Addis Ababa University Libraries Electronic Thesis and Dissertations: AAU-ETD! >
College of Health Sciences >
Thesis - Medical Microbiology >
Please use this identifier to cite or link to this item:
|Title: ||GENOMIC CHARACTERIZATION OF HIV-1 ISOLATES FROM ETHIOPIAN PATIENTS: BASELINE STUDIES ON ANTIRETROVIRAL DRUG RESISTANCE AND SUB-TYPE VARIATIONS|
|Authors: ||WOLDAREGAY, ERKU|
|Advisors: ||Dr. Yohannes Mengistu|
|Copyright: ||May-2011 |
|Date Added: ||11-May-2012 |
Several studies were conducted in the past two decades in Ethiopia to understand the genotypic
characteristics of prevalent HIV subtypes in the country. However, the majority of the information
gathered relied on sequencing of the envelope and a certain portion of the gag gene fragments of the HIV
genome. Considering the facts that relying only on sequencing of the env and gag regions has inherent
inadequacy in characterizing HIV subtypes, that long time has elapsed since genomic characterization of
HIV isolates in Ethiopia, that non-C subtypes have expanded in the neighboring countries, and that a new
selective pressure coming from the recently initiated ARV treatment has been introduce into current HIV
isolates circulating in the country as a consequence of which newHIV subtypes might be currently
circulating in the country, this PhD study was, therefore, undertaken to understand the genomic
characteristics of HIV isolates from ARV drug-naïve and drug-experienced persons by sequencing the
protease (PR) and reverse transcriptase (RT) genes using standard RT-PCR/PCR amplification and
genome sequencing protocols. In addition, two test evaluation studies were carried out: one on use of Inhouse
brewed genotyping system, and the other on use of DBS as source of specimen for genotypic HIV
drug resistance monitoring.
Materials and Methods
Different set of criteria were used to select eligible HIV-infected people to take part in three study groups:
recently infected drug-naïve antenatal care (ANC) attendee pregnant women (recent drug-naïve),
chronically infected drug-naïve but who were eligible to start ART (chronic drug-naïve), and chronically
infected heavily treated patients (chronic drug-experienced). For drug resistance assays, list of mutations
from three algorithms were utilized. For evaluations of the performance of In-house genotyping system
and DBS as source sample, the commercial genotyping system ViroSeqTM and plasma samples,
respectively, were used as standards. Mean nucleotide similarities of paired sequences generated from Inhouse/
ViroSeqTM genotyping system and DBS/plasma were determined on pairwise alignment sequences
using ClustalW multiple alignment program in BioEdit Version 22.214.171.124 software. For other purposes,
FASTA-formatted nucleotide and amino acid sequences were aligned and analyzed using various on-line
and off-line software. Descriptive, 2 test, and Student t test statistical data analyses were used as
appropriate, with significant level of 0.05.
Among chronically infected participant groups, 74% of drug-naïve and 84% of drug-experienced
participants were in the advanced WHO clinical stages III and IV at baseline. Nearly 92% of chronic
drug-naïve patients had CD4+ count of <200 cells/μL, and 70% had over 100,000 copies/ml of plasma
viral RNA. Mean baseline CD4+ T cell count for chronic drug-experienced patients was ~114 cells/μL.
Moreover, mean plasma viral RNA and mean CD4+ T cell counts for these patients at the time of
genotyping were 289,128 copies/ml and 238 cells/μL, respectively, after median treatment duration of 42
months. The performance of the In–house genotyping system relative to the performance of the
commercial ViroSeqTM genotyping system both in amplifying and genotyping the specimens in this study
was 100%. It had also mean concordance value of 98.72% at the nucleotide level. Similarly, the
performance of DBS as specimen for drug resistance genotyping relative to the performance of plasma
specimens was 100% for both amplification and genotyping. The mean nucleotide concordance rate
between paired DBS/plasma sequences was 98.82%.
The prevalence of transmitted drug resistance mutations was 4.9% among recent drug-naïve persons.
Among chronically infected drug-naïve patients, 22 major resistance mutations were detected: eight
against PIs, eight against NRTIs, and 6 against NNRTIs. Virological treatment failure among treatmentexperienced
patients was 31%; 70% of successfully genotyped viremic specimens harbored at least one
confirmed major drug resistance mutation from any class. Drug class-wide mutations were 3% against
PIs, 62% against NRTIs, 68% against NNRTIs, 61% against any NRTI and NNRTI double class
mutations, and 3% against all the three classes of drugs in use. Nearly 37% of D4T- and 20% of AZTcontaining
initial regimens were responsible for selection of at least one NRTI resistance mutations. Of
the initial drug regimens, ~27% of treated patients with NVP-containing regimens or ~55% of NVPexposed
patients with genotyped viral isolates harbored at least one major NNRTI resistance mutation.
Likewise, ~29% of all EFV-treated patients or 96% of EFV-exposed patients with genotyped viral isolates
had at least one NNRTI resistance mutation.
With regards to subtype distribution among the study participants, ~96% of the sequences tested were of
subtype C both at PR and RT genes. Sequences from two individuals (1%) were identified as subtype B,
while the remaining 3% were subtyped, at least by one genotyping algorithm, as mosaic forms at the PR
and RT genes; including four (2%) BCs, one (0.5%) CRF_02AG (or BG? or ABG?), and one (0.5%) A1D.
Analysis of genome diversity among the three groups of study participants showed that while certain non
resistance mutations had high prevalence in all the three groups, others showed differential occurrences.
Synonymous/non-synonymous substitution analysis and test of Shannon Entropy have also shown genome
diversity differences among the three groups with distinct positional patterns, diversity at certain patches
in the RT region being most prominent.
Discussions and Conclusions
Among chronically infected patients, the high baseline viral RNA load, low CD4+ T cell count, and more
persons in clinical stages III and IV during treatment initiation show that treatment was started late after
the patients’ virological, immunological and clinical conditions have already been deteriorated. These
poor baseline conditions were probably reflected by the high viral load values and low CD4+ T cell
counts observed after treatment period of over three years among heavily treated patients. The
transmitted drug resistance mutation rate documented in this study was in the WHO’s ‘low prevalence
category’. However, more surveillance works of greater scope are required in terms of both geographic
coverage and study population in order to understand the current prevalence of transmitted drug
resistance. The detection of 22 drug resistance mutations among drug-naïve patients starting ART
indicates the need for undertaking resistance testing before initiating therapy whenever settings allow.
Among drug experienced patients, the most important drugs which were associated with NRTI and NNRTI
resistance mutations were D4T and EFV, respectively. Suboptimal therapy could be the source of
resistance mutations against these two drugs. It could also be due to impaired absorption of the two
drugs, as the pharmacokinetics of ART drugs have not been investigated under the settings of Ethiopian
patients. This therefore calls for the need to conduct therapeutic drug monitoring (TDM) at least in
research settings on Ethiopian patients taking the various drugs. A lot of so-called non-resistance
mutations occurred at positions known to confer resistance. Since most resistance/non-resistance
mutation identification was done based on subtype B genetic background, and since virus variants behave
differently under different settings, it would be of great benefit if phenotypic drug resistance profiles of
isolates from Ethiopia with these non-resistance mutations are investigated.
With regards to subtype distribution, the overwhelming majority of the isolates belonged to subtype C at
the pol region investigated in this study, showing that still this subtype has not been replaced by any other
in Ethiopia. However, variants with mosaic genomic regions have been detected, showing the possibility
of finding subtypes other than C, particularly recombinant forms, if further sequencing is made from
genomic region larger than the PR and RT regions. Because of discrepancies between various genotyping
algorithms in assigning subtypes, and because none of these mosaic forms had exact similarity with those
previously known subtype identity, the mosaic variants identified in this study might be considered as
Unique Recombinant Forms (URFs). Worth noting, however, is the population sequencing method used in
this study might have masked presence of minority populations both in terms of variants with drug
resistance mutations and those with different subtype profiles. In the latter case, use of only the short pol
region might have contributed to the ambiguity of subtype identification in the non-C isolates detected in
|Appears in:||Thesis - Medical Microbiology|
Items in the AAUL Digital Library are protected by copyright, with all rights reserved, unless otherwise indicated.