Kim Thys

Acquired resistance of Mycobacterium tuberculosis to bedaquiline.

Bedaquiline (BDQ), an ATP synthase inhibitor, is the first drug to be approved for treatment of multi-drug resistant tuberculosis in decades. In vitro resistance to BDQ was previously shown to be due to target-based mutations. Here we report that non-target based resistance to BDQ, and cross-resistance to clofazimine (CFZ), is due to mutations in Rv0678, a transcriptional repressor of the genes encoding the MmpS5-MmpL5 efflux pump. Efflux-based resistance was identified in paired isolates from patients treated with BDQ, as well as in mice, in which it was confirmed to decrease bactericidal efficacy. The efflux inhibitors verapamil and reserpine decreased the minimum inhibitory concentrations of BDQ and CFZ in vitro, but verapamil failed to increase the bactericidal effect of BDQ in mice and was unable to reverse efflux-based resistance in vivo. Cross-resistance between BDQ and CFZ may have important clinical implications.

Minor variant detection in amplicons using 454 massive parallel pyrosequencing: experiences and considerations for successful applications

Ultra-deep sequencing (UDS) of amplicons is a major application for next-generation sequencing technologies, even more so for the 454 Genome Sequencer FLX. Especially for this application, errors that might be introduced during any of the sample processing or data analysis steps should be avoided or at least recognized, as they might lead to aberrant sequence variant calling. Since 454 pyrosequencing relies on PCR-driven target amplification, it is key to differentiate errors introduced during the amplification step from genuine minority variants. Thereto, optimal primer design is imperative because primer selection, primer dimer formation, and nonspecific binding may all affect the quality and outcome of amplicon-based deep sequencing. Also, other intrinsic PCR characteristics including amplification drift and the formation of secondary structures may influence sequencing data quality. We illustrate these phenomena using real life case studies and propose experimental and analytical evidence-based solutions for effective practice. Furthermore, because accuracy of the DNA polymerase is vital for reliable UDS results, a comparative analysis of error profiles from seven different DNA polymerases was performed and experimentally assessed in parallel by 454 sequencing. Finally, intra and interrun variability evaluation of the 454 sequencing protocol revealed highly reproducible results in amplicon-based UDS.

High-Throughput Analysis of Human Cytomegalovirus Genome Diversity Highlights the Widespread Occurrence of Gene-Disrupting Mutations and Pervasive Recombination

ABSTRACT Human cytomegalovirus is a widespread pathogen of major medical importance. It causes significant morbidity and mortality in immunocompromised individuals, and congenital infections can result in severe disabilities or stillbirth. Development of a vaccine is prioritized, but no candidate is close to release. Although correlations of viral genetic variability with pathogenicity are suspected, knowledge about the strain diversity of the 235-kb genome is still limited. In this study, 96 full-length human cytomegalovirus genomes from clinical isolates were characterized, quadrupling the amount of information available for full-genome analysis. These data provide the first high-resolution map of human cytomegalovirus interhost diversity and evolution. We show that cytomegalovirus is significantly more divergent than all other human herpesviruses and highlight hot spots of diversity in the genome. Importantly, 75% of strains are not genetically intact but contain disruptive mutations in a diverse set of 26 genes, including the immunomodulatory genes UL40 and UL111A. These mutants are independent of culture passage artifacts and circulate in natural populations. Pervasive recombination, which is linked to the widespread occurrence of multiple infections, was found throughout the genome. The recombination density was significantly higher than those of other human herpesviruses and correlated with strain diversity. While the overall effects of strong purifying selection on virus evolution are apparent, evidence of diversifying selection was found in several genes encoding proteins that interact with the host immune system, including UL18, UL40, UL142, and UL147. These residues may present phylogenetic signatures of past and ongoing virus-host interactions. IMPORTANCE Human cytomegalovirus has the largest genome of all viruses that infect humans. Currently, there is a great interest in establishing associations between genetic variants and strain pathogenicity of this herpesvirus. Since the number of publicly available full-genome sequences is limited, knowledge about strain diversity is highly fragmented and biased toward a small set of loci. Combined with our previous work, we have now contributed 101 complete genome sequences. We have used these data to conduct the first high-resolution analysis of interhost genome diversity, providing an unbiased and comprehensive overview of cytomegalovirus variability. These data are of major value to the development of novel antivirals and a vaccine and to identify potential targets for genotype-phenotype experiments. Furthermore, these data have enabled a thorough study of the evolutionary processes that have shaped cytomegalovirus diversity.

Ultra-deep sequencing of HIV-1 reverse transcriptase before start of an NNRTI-based regimen in treatment-naive patients

There are conflicting data on the impact of low frequency HIV-1 drug-resistant mutants on the response of first-line highly active antiretroviral therapy (HAART), more specifically containing a NNRTI. As population sequencing does not detect resistant viruses representing less than 15-25% of the viral population, more sensitive techniques have been developed but still need clinical validation. We evaluated ultra-deep sequencing (UDPS), recently more available and affordable, as a tool for the detection of HIV-1 minority species carrying drug resistant mutation (DRM) in a clinical setting. A retrospective analysis of the reverse transcriptase (RT) gene of plasma HIV-1 from 70 patients starting a NNRTI based regimen was performed. Minority populations were defined as representing > 1% and < 20% of the total viral population. Using UDPS, we could not confirm an association between the presence of low minority variants harbouring RT mutations at the start of therapy and primary or secondary therapeutic failure.

Quasispecies Analysis of JC Virus DNA Present in Urine of Healthy Subjects

JC virus is a human polyomavirus that infects the majority of people without apparent symptoms in healthy subjects and it is the causative agent of progressive multifocal leucoencephalopathy (PML), a disorder following lytic infection of oligodendrocytes that mainly manifests itself under immunosuppressive conditions. A hallmark for JC virus isolated from PML-brain is the presence of rearrangements in the non-coding control region (NCCR) interspersed between the early and late genes on the viral genome. Such rearrangements are believed to originate from the archetype JC virus which is shed in urine by healthy subjects and PML patients. We applied next generation sequencing to explore the non-coding control region variability in urine of healthy subjects in search for JC virus quasispecies and rearrangements reminiscent of PML. For 61 viral shedders (out of a total of 254 healthy subjects) non-coding control region DNA and VP1 (major capsid protein) coding sequences were initially obtained by Sanger sequencing. Deletions between 1 and 28 nucleotides long appeared in ∼24.5% of the NCCR sequences while insertions were only detected in ∼3.3% of the samples. 454 pyrosequencing was applied on a subset of 54 urine samples demonstrating the existence of JC virus quasispecies in four subjects (∼7.4%). Hence, our results indicate that JC virus DNA in urine is not always restricted to one unique virus variant, but can be a mixture of naturally occurring variants (quasispecies) reflecting the susceptibility of the non-coding control region for genomic rearrangements in healthy individuals. Our findings pave the way to explore the presence of viral quasispecies and the altered viral tropism that might go along with it as a potential risk factor for opportunistic secondary infections such as PML.

Deep-Sequencing Analysis of the Gene Encoding the Hepatitis C Virus Nonstructural 3–4A Protease Confirms a Low Prevalence of Telaprevir-Resistant Variants at Baseline and the End of the REALIZE Study

BACKGROUND Population sequencing (PS) has shown that telaprevir-resistant variants are not typically detectable at baseline (prevalence, ≤5% of patients), and most variants present at the time of treatment failure are no longer detectable at the end of the study. METHODS To gain insight into the evolution of telaprevir-resistant variants, their baseline prevalence and persistence after treatment was investigated using a more sensitive, deep-sequencing (DS) technique in a large number of treatment-experienced patients from the REALIZE study who were infected with hepatitis C virus genotype 1. RESULTS Before treatment initiation, telaprevir-resistant variants (T54A, T54S, or R155K in 1%-2% of the viral population) were detected by DS in a fraction (2%) of patients for whom PS failed to detect resistance; these variants were not necessarily detected at the time of treatment failure. Of 49 patients in whom telaprevir-resistant variants were detected by PS at the time of treatment failure but not at the end of the study, DS revealed the presence of variants (V36A/L/M, T54S, or R155K in 1%-36% of the viral population) in 16 patients (33%) at the end of the study. CONCLUSIONS Similar to PS findings, DS analysis revealed that the frequency of telaprevir-resistant variants before treatment was also low, and variants detected at the time of treatment failure were no longer detectable in the majority of patients during follow-up.

Performance assessment of the Illumina massively parallel sequencing platform for deep sequencing analysis of viral minority variants.

Massively parallel sequencing (MPS) technology has opened new avenues to study viral dynamics and treatment-induced resistance mechanisms of infections such as human immunodeficiency virus (HIV) and hepatitis C virus (HCV). Whereas the Roche/454 platform has been used widely for the detection of low-frequent drug resistant variants, more recently developed short-read MPS technologies have the advantage of delivering a higher sequencing depth at a lower cost per sequenced base. This study assesses the performance characteristics of Illumina MPS technology for the characterization of genetic variability in viral populations by deep sequencing. The reported results from MPS experiments comprising HIV and HCV plasmids demonstrate that a 0.5-1% lower limit of detection can be achieved readily with Illumina MPS while retaining good accuracy also at low frequencies. Deep sequencing of a set of clinical samples (12 HIV and 9 HCV patients), designed at a similar budget for both MPS platforms, reveals a comparable lower limit of detection for Illumina and Roche/454. Finally, this study shows the possibility to apply Illuminas paired-end sequencing as a strategy to assess linkage between different mutations identified in individual viral subspecies. These results support the use of Illumina as another MPS platform of choice for deep sequencing of viral minority species.

VirVarSeq: a low-frequency virus variant detection pipeline for Illumina sequencing using adaptive base-calling accuracy filtering

MOTIVATION In virology, massively parallel sequencing (MPS) opens many opportunities for studying viral quasi-species, e.g. in HIV-1- and HCV-infected patients. This is essential for understanding pathways to resistance, which can substantially improve treatment. Although MPS platforms allow in-depth characterization of sequence variation, their measurements still involve substantial technical noise. For Illumina sequencing, single base substitutions are the main error source and impede powerful assessment of low-frequency mutations. Fortunately, base calls are complemented with quality scores (Qs) that are useful for differentiating errors from the real low-frequency mutations. RESULTS A variant calling tool, Q-cpileup, is proposed, which exploits the Qs of nucleotides in a filtering strategy to increase specificity. The tool is imbedded in an open-source pipeline, VirVarSeq, which allows variant calling starting from fastq files. Using both plasmid mixtures and clinical samples, we show that Q-cpileup is able to reduce the number of false-positive findings. The filtering strategy is adaptive and provides an optimized threshold for individual samples in each sequencing run. Additionally, linkage information is kept between single-nucleotide polymorphisms as variants are called at the codon level. This enables virologists to have an immediate biological interpretation of the reported variants with respect to their antiviral drug responses. A comparison with existing SNP caller tools reveals that calling variants at the codon level with Q-cpileup results in an outstanding sensitivity while maintaining a good specificity for variants with frequencies down to 0.5%. AVAILABILITY The VirVarSeq is available, together with a users guide and test data, at sourceforge: http://sourceforge.net/projects/virtools/?source=directory.

Sensitive testing of plasma HIV-1 RNA and Sanger sequencing of cellular HIV-1 DNA for the detection of drug resistance prior to starting first-line antiretroviral therapy with etravirine or efavirenz

OBJECTIVES This study investigated strategies that may increase the yield of drug resistance testing prior to starting antiretroviral therapy (ART), and whether transmitted and polymorphic resistance-associated mutations (RAMs) correlated with virological outcomes. METHODS We carried out retrospective testing of baseline samples from patients entering the SENSE trial of first-line ART in Europe, Russia and Israel. Prior to randomization to etravirine or efavirenz plus two nucleos(t)ide reverse transcriptase inhibitors (NRTIs), plasma samples underwent routine Sanger sequencing of HIV-1 RT and protease ((plasma)SS) in order to exclude patients with transmitted RAMs. Retrospectively, Sanger sequencing was repeated with HIV-1 DNA from baseline peripheral blood mononuclear cells (PBMCSS); baseline plasma samples were retested by allele-specific PCR targeting seven RT RAMs (AS-PCR) and ultra-deep RT sequencing (UDS). RESULTS By (plasma)SS, 16/193 (8.3%) patients showed ≥ 1 transmitted RAM affecting the NRTIs (10/193, 5.2%), non-nucleoside reverse transcriptase inhibitors (4/193, 2.1%) or protease inhibitors (2/193, 1.0%). No additional RAMs were detected by AS-PCR (n = 152) and UDS (n = 24); PBMCSS (n =  91) yielded two additional samples with one RAM each. Over 48 weeks, 4/79 (5.1%) patients on etravirine and 7/78 (9.0%) on efavirenz experienced virological failure; none had baseline RAMs. Conversely, 11/79 (13.9%) patients randomized to etravirine had one polymorphic RAM from the etravirine score in baseline plasma (V90I, V106I or E138A), without any impact on virological outcomes. CONCLUSIONS The detection of resistance increased marginally with PBMC testing but did not increase with sensitive plasma testing. A careful consideration is required of the cost-effectiveness of different strategies for baseline HIV drug resistance testing.

HIV-1 nucleotide mixture detection in the virco(®)TYPE HIV-1 genotyping assay: a comparison between Sanger sequencing and 454 pyrosequencing.

HIV-1 Protease (PR) and Reverse Transcriptase (RT) genotyping is well established for the management of antiretroviral (ARV) drug therapy, as it is able to detect gene mutations encoding resistance to ARV compounds or drug classes, that are associated with reduced drug susceptibility (i.e. phenotype). A correct phenotypic interpretation from the derived PR-RT genotype (i.e. virtual phenotype), requires a well characterized geno-phenotype correlative database and appropriate statistical predictive models. The applicability of the virtual phenotype for the patient, will, however, not only depend on the accuracy of the statistical models and the database they rely on, but also depend largely on the sequence information that is provided. Since HIV-1 evolves as a complex of closely related but non-identical viral genomes (i.e. quasispecies) it is crucial that the sequencing method used, is able to characterize most of the genetic mixtures that make up the different quasispecies within a single patient. US regulatory agencies require that developers of HIV-1 genotyping assays, determine and report the HIV-1 mixture detection level of their assay. Hence, the mixture scoring sensitivity of the population-based Sanger sequencing method, along with the defined mixture scoring rules, used to drive the virco(®)TYPE HIV-1 virtual phenotype, was investigated by comparing it to the 454 pyrosequencing technique, which is able to generate the complete viral population sequence. To this end the PR-RT coding sequence of 20 clinical isolates was determined by both sequencing methodologies. The genotyping assay which feeds the virco(®)TYPE HIV-1 virtual phenotype was able to call automatically 97.5% (i.e. 268 mixtures) and 95.3% (i.e. 326 mixtures) of the mixtures that were present between 25 and 75% and between 20 and 80% in the viral population, as detected by 454. From the not called mixtures, all but one did present a mixture sequence in the Sanger DNA chromatograms, however, with a peak surface area for the second peak that was below the threshold setting for automatic mixture calling in the basecaller software (i.e. 25%). Viral loads ranged from 470 to 629,000 copies/mL and exerted no effect on the mixture calling relationship between both sequencing methodologies (R(2)=0.92). In some occasions (i.e. 55 mixtures) the genotyping assay would detect automatically mixtures that were present below 20% in the viral population, when measured by 454. Hence, the mixture scoring sensitivity of the automated high throughput virco(®)TYPE HIV-1 genotyping assay is currently set at 97.5% and 95.3%, for mixtures present at 25 and 20% in the viral population and may identify occasionally mutations that are present at lower frequencies. These findings were not influenced by the viral load of the examined samples.