Identification of the K103N resistance mutation in Ugandan women receiving nevirapine to prevent HIV-1 vertical transmission

Objective: A recent trial in Uganda demonstrated that a simple, inexpensive regimen of nevirapine (NVP) prophylaxis can dramatically reduce HIV-1 vertical transmission risk. In this regimen, women receive a single dose of NVP at the onset of labor and infants receive a single dose of NVP within 72 h of birth. The objective of this study was to determine whether HIV-1 variants with NVP resistance mutations were selected in Ugandan women who received this regimen in the Phase I/II trial HIVNET 006.

Cited Here...: Reverse transcriptase (RT) sequences from plasma HIV-1 were analyzed from 15 women 6 weeks after NVP dosing. RT sequences from plasma collected prior to NVP dosing were also analyzed.

Cited Here...: The K103N NVP resistance mutation was detected 6 weeks after NVP administration in three (20%) out of 15 women (95% confidence interval, 0-40%). Pre-dose samples were available from two of the three women; both pre-dose samples lacked the mutation. Other NVP resistance mutations were absent from all 15 women. Women with the K103N mutation had a longer median NVP elimination half-life, decreased median oral clearance, and increased median area under the concentration time curve than those without the mutation. An evaluable sample was obtained from one of these three women 33 months after delivery; the K103N mutation was not detected in that sample.

Conclusions: This preliminary study demonstrates that HIV-1 with the RT K103N mutation can be detected in some Ugandan women following a single dose of NVP. This suggests that non-nucleoside RT inhibitor resistance may be selected in some people by single dose NVP prophylaxis. Pharmacokinetic data suggested that a more prolonged exposure to NVP after dosing may favor selection of NVP-resistant HIV-1.

Introduction

Nevirapine (NVP) is a highly potent non-nucleoside inhibitor of HIV-1 reverse transcriptase (RT). A recent clinical trial in Uganda, HIVNET 012, demonstrated that a simple and inexpensive regimen of NVP given to pregnant women and their infants can significantly reduce the rate of HIV-1 vertical transmission ^[1]. Women in HIVNET 012 received a single 200 mg dose of NVP at the onset of labor, and infants received a single 2 mg/kg dose of NVP within 72 h of birth. The NVP regimen in HIVNET 012 was more effective than a short course of zidovudine prophylaxis starting in labor, and was simpler and considerably less expensive. For these reasons, the HIVNET 012 NVP regimen is attractive for use in developing countries. The efficacy of single dose NVP in preventing HIV-1 vertical transmission is consistent with its rapid oral absorption and placental transfer, and its long half-life in pregnant women during labor and infants (median t_½, 61.3 h in women and 46.5 h in infants) ^[2,3].

When NVP is administered as a single drug for treatment of HIV-1 disease, HIV-1 variants with NVP resistance mutations in the RT gene are rapidly selected ^[4]. It is not known whether NVP-resistant HIV-1 variants can be detected following a single dose of NVP for prevention of HIV-1 vertical transmission. To address this question, we analyzed HIV-1 from women who were enrolled in the Ugandan Phase I/II trial, HIVNET 006 ^[3]. In HIVNET 006, 21 women received a single 200 mg dose of NVP at the onset of labor, the same regimen given to women in HIVNET 012. Plasma HIV-1 was analyzed to assess whether HIV-1 with NVP resistance mutations were detectable in samples collected 6 weeks after NVP administration.

Materials and methods

Plasma samples (0.2-0.5 ml) collected in HIVNET 006 were analyzed for evidence of NVP resistance mutations. Analysis of HIV-1 RT sequences (amino acids 88-200) was performed with the PE Biosystems HIV Genotyping System Prt/5′RT (PE Biosystems, Foster City, California, USA) as described previously ^[5]. RT sequences were confirmed by sequencing both DNA strands. Controls used to monitor for and prevent DNA contamination in the analysis have been described previously ^[6]. For some samples, the standard amplification procedure provided insufficient DNA for sequencing. Those samples were further amplified in a nested PCR using additional primers. HIV-1 subtypes in the RT region (336 nucleotides) were determined by phylogenetic methods using subtype A-J reference sequences recommended for HIV subtyping ^[7]. Subtype assignments were based on a neighbor-joining tree constructed with 500 bootstrap replications essentially as described ^[5]. The programs DNAdist, NEIGHBOR, and CONSENSE of PHYLIP v3.572.0 were used for analysis. The Genbank accession numbers for the nucleotide sequences from the 15 women are AF239197-AF239211.

Plasma HIV RNA levels (viral loads) were determined in the HIVNET 006 trial using the Roche AMPLICOR Monitor kit ^[3]. Twelve plasma samples were obtained from each woman for NVP concentration during the week following NVP dosing. NVP plasma concentrations were determined by high pressure liquid chromatography ^[2]. NVP pharmacokinetic parameters were determined using non-compartmental methods, as reported previously ^[2-3]. Mann-Whitney Rank Sum Tests were used to compare median NVP pharmacokinetic parameters, log transformed viral loads, and CD4 counts in women who developed resistance with those who did not.

Results

Plasma samples collected 6 weeks following NVP administration (6 weeks after delivery) were available from 18 of the 21 women enrolled in HIVNET 006. These women received a single 200 mg dose of NVP at the onset of labor. All of the women were antiretroviral drug naive, and none received any antiretroviral drugs post-partum. HIV-1 RT from 15 of the 18 samples was successfully amplified and sequenced. The 15 women analyzed included all four women whose infants were HIV-1 infected in HIVNET 006.

RT sequences were analyzed for the presence of mutations associated with NVP resistance ^[7]. Those mutations included A98G, L100I, K103N, V106A, V108I, Y181C, Y188C, and G190A. Analysis revealed the presence of the K103N mutation in samples from three of the 15 women, including one whose infant was HIV-1 infected. A mixture of lysine and asparagine was present at position 103 in two of the samples. None of the other NVP resistance mutations were detected in any of the samples.

Samples collected prior to NVP administration were available from two of the three women with the K103N mutation detected in the 6 week samples. Comparison of RT sequences from the pre-NVP and 6 week samples from two of these women confirmed that the paired samples were obtained from the same individuals. This analysis was not possible for the third woman with the K103N mutation, as the 6 week sample was the only sample available for analysis. The K103N mutation was not detected in either of the pre-NVP samples. The K103N mutation was also not detected in samples from 27 antiretroviral drug-naive Ugandan adults analyzed previously ^[5], suggesting a low prevalence of this mutation in Ugandan adults prior to NVP exposure. An evaluable sample was obtained from one of the three women 33 months after delivery. Unlike the 6 week post-delivery sample from this woman, which revealed only asparagine at codon 103, the sample obtained 33 months post-delivery revealed only the wild-type amino acid lysine at that position.

HIV-1 RNA levels (viral loads) were measured for each of the women in HIVNET 006 prior to NVP administration, and at 7 days and 6 weeks after NVP administration ^[3]. There was no significant difference between the viral loads or CD4 cell counts prior to NVP administration among the three women who developed the K103N mutation compared to the 12 women who did not. In addition, changes in HIV RNA copy number observed at 7 days and 6 weeks following NVP administration were similar for the three women with the K103N mutation as compared to the 12 women without the mutation (Fig. 1).

Fig. 1

NVP pharmacokinetic parameters were determined in the women participating in the HIVNET 006 trial, as previously reported ^[3]. When the three women who developed the K103N mutation were compared to the 12 women who did not, median peak NVP concentrations were not significantly different (Table 1). However, the women with the resistance mutation had a longer median elimination half-life, decreased median oral clearance, and increased median area under the concentration time curve.

Table 1

The HIV-1 subtype of the RT region in each woman was determined by phylogenetic analysis. Ten of the 15 women had subtype D HIV-1 and five had subtype A HIV-1. Among the women with the K103N mutation, two had subtype D HIV-1 and one had subtype A HIV-1.

Discussion

This study demonstrates the potential for NVP-resistant HIV-1 to be selected in women receiving NVP prophylaxis to prevent HIV-1 vertical transmission. We detected the K103N mutation in three of 15 women 6 weeks following a single 200-mg NVP dose administered during labor. It should be noted that the methods used for genotypic analysis in this study involve bulk sequencing of PCR products; these methods typically do not detect HIV-1 variants that represent a minor proportion of the viral population. Detection of NVP-resistant variants in this setting is not unexpected. First, HIV-1 variants containing a single mutation sufficient for drug resistance are believed to exist in the viral population of every patient even prior to drug exposure ^[8]. For example, the Y181C NVP resistance mutation is estimated to be present in approximately 1 per 1000 copies/ml of plasma HIV-1 in drug-naive individuals ^[4]. Second, NVP is extremely potent and has a long half-life. Replication of wild-type, NVP-sensitive virus is rapidly inhibited, permitting selection of NVP-resistant variants. Rapid selection of drug resistant HIV-1 may also occur during strategic or unplanned interruption of treatment with drugs like NVP or efavirenz that have a long half-life and low genetic barrier to resistance. In this study, women with the K103N mutation had longer median NVP elimination half-life, decreased median oral clearance, and increased median area under the NVP concentration time curve than women without the mutation. While these data are limited due to the small number of women studied, they suggest that a more prolonged maternal exposure to NVP after dosing may favor selection of NVP-resistant variants.

HIV-1 variants with resistance mutations were detected more frequently in HIVNET 006 than in studies using zidovudine for perinatal prophylaxis among antiretroviral-naive women. For example, in ACTG 076, genotypic evidence of zidovudine resistance developed in only one of 39 women at delivery after antepartum zidovudine treatment ^[9]. It is important to recognize, however, that ACTG 076 enrolled women who were relatively healthy (CD4 count > 200 × 10⁶/l; median HIV-1 RNA, 5700 copies/ml), whereas the 15 women from HIVNET 006 analyzed in this report had relatively advanced HIV-1 infection (median HIV-1 RNA, 55 108 copies/ml), which may have favored selection of resistant variants. Other factors, such as maternal post-partum treatment, the timing of sample collection for resistance testing, and the number and type of mutations associated with resistance to different antiretroviral drugs, differ from study to study, making it difficult to compare the rates of resistance observed with different antiretroviral regimens.

It is interesting to note that the K103N mutation was detected in our cohort, whereas the Y181C mutation was not. In a study from the USA, among patients receiving NVP monotherapy 79% developed the Y181C NVP resistance mutation whereas only 33% developed the K103N mutation ^[10]. Other NVP resistance mutations were also identified in that study. Our finding of the K103N mutation in antiretroviral drug-naive Ugandan women with non-subtype B HIV-1 who received a single dose of NVP suggests that different mutations may be selected during antiretroviral therapy in different HIV-1 subtypes.

This report demonstrates that the K103N NVP resistance mutation can be detected in women receiving NVP to prevent HIV-1 perinatal transmission. The clinical significance of this finding is not known. There is currently no evidence that resistance to NVP will affect clinical progression of HIV-1 infection. There is also no evidence that post-partum selection of NVP-resistant variants will affect the risk of HIV-1 sexual transmission or that of mother-to-child transmission in the setting of NVP prophylaxis. Transmission in the intrapartum period would not be affected with the first use of NVP, since there would be insufficient time for selection of resistant variants. While post-partum transmission (e.g. via breastfeeding) could be affected, recent studies demonstrate that most resistance mutations in HIV-1 disappear rapidly after discontinuation of antiretroviral drugs ^[11,12]. Supporting this, we found that the K103N mutation was no longer detectable in one woman with a sample available 33 months after delivery. If NVP-resistant HIV-1 were rapidly replaced by wild-type, NVP-sensitive HIV-1 in the post-partum period, there would be less potential for NVP-resistant HIV-1 to be transmitted to infants or to other adults. It is also likely that NVP prophylaxis would remain effective for interruption of intra-partum transmission in subsequent pregnancies, if NVP-resistant HIV-1 variants were rapidly replaced by wild-type, NVP-sensitive virus.

Regarding the potential impact on women's health, it is important to recognize that antiretroviral treatment options are currently extremely limited in many countries where the NVP prophylactic regimen may be implemented. If treatment options were expanded in the future in those countries, patients with NVP-resistant HIV-1 could still be offered effective treatment regimens that do not include NVP or other non-nucleoside RT inhibitors. Women receiving NVP-prophylaxis during labor could also begin treatment for their HIV-1 infection during the immediate postnatal period, thereby decreasing the potential for selection of NVP-resistant HIV-1. In the USA the NVP regimen is now recommended as one of the options for prevention of perinatal transmission in women who have not received antiretroviral therapy during pregnancy ^[13]. Initiation of highly active antiretroviral therapy in such women during the immediate postnatal period to suppress viral replication would probably prevent the development of NVP resistance.

The preliminary nature of this report should be noted. Future research is needed to determine if these findings are confirmed in larger studies, to assess whether replacement with wild-type, NVP-sensitive virus occurs rapidly in the absence of treatment, and to evaluate whether NVP-resistant HIV-1 also emerges in infants receiving NVP prophylaxis. The potential risks and concerns raised in this report must be balanced against simplicity, cost-effectiveness, and documented efficacy of the NVP prophylactic regimen. This regimen has the potential to reduce HIV-1 perinatal transmission by 47% or more in settings where other prophylactic regimens are impractical and treatment options are extremely limited. If quickly implemented, this regimen could prevent HIV-1 transmission in 800 infants a day and prevent millions of infants from becoming HIV-1 infected over the next decade.

Acknowledgements

The authors acknowledge the assistance of M. Allen (Protocol Specialist, Family Health International). The authors thank E. Piwowar-Manning, C. Ducar, and the laboratory staff in Uganda for assistance with sample processing. The authors also thank E. Shulse and the PE Biosystems Genotyping Team for helpful discussions and for providing reagents used in this study.