High Viral Diversity and Mixed Infections in Cerebral Spinal Fluid From Cases of Varicella Zoster Virus Encephalitis

Abstract

Background: Varicella zoster virus (VZV) may cause encephalitis, both with and without rash. Here we investigate whether viruses recovered from the central nervous system (CNS; encephalitis or meningitis) differ genetically from those recovered from non-CNS samples.

Methods: Enrichment-based deep sequencing of 45 VZV genomes from cerebral spinal fluid (CSF), plasma, bronchoalveolar lavage (BAL), and vesicles was carried out with samples collected from 34 patients with and without VZV infection of the CNS.

Results: Viral sequences from multiple sites in the same patient were identical at the consensus level. Virus from vesicle fluid and CSF in cases of meningitis showed low-level diversity. By contrast, plasma, BAL, and encephalitis had higher numbers of variant alleles. Two CSF-encephalitis samples had high genetic diversity, with variant frequency patterns typical of mixed infections with different clades.

Conclusions: Low viral genetic diversity in vesicle fluid is compatible with previous observations that VZV skin lesions arise from single or low numbers of virions. A similar result was observed in VZV from cases of VZV meningitis, a generally self-limiting infection. CSF from cases of encephalitis had higher diversity with evidence for mixed clade infections in 2 cases. We hypothesize that reactivation from multiple neurons may contribute to the pathogenesis of VZV encephalitis.

Figure 1.

Network phylogeny identifies 2 putative interclade recombinant varicella zoster virus (VZV) genomes. Phylogenetic network of 55 VZV genomes, including 45 from this study and 10 representative the 6 major geographic clades (numbered 1–6). CSF1 and CSF2 represent putative interclade recombinants. Sample types are color coded. Abbreviations: BAL, bronchoalveolar lavage; CSF, cerebrospinal fluid; PLAS, plasma; SPU, sputum; VES, vesicle.

Figure 2.

Intrahost nucleotide diversity (π) of varicella zoster virus (VZV) across compartments. Nucleotide diversity estimates of the VZV population in each individual sample is shown, grouped by sampled compartment. Median, interquartile range, and maximum-minimum range are marked, as is a dotted red line denoting nucleotide diversities that are 2 SD from the mean (calculated from the nucleotide diversities of all samples). Outlier CSF (red) and plasma (orange) samples are shown. The blue symbol represents the sum diversity of individual vesicles from patient D. Abbreviations: BAL, bronchoalveolar lavage; CSF, cerebrospinal fluid; PLAS, plasma.

Figure 3.

Total counts of filtered variant alleles per sample sequenced in this study. Sample identities correspond to samples shown in Table 1. Multiple-sampled patient are identified in parentheses (A–F). Abbreviations: BAL, bronchoalveolar lavage; CSF, cerebrospinal fluid; PLAS, plasma; SPU, sputum; VES, vesicle.

Figure 4.

The genome-wide distribution and frequency scatter plots for all variant alleles in multiple samples from patients A and B. The x axis denotes genome position and y axis denotes the frequency of each variant allele (black circle). The number of variant alleles are binned at 1% frequencies and are displayed in a horizontal bar chart on the right side. Abbreviation: BAL, bronchoalveolar lavage.

Figure 5.

Evidence of varicella zoster virus superinfections in cerebrospinal fluid and plasma. A–C, The genome-wide distribution and frequency of all variant alleles shown as a scatter plot (left). X axis denotes genome position and y axis denotes the frequency of each variant allele (black circle). The number of variant alleles are binned at 1% frequencies and are displayed in a horizontal bar chart (right side). A normally distributed subpopulation of alleles (dotted red box) is suggestive of a mixed infection. Subpopulations are shown in the network phylogenies (right) (major sequence, blue circle; minor sequence, red circle). Abbreviations: CSF, cerebrospinal fluid; PLAS, plasma.