramsper kilogramper dose administeredevery 6 hours.
Table II. Changes in sensitivity of HIV-1 isolates from patients receiving ddI Patient No,
Age at entry (yr)
I-1 I-2 I-3 I-4 I-5 I-6 I-7
4,8 8.5 5.4 12.4 7.5 2.7 8.4
Dose of ddl* 40-120 90 90-120 90 40-120 90-120 90
Weeks on study
AZT CNso (~mol/L) pre/post
dclC CNso (grnol/L) pre/post
ddl C N 5 0 (prnol/L) pre/post
Codon 74 genotype pre/post
66 78 92 63 49 50 69
3.35/1.50 0.13/0.04 0.09/0.09 0.67/0.06 0.15/0.05 0.04/0.04 0.09/0,06
0.30/2.24 0.67/2.20 0.20/2.24 0.09/0.09 0.75/1.12 0.09/0.09 0.20/0.45
2.7,/6.3 7.5/10,6 4.5/10.6 8.9/8.9 2.2/6.3 4.5/4.5 5.3/8.9
WT/MUT WT/MIX WT/MUT WT/MIX WT/MIX WT/WT WT/MIX
Pre, Pretherapy;post, posttherapy; WT, wild-typesequence;MIX, mixedsequence;MUT, mutantsequence.
*The dose of ddl is presentedas milligramsper square meter per dose administeredevery 8 hours. of these changes is small, analysis of paired values suggests that these decreases in CNs0 are statistically significant (p = 0.04 by the Wilcoxon sign rank test). In addition, in five of seven post-ddl isolates, the CN50 for AZT decreased, consistent with previous observations suggesting increased AZT sensitivity in AZT-resistant isolates that acquire this ddI-associated mutation] Because the phenotypic analysis of drug sensitivity of HIV-1 is labor intensive, time-consuming, and expensive, and in particular because the magnitude of changes in sensitivity to ddI of posttherapy isolates was invariably small, we sought to apply a rapid test to detect the presence of AZT- or ddI-associated mutations in these isolates. The mutations resulting in a change in the codon for amino acid 215 from threonine to phenylalanine or tyrosine are among the earliest and most persistent AZT-associated mutations observed in clinical isolates, and they appear to be consistently associated with AZT resistance. 19, 2o The mutation that was initially associated with ddI resistance results in a change in amino acid 74 from leucine to valine.7 We therefore modified selective amplification procedures for the identification of mutations in these codons to provide molecular data complementary to the phenotypic sensitivity results described above. We optimized the selective step of this PCR assay to
maximize its specificity, so the conditions used in this study are more stringent than those originally described, particularly for amplification of the codon 215 sequence], 12 To assess the sensitivity and specificity of this assay under these conditions, we amplified serial dilutions of plasmid DNA containing the 1.7-kilobase reverse transcriptase insert determined by sequence analysis to have wild-type or mutant codons at positions 74 and 215 ~7 (Fig. 1). The limits of detection were 2 X 105 copies for codon 215 for both the wildtype and mutant sequence, and 2 x 104 copies for codon 74 for both the wild-type and mutant sequence. The specificity of these reactions was tested at high concentrations of template DNA (Fig. 2). At concentrations up to 2 • 101~copies per 100 ~1 reaction volume, these amplification conditions were highly specific for both sequences. At this concentration of template for the codon 215 sequence, there was no amplification of mutant DNA by wild-type primers and a minimally visible band from amplification of wild-type DNA by mutant primers. Similarly, at this concentration of template for the codon 74 sequence, there was no amplification of wild-type DNA by mutant primers and a minimally visible band from amplification of mutant DNA by wild-type primers. In mixing experiments of mutant and wild-type plasmid DNA, it was determined that amplification of mutant or wild-type DNA
The Journal of Pediatrics Volume 123, Number 1
Husson et al.
13
Fig. t. Sensitivity•fse•ectivePCRassayf•rc•d•ns74and2l5.Tenf•lddiluti•ns•fp•asmidDNAc•ntainingwild-type (WT) or mutant (Mut) sequence were amplified to determine the lower limit of detection of this selective PCR assay: a, 2 • 10 9 copies; b, 2 • 108; c, 2 • 107; d, 2 X 106; e, 2 • 105;f, 2 • 104.
Fig. 2. Specificity of selective PCR assay for codons 74 and 215. Tenfold dilutions of plasmid DNA containing wild-type (WT) or mutant (Mut) sequence were amplified with primers specific for the wild-type (W) or the mutant (M) sequence to determine the specificity of these primers at high concentrations of template DNA. a, 2 X 101~ copies; b, 2 x 109; c, 2X lOg.
Fig. 3. Tw••stepse•ectiveamp•ificati•n•fpr•vira•DNAfr•mcu•tures•fpatientis••ates•fHIV-•bef•re(pre)andafter (post) therapy with ddI or with AZT plus ddC. Each pretherapy and posttherapy isolate was subjected to primary nonselective PCR assay followed by selective PCR assay with primers specific for the wild-type sequence (W) or for the mutant sequence (M) for codon 74 for isolates from patients treated with ddI, or for codon 215 for isolates from patients treated with AZT plus ddC. depended primarily on the absolute concentration of the t e m p l a t e molecule r a t h e r t h a n on the ratio of m u t a n t to wild-type template ( d a t a not shown). Thus, under these conditions, the selective amplification was sensitive and specific for a wide r a n g e of input t e m p l a t e D N A at codons 74 and 215. This r a n g e encompassed the a m o u n t of input t e m p l a t e D N A in the selective P C R assay u n d e r the conditions t h a t we used; this assay was therefore both sensitive
and specific for the detection of wild-type and m u t a n t codons at positions 74 a n d 215 in the clinical isolates evaluated in this study. This selective P C R assay did not, however, provide quantitative information on the proportion of strains t h a t contain the m u t a t i o n of interest. The results of this two-step, selective P C R assay in the detection of m u t a n t and wild-type sequence in codon 215 in isolates obtained before a n d after therapy with A Z T plus
14
Husson et al.
ddC are shown in Fig. 3, a. All six pretherapy isolates had only the wild-type sequence present, whereas four (67%) of six posttherapy isolates had the mutant sequence present, including two in which the population was mixed and two in which only the mutant codon was detected. The presence of the codon 215 mutant sequence was associated with decreased sensitivity to AZT in all four isolates in which it was observed. As summarized in Table I, the two posttherapy isolates in which only the mutant eodon 215 was detected had the greatest increase in CNs0 (71 and 625 times the pretherapy CNso) and the highest absolute CNs0 (5.0 and 25.0 ~mol/L), and those in which this codon was a mixture of mutant and wild type had intermediate changes in CNs0 (7.4 and 17.2 times the pretherapy CNs0) and absolute posttherapy CNso (0.67 and 2.24 #mol/L). In the posttherapy isolates in which only the wild-type codon was detected, the CNso was in the same range as that of pretherapy isolates (0.09 and 0.13 ~mol/L). The results of selective PCR assay in the detection of the presence of mutant and wild-type bases in codon 74 in preand post-ddI isolates is shown in Fig. 3, b. All seven pre-ddI therapy isolates had only the wild-type sequence present in this position, whereas six (86%) of seven posttherapy isolates had the mutant sequence present, including three in which the population was mixed and three in which only the mutant codon was detected. As shown in Table II, examination of these results in relation to the phenotypic results revealed that the CNso increased in five post-ddI isolates in which the mutant base was detected and was unchanged in the sixth. There was no change in CNs0 in the post-ddI isolate in which only wild-type sequence was detected. Although large changes in ddC sensitivity were not observed in isolates from patients in either study arm, small decreases in ddC sensitivity (_
The Journal of Pediatrics July 1993
phenotypic analysis were supported by the presence of the codon 215 mutation in HIV-I reverse transcriptase that is thought to be most consistently associated with AZT resistance 19, 20 in all four isolates in which phenotypic resistance was observed. In contrast, large changes in sensitivity to ddI were not observed, despite the presence of the codon 74 mutation reported to be associated with decreased susceptibility to ddI in six of seven isolates. Similarly, large changes in ddC sensitivity were not observed, although small increases in the ddC CNso occurred in five posttherapy isolates from children receiving ddI and in two isolates from children receiving both AZT and ddC. A notable result of this study is the failure of alternating combination therapy to prevent the emergence of high-level resistance to AZT, similar to our findings in a study of HIV-1 isolates from adults treated with a different schedule of AZT alternating with ddC. ~7 It is possible that this AZT-ddC regimen may have slowed the development of AZT resistance, or that other combination regimens may slow the emergence of drug resistance more effectively; this question can be answered only in prospective comparative trials. The ability of AZT- and ddl-related mutations to coexist in vivo in the virus population of individual patients, however, appears to make it unlikely that combinations of these nucleoside analogs will prevent the emergence of viral resistance in vivo. An interesting and potentially important finding in this study is the high frequency of the codon 74 ddl mutation in children treated with ddI as a single agent. Indeed, the high proportion of children whose virus contained the AZT- or ddI-related mutations in this study is striking. This result indicates that therapy-induced ddI-associated mutations are not rare in children. In addition to having been treated for relatively long periods, however, all these children had symptoms and advanced HIV infection, conditions believed to increase the probability that resistance to A Z T will develop.2t,22 This finding underscores the urgent need for carefully designed prospective studies to determine (1) the frequency with which viral resistance, therapy-induced mutations, or both occur in different populations of children infected with HIV-1 who are receiving antiretroviral therapy and (2) the clinical significance of these findings. By adjusting conditions for selective amplification, we were able to obtain highly specific results for the detection of mutant and wild-type viral sequences in cell lysates for a wide range of input template concentrations. Although these conditions limit the sensitivity of this amplification, its use as the second step in a nested PCR strategy, in which the primary amplification step is highly sensitive, results in a genotypic assay that is both sensitive and specific when applied to clinical samples. In this study, amplification of proviral D N A from crude cell lysates of HIV cultures was successful in 100% of the 26 pretherapy and posttherapy
The Journal o f Pediatrics Volume 123, Number 1
isolates evaluated. This approach may be particularly valuable for rapid screening for drug resistance in treatment research or clinical settings. The early appearance and persistence of the codon 215 mutation, together with its association with phenotypic A Z T resistance, are supported by our identification of this mutation in all four isolates in which A Z T resistance was detected in vitro. Although other mutations may contribute to A Z T resistance in clinical isolates, their presence in AZT-resistant isolates is more variable, 1~ 19.20 so the detection of this single mutation may be sufficiently sensitive for use in settings in which large numbers of samples will be analyzed. Alternatively, selective amplification of additional AZT-associated mutations may provide more comprehensive detection of AZT-resistant variants in clinical HIV-1 isolates. The codon 74 mutation was the only mutation initially reported to be associated with dd[ administration and decreased sensitivity to ddI, although a second ddI-associated mutation at codon 184 was recently described. 23 The small changes in sensitivity associated with ddI therapy observed in this and other PBMC-based assays, 24 and the inherent variability of these assays, may limit the utility of phenotypic testing for resistance to ddI. This limitation results in particular problems in the assessment of the sensitivity of individual HIV-1 isolates to ddI. Thus the identification of the mutant genotypes associated with A Z T and ddI therapy in clinical HIV-1 isolates may be useful for studies in which the investigators attempt to determine whether these changes are clinically important. A t present, the treatment of H I V infection is limited by the lack of therapeutic alternatives to the nucleoside analogs A Z T , ddC, and ddI. Viral resistance to these agents may be a major limitation to their ability to prevent disease progression in HIV-infected children. Our findings indicate that combination therapy with A Z T and ddC does not prevent the emergence of A Z T resistance in vivo. In addition, our results, and those of others, suggest that there are differences in the degree of viral resistance that occurs with mutations associated with different nucleoside analogs. These results underscore the need for studies to determine the clinical importance of the resistance mutations that occur during treatment and for new strategies to overcome or prevent the emergence of viral resistance to nucleoside analogs in vivo. REFERENCES 1. Fischl MA, Richmond DD, Grieco MH, et al. The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex: a double-blind, placebo-controlled trial. N Engl J Med 1987;317:185-91. 2. Pizzo PA, Eddy J, Falloon J, et al. Effect of continuous intravenous infusion zidovudine (AZT) in children with symptomatic HIV infection. N Engl J Med 1988;319:889-96.
H u s s o n et al.
15
3. Fischl MA, Richmann DD, Causey DM, et al. Prolonged zidovudine therapy in patients with AIDS and advanced AIDSrelated complex. N Engl J Med 1989;262:2405-10. 4. Richman DD, Fischl MA, Grieco MH, et al. The toxicity of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex: a double-blind, placebo-controlled trial. N Engl J Med 1987;317:192-7. 5. McKinney RE, Maha MA, Connor EM, et al. A multicenter trial of oral zidovudine in children with advanced human immunodeficiency virus disease. N Engl J Med 1991;324:101825. 6. Larder BA, Darby G, Richman DD. HIV with reduced sensitivity to zidovudine (AZT) isolated during prolonged therapy. Science 1989;243:1731-4. 7. St. Clair MH, Martin JL, Tudor-Williams G, et al. A single mutation in HIV-1 reverse transcriptase confers dideoxyinosine resistance and collateral sensitivity to zidovudine. Science 1991;91:1557-9. 8. Tudor-Williams G, St. Clair M, McKinney R, et al. HIV-1 sensitivity to zidovudine and clinical outcome in children. Lancet 1992;339:15-9. 9. Larder BA, Kemp SD. Multiple mutations in HIV-1 reverse transcriptase confer high-level resistance to zidovudine (AZT). Science 1989;246:1155-8. 10. Kellam P, Boucher C, Larder B. Fifth mutation in human immunodeficiency virus type 1 reverse transcriptase contributes to the development of high-level resistance to zidovudine. Proc Natl Acad Sci USA 1992;89:1934-8. 11. Kohlstaedt L, Wang J, Friedman J, Rice P, Steitz T. Crystal structure at 3.5 A resolution of HIV-1 reverse transcriptase complexed with an inhibitor. Science 1992;256: 1783-90. 12. Larder B, Kellam P, Kemp S. Zidovudine resistance predicted by direct detection of mutations in DNA from HIV-infected lymphocytes. AIDS 1991;5:137-44. 13. Gingeras T, Prodanovich P, Latimer T, et al. Use of self-sustained sequence replication amplification reaction to analyze and detect mutations in zidovudine-resistant human immunodeficiency virus. J Infect Dis 1991;164:1066-74. 14. Richman D, Guatelli J, Grimes J, Tsiatis A, Gingeras T. Detection of mutations associated with zidovudine resistance in human immunodeficiency virus by use of the polymerase chain reaction. J Infect Dis t991;164:1075-81. 15. Pizzo PA, Butler KM, Balis F, et al. Dideoxycytidine alone and in an alternating schedule with zidovudine (AZT) in children with symptomatic HIV infection: a pilot study. J PEDIATR 1990;117:799-808. 16. Butler KM, Husson RN, Balis FM, et al. Dideoxyinosine in children with symptomatic human immunodeficiency virus infection. N Engl J Med 1991;324:137-44. 17. Shirasaka T, Yarchoan R, O'Brien M, et al. Changes in drug sensitivity of human immunodeficiency virus type 1 during therapy with azidothymidine, dideocytidine, of dideoxyinosine: an in vitro comparative study. Proc Natl Acad Sci USA 1993;90:562-6. 18. Leland D, French M. Virus isolaticm and identification. In: Lennette E, Halonen P, Murphy F, eds. Laboratory diagnosis of infectious diseases; vol 2. Viral, rickettsial and chlamydial diseases. New York: Springer-Verlag, 1988:39-59. 19. Boucher C, O'Sullivan E, Mulder J, et al. Ordered appearance of zidovudine resistance mutations during treatment of 18 human immunodeficiency virus-positive subjects. J Infect Dis 1992;165:105-10. 20. Mayers D, McCutchan F, Sanders-Buell E, et al. Character-
16
H u s s o n et al.
The Journal o f Pediatrics July 1993
ization of HIV isolates arising after prolonged zidovudine therapy. J Acquir Immune Defic Syndr 1992;5:749-59. 21. Richman D, Grimes J, Lagakos S. Effect of stage of disease and drug dose on zidovudine susceptibilities of isolates of human immunodeficiency virus. J Acquir Immune Defic Syndr 1990; 3:743-6. 22. Edlin B, St. Clair M, Pitha P, et al. In-vitro resistance to zidovudine and alpha-interferon in HIV-1 isolates from patients: correlations with treatment duration and response. Ann Intern Med 1992;117:457-60.
23. Gu Z, Gao Q, Li X, Parniak MA, Wainberg M. Novel mutation in the human immunodeficiency virus type 1 reverse transcriptase gene that encodes cross-resistance to 2' ,3 '-dideoxyinosine and 2' 3'-dideoxycytidine. J Virol 1992;66:7128-35. 24. McLeod G, McGrath J, Ladd E, Hammer S. Didanosine and zidovudine resistance patterns in clinical isolates of human immunodeficiency virus type 1 as determined by a replication endpoint concentration assay. Antimicrob Agents Chemother 1992;36:920-5.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the 1993 issues of THE JOURNAL OF PEDIATRICS are available to subscribers (only) from the Publisher, at a cost of $63.50 for domestic, $83.96 for Canadian, and $79.50 for international subscribers, for Vol. 122 ( J a n u a r y - J u n e ) and Vol. 123 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. T h e binding is durable buckram, with the Journal name, volume number, and year s t a m p e d in gold on the spine. P a y m e n t m u s t accompany all orders. C o n t a c t Mosby, Subscription Services, 11830 Westline Industrial Dr., St. Louis, M O 63146-3318, U S A / 8 0 0 - 4 5 3 - 4 3 5 1 , or 314-453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.
The development of nucleoside analog for the treatment of infection with human immunodeficiencyvirus has resulted in improved survival and decreased morbidity rates for individuals infected with this virus. 1-3The use of these agents, however, is limited by their toxicity and by their lack of long-term efficacy in preventing progression of HIV disease in many patients. 25 The development of viral resistance to Submitted for publication Dec. 30, 1992; accepted March 3, 1993. Reprint requests: Robert N. Husson, MD, Division of Infectious Diseases, Children's Hospital, 300 Longwood Ave., Boston, MA 02115. *Now at Children's Hospital, Harvard Medical School, Boston, Massachusetts. 9/25/46919
these agents, initially described for zidovudine6 and subsequently for didanosine,7 in clinical isolates of HIV- 1 may be a significant factor in the incomplete efficacy of these agents in vivo. Indeed, the emergence of HIV-1 isolates with decreased susceptibility to AZT has been linked to deteriSee related article, p. 1. oration in clinical status in children being treated with thisagent. ~ The occurrence of viral resistance to nucleoside analogs has led to changes in approaches to the development and clinical use of antiretroviral agents. It is now essential for new antiretroviral agents to be assessed early in their devel-
10
H u s s o n et al.
AZT CN5o ddC dd[ HIV- 1 PBMC PCR TCID5o
3' -Azido-2',3'-dideoxythymidine(zidovudine) Drug concentration required to give negative results (no p24 gag protein) in 50% of culture wells 2 ' ,3 ' -Dideoxycytidine (zalcitabine) 2 ',3 '-Oideoxyinosine (didanosine) Human immunodeficiency virus type 1 Peripheral blood mononuclear cell Polymerase chain reaction Median (50%) tissue culture infective dose
opment for their ability to induce viral resistance in vitro and in initial clinical studies. Treatment approaches increasingly focus on combination therapies in an attempt to enhance antiviral activity and prevent or slow the development of resistance. The identification of mutations in HIV-I reverse transcriptase associated with the development of the resistant phenotype, 7, 9, m together with new information on the three-dimensional structure of this enzyme, Ll provides insight into the molecular basis of this resistance and have led to the development of m o l e c u l a r techniques for the rapid identification of resistance mutations in clinical isolates of HIV-1. 7' 12-t4 Although A Z T resistance occurs frequently in adults receiving long-term therapy with this agent, its incidence in children, the factors that contribute to its occurrence, and its clinical implications are less well understood. Resistance to ddI is less well characterized, and its frequency in both adults and children receiving long-term therapy, as well as its clinical significance, is unknown. In this study, we evaluated the occurrence of resistance to A Z T in HIV- 1 isolates from children receiving long-term therapy with a sequential combination of A Z T and zalcitabine, and the occurrence of resistance to ddI in HIV-1 isolates from children receiving long-term therapy with this drug as a single agent. In addition, we optimized a selective polymerase chain reaction technique to identify in these isolates a mutation known to be highly associated with A Z T resistance, and the one previously described ddI-associated mutation. METHODS
Nucleosides. The A Z T was purchased from Sigma Chemical Co. (St. Louis, Mo.); ddC and ddI were provided by the Developmental Therapeutics Program, Division of Cancer Treatment, National Cancer Institute (Bethesda, Md.). Isolation and titration of pretherapy and posttherapy clinical isolates of HIV-I. Patients were enrolled in antiretroviral therapy protocols incorporating monthly blood sampling in the Pediatric Branch, National Cancer Institute.I5, l6 Informed consent was obtained from the patient's parent or guardian before enrollment. The AZT-ddC treatment regimen consisted of 8 weeks of ddC followed by 4
The Journal of Pediatrics July 1993
weeks without therapy, after which an alternating schedule of 3 weeks of AZT followed by 1 week of ddC continued throughout the study period. In the alternating regimen, the AZT dose was 180 m g / m 2 per dose administered orally every 6 hours, and the ddC dose was 0.02 to 0.04 mg/kg per dose administered orally every 6 hours. None of the patients enrolled in this study had been treated previously with antiretroviral agents. The ddI treatment regimen consisted of ddI administered orally as monotherapy throughout the study period, at doses ranging from 40 to 120 m g / m 2 per dose administered every 8 hours. Two patients (Nos. I- 1 and 1-4) whose HIV-1 isolates were evaluated in this study had been treated previously with AZT-containing regimens, including one (No. 1-4) who had been treated with the alternating AZT-ddC regimen. All pretherapy HIV-1 isolates were obtained from previously frozen nonfractionated peripheral blood mononuclear cells remaining after flow cytometric analysis of T-cell subsets. Posttherapy isolates were obtained from frozen or fresh PBMCs. Patients' cells were cultured in the presence of HIV-1 seronegative donor phytohemagglutinin-stimulated PBMCs as previously describedJ 7 When the p24 gag protein concentration in the culture supernatant was >2 ng/ml by radioimmunoassay (Du Pont Co., Wilmington, Del.), the supernatant was collected and stored in aliquots at - 7 0 ~ C. These primary virus isolates were titered in quadruplicate by culturing fivefold serial dilutions of virus with 106 phytohemagglutinin-stimulated PBMCs. The supernatant was tested on day 7 for p24 gag protein by radioimmunoassay, and the median tissue culture infective dose was determined by the Karber method3 s Each pair of pretherapy and posttherapy HIV-1 isolates was titered in the same experiment with the same batch of PBMCs to minimize the variation inherent in the use of PBMCs as target cells. Sensitivity testing of HIV-1 clinical isolates. Titered primary virus isolates of pretherapy and posttherapy HIV-I isolates were diluted in medium to a concentration of approximately I TCID50/~I, and 20 TCIDs0 was then added to 106 phytohemagglutinin-stimulated PBMCs in a 24-well plate in the presence of serial dilutions of AZT, ddC, or ddI. Isolates were cultured in the continuous presence of drug, and a 50% medium change with drug-containing medium was performed on day 3 or 4. On day 7, the culture supernatant was tested for p24 gag protein by radioimmunoassay (Du Pont). The sensitivity of each strain was defined as the drug concentration required for 50% of the wells to contain no p24 gag protein, as previously described) 7 The drug concentrations used in these assays were, for AZT: 25, 5.0, 1.0, 0.2, and 0.04 ~mol/L; for ddC: 5.0, 1.0, 0.2, 0.04, and 0.008/~mol/L; and for ddl: 25, 12.5, 6.3, 3.1, 1.6, and 0.8 ~mol/L. Each assay was performed in
The Journal o f Pediatrics Volume 123, Number 1
quadruplicate, and each pair of pretherapy and posttherapy HIV-1 isolates was tested in the same experiment by using the same batch of PBMCs. In addition, in each sensitivity assay, a limited titration was performed to verify the accuracy of the initial titration.
Selective amplification of wild-type and mutant proviral DNA. Each pretherapy and posttherapy HIV-1 primary isolate was cultured in the absence of drug with phytohemagglutinin-stimulated target PBMCs. Once the culture supernatant contained p24 gag protein, the cells were harvested, washed with phosphate-buffered saline solution, and lysed by incubation of cells in lysis buffer (KC1, 50 mmol/L; Tris-HCl [pH 8.3], 10 retool/L; MgC12, 2 retool/L; a nonionic detergent [Nonidet P-40; Sigma Chemical Co,, St. Louis, Mo.], 0.045% [vol/vol]; and proteinase K, 100 #g/ ml) at 55 ~ C for 60 minutes, followed by incubation for 12 minutes at 95 ~ C to inactivate proteinase K. Cell lysates were subjected to two-step PCR assay. The first amplification step used previously described primer pairs and reaction conditions to amplify a 1.7-kilobase fragment containing the HIV-1 reverse transcriptase. 9 The product of this amplification was then diluted 10-fold, and 1 to 10 #1 of this dilution was then subjected to selective PCR assay to identify wild-type or mutant sequences in the reverse transcriptase coding region. For codon 215, AZTrelated mutations in two bases result in a change from threonine to tyrosine or phenylalanine; for codon 74, the ddl-related mutation in a single base results in a change from leucine to valine. Previously published primer pairs 12 were used with modified conditions to enhance the specificity of these reactions. For detection of the codon 215 wildtype or mutant sequence, reaction components in a final volume of 100 #1 were as follows: KC1, 25 retool/L; Tris-HC1 (pH 8.3), 10 mmol/L; MgCIz, 1.5 mmol/L; deoxyadenosine triphosphate, deoxyguanosine triphosphate, deoxycytidine triphosphate, and deoxythymidine triphosphate, 0.25 mmol/L each; sonicated salmon sperm DNA, 10 ~g/ml; and 50 pmol of each primer. The amplification was performed in a thermal cycler (Perkin-Elmer/ Cetus, Norwalk, Conn.), at temperatures of 94 ~ C for 1 minute, 55 ~ C for 30 seconds, and 72 ~ C for 30 seconds, for 30 cycles. For detection of the codon 74 wild-type or mutant sequence, the conditions were identical except that the annealing temperature was 47 ~ C. The sensitivity and specificity of the selective amplification step under these conditions were determined by using plasmid ERS200E1, a plasmid containing the reverse transcriptase gene from a clinical isolate that had been sequenced and found to be a wild type at both the 74 and 215 codons, and plasmid ERS203L2-8, a plasmid containing the reverse transcriptase gene from a clinical isolate that had
H u s s o n et al.
11
been sequenced and found to contain both the codon 215 AZT-associated mutations and the codon 74 ddI-associated mutation. ~7 Copy numbers of the reverse transcriptase template were calculated from the absorbance at 260 nm of the purified plasmid DNA. Amplification products were electrophoresed through 1.2% agarose, stained with ethidium bromide, and visualized with ultraviolet light. Each pretherapy and posttherapy lysate was amplified with primers to detect the wild type or mutant base in parallel reactions in the same experiment, and negative controls were included in each experiment. Selective amplification results were categorized as wild type if there was a clearly visible band of the correct size resulting from amplification with the wild-type primers and no clearly visible band resulting from amplification with the mutant primers, m i x e d if bands of the appropriate size were clearly visible after amplification with wild-type and mutant primers, and m u t a n t if there was a clear band of the correct size resulting from amplification with the mutant primers and no clearly visible band resulting from amplification with the wild-type primers. RESULTS Pretherapy and posttherapy HIV- 1 strains were obtained from six patients treated with an alternating regimen of A Z T and ddC, and from seven patients treated with ddI as a single agent. All patients had symptomatic HIV infection (Centers for Disease Control and Prevention class P-2). The median CD4 cell count at the time of study entry of the patients treated with AZT plus ddC was 744 cells/ram 3 (range, 145 to 1944); of the patients treated with ddI, it was 265 cells/mm 3 (range, 0 to 1544). At the time the posttherapy isolate was obtained, the patients receiving A Z T with ddC had been treated on this regimen for a median of 97 weeks (range, 81 to 128), and the patients receiving ddI had been treated with this agent for a median of 66 weeks (range, 49 to 92). The results of sensitivity testing for AZT, ddC, and ddI for each pair of isolates are shown in Table I. Substantial increases in CN50 for AZT were seen in four of six posttherapy isolates from patients receiving AZT plus ddC, indicating the development of decreased sensitivity of these isolates to AZT. In contrast, with the possible exceptions of patients AC-1 and AC-3, appreciable changes in sensitivity to ddC or ddI in isolates from patients on this regimen were not observed. Among the posttherapy isolates from patients receiving ddI monotherapy, no substantial changes in CNs0 for AZT, ddC, or ddI were observed (Table II). Despite the lack of marked increase in CNso after ddI therapy, five of seven posttherapy isolates did show small increases in CNs0 of 1.4 to 2.9 times the pretherapy value. Although the magnitude
12
Husson et al.
The Journal of Pediatrics July 1993
Table I. Changes in sensitivity of HIV-I isolates from patients receiving AZT with ddC Patient No. AC-1 AC-2 AC-3 AC-4 AC-5 AC-6
Age at entry (yr) 5.5 1.3 0,7 10.7 12.2 3.8
AZT-ddC close* 180/0.02-0.04 180/0,03 180/0.04 180/0.01-0.03 180/0.02-0.03 180/0.02-0.03
Weeks on study
AZT CNso (pmol/L) Pre/post
ddC CN50 (pmol/L) Pre/post
ddl C N 5 0 (pmol/L) Pre/post
Codon 245 genotype Pre/post
92 101 80 82 103 128
0.13/2.24 0.04/0,13 0.04/25.0 0.09/0.67 0.07/5.0 0.09/0.09
0.67/2.24 0.45/0.45 0.06/0.45 0.09/0.09 0,30/0.30 0,47/0.30
5.3/8,9 3.8/7,5 5.3/15 8.9/8.9 3.1 / 3.1 2.1/2.2
WT/MIX WT/WT WT/MUT WT/MIX WT/MUT WT/WT
Pre, Pretherapy;post, posttherapy; WT, wild-typesequenee;MIX, mixedsequence;MUT, mutantsequence.
*Thedoseof AZT is presentedas milligramsper squaremeterper dose administeredevery6 hours;the doseof ddC is presentedas milligramsper kilogramper dose administeredevery 6 hours.
Table II. Changes in sensitivity of HIV-1 isolates from patients receiving ddI Patient No,
Age at entry (yr)
I-1 I-2 I-3 I-4 I-5 I-6 I-7
4,8 8.5 5.4 12.4 7.5 2.7 8.4
Dose of ddl* 40-120 90 90-120 90 40-120 90-120 90
Weeks on study
AZT CNso (~mol/L) pre/post
dclC CNso (grnol/L) pre/post
ddl C N 5 0 (prnol/L) pre/post
Codon 74 genotype pre/post
66 78 92 63 49 50 69
3.35/1.50 0.13/0.04 0.09/0.09 0.67/0.06 0.15/0.05 0.04/0.04 0.09/0,06
0.30/2.24 0.67/2.20 0.20/2.24 0.09/0.09 0.75/1.12 0.09/0.09 0.20/0.45
2.7,/6.3 7.5/10,6 4.5/10.6 8.9/8.9 2.2/6.3 4.5/4.5 5.3/8.9
WT/MUT WT/MIX WT/MUT WT/MIX WT/MIX WT/WT WT/MIX
Pre, Pretherapy;post, posttherapy; WT, wild-typesequence;MIX, mixedsequence;MUT, mutantsequence.
*The dose of ddl is presentedas milligramsper square meter per dose administeredevery 8 hours. of these changes is small, analysis of paired values suggests that these decreases in CNs0 are statistically significant (p = 0.04 by the Wilcoxon sign rank test). In addition, in five of seven post-ddl isolates, the CN50 for AZT decreased, consistent with previous observations suggesting increased AZT sensitivity in AZT-resistant isolates that acquire this ddI-associated mutation] Because the phenotypic analysis of drug sensitivity of HIV-1 is labor intensive, time-consuming, and expensive, and in particular because the magnitude of changes in sensitivity to ddI of posttherapy isolates was invariably small, we sought to apply a rapid test to detect the presence of AZT- or ddI-associated mutations in these isolates. The mutations resulting in a change in the codon for amino acid 215 from threonine to phenylalanine or tyrosine are among the earliest and most persistent AZT-associated mutations observed in clinical isolates, and they appear to be consistently associated with AZT resistance. 19, 2o The mutation that was initially associated with ddI resistance results in a change in amino acid 74 from leucine to valine.7 We therefore modified selective amplification procedures for the identification of mutations in these codons to provide molecular data complementary to the phenotypic sensitivity results described above. We optimized the selective step of this PCR assay to
maximize its specificity, so the conditions used in this study are more stringent than those originally described, particularly for amplification of the codon 215 sequence], 12 To assess the sensitivity and specificity of this assay under these conditions, we amplified serial dilutions of plasmid DNA containing the 1.7-kilobase reverse transcriptase insert determined by sequence analysis to have wild-type or mutant codons at positions 74 and 215 ~7 (Fig. 1). The limits of detection were 2 X 105 copies for codon 215 for both the wildtype and mutant sequence, and 2 x 104 copies for codon 74 for both the wild-type and mutant sequence. The specificity of these reactions was tested at high concentrations of template DNA (Fig. 2). At concentrations up to 2 • 101~copies per 100 ~1 reaction volume, these amplification conditions were highly specific for both sequences. At this concentration of template for the codon 215 sequence, there was no amplification of mutant DNA by wild-type primers and a minimally visible band from amplification of wild-type DNA by mutant primers. Similarly, at this concentration of template for the codon 74 sequence, there was no amplification of wild-type DNA by mutant primers and a minimally visible band from amplification of mutant DNA by wild-type primers. In mixing experiments of mutant and wild-type plasmid DNA, it was determined that amplification of mutant or wild-type DNA
The Journal of Pediatrics Volume 123, Number 1
Husson et al.
13
Fig. t. Sensitivity•fse•ectivePCRassayf•rc•d•ns74and2l5.Tenf•lddiluti•ns•fp•asmidDNAc•ntainingwild-type (WT) or mutant (Mut) sequence were amplified to determine the lower limit of detection of this selective PCR assay: a, 2 • 10 9 copies; b, 2 • 108; c, 2 • 107; d, 2 X 106; e, 2 • 105;f, 2 • 104.
Fig. 2. Specificity of selective PCR assay for codons 74 and 215. Tenfold dilutions of plasmid DNA containing wild-type (WT) or mutant (Mut) sequence were amplified with primers specific for the wild-type (W) or the mutant (M) sequence to determine the specificity of these primers at high concentrations of template DNA. a, 2 X 101~ copies; b, 2 x 109; c, 2X lOg.
Fig. 3. Tw••stepse•ectiveamp•ificati•n•fpr•vira•DNAfr•mcu•tures•fpatientis••ates•fHIV-•bef•re(pre)andafter (post) therapy with ddI or with AZT plus ddC. Each pretherapy and posttherapy isolate was subjected to primary nonselective PCR assay followed by selective PCR assay with primers specific for the wild-type sequence (W) or for the mutant sequence (M) for codon 74 for isolates from patients treated with ddI, or for codon 215 for isolates from patients treated with AZT plus ddC. depended primarily on the absolute concentration of the t e m p l a t e molecule r a t h e r t h a n on the ratio of m u t a n t to wild-type template ( d a t a not shown). Thus, under these conditions, the selective amplification was sensitive and specific for a wide r a n g e of input t e m p l a t e D N A at codons 74 and 215. This r a n g e encompassed the a m o u n t of input t e m p l a t e D N A in the selective P C R assay u n d e r the conditions t h a t we used; this assay was therefore both sensitive
and specific for the detection of wild-type and m u t a n t codons at positions 74 a n d 215 in the clinical isolates evaluated in this study. This selective P C R assay did not, however, provide quantitative information on the proportion of strains t h a t contain the m u t a t i o n of interest. The results of this two-step, selective P C R assay in the detection of m u t a n t and wild-type sequence in codon 215 in isolates obtained before a n d after therapy with A Z T plus
14
Husson et al.
ddC are shown in Fig. 3, a. All six pretherapy isolates had only the wild-type sequence present, whereas four (67%) of six posttherapy isolates had the mutant sequence present, including two in which the population was mixed and two in which only the mutant codon was detected. The presence of the codon 215 mutant sequence was associated with decreased sensitivity to AZT in all four isolates in which it was observed. As summarized in Table I, the two posttherapy isolates in which only the mutant eodon 215 was detected had the greatest increase in CNs0 (71 and 625 times the pretherapy CNso) and the highest absolute CNs0 (5.0 and 25.0 ~mol/L), and those in which this codon was a mixture of mutant and wild type had intermediate changes in CNs0 (7.4 and 17.2 times the pretherapy CNs0) and absolute posttherapy CNso (0.67 and 2.24 #mol/L). In the posttherapy isolates in which only the wild-type codon was detected, the CNso was in the same range as that of pretherapy isolates (0.09 and 0.13 ~mol/L). The results of selective PCR assay in the detection of the presence of mutant and wild-type bases in codon 74 in preand post-ddI isolates is shown in Fig. 3, b. All seven pre-ddI therapy isolates had only the wild-type sequence present in this position, whereas six (86%) of seven posttherapy isolates had the mutant sequence present, including three in which the population was mixed and three in which only the mutant codon was detected. As shown in Table II, examination of these results in relation to the phenotypic results revealed that the CNso increased in five post-ddI isolates in which the mutant base was detected and was unchanged in the sixth. There was no change in CNs0 in the post-ddI isolate in which only wild-type sequence was detected. Although large changes in ddC sensitivity were not observed in isolates from patients in either study arm, small decreases in ddC sensitivity (_
The Journal of Pediatrics July 1993
phenotypic analysis were supported by the presence of the codon 215 mutation in HIV-I reverse transcriptase that is thought to be most consistently associated with AZT resistance 19, 20 in all four isolates in which phenotypic resistance was observed. In contrast, large changes in sensitivity to ddI were not observed, despite the presence of the codon 74 mutation reported to be associated with decreased susceptibility to ddI in six of seven isolates. Similarly, large changes in ddC sensitivity were not observed, although small increases in the ddC CNso occurred in five posttherapy isolates from children receiving ddI and in two isolates from children receiving both AZT and ddC. A notable result of this study is the failure of alternating combination therapy to prevent the emergence of high-level resistance to AZT, similar to our findings in a study of HIV-1 isolates from adults treated with a different schedule of AZT alternating with ddC. ~7 It is possible that this AZT-ddC regimen may have slowed the development of AZT resistance, or that other combination regimens may slow the emergence of drug resistance more effectively; this question can be answered only in prospective comparative trials. The ability of AZT- and ddl-related mutations to coexist in vivo in the virus population of individual patients, however, appears to make it unlikely that combinations of these nucleoside analogs will prevent the emergence of viral resistance in vivo. An interesting and potentially important finding in this study is the high frequency of the codon 74 ddl mutation in children treated with ddI as a single agent. Indeed, the high proportion of children whose virus contained the AZT- or ddI-related mutations in this study is striking. This result indicates that therapy-induced ddI-associated mutations are not rare in children. In addition to having been treated for relatively long periods, however, all these children had symptoms and advanced HIV infection, conditions believed to increase the probability that resistance to A Z T will develop.2t,22 This finding underscores the urgent need for carefully designed prospective studies to determine (1) the frequency with which viral resistance, therapy-induced mutations, or both occur in different populations of children infected with HIV-1 who are receiving antiretroviral therapy and (2) the clinical significance of these findings. By adjusting conditions for selective amplification, we were able to obtain highly specific results for the detection of mutant and wild-type viral sequences in cell lysates for a wide range of input template concentrations. Although these conditions limit the sensitivity of this amplification, its use as the second step in a nested PCR strategy, in which the primary amplification step is highly sensitive, results in a genotypic assay that is both sensitive and specific when applied to clinical samples. In this study, amplification of proviral D N A from crude cell lysates of HIV cultures was successful in 100% of the 26 pretherapy and posttherapy
The Journal o f Pediatrics Volume 123, Number 1
isolates evaluated. This approach may be particularly valuable for rapid screening for drug resistance in treatment research or clinical settings. The early appearance and persistence of the codon 215 mutation, together with its association with phenotypic A Z T resistance, are supported by our identification of this mutation in all four isolates in which A Z T resistance was detected in vitro. Although other mutations may contribute to A Z T resistance in clinical isolates, their presence in AZT-resistant isolates is more variable, 1~ 19.20 so the detection of this single mutation may be sufficiently sensitive for use in settings in which large numbers of samples will be analyzed. Alternatively, selective amplification of additional AZT-associated mutations may provide more comprehensive detection of AZT-resistant variants in clinical HIV-1 isolates. The codon 74 mutation was the only mutation initially reported to be associated with dd[ administration and decreased sensitivity to ddI, although a second ddI-associated mutation at codon 184 was recently described. 23 The small changes in sensitivity associated with ddI therapy observed in this and other PBMC-based assays, 24 and the inherent variability of these assays, may limit the utility of phenotypic testing for resistance to ddI. This limitation results in particular problems in the assessment of the sensitivity of individual HIV-1 isolates to ddI. Thus the identification of the mutant genotypes associated with A Z T and ddI therapy in clinical HIV-1 isolates may be useful for studies in which the investigators attempt to determine whether these changes are clinically important. A t present, the treatment of H I V infection is limited by the lack of therapeutic alternatives to the nucleoside analogs A Z T , ddC, and ddI. Viral resistance to these agents may be a major limitation to their ability to prevent disease progression in HIV-infected children. Our findings indicate that combination therapy with A Z T and ddC does not prevent the emergence of A Z T resistance in vivo. In addition, our results, and those of others, suggest that there are differences in the degree of viral resistance that occurs with mutations associated with different nucleoside analogs. These results underscore the need for studies to determine the clinical importance of the resistance mutations that occur during treatment and for new strategies to overcome or prevent the emergence of viral resistance to nucleoside analogs in vivo. REFERENCES 1. Fischl MA, Richmond DD, Grieco MH, et al. The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex: a double-blind, placebo-controlled trial. N Engl J Med 1987;317:185-91. 2. Pizzo PA, Eddy J, Falloon J, et al. Effect of continuous intravenous infusion zidovudine (AZT) in children with symptomatic HIV infection. N Engl J Med 1988;319:889-96.
H u s s o n et al.
15
3. Fischl MA, Richmann DD, Causey DM, et al. Prolonged zidovudine therapy in patients with AIDS and advanced AIDSrelated complex. N Engl J Med 1989;262:2405-10. 4. Richman DD, Fischl MA, Grieco MH, et al. The toxicity of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex: a double-blind, placebo-controlled trial. N Engl J Med 1987;317:192-7. 5. McKinney RE, Maha MA, Connor EM, et al. A multicenter trial of oral zidovudine in children with advanced human immunodeficiency virus disease. N Engl J Med 1991;324:101825. 6. Larder BA, Darby G, Richman DD. HIV with reduced sensitivity to zidovudine (AZT) isolated during prolonged therapy. Science 1989;243:1731-4. 7. St. Clair MH, Martin JL, Tudor-Williams G, et al. A single mutation in HIV-1 reverse transcriptase confers dideoxyinosine resistance and collateral sensitivity to zidovudine. Science 1991;91:1557-9. 8. Tudor-Williams G, St. Clair M, McKinney R, et al. HIV-1 sensitivity to zidovudine and clinical outcome in children. Lancet 1992;339:15-9. 9. Larder BA, Kemp SD. Multiple mutations in HIV-1 reverse transcriptase confer high-level resistance to zidovudine (AZT). Science 1989;246:1155-8. 10. Kellam P, Boucher C, Larder B. Fifth mutation in human immunodeficiency virus type 1 reverse transcriptase contributes to the development of high-level resistance to zidovudine. Proc Natl Acad Sci USA 1992;89:1934-8. 11. Kohlstaedt L, Wang J, Friedman J, Rice P, Steitz T. Crystal structure at 3.5 A resolution of HIV-1 reverse transcriptase complexed with an inhibitor. Science 1992;256: 1783-90. 12. Larder B, Kellam P, Kemp S. Zidovudine resistance predicted by direct detection of mutations in DNA from HIV-infected lymphocytes. AIDS 1991;5:137-44. 13. Gingeras T, Prodanovich P, Latimer T, et al. Use of self-sustained sequence replication amplification reaction to analyze and detect mutations in zidovudine-resistant human immunodeficiency virus. J Infect Dis 1991;164:1066-74. 14. Richman D, Guatelli J, Grimes J, Tsiatis A, Gingeras T. Detection of mutations associated with zidovudine resistance in human immunodeficiency virus by use of the polymerase chain reaction. J Infect Dis t991;164:1075-81. 15. Pizzo PA, Butler KM, Balis F, et al. Dideoxycytidine alone and in an alternating schedule with zidovudine (AZT) in children with symptomatic HIV infection: a pilot study. J PEDIATR 1990;117:799-808. 16. Butler KM, Husson RN, Balis FM, et al. Dideoxyinosine in children with symptomatic human immunodeficiency virus infection. N Engl J Med 1991;324:137-44. 17. Shirasaka T, Yarchoan R, O'Brien M, et al. Changes in drug sensitivity of human immunodeficiency virus type 1 during therapy with azidothymidine, dideocytidine, of dideoxyinosine: an in vitro comparative study. Proc Natl Acad Sci USA 1993;90:562-6. 18. Leland D, French M. Virus isolaticm and identification. In: Lennette E, Halonen P, Murphy F, eds. Laboratory diagnosis of infectious diseases; vol 2. Viral, rickettsial and chlamydial diseases. New York: Springer-Verlag, 1988:39-59. 19. Boucher C, O'Sullivan E, Mulder J, et al. Ordered appearance of zidovudine resistance mutations during treatment of 18 human immunodeficiency virus-positive subjects. J Infect Dis 1992;165:105-10. 20. Mayers D, McCutchan F, Sanders-Buell E, et al. Character-
16
H u s s o n et al.
The Journal o f Pediatrics July 1993
ization of HIV isolates arising after prolonged zidovudine therapy. J Acquir Immune Defic Syndr 1992;5:749-59. 21. Richman D, Grimes J, Lagakos S. Effect of stage of disease and drug dose on zidovudine susceptibilities of isolates of human immunodeficiency virus. J Acquir Immune Defic Syndr 1990; 3:743-6. 22. Edlin B, St. Clair M, Pitha P, et al. In-vitro resistance to zidovudine and alpha-interferon in HIV-1 isolates from patients: correlations with treatment duration and response. Ann Intern Med 1992;117:457-60.
23. Gu Z, Gao Q, Li X, Parniak MA, Wainberg M. Novel mutation in the human immunodeficiency virus type 1 reverse transcriptase gene that encodes cross-resistance to 2' ,3 '-dideoxyinosine and 2' 3'-dideoxycytidine. J Virol 1992;66:7128-35. 24. McLeod G, McGrath J, Ladd E, Hammer S. Didanosine and zidovudine resistance patterns in clinical isolates of human immunodeficiency virus type 1 as determined by a replication endpoint concentration assay. Antimicrob Agents Chemother 1992;36:920-5.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the 1993 issues of THE JOURNAL OF PEDIATRICS are available to subscribers (only) from the Publisher, at a cost of $63.50 for domestic, $83.96 for Canadian, and $79.50 for international subscribers, for Vol. 122 ( J a n u a r y - J u n e ) and Vol. 123 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. T h e binding is durable buckram, with the Journal name, volume number, and year s t a m p e d in gold on the spine. P a y m e n t m u s t accompany all orders. C o n t a c t Mosby, Subscription Services, 11830 Westline Industrial Dr., St. Louis, M O 63146-3318, U S A / 8 0 0 - 4 5 3 - 4 3 5 1 , or 314-453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.