Deep Sequencing of Hepatitis C Virus Reveals Genetic Compartmentalization in Cerebrospinal Fluid From Cognitively Impaired Patients

Damien C. Tully; Simon Hjerrild; Peter D. Leutscher; Signe G. Renvillard; Colin B. Ogilvie; David J. Bean; Poul Videbech; Todd M. Allen; Jane A. McKeating; Nicola F. Fletcher


Liver International. 2016;36(10):1418-1424. 

In This Article

Materials and Methods

Study Subjects

Matched plasma and CSF was obtained from six patients enrolled in a study of cerebral function in hepatitis C infection. Patients were from an outpatient clinic at Aarhus University Hospital, Denmark. The patients were Caucasian, viraemic, non-cirrhotic patients infected with HCV genotypes 1 and 3 and did not have any significant neurological or psychotic disorders, previous history of head trauma, severe liver fibrosis or cirrhosis, history of drug abuse within the past two years, HIV or hepatitis B virus (HBV) co-infection or prior interferon treatment. Plasma was obtained from six non-HCV-infected age-, sex- and education-matched subjects free of somatic and psychiatric morbidity (Table S1). Four of six HCV-infected patients had significant fatigue (FSS average score ≥5) and cognitive dysfunction. A lumbar puncture was performed on the HCV-infected patients and peripheral blood-derived plasma was collected from all HCV-infected and control subjects.

Neuropsychological Testing and Fatigue Assessment

Fatigue was assessed using the Fatigue Severity Scale (FSS). The psychometric properties of the FSS has been validated in HCV-infected subjects, and in accordance with Lerdal et al.[8] we defined clinically significant fatigue as an average item score of 5 or above. Neuropsychological testing included assessment of the following cognitive domains: Visuomotor speed (Symbol digit modalities test (SDMT) and trail making test A + B); attention (D2 test of attention and Stroop's test); perceptual organization (WAIS III matrix reasoning and Street completion test); memory and learning (WMS III logical memory I and II, Rey complex figure test (RCFT) and Rey auditory verbal learning test (RAVLT)), and executive function (WAIS III letter number sequencing, verbal fluency (s and d words, animals, street items) and Wisconsin card sorting test (WCST)). Performance was calculated as a z score based on the performance of the healthy controls, and z scores from each individual test were added to yield a domain z score. Subsequently, clinically significant cognitive impairment was defined as a cognitive performance ≥1 SD below healthy controls in ≥2 domains.[9]

Cytokine Measurement

Following collection of CSF, microscopy was performed to identify the presence of any contaminating cells. Patients 1–5 had no detectable contaminating cells in the CSF, and patient 6 had a low level of detectable leucocytes in the CSF (6 cells/10−6). To further eliminate the presence of contaminating peripheral cells, the CSF was filtered through a 0.2-μm filter before RNA was prepared or cytokine levels were measured by Luminex analysis. Cytokines were measured using a Luminex bead array consisting of a human 27-plex cytokine kit (Bio-Rad Laboratories Inc., Hercules, CA, USA) which included TNF-α, MCP-1, MIP-1α, IL-8, IL-1β, IL-6, VEGF, IL-1ra, IL-10, IL-2, IL-4, IL-5, IL-7, IL-9, IL-12 (p70), IL-13, IL-15, IL-17, IFN-γ, IP-10, G-CSF, GM-CSF, RANTES, FGF-basic, Eotaxin, PDGF-bb, and MIP-1β. All samples were assayed concurrently in duplicate.


All participants provided written informed consent, and the study was conducted in accordance with the ethical guidelines of the World Medical Association's Declaration of Helsinki. The study was approved by The Central Denmark Region Committees on Health Research Ethics (2008-0096), Danish Data Protection Agency (2009-41-2226) and monitored by the Good Clinical Practice Unit at Aarhus University Hospital. identifier: NCT-00788918.

RNA Amplification and Deep Sequencing

RNA was isolated from 140 μl of plasma or CSF using the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany). An amplicon covering Core to NS2 was generated using the SuperScript III One-Step RT-PCR System with Platinum Taq DNA Polymerase High Fidelity (Invitrogen, Carlsbad, CA, USA) (primers available on request). Thermal cycling parameters were: 55°C for 30 min, 94°C for 2 min, followed by 40 cycles of 94°C for 15 s, 58°C for 30 s and 68°C for 4 min, with a final extension step at 68°C for 10 min. Purified PCR products were pooled in equimolar ratios, fragmented using the transposon-mediated Nextera DNA Sample Prep Kit (Illumina, San Diego, CA, USA) and purified with a DNA Clean and Concentrator Kit (Zymo Research, Freiburg, Germany). Adapters and barcodes were added by limited-cycle PCR (Nextera, Florida, USA). Small fragments were removed following Roche's recommended size selection protocol and quantified using a Promega Quantiflor-ST fluorometer. After quantification, barcoded samples were pooled at a final concentration of 107 molecules/μl to create the library for sequencing on a GS Junior 454 Genome Sequencer (Roche, Dublin, Ireland). Emulsion PCR, breaking, and DNA sequencing were performed according to the manufacturer's protocols for Lib-L (Roche). Genetic compartmentalization was assessed using the Slatkin–Maddison test[10] and the non-parametric test for panmixia described by Hudson et al.[11] with 10 000 permutations to generate a P value.

In the event that an amplification using the above strategy was unsuccessful, an alternative PCR amplicon-based ultra-deep sequencing approach was used to analyse a region encompassing the hypervariable region (HVR) in the N-terminus of E2 (primers available on request). Thermal cycling conditions were: 55°C for 30 min, 94°C for 2 min, followed by 40 cycles of 94°C for 15 s, 52°C for 30 s and 68°C for 30 s, with a final extension step at 68°C for 5 min. Sequencing was performed as described above except that Emulsion PCR, enrichment, breaking and DNA sequencing were performed according to the GS Junior FLX Titanium Series manuals for Lib-A (Roche).

Detection of minor HCV variants required a targeted coverage 250-fold per site. In this study, plasma and CSF samples were sequenced to an average depth of 354 and 314 reads per site respectively. The raw sequence outputs ('reads') were assembled by VICUNA de novo assembler software[12] and corrected for systematic 454 sequencing errors using the previously developed RC454 and V-Phaser algorithms.[13,14]

Statistical Analyses

Principal component and hierarchical clustering analysis were performed using the JMP Pro version 12 statistical package.