Breast-Milk Shedding of Drug-Resistant HIV-1 Subtype C in Women Exposed to Single-Dose Nevirapine

Transcription

1 BRIEF REPORT Breast-Milk Shedding of Drug-Resistant HIV-1 Subtype C in Women Exposed to Single-Dose Nevirapine Esther J. Lee, 1 Rami Kantor, 1 Lynn Zijenah, 6 Wayne Sheldon, 3 Lynda Emel, 3 Patrick Mateta, 6 Elizabeth Johnston, 1 Jennifer Wells, 6 Avinash K. Shetty, 6 Hoosen Coovadia, 7 Yvonne Maldonado, 2 Samuel Adeniyi Jones, 4 Lynne M. Mofenson, 5 Christopher H. Contag, 2 Mary Bassett, 6 and David A. Katzenstein, 1 for the HIVNET 023 Study Team a 1 Center for AIDS Research, Division of Infectious Disease, and 2 Department of Pediatrics, Stanford University School of Medicine, Stanford, California; 3 The Statistical Center for HIV/AIDS Research and Prevention, Fred Hutchinson Cancer Research Center, Seattle, Washington; 4 Division of AIDS, National Institute of Allergy and Infectious Diseases, and 5 National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland; 6 Zimbabwe AIDS Prevention Project and University of Zimbabwe School of Medicine, Harare, Zimbabwe; 7 Nelson Mandela School of Medicine, University of Natal, Durban, South Africa Single-dose nevirapine reduces intrapartum human immunodeficiency virus 1 type (HIV-1) transmission but may also select for nonnucleoside reverse-transcriptase inhibitor (NNRTI) resistance in breast milk (BM) and plasma. Among 32 Zimbabwean women, median 8-week postpartum plasma and BM HIV-1 RNA levels were 4.57 and 2.13 log 10 copies/ ml, respectively. BM samples from women with laboratorydiagnosed mastitis (defined as elevated BM Na + levels) were 5.4-fold more likely to have HIV-1 RNA levels above the median. BM RT sequences were not obtained for 12 women with BM HIV-1 RNA levels below the lower limit of detection of the assay used. In 20 paired BM and plasma samples, 65% of BM and 50% of plasma RT sequences had NNRTI-resistance mutations, with divergent mutation patterns. In 2003, 1600,000 infants were infected with HIV-1 through mother-to-child transmission (MTCT) in sub-saharan Africa [1]. In resource-constrained countries, most HIV-1 positive Received 13 October 2004; accepted 6 May 2005; electronically published 23 August Presented in part: 10th Conference on Retroviruses and Opportunistic Infections, Boston, February 2003 [abstract 799]; 11th Conference on Retroviruses and Opportunistic Infections, San Francisco, 8 11 February 2004 [abstract 38]. Potential conflicts of interest: none reported. a Study group members are listed after the text. Reprints or correspondence: Dr. David Katzenstein, Div. of Infectious Disease, Stanford University School of Medicine, 300 Pasteur Dr. S-169, Stanford, CA (davidkk@stanford.edu). The Journal of Infectious Diseases 2005; 192: by the Infectious Diseases Society of America. All rights reserved /2005/ XX$15.00 women breast-feed their infants, and up to one-third of MTCT is due to breast-feeding [2]. Mastitis, or breast-tissue inflammation, which has been linked to elevated Na + levels in breast milk (BM), has been associated with an increase in HIV-1 RNA levels in BM and a higher risk of MTCT [3]. Single-dose nevirapine (SD-NVP) given to women during labor has been shown to be an effective and low-cost intervention for the reduction of MTCT in resource-constrained countries [4]. However, nonsuppressive treatment with NVP, as with other nonnucleoside reverse-transcriptase inhibitors (NNRTIs), may result in the rapid acquisition of high-level drug resistance in plasma HIV-1 RNA [5 7]. Although several studies have characterized the effects of SD- NVP on HIV-1 in plasma, the effects of SD-NVP on BM HIV- 1 RNA shedding are not known. In women in Zimbabwe who were exposed to SD-NVP as part of the clinical study HIVNET 023 [8], we quantified HIV-1 RNA levels in plasma and BM, correlated BM Na + levels and viral shedding, and identified NNRTI-related resistance mutations in both compartments at 8 weeks postpartum. Participants, materials, and methods. HIVNET 023, a randomized, open-label, pharmacokinetic phase 1/2 trial, was conducted to study the safety and trough concentrations of NVP in breast-feeding infants between April 2000 and October 2001 in Chitungwiza, Zimbabwe. NVP tablets were dispensed to 32 women after 36 weeks of gestation, to be taken orally at the onset of labor. Maternal clinical data on plasma and other enrollment characteristics were obtained before NVP was taken by the participants, and plasma and left and right BM samples were collected at 8 weeks postpartum. HIV-1 RNA in plasma was quantified in the parent study; measurement of BM HIV- 1 RNA and Na + levels and sequencing to identify NNRTI resistance were secondary objectives. Informed consent was received from all participants, and the protocol was approved by the US Department of Health and Human Services, the Biomedical Research Institute of Zimbabwe, the Medical Research Council and Medicines Control Authority of Zimbabwe, and the Stanford University Panel on Human Subjects [8]. Financial support: Doris Duke Distinguished Clinical Research Award (to D.A.K.); Doris Duke Charitable Foundation (to E.J.L., R.K., and E.J.). The clinical study (HIVNET 023) was supported by the HIV Clinical Trials Network, which is sponsored by the National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health (NIH). In addition, the clinical study was supported by the HIV Prevention Trials Network, which is sponsored by the NIAID, the National Institute of Child Health and Human Development, the National Institute of Mental Health, the National Institute on Drug Abuse, and the Office of AIDS Research, NIH, Department of Health and Human Services. Funding and research reagents were provided by Boehringer Ingelheim, Bayer, and Roche Molecular Systems.

2 Plasma and whole BM samples were frozen at 70 C within 4 h of collection. Plasma HIV-1 RNA was quantified by use of the Roche Amplicor HIV-1 Monitor Test (version 1.5; Roche Molecular Diagnostics), which has a lower limit of detection of 400 copies/ml. Whole BM samples (1 4 ml) were centrifuged at 300 g for 10 min, to isolate lactoserum, and BM HIV- 1 RNA was quantified by use of the Roche Amplicor HIV-1 Monitor Test, modified for ultrasensitive detection (lower limit of detection, 50 copies/ml). The ultracentrifuged lactoserum was assayed for BM Na + level (Synchron LX20; Beckman Coulter) at Stanford Clinical Laboratories. BM Na + levels 112 mmol/l (3 SDs above the mean level in BM from HIV-1 negative women [5 6 mmol/ L]) were defined as elevated, and the women were considered to have laboratory-diagnosed mastitis [3]. Plasma RNA was extracted by use of the TruPrep Viral RNA Kit (Visible Genetics/Bayer). BM HIV-1 RNA was extracted by use of the Roche Amplicor HIV-1 Monitor Test. RT positions were sequenced by use of the TRUGENE HIV-1 Genotyping Kit (version 1.5; Visible Genetics/Bayer) with RT polymerase chain reaction (PCR) research-use-only primers, optimized for non subtype B virus. The GenBank accession numbers for the plasma and BM RT sequences are AY AY998373, AY AY901253, and AY AY Mutations were defined as differences from the consensus B amino acid reference sequence that are associated with NNRTI resistance (L100I, K101EP, K103NS, V106AM, V108I, Y181CI, Y188CLH, and G190ASE) [9]. Mixtures detected at 120% were considered to be mutations, in accordance with TRUGENE program algorithms. For noncorrelation analysis, HIV-1 RNA levels below the lower limit of detection were assigned a value of 399 copies/ ml for plasma and 49 copies/ml for BM. All HIV-1 RNA levels were log 10 transformed. The 2-sided Wilcoxon rank sum test was used to analyze the relationship between the plasma NNRTI-resistance mutation status of the participants and their baseline characteristics. Spearman s rank correlation was used to assess the level and statistical significance of associations between selected study variables. The odds ratios (ORs) for noncorrelated data were calculated by standard x 2 methodology; for correlated data, the generalized estimating equation was used. Statistical analysis was conducted by use of SAS software (version 8.2; SAS Institute) at the Statistical Center for HIV/AIDS Research and Prevention, Fred Hutchinson Cancer Research Center. Results. Thirty-two HIV-1 infected pregnant women were enrolled in HIVNET 023 in Zimbabwe. Although all women received 200 mg of SD-NVP at labor, the first 4 women who participated in the early stage of the study also received 300 mg of zidovudine twice daily, starting at 36 weeks of gestation. Two women received 2 doses of NVP, because of false labor. Plasma 8-week postpartum HIV-1 RNA levels were 400 copies/ml in 29 (91%) of the 32 women, with a median of 4.57 log 10 copies/ml and an interquartile range (IQR) of log 10 copies/ml. At least 1 BM sample was available from all 32 women (30 from the left breast and 32 from the right breast), and quantifiable HIV-1 RNA levels ( 50 copies/ml) were detected in at least 1 BM sample from 23 (72%) of the 32 women. The median HIV-1 RNA level for all 62 BM samples was 2.20 log 10 copies/ml; the median HIV-1 RNA level in left BM samples was 2.20 log 10 copies/ml (IQR, log 10 copies/ml); and the median HIV-1 RNA level in right BM samples was 2.18 log 10 copies/ml (IQR, log 10 copies/ ml). The median HIV-1 RNA level for the averages of each woman s left and right BM samples was 2.13 log 10 copies/ml (IQR, log 10 copies/ml). A correlation between the plasma and the average BM HIV-1 RNA levels was observed ( rs p 0.44; P p.036). Lactoserum electrolyte concentration was tested in 56 BM samples from 30 women. Elevated Na + levels (112 mmol/l) were identified in 24 (43%) of 56 BM samples from 17 (57%) of the 30 women. BM samples with elevated Na + levels were 5.4-fold (95% confidence interval [CI], fold; generalized estimating equation) more likely to have HIV-1 RNA levels above the study median than were BM samples with lower Na + levels. All plasma and BM RT sequences were most closely related to those of the HIV-1 subtype C reference strain, and RT sequences (BM and plasma) from the same woman were closer to one another than they were to sequences from other women (figure 1). Plasma viruses from all 32 women at 8 weeks postpartum were amplifiable (table 1). Eleven (34%) of the 32 women had at least 1 NNRTI-resistance mutation that was not present before exposure to NVP (data not shown). The most common NNRTI-resistance mutation was K103N, which was present in 9 (82%) of the 11 RT sequences with NNRTI-resistance mutations. The women with NNRTI-resistance mutations had baseline CD4 cell counts (median, cells/ml; IQR, cells/ml) that were significantly lower than those of the women without NNRTI-resistance mutations (median, 408 cells/ml; IQR, cells/ml) ( P p.034, Wilcoxon rank sum test). There were no significant associations between the selection of resistance mutations in plasma and other enrollment characteristics. BM viruses from 20 (63%) of the 32 women were amplifiable. Of these 20, 13 (65%) had at least 1 NNRTI-resistance mutation in 1 or both of the left or right BM RT sequences, most frequently K103N (11/13 [85%]). Of the 14 women from whom both left and right BM RT sequences were available, 10 (71%) had an NNRTI-resistance mutation in 1 or both of them. Of the 20 women from whom both a plasma sequence and 000 JID 2005:192 (1 October) BRIEF REPORT

3 Figure 1. Neighbor-joining phylogenetic tree of paired 8-week postpartum plasma and breast milk (BM) reverse-transcriptase (RT) sequences from 32 HIV-1 subtype C infected Zimbabwean women who participated in HIVNET 023. The tree was obtained by use of Tajima and Nei distance estimation in TREECON (version 1.3b) [10]. The name of each sequence is constructed with the participant number (1 32), the source of the sequence (plasma or BM), and, for BM samples, the specific breast (left [L] or right [R]). A distance scale of 2% and 1000-replicate bootstrap values of 170% are shown. The subtype B reference is isolate HXB2 (GenBank accession no. K03455), and the subtype C reference is isolate C2220 (GenBank accession no. U46016). The medians and interquartile ranges (IQRs) of nucleic acid genetic distances within each woman between plasma and left BM RT sequences, plasma and right BM RT sequences, and right and left BM RT sequences were 1.1% (1.0% 1.9%), 1.3% (0.8% 2.2%), and 0.7% (0.5% 1.1%), respectively ( P 1.05, for all; Wilcoxon signed rank test). The medians and IQRs of nucleic acid genetic distances between women comparing plasma RT sequences, all BM RT sequences, and 1 BM RT sequence per woman were 5.7% (4.9% 6.7%), 5.7% (4.8% 6.6%), and 5.8% (5.1% 6.7%), respectively ( P 1.05, for all; Wilcoxon signed rank test). Genetic distances were determined by use of Syn-SCAN (with the weighted scoring matrix) [11]. at least 1 BM sequence were available, 10 (50%) had an NNRTIresistance mutation in the plasma sequence, and 13 (65%) had an NNRTI-resistance mutation in at least 1 BM sequence. Comparison of drug-resistance patterns between plasma and BM RT sequences within individual women showed that 10 (50%) had the same pattern with respect to NNRTI resistance, including 6 (30%) women who had no resistance mutations in either compartment (women 1 6 in table 1) The OR of a detectable NNRTIresistance mutation in either the left or the right BM sequence was estimated to be 13.5-fold (95% CI, fold) higher in women with plasma NNRTI-resistance mutations than in women without plasma NNRTI-resistance mutations. Five women had mutations in plasma or BM RT sequences at positions associated with resistance to nucleoside RT inhibitors. Two women had unique polymorphisms at NNRTI-resistance positions; these polymorphisms are not known to contribute to resistance (table 1). Four infants received a diagnosis of HIV-1 infection (defined as 2 positive PCR assays for HIV-1 RNA). Two of these infants had their first positive test at birth (in utero infection), and 1 of these infants was HIV-1 RNA negative at 2 weeks of age but was positive at 8 weeks of age (intrapartum/early BM infection). The fourth infant was HIV-1 RNA negative at 24 weeks of age but was positive at 32 weeks of age. This infant was likely infected through breast-feeding; however, the plasma and BM samples obtained from the mother at 8 and 24 weeks postpartum contained wild-type virus. Discussion. Measurement of BM viral shedding by a sensitive PCR assay showed that a majority (23/32 [72%]) of the HIV-1 subtype C infected women in the present study who were breast-feeding their infants had detectable virus in 1 or both BM samples obtained 8 weeks postpartum. Higher BM HIV-1 RNA levels were associated with laboratory-diagnosed mastitis, defined as a BM Na + level 112 mmol/l; this was identified in at least 1 BM sample from 17 (57%) of 30 women. Although this frequency is higher than those observed in other studies in Africa [3], complaints of clinical mastitis were rare in our cohort. However, BM samples with elevated Na + levels were 5.4-fold (95% CI, fold; generalized estimating equation) more likely to have HIV-1 RNA levels above the study median than were BM samples with lower Na + levels. Further investigations are warranted to determine causes (such as mastitis and breast inflammation) of elevated BM Na + levels that potentially drive BM viral shedding. The reported prevalences of NNRTI resistance in plasma RT sequences have been 20% in studies in Uganda at 6 8 weeks postpartum (subtypes A and D) [6, 7] and 32% in a study in Thailand at 10 days postpartum (CRF01_AE) [5]. The present results 34% prevalence at 8 weeks postpartum support the findings of recent studies that have indicated that the incidence of postpartum NNRTI resistance in plasma may be high-

4 Table 1. Drug-resistance mutations in plasma ( n p 32) and breast milk (BM; n p 20) reverse transcriptase (RT) sequences from HIV-1 subtype C infected Zimbabwean women participating in HIVNET 023. Participant Plasma at 8 weeks postpartum BM at 8 weeks postpartum Left Right 1 wt wt wt 2 a wt wt wt 3 wt wt wt 4 wt wt wt (V75VI) 5 wt wt * 6 wt (K103R) wt (K103R) * 7 K103KN K103KN (Y188DY) K103KN 8 K103KN K103KN K103KN 9 K103KN K103KN K103KN 10 K103KN, G190GA K103KN, G190GA K103KN, G190GA 11 a wt wt K103KN 12 wt (E44K), (A62P) * K103KN 13 wt K103KN * 14 wt K103KN * 15 b K103KN * wt (T215TS) 16 K103KN K103N, Y188YC K103N, Y188YC 17 K103KN, Y188YC wt K103KN 18 a K103KN, Y181YC wt K103KN 19 K103KN, V106IM V106VA wt 20 Y181YC K103KN wt 21 a (A98AT), (K101KN), * * V106VA, K190AG, (T69TS) 22 wt (V118I) * * 23 wt * * 24 wt * * 25 wt * * 26 wt * * 27 wt M * 28 wt * * 29 b wt * * 30 wt (A98S) * * 31 wt * * 32 wt M * NOTE. Mutations in RT sequences at drug resistance associated sites that are not known to be associated with resistance to nonnucleoside RT inhibitors (NNRTIs) are in parentheses; drug-resistance mutations associated with resistance to nucleoside RT inhibitors are in parentheses and italicized. Women with 1 or both BM RT sequences available are denoted 1 20, and the remaining women, from whom BM virus could not be amplified, are denoted Asterisks indicate that a sequence was not obtained because the virus was not amplifiable. M, missing samples; wt, wild type. a Exposed to zidovudine during gestation before receiving single-dose nevirapine. b Exposed to a second dose of nevirapine, because of false labor. er in HIV-1 subtype C than in other subtypes [12, 13]. In the present study, of the 29 women for whom baseline CD4 cell counts were available, the 10 with plasma NNRTI-resistance mutations had CD4 cell counts that were significantly lower than those of the 19 women without plasma NNRTI-resistance mutations. RT sequences could not be generated from the BM samples from 12 of the 32 women, primarily those with lower plasma HIV-1 RNA levels and with BM HIV-1 RNA levels that were below the lower limit of detection. Of the 20 women with paired plasma and BM samples, NNRTI-resistant virus was detected in 50% of the plasma and 65% of the BM samples. One-half of these paired samples demonstrated different NNRTI-resistance mutations, including onequarter of the paired samples in which an NNRTI-resistance mutation was found in only 1 compartment. Of the 14 paired left and right BM samples, 5 (36%) had different NNRTI-resistance mutation patterns (table 1), with resistance in one sample but not in the other. Given that baseline plasma RT sequences did not contain NNRTI-resistance mutations, the selection of resistant virus in plasma and BM after exposure to SD-NVP was 000 JID 2005:192 (1 October) BRIEF REPORT

5 likely the result of differences in viral replication and selective pressure within plasma and BM, where decreasing NVP levels are maintained for weeks [14]. One study comparing env genotypes between BM and plasma identified distinct variants in each compartment [15]. Although the small sample in our study limits broad conclusions, it does provide evidence of inter- and intracompartmental differences in the selection and persistence of NNRTI-resistance mutations. The observation of drug-resistant virus in a high proportion of SD-NVP exposed women who breast-fed their infants is of concern. Despite the high prevalence of NNRTI-resistance mutations observed in BM RT sequences from the women in the present study, the mother of the only infant for whom infection was definitively transmitted through breast milk had wild-type virus found at 8 and 24 weeks postpartum. Infant prophylaxis by use of daily NVP is under investigation as a strategy to preserve safe breast-feeding among HIV-1 infected women [6]; evaluation of drug-resistant virus in maternal plasma and BM, and the association between resistance and BM transmission, is planned as part of the study. Additional interventions to reduce mastitis and other factors that drive viral shedding and strategies to mitigate the selection of drug-resistant virus are warranted, to increase the safety of breast-feeding and prevent BM transmission of HIV-1. The HIVNET 023 Study Team. In addition to the authors of the present article, members of the study team include Charles Chiedza Maponga, Norbert Nyoni, Lynda Stranix, Tsungai Chipato, Nhamo Gonah, and David Hill (Zimbabwe AIDS Prevention Project, University of Zimbabwe and Stanford University); Kathy George (protocol specialist, Family Health International); Tom Fleming (protocol statistician, Statistical Center for HIV/AIDS Research and Prevention, Fred Hutchinson Cancer Research Center); Scharla Estep (protocol pharmacist); Maria Gigliotti and Pat Robinson (pharmaceutical sponsors, Boehringer Ingelheim Pharmaceuticals); and Stephen Day and Dean Winslow (pharmaceutical sponsors, Visible Genetics/Bayer). Acknowledgments We thank the study staff in Zimbabwe, including Agnes Munhenga, Sylvia Jena, Sabina Chiwara, Jane Chirwa, Rosemary Maredza, Zororai Muchabaiwa, Cathrine Kagona, Lucy Banda, Patricia Sane, Andrea von Lieven, and Jennifer Welles, as well as the Batanai support group. References 1. UNAIDS. UNAIDS questions and answers. Geneva: UNAIDS World Health Organization, Datta P, Embree JE, Kreiss JK, et al. Mother-to-child transmission of human immunodeficiency virus type 1: report from the Nairobi Study. J Infect Dis 1994; 170: Semba RD, Kumwenda N, Hoover DR, et al. Human immunodeficiency virus load in breast milk, mastitis, and mother-to-child transmission of human immunodeficiency virus type 1. J Infect Dis 1999; 180: Jackson JB, Musoke P, Fleming T, et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: 18-month follow-up of the HIVNET 012 randomised trial. Lancet 2003; 362: Jourdain G, Ngo-Giang-Huong N, Le Coeur S, et al. Intrapartum exposure to nevirapine and subsequent maternal responses to nevirapine-based antiretroviral therapy. N Engl J Med 2004; 351: Jackson JB, Becker-Pergola G, Guay LA, et al. Identification of the K103N resistance mutation in Ugandan women receiving nevirapine to prevent HIV-1 vertical transmission. AIDS 2000; 14:F Eshleman SH, Mracna M, Guay LA, et al. Selection and fading of resistance mutations in women and infants receiving nevirapine to prevent HIV-1 vertical transmission (HIVNET 012). AIDS 2001; 15: Shetty AK, Coovadia HM, Mirochnick MM, et al. Safety and trough concentrations of nevirapine prophylaxis given daily, twice weekly, or weekly in breast-feeding infants from birth to 6 months. J Acquir Immune Defic Syndr 2003; 34: Los Alamos HIV Databases, Available at: Van de Peer Y, De Wachter R. TREECON for Windows: a software package for the construction and drawing of evolutionary trees for the Microsoft Windows environment. Comput Appl Biosci 1994; 10: Gonzales MJ, Dugan JM, Shafer RW. Synonymous-non-synonymous mutation rates between sequences containing ambiguous nucleotides (Syn-SCAN). Bioinformatics 2002; 18: Eshelman SH, Hoover D, Chen S, et al. Comparison of nevirapine resistance in women with subtype C compared with subtype A and after single-dose NVP [abstract 799]. In: Program and abstracts of the 12th Conference on Retrovirus and Opportunistic Infections (Boston) Martinson N, Morris L, Gray G, et al. HIV resistance and transmission following single dose nevirapine in a PMTCT cohort [abstract 38]. In: Program and abstracts of the 11th Conference on Retroviruses and Opportunistic Infections (San Francisco) Musoke P, Guay LA, Bagenda D, et al. A phase I/II study of the safety and pharmacokinetics of nevirapine in HIV-1-infected pregnant Ugandan women and their neonates (HIVNET 006). AIDS 1999; 13: Becquart P, Chomont N, Roques P, et al. Compartmentalization of HIV-1 between breast milk and blood of HIV-infected mothers. Virology 2002; 300: