High Prevalence of HIV Low Abundance Drug-Resistant Variants in a Treatment-Naive Population in North Rift Kenya

Abstract

The advent of antiretroviral treatment (ART) has resulted in a dramatic reduction in AIDS-related morbidity and mortality. However, the emergence and spread of antiretroviral drug resistance (DR) threaten to negatively impact treatment regimens and compromise efforts to control the epidemic. It is recommended that surveillance of drug resistance occur in conjunction with scale-up efforts to ensure that appropriate first-line therapy is offered relative to the resistance that exists. However, standard resistance testing methods used in Sub-Saharan Africa rely on techniques that do not include low abundance DR variants (LADRVs) that have been documented to contribute to treatment failure. The use of next generation sequencing (NGS) has been shown to be more sensitive to LADRVS. We have carried out a preliminary investigation using NGS to determine the prevalence of LDRVS among a drug-naive population in North Rift Kenya. Antiretroviral-naive patients attending a care clinic in North Rift Kenya were requested to provide and with consent provided blood samples for DR analysis. DNA was extracted and amplified and nested PCR was conducted on the pol RT region using primers tagged with multiplex identifiers (MID). Resulting PCR amplicons were purified, quantified, and pyrosequenced using a GS FLX Titanium PicoTiterPlate (Roche). Valid pyrosequencing reads were aligned with HXB-2 and the frequency and distribution of nucleotide and amino acid changes were determined using an in-house Perl script. DR mutations were identified using the IAS-USA HIV DR mutation database. Sixty samples were successfully sequenced of which 26 were subtype A, 9 were subtype D, 2 were subtype C, and the remaining were recombinants. Forty-six (76.6%) had at least one drug resistance mutation, with 25 (41.6%) indicated as major and the remaining 21 (35%) indicated as minor. The most prevalent mutation was NRTI position K219Q/R (11/46, 24%) followed by NRTI M184V (5/46, 11%) and NNRTI K103N (4/46, 9%). Our use of NGS technology revealed a high prevalence of LADRVs among drug-naive populations in Kenya, a region with predominantly non-B subtypes. The impact of these mutations on the clinical outcome of ART can be ascertained only through long-term follow-up.

Introduction

In Kenya the availability of antiretroviral therapy (ART) is increasing as well as in other low-income and middle-income countries. As ART use continues to be scaled up, there is mounting evidence suggesting that drug resistance (DR) will develop and increase over time.¹ A survey performed in Kampala between 2009 and 2010 showed that the prevalence of transmitted drug resistance was 8.6%.² It is highly recommended by the World Health Organization (WHO) that surveillance of drug resistance occurs in conjunction with scale-up efforts to ensure that appropriate first-line therapy is offered relative to the resistance that exists.³

It is believed that surveillance will maximize the utility of first-line therapy and help minimize the cost of providing ART, thereby sustaining current antiretroviral drug programs. However, current standard genotypic resistance testing methods used in surveillance programs rely on Sanger sequencing (SS), a method that has a detection limit in the neighborhood of 20% of the virus quasispecies.⁴ Increased rates of virological failure to ART regimens, especially nonnucleoside reverse transcriptase inhibitors (NNRTIs), have been noticed despite no evidence of DR mutations by SS at baseline. In fact, many studies have shown a low abundance of DR variants (LADRVs) at frequencies less than 20% in both ART-naive and heavily ART-treated subjects.^5,6

Furthermore, it has been noted that these LADRVs can increase and outcompete wild-type strains under drug selection pressure leading to treatment failure.⁷ Due to limitations of Sanger sequencing and the need for low-cost genome sequencing, there has been a revolution in the large-scale genomics field. To date, three major next-generation sequencing (NGS) technologies—Roche 454 Life Science FLX (454), Illumina (Solexa), and Ion PGM (Life Technologies)—have been commercialized.⁸ These technologies share the paradigm of massive, parallel, clonal analysis of DNA templates with high data throughput. One application of these technologies is ultradeep pyrosequencing (UDPS), which makes it possible to identify LADRVs not detectable by the standard Sanger sequencing genotypic technique. Various studies from North America and Europe have shown that UDPS could identify LADRVs at frequencies as low as 0.05% of the entire viral population and enable detailed coverage of rare HIV DR variants.⁹

The majority of LADRV studies have been carried in Europe and North America, regions predominantly infected with HIV subtype B viruses. Less information is available from Sub-Saharan Africa, a region in which non-B subtypes are prevalent and which reportedly has the highest projected rate of emerging transmitted HIV drug resistance.⁴ The aim of our study was to survey the prevalence of LADRVS among treatment-naive populations in North Rift Kenya using NGS techniques to inform on current status of HIV DR in the country.

Materials and Methods

Between September 2009 and October 2011, patients who were ART naive according to WHO guidelines were consecutively enrolled at the Moi Teaching and Referral Hospital, Eldoret. After informed consent was obtained, a standardized questionnaire was administered to assess demographic, epidemiological, clinical, and treatment information. Remnant blood samples from clinical analyses were acquired and DNA was extracted and amplified as per the procedures previously outlined.¹⁰ Briefly, viral RNA was reverse transcribed and amplified according to the manufacturer's directions using the QIAGEN one-step RT-PCR kit (QIAGEN, Canada). Reverse transcriptase (RT) primers were TGAARGAITGYACTGARAGRCAGGCTAAT and CCTCATTYTTGCATAYTTYCCTGTT with cycling conditions of 50°C 40 min one cycle in the first step, 95°C 15 min one cycle in the second step, 35 cycles at 95°C 30 s, 53°C 30 s, and 72°C 2 min 30 s in the third step, one cycle at 72°C 10 min in the fourth cycle, and one cycle at 4°C as a final extension. Nested PCR was conducted using fusion primers with forward primers tagged with multiplex identifiers (MID). The sequences of the MID-tagged primers are shown in Table 1. All MID-tagged forward fusion primers consisted of a forward primer adaptor sequence (5′-CGTATCGCCTCCCTCGCGCCA-3′) and a reverse adaptor (5′-CTATGCGCCTTGCCAGCCCGC-3′).

Table 1.

Primer Sequences Used in the Analysis

454-1F-MID-X	Degr_Seg1_F	2251∼2270	TCCCTCARATCACTCTTTGG	462 bp	Seg 1 covers the whole PR
454-1R	Degr_Seg2_R	2692∼2713	GGRTTTTYAGGCCCAATTTT
454-2F-MID-X	Degr_Seg2_F	2583∼2602	TKAAAG CCAGGRATGGATGG	390 bp	Cover aa 1∼140 in RT
454-2R	Degr_Seg3_R	2954–2973	TCCCTGGTGTCTCATTGTTT
454-3F-MID-X	Degr_Seg3_F	2870∼2890	AGTACTRGATGTGGGWGATGC	400 bp	Cover aa 107∼240 in RT

Primers for nested PCR (preoptimized tagged primers).

PR, protease; RT, reverse transcriptase.

All nested PCR procedures were performed using common reaction conditions at an annealing temperature of 58.9°C. The resulting PCR amplicons were purified, quantified, and pyrosequenced using 1/16 the capacity of a full GS FLX Titanium PicoTiterPlate. Reads that passed the quality control software, which were of sufficient read length to cover the amplicon and could be successfully mapped to the HXB-2 reference sequence, underwent further analysis.

Sequence and phylogenetic analysis

The valid pyrosequencing reads were aligned with HXB-2 and the frequency and distribution of nucleotide and AA changes were determined using an in-house Perl script.¹⁰ Drug resistance mutations were identified using the IAS-USA HIV DR mutation database with a threshold of 0.1% of sequences set to show the same mutation.

Phylogenetic analysis was performed on the group of pyrosequencing reads containing the TDR mutation of interest, and subtype determination was done with reference sequences.

Results

Sequence analysis and drug-resistant mutations (DRMs)

Sixty samples from ARV-naive patients were successfully sequenced of which 25 were subtypes A, 11 were subtype D, 1 was subtype C, and the remaining were recombinants. Forty-six (76.6%) had at least one drug resistance mutation, with 25 (41.6%) indicated as major and the remaining 21 (35%) indicated as minor. Thirty-one of the 60 patients (51.67%) had the RT mutations. Nineteen patients (31%) had established NRTI-resistant mutations. The most prevalent mutation was NRTI position K219Q/R (11 of 46, 24%) followed by NRTI M184V/I (5 of 46, 11%) and NNRTI K103N (4 of 46, 9%). Table 2 shows the profiles of the patients, the DRM frequencies, and the subtypes.

Table 2.

Drug Resistance Mutation Frequencies Observed in the Antiretroviral Drug-Naive Population in North Rift Kenya

Sample ID	Gender	Age	Mutations and percentages	Subtype
NLHG-72	Female	39	V82A 1.03	D
NLHG-77	Male	51	M184V 1.75, N88D1.14	A1
NLHG-78	Female	28	K103N 7.14	A1
NLHG-79	Female	28	K219R 2.22, L24I 5.88I, M184V 1.79, G190E 2.22	C
NLHG-80	Male	33	Y188H 1.21, I54T 1.07, K219Q 2.13, K219E 1.06	D
NLHG-83	Male	45	K219R 1.05, N83D 1.66, K219E1.05	A1
NLHG-84	Male	36	K219R 1.72, F77L 2.21, G190E 36.29, K219Q 3.86, K219E 1.72	AE/D
NLHG-85	Female	46	L1004.93I, L24I1.75, K101E1.75	A2/D
NLHG-88	Male	30	M46L 1.66, Y188H 1.22, V82A 1.07, F53Y 1.34, G73S 1.07	A1
NLHG-89	Female	29	M184V 1.08, G190E 1.08	A1
NLHG-90	Female	54	I541.27 T, M184V 1.95	AE/D
NLHG-91	Female	44	Y188H 1.64, L24I 2, N88S1.92, I84V 1.92, I50V 1.89, Y181C 1.61	D
NLHG-92	Female	27	K101E 2.04, K219Q 7.06	B
NLHG-94	Female	41	K101E 50	A/AE
NLHG-95	Female	30	K65R 1.12, I47V 1.47	A1
NLHG-96	Female	52	F77L 2.59, D67G 1.72	A1
NLHG-97	Female	30	K219R 1.49, L100 4.23I, K65R 1.11, V32I 4.17, K219Q 7.46, K219E 2.99	A1
NLHG-99	Female	39	M184I3.28, N83D 1.39, N88S 1.39, I50V 2.78, I54T 2.78, D67G 16.67, V82A1.39	A1
NLHG-100	Male	40	K219Q 1.33	A1
NLHG-101	Female	39	K101E 4.86	A1
NLHG-103	Male	63	M184V 3.7, F53L 1.54	A1
NLHG-105	Female	58	D67N 8.33, T69D8.33	A1
NLHG-107	Female	25	K101E 4.44, N83D 2.86	A1
NLHG-110	Male	39	D30N 7.14, I85V 7.14	D
NLHG-111	Male	32	G190E 1.48, K101E 11.11, N88D 3.33, K70R 11.11, T215Y 1.19	A1
NLHG-114	Female	28	N88D1.36, K219Q 3.23, L74V 3.03	A1
NLHG-115	Female	36	K219Q 9.09	A1/B/D
NLHG-117	Male	30	K219Q 3.03	A1
NLHG-118	Female	26	K219Q 5.1	D
NLHG-119	Male	36	F77L 4.35, K103N 14.29	A1
NLHG-121	Female	30	K219Q1.23	D
NLHG-172	Female	33	G73S1.37	A1
NLHG-122	Female	36	I54T 7.14, I47V 7.14	D
NLHG-126	Female	66	K103N 71.43	A1
NLHG-128	Female	8	N83D 1.42, K219Q 13.64, M46I 1.41	D
NLHG-129	Female	33	D67G 3.12, N88D 1.45	A1
NLHG-133	Female	29	L100 5.26I, K219Q 10.09	A1/D
NLHG-173	Female	25	K219Q 3.33	A1/D
1NLHG-35	Male	33	M41L 2.5, M184I 1.42, L210W 2.56, T215Y 2.56	A1
NLHG-149	Female	38	K219R 8.16 K65R 1.54, K101E 1.49, K70R 1.46	A1
NLHG-175	Female	27	K103S 100, M184V 95.35, F53L 9.09, V106M 100, K70E 100	C
NLHG-150	Female	32	F77L 1.16, K101E 1.16	D
NLHG-157	Male	46	K103S 1.9, N88D 4.41, K219Q 9.92, K103N 84.81, V82A 1.43, G73S 1.43, I47V 1.43	A1/D
NLHG-159	Female	34	M184V 100, Y181V 100, N88D 5	A1
NLHG-162	Female	29	F53L1.64, N88D 1.67, K103N 3.41	A1
NLHG-164	Female	0	F53L3.77, N88D 1.85, I47V 1.85, M46I 1.85	A1

Discussion

Our use of NGS technology revealed a high prevalence of low abundance drug-resistant variants among the drug-naive populations of North Rift Kenya. To our knowledge this is one of the rare instances in which NGS technology has been utilized to determine the status of HIV drug resistance in HIV-infected populations in the country. Our earlier studies of HIV DR on drug-naive antenatal clinic attendees in the region using direct Sanger sequencing had revealed a prevalence of transmitted drug resistance mutations of only 3.2%.¹¹ This high outcome with the use of NGS is in concordance with results observed elsewhere in Africa with a prevalence of 80% among drug-naive populations in Zambia infected by HIV subtype C,¹² thus confirming the high sensitivity of this technique for the detection and quantification of DRMs.

The most prevalent variants observed in our study were the mutants carrying the thymidine analogue mutations (TAM) K219Q/R and M184V/I and none with TAM-K103N. Viruses with K219Q mutations have been noted to evolve rapidly to zidovudine (AZT) resistance and show high replicative fitness in the presence of AZT. However, the M184V mutation confers high-level resistance to lamivudine (3TC), a key backbone to first-line antiretroviral treatment regimens in Kenya. The M184I mutation has been noted to be the first to appear, but is quickly replaced by the M184V mutation since this mutation has a greater ability to induce a higher replicative capacity.¹³

In the WHO drug resistance reports an increase of transmitted drug-resistant variants has been observed in Sub-Saharan Africa over time with the most commonly observed DRMs being M184V and K103N.¹⁴ The K103N mutation in particular was noted in more than half of HIV-infected patients presenting with NNRTI resistance. Despite such evidence of increasing rates of transmitted and acquired NNRTI resistance, efavirenz or nevirapine is still a key component in first-line ART in Africa.

It should be noted that our assessment of the prevalence of each mutation was population based and we did not assemble reads into variants (haplotypes). Our sequence read lengths were short (about 400 bps) and we used three amplicons to cover the whole protease and only 1–240 amino acids of the reverse transcriptase gene. Our approach could therefore not determine if multiple mutations were on the same amplicon.

Nevertheless, the outcome of our study denotes widespread low abundance drug-resistant strains in regions with non-B subtypes. Our use of NGS technology revealed a high prevalence of LADRVs among drug-naive populations in Kenya. This scenario might have been the result of the transmission of drug-resistant viruses from partners infected with the resistant virus or selection as a result of undisclosed use of ART. Nevertheless, this calls for the use of feasible next generation sequencing technologies for the surveillance of HIV drug resistance at the population level to reliably detect and monitor emerging drug resistance patterns that may impact ART. Similarly, they will be a need for a continued follow-up of persons with LADRVS to determine clinical impact and help guide therapies for drug-naive populations.

Sequence Data

The sequence data have been deposited at GenBank under accession number SRP053141.

Footnotes

Acknowledgments

We wish to thank the directors and staff of AMPATH clinics and laboratories in Eldoret, Kenya, the staff of the Kenya Medical Research Institute (KEMRI) in Nairobi, and HIV laboratory personnel of the National HIV and Retrovirology Laboratories (NHRL), Ottawa for their technical support and cooperation. This study was funded by the Canadian Commonwealth Scholarship Program (CCSP) of the Canada Ministry of Foreign Affairs and International Trade (DFAIT).

Author Disclosure Statement

No competing financial interests exist.

References

Steegen

, Stanley

, Kenny

, et al.: Effectiveness of antiretroviral therapy and development of drug resistance in HIV-1 infected patients in Mombasa, Kenya. AIDS Res Ther, 2009; 16(6):12.

Lihana

, Khamadi

, Lubano

, et al.: HIV type 1 subtype diversity and drug RS among HIV type 1-infected Kenyan patients initiating antiretroviral therapy. AIDS Res Hum Retroviruses, 2009; 25(12):1211–1217.

Ndembi

, Hamers

, Sigaloff

, et al.: Transmitted antiretroviral drug resistance among newly HIV-1 diagnosed young individuals in Kampala. AIDS, 2011; 25:905–910.

Bennett

, Myatt

, Bertagnolio

, et al.: Recommendations for surveillance of transmitted HIV drug resistance in countries scaling up antiretroviral treatment. Antivir Ther, 2008; 13(Suppl 2):2.

World Health Organization and Joint United Nations Programme on HIV/AIDS: Global HIV/AIDS response: Epidemic update and health sector progress towards universal access: progress report 2011. Available at www.who.int/hiv/pub/progress report2011/hiv. Full report 2011, pdf.

Paredes

, Lalama

, Ribaudo

, et al.: Pre-existing minority drug-resistant HIV-1 variants, adherence, and risk of antiretroviral treatment failure. J Infect Dis, 2010; 201(5):662–671.

Dudley

, Chin

, Bimber

, et al.: Low-cost ultra-wide genotyping using Roche/454 pyrosequencing for surveillance of HIV drug resistance. PLoS One, 2012; 7(5):e36494.

Bennett

, Camacho

, Otelea

, et al.: Drug RS mutations for surveillance of transmitted HIV-1 drug-RS: 2009 update. PLoS One, 2009; 4:e4724.

, Masse

, Tyler

, et al.: HIV drug resistance surveillance using pooled pyrosequencing. PLoS One, 2010; 5:e9263.

10.

, Paredes

, Ribaudo

, et al.: Low frequency HIV-1 drug resistance mutations and risk of NNRTI-based antiretroviral treatment failure: A systematic review and pooled analysis. JAMA, 2011; 305:1327–1335.

11.

Kiptoo

, Ichimura

, Wembe

, et al.: Prevalence of nevirapine-associated resistance mutations after single dose prophylactic treatment among antenatal clinic attendees in North Rift Kenya. AIDS Res Hum Retroviruses, 2008; 24(12):1555–1559.

12.

Gonzalez

, Tully

, Gondwe

, and Wood

: Low-abundance resistant mutations in HIV-1 subtype C antiretroviral therapy-naive individuals as revealed by pyrosequencing. Current HIV Res, 2013; 11:43–49.

13.

García-Lerma

, MacInnes

, Bennett

, et al.: Transmitted human immunodeficiency virus type 1 carrying the D67N or K219Q/E mutation evolves rapidly to zidovudine resistance in vitro and shows a high replicative fitness in the presence of zidovudine. J Virol, 2004; 78:7545–7552.

14.

WHO Drug resistance report 2012, available as pdf-report at www.who.int.