https://orcid.org/0000-0001-6836-1468
Figures
Abstract
Viruses can infect the brain in individuals with and without HIV-infection: however, the brain virome is poorly characterized. Metabolic alterations have been identified which predispose people to substance use disorder (SUD), but whether these could be triggered by viral infection of the brain is unknown. We used a target-enrichment, deep sequencing platform and bioinformatic pipeline named “ViroFind”, for the unbiased characterization of DNA and RNA viruses in brain samples obtained from the National Neuro-AIDS Tissue Consortium. We analyzed fresh frozen post-mortem prefrontal cortex from 72 individuals without known viral infection of the brain, including 16 HIV+/SUD+, 20 HIV+/SUD-, 16 HIV-/SUD+, and 20 HIV-/SUD-. The average age was 52.3 y and 62.5% were males. We identified sequences from 26 viruses belonging to 11 viral taxa. These included viruses with and without known pathogenic potential or tropism to the nervous system, with sequence coverage ranging from 0.03 to 99.73% of the viral genomes. In SUD+ people, HIV-infection was associated with a higher total number of viruses, and HIV+/SUD+ compared to HIV-/SUD+ individuals had an increased frequency of Adenovirus (68.8 vs 0%; p<0.001) and Epstein-Barr virus (EBV) (43.8 vs 6.3%; p=0.037) as well as an increase in Torque Teno virus (TTV) burden. Conversely, in HIV+ people, SUD was associated with an increase in frequency of Hepatitis C virus, (25 in HIV+/SUD+ vs 0% in HIV+/SUD-; p=0.031). Finally, HIV+/SUD- compared to HIV-/SUD- individuals had an increased frequency of EBV (50 vs 0%; p<0.001) and an increase in TTV viral burden, but a decreased Adenovirus viral burden. These data demonstrate an unexpectedly high variety in the human brain virome, identifying targets for future research into the impact of these taxa on the central nervous system. ViroFind could become a valuable tool for monitoring viral dynamics in various compartments, monitoring outbreaks, and informing vaccine development.
Citation: Dang X, Hanson BA, Orban ZS, Jimenez M, Suchy S, Koralnik IJ (2024) Characterization of the brain virome in human immunodeficiency virus infection and substance use disorder. PLoS ONE 19(4): e0299891. https://doi.org/10.1371/journal.pone.0299891
Editor: Yanpeng Li, Shanghai Public Health Clinical Center, Fudan University, CHINA
Received: September 5, 2023; Accepted: February 17, 2024; Published: April 17, 2024
Copyright: © 2024 Dang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are available on NCBI BioProject (Accession number: PRJNA1068104).
Funding: This publication was made possible from NIH funding NIDA Avant Garde DP1 DA048493 to IJK; along with shared resources from NIH funding through the NIMH and NINDS by the following grants: Manhattan HIV Brain Bank (MHBB): U24MH100931 Texas NeuroAIDS Research Center (TNRC): U24MH100930 National Neurological AIDS Bank (NNAB): U24MH100929 California NeuroAIDS Tissue Network (CNTN): U24MH100928 Data Coordinating Center (DCC): U24MH100925 The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Igor Koralnik receives royalties from UpToDate chapters on CND mass lesions in HIV+ individuals, Progressive Multifocal Leukoencephalopathy and Neurologic Manifestations of COVID-19 the declared competing interest for IJK does not alter our adherence to PLOS ONE policies on sharing data and materials
Introduction
Drug use has been linked with the human immunodeficiency virus (HIV) since the beginning of the HIV pandemic. Drug usage is also associated with a higher risk of other viral infections, but little is known about the influence of viruses on substance use disorder (SUD). Therefore, whether viral infections of the brain can trigger an immune, metabolic, regulatory, or chemical alteration that may contribute to drug use in certain individuals, is unknown.
Metagenomic studies in HIV/AIDS and substance use disorder (SUD) research have mainly focused on characterizing HIV clades in various populations and understanding the role of the gut microbiome [1,2].
Conversely, there has been limited emphasis on studying viruses other than HIV and a few selected viral pathogens associated with opportunistic infections. Furthermore, accessing brain areas of interest or cerebrospinal fluid (CSF) samples for targeted viromics studies in the central nervous system (CNS) represents another major obstacle. Therefore, whether the brain virome differs in HIV-infected individuals with and without SUD is unknown.
Deep sequencing (also called Next Generation sequencing; NGS) has played an essential role in identifying viral sequences. However, due to the enormous imbalance between size and abundance of human genomic DNA and RNA and viral nucleic acids, enrichment of viral targets is necessary prior to sequencing. In previous studies, researchers have attempted to concentrate viral particles from input samples through sedimentation [3,4]. However, such methods are restricted to limited cell free specimens, such as feces and plasma, and cannot be used in solid organs. Moreover, viral particle concentration can only identify abundant and actively replicating viruses of a certain density, while species with a low viral load or those that cause latent infections, as well as those that are integrated in the human genome, will be neglected.
To perform a comprehensive evaluation of the human virome, we developed an unbiased, target-enrichment deep-sequencing platform named ViroFind [5]. Because Virofind enriches viral sequences rather than virions, it is appropriate as a general method for human viromic studies using any type of bio-specimen. In a pilot study, ViroFind was capable of enriching viral sequences up to 127-fold from human brain samples compared to deep sequencing alone [5]. This method also allowed us to identify viral sequences in heart samples of patients with myocarditis [6]. We used ViroFind to explore the virome from post-mortem brain samples of individuals with and without HIV and SUD in order to establish a better understanding of which viruses are present in the CNS of people who are affected with these conditions.
Material and methods
Clinical samples
A total of 72 frozen prefrontal cortex samples of HIV infected or uninfected individuals with or without SUD were obtained from participating National NeuroAIDS Tissue Consortium (NNTC) centers [7]: Manhattan HIV Brain Bank (MHBB) at Mt Sinai Hospital, NY (n = 39); Brain banks at University of Texas Medical Branch at Galveston (UTMB, n = 20); California NeuroAIDS Tissue Network (CNTN) at University of California, San Diego (UCSD, n= 10); National Neurological AIDS Bank (NNAB) at University of California, Los Angeles (UCLA, n = 3) in different batches, by August 2021. Samples were categorized in four different cohorts by the NNTC staff: HIV+/SUD+, HIV+/SUD-, HIV-/SUD+ and HIV-/SUD-. All samples were processed through the ViroFind pipeline. The NNTC database was accessed on 6/29/2023 to obtain data related to Tables 1 and 2. At no time did the authors have access to information that could identify individual participants during or after data collection.
IRB approval
This study was approved by Northwestern University Institutional Review Board STU#: 00211556. All samples used in this study were archival deidentified post-mortem brain samples obtained from the National Neuro-AIDS Tissue Consortium repository (NNTC; https://nntc.org/), and informed consent was not applicable.
ViroFind design
NNTC prefrontal cortex brain samples of HIV/SUD individuals were kept at -80C prior to processing. Nucleic acid extraction with QIAGEN AllPrep DNA/RNA/miRNA Universal Kit (Cat #: 80224). RNA was reverse transcribed to cDNA with NEBNext Ultra II RNA First Strand Synthesis Module (E7771L) and NEBNext Ultra II Non-Directional RNA Second Strand Synthesis Module (E6111L) from New England Biolabs. Combined genomic DNA (200ng) and cDNA (200ng) from the same sample were then processed through the ViroFind protocol which is an in-solution target-enrichment platform for virus detection and discovery. Pre-hybridization library preparation was performed with the Agilent SureSelect XT HS2 kit (G9983D). Briefly, samples were sonicated to 150-200bp fragments which were ligated to pre-PCR primer adapters on the 3’ ends followed by first round amplification with pre-PCR adapters followed by indexing adapter and post-PCR amplification indexing adapters used at 2X volume to accommodate larger starting nucleic acid concentration (400ng total). All other steps are performed following the kit protocols. NGS library preparation was followed by target enrichment which is achieved through hybridization of viral DNA fragments to biotinylated RNA probes allowing for positive selection of viral sequences. The hybridization was performed as the kit instructed with the following procedure: 95°C 5 min, 65°C 10 min, 65°C 1 min (add 13ul of RNase Blocker solution, probe and SureSelect Fast Hybridization Buffer mixture during this segment), then perform 60 cycles of 65°C 1 min, 37°C 3 sec. After hybridization cycles, the sample was kept at 65°C on hold for a short waiting time or 21°C for overnight hold. The ViroFind probe library targets 561 unique viral genomes with a mean genomic coverage of 81.04% and is comprised of 131,706 unique 125mer probes. Viral species covered include viral taxa known to infect humans and those with zoonotic potential from all Baltimore classifications. Viral species, NCBI accession number, and % of sequence covered by probes are listed in S1 Table. Positively selected viral sequences are isolated with magnetic streptavidin beads and subsequently amplified then isolated through Amplipure magnetic silica. ViroFind libraries were sequenced by NovoGene Paired-end deep sequencing for virus enriched samples (length of reads 150 bp) was performed on the Illumina HiSeq platform.
Sequencing data analysis
Paired-end deep sequencing for virus enriched samples (length of reads 150 bp) was performed on the Illumina HiSeq platform. Following demultiplexing, each sample had about 20-30 million total reads which were run through the ViroFind analysis pipeline v2.0 [8]. Raw sequencing data was quality checked and filtered for poor quality and low complexity. Reads with overall q-scores <20 and length <50 bp were removed using Skewer (‘-q 20 -I 50 -m pe -z -o’) [9]. Repetitive sequences were removed using PRINSEQ++ (-lc_entropy 60) [10]. Reads which aligned to the reference human genome (hg38, EBV removed) by BWA MEM were subsequently discarded (-M -k 50) [11]. Thereafter, we aligned the remaining reads to a database containing all viral genomes available from the NCBI database (accessed 11/16/2020) using BWA (bwa aln -l 32 -k 2; bwa samse -n 1000). Identified viral regions were matched with gene descriptions corresponding to viral references from NCBI using the BEDOPS program and in-house script [12]. PCR duplicates were marked and removed using the PICARD tool [13]. Finally, the reads from identified viruses were assembled into larger contiguous sequences using SPAdes denovo assembler (-m 10) [14]. Consensus sequence and FASTQ files were generated for reads mapping to different viruses using SAMTOOLS [15].Results for all the samples were evaluated and complex heatmaps were generated using R script as previously described [16–18].
Curation and thresholding
Data were manually curated as follows: Any viral taxa which included repetitive regions only were removed. Viral taxa which are not known to infect humans, and those with fewer than 10 reads identified were manually assessed through NCBI basic local alignment search tool (blast) to confirm the correct attribution, reads which did not blast to the identified virus were removed.
We conducted a sensitivity analysis to establish an appropriate threshold for our virome data. Initially, viruses identified with fewer than 10 reads were subject to manual confirmation to ensure accuracy. Multiple thresholding analyses were performed, ranging from 1 to 10 reads, to identify the minimum read count at which our qualitative results remained consistent. Notably, our analyses revealed no statistically significant differences between the thresholds set at 5 and 10 reads. Viral taxa with 5 or fewer reads per sample were subsequently excluded.
Statistical analysis
Summary demographic data, substance use frequency, and presence of each viral taxa are presented as number of patients (percentage/frequency). Normality was determined by Kolmogorov-Smirnov test using GraphPad Prism 9.4.1 (GP). Normally distributed variables are presented as mean (standard deviation/SD); non-normally distributed variables with median (interquartile range/IQR). Between group ages were compared by one-way ANOVA for normally distributed values with Tukey’s post hoc analysis to determine which groups were statistically dissimilar using GP. Fisher’s Exact testing for independence was performed to compare qualitative data for sex, race, ethnicity, SUD usage, and viral presence using either GP or R-Studio for R. Comparisons of non-normally distributed viral burden data was performed using Mann-Whitney U testing on GP. Differences were considered statistically significant at p-values ≤ 0.05.
Results and discussion
Samples
The frozen prefrontal cortex samples (n=72) of HIV infected or uninfected individuals with or without SUD were obtained from participating Neuro-AIDS National Tissue Consortium (NNTC) centers [7] in different batches. Samples were categorized in four different cohorts (HIV+/SUD+, HIV+/SUD-, HIV-/SUD+, and HIV-/SUD-), which were processed through the ViroFind deep sequencing library preparation and bioinformatics pipeline obtaining a median of 3.26 (1.2-8.5) million reads per sample following quality and size filtering, the overall median number of quality controlled and curated viral reads per sample was 9 (0.7-465,281) reads per million (rPM).
Study subject demographics
The demographics of the study subject populations are shown in Table 1.
Overall, the average age was 52.3 y, and HIV+ were approximately a decade younger than HIV- individuals.
There was a predominance of males compared to females (62.5 vs 37.5%) without difference between the groups. However, there was a higher frequency of blacks in the SUD+ groups, either HIV+ or HIV-.
The SUD characteristics for SUD+ subjects are shown in Table 2.
Attribution of individuals to the SUD+ group was performed by the NNTC staff. Qualitative drug use data was collated from the NNTC files and consisted of multiple urine screening and neuropsychiatric evaluations. Overall, the median number of substances used by SUD+ individuals was 1 and was higher in HIV+ compared to HIV- individuals (3 vs 1; p<0.001). The most frequently used drugs were opiates (53.1%), and only cocaine was more frequently used by HIV+ than HIV- individuals (68.8 vs 13.3%; p = 0.003).
Viral species detected in the prefrontal cortex of study subjects
We detected nucleic acid from 26 viruses belonging to 11 viral taxa as shown in heatmaps comparing the virome according to HIV infection status (Fig 1), SUD status (Fig 2), SUD status in HIV+ individuals (Fig 3), SUD status in HIV-negative individuals (Fig 4), HIV status in SUD+ individuals (Fig 5) and HIV status in SUD-negative individuals (Fig 6).
Computed heatmap showing all viral taxa identified by ViroFind in-house pipeline with purple log2 gradient scale indicating the raw number of viral reads. The frequency of each viral species for 36 HIV+ individuals (red) and 36 HIV- individuals (blue) as well as the raw mean read count on a log scale for both groups are shown.
Computed heatmap showing all viral taxa identified by ViroFind in-house pipeline with purple log2 gradient scale indicating the number of raw viral reads. The frequency of each viral species for 32 SUD+ individuals (red) and 40 SUD- individuals (blue) as well as the raw mean read count on a log scale for both groups are shown.
Computed heatmap showing all viral taxa identified by ViroFind in-house pipeline with purple log2 gradient scale indicating the number of raw viral reads. The frequency of each viral type of 16 HIV+/SUD+ individuals (red) and 20 HIV+/SUD- individuals (blue), as well as the raw mean read count on a log scale for both groups are shown.
Computed heatmap showing all viral taxa identified by ViroFind in-house pipeline with purple log2 gradient scale indicating the number of raw viral reads. The frequency of each viral species for 16 HIV-/SUD+ individuals (red) and 20 HIV-/SUD- individuals (blue) as well as the raw mean read on a log scale count for each viral species for both groups are shown.
Computed heatmap showing all viral taxa identified by ViroFind in-house pipeline with purple log2 gradient scale. Indicating the raw number of viral reads. The frequency of each viral species for 16 HIV+/SUD+ individuals (red) and 16 HIV-/SUD+ individuals (blue) as well as the raw mean read on a log scale count for each viral species for both groups are shown.
Computed heatmap showing all viral taxa identified by ViroFind in-house pipeline with purple log2 gradient scale indicating the number of raw viral reads. The frequency of each viral species for 20 HIV+/SUD- individuals (red) and 20 HIV-/SUD- individuals (blue) as well as the raw mean read on a log scale count for each viral species for both groups are shown.
We explored the prevalence of viral nucleic acids in HIV + or – individuals (Fig 1). The median number of viruses (excluding HIV) detected in the brain sample of each subject was 4 in HIV+ and 3 in HIV- individuals, which was significantly different (p<0.001). In addition, the HIV+ compared to the HIV- group had increased frequency of nucleic acids from Adenovirus (58.3 vs 13.9%; p<0.001), Epstein Barr virus (EBV) (47.2 vs 2.8%; p<0.001) and HIV (36.1 vs 0%; p<0.001). Furthermore, quantitative analyses of the overall viral burden (excluding HIV) per subject showed that it was higher in the HIV+ than in HIV- group (13.5 vs 7.0 viral reads/million [rPM]; p=0.003). Individual viruses with higher burden in HIV+ than in HIV- individuals (excluding HIV) included Adenovirus, and Torque teno virus. (Table 3).
We then explored the prevalence of viral nucleic acids in SUD+ or – individuals (Fig 2). The median number of viruses detected in the brain sample of each subject was 3 in both SUD+ and SUD- individuals, HIV inclusive, which was not significantly different. However, the SUD+ compared to the SUD- group had increased frequency of nucleic acids for Hepatitis C virus (HCV) (12.5 vs 0%; p=0.035). Furthermore, quantitative analyses of the overall viral burden per subject showed that there was no significant difference in the SUD+ than in SUD- group (8.6 vs 9.7 rPM; p<0.757). (Table 4).
We then explored the prevalence of viral nucleic acids in HIV+/SUD + or HIV+/SUD- individuals (Fig 3). There was no significant difference in the median number of viruses detected in the brain sample of each subject between the HIV+/SUD+ group and the HIV+/SUD- group, HIV inclusive. However, the HIV+/SUD+ group compared to the HIV+/SUD- group had an increased incidence of nucleic acids from HCV (25 vs 0%; p=0.010). Furthermore, quantitative analyses of the overall viral burden per subject showed no significant difference between the two groups. (Table 5).
We then qualitatively explored the prevalence of viral nucleic acids in HIV-/SUD+ and HIV-/SUD- individuals (Fig 4). There was no significant difference in the median number of viruses or of any individual virus detected in the brain sample of each subject between the HIV-/SUD+ and HIV-/SUD- groups. Furthermore, quantitative analyses of the overall viral burden or of the burden of individual viruses showed no significant differences between the two groups (Table 6).
We then explored the prevalence of viral nucleic acids in HIV+/SUD+ or HIV-/SUD+ individuals (Fig 5). There was a significantly higher median number of viruses (excluding HIV) in the HIV+/SUD+ group compared to the HIV-/SUD+ group (4 vs 3; p=0.005). In addition, the HIV+/SUD+ group had increased frequency of nucleic acids from Adenovirus (68.8 vs 0%, p<0.001) and EBV compared to the HIV-/SUD+ group (43.8 vs 6.3%; p=0.037). Furthermore, quantitative analyses of the overall viral burden per subject (excluding HIV) showed that there were no significant differences overall between the HIV+/SUD+ than in HIV-/SUD+ groups. Of all viruses detected, Torque teno virus had a significantly higher viral burden in HIV+/SUD+ than in HIV-/SUD+ individuals (Table 7).
Finally, we explored the prevalence of viral nucleic acids in HIV+/SUD- or HIV-/SUD- individuals (Fig 6). There was a significant difference in the median number of viruses (excluding HIV) detected in the brain sample of each subject in HIV+/SUD- compared to the HIV-/SUD- groups (4 vs 3, p=0.035). In addition, the HIV+/SUD- compared to the HIV-/SUD- group had increased frequency of EBV (50 vs 0%; p<0.001). Furthermore, quantitative analyses showed an increase in the overall viral burden (excluding HIV) per subject in the HIV+/SUD- compared to the HIV-/SUD- group (14.0 vs 6.0 median viral rPM; p=0.008). Adenovirus viral burden higher in the HIV-/SUD- group relative to HIV+/SUD- individuals (p=0.003), whereas HIV+/SUD- had a higher burden of Torque teno viruses (p=0.017). (Table 8).
Overall, the median % (IQR) sequence coverage of the viral genomes for all samples was 3.22% (0.84-8.07), with a range of 0.03-99.73%. Sequence coverage of > 90% are shown in S1 Fig and have been deposited in GenBank. To determine if the presence of HCV and HBV in the brain was associated with known infection by those viruses, we reviewed the available serologic and virologic data in the NNTC database. Of 7 individuals who had detectable HBV in their brain by ViroFind, only two had been tested for HBV infection during their lifetime, including 1 with positive and 1 with negative results. Conversely, of 65 individuals with undetectable HBV by ViroFind, 35 had been tested for HBV infection during their lifetime, including 3 with positive and 32 with negative results. Therefore, there was no association between ViroFind testing of prefrontal cortex samples and previously documented HBV infection (Fisher’s exact test p= 0.207).
However, of 4 individuals who had detectable HCV by ViroFind, all had been tested for HCV infection during their lifetime and found to be positive. Conversely, of 68 individuals with undetectable HCV by ViroFind, 39 had been tested for HCV infection during their lifetime including 17 who were positive and 22 who were negative. Therefore, there was an association between ViroFind prefrontal cortex results and previously documented HCV infection (Fisher’s exact test p=0.049).
Qualitative and quantitative diversity of the brain virome
Although the brain has long been considered a sterile environment, our unbiased method for virus detection reveals an unexpectedly large number of DNA and RNA viral sequences in the prefrontal cortex of individuals with and without HIV-infection or SUD. Although none of the samples came from individuals with known active viral encephalitis, we detected several viral species with known pathogenic potential in the nervous system (eg. Herpes Simplex, Cytomegalovirus, JC polyomavirus), species with known pathogenic potential outside of the nervous system (eg Adenovirus, HCV, HBV, BK polyomavirus), as well as species with no defined pathogenic potential (eg Torque teno virus, Pegivirus, adeno-associated viruses). We even detected sequences of Simian Polyomavirus 40 in a few samples, a virus thought to be restricted to monkeys, which was inoculated unwittingly to millions of people in the US in the late fifties as a contaminant of polio vaccines [19].
In addition, ViroFind provides a semi-quantitative evaluation of the viral burden, allowing comparisons between groups for the entire virome as well as individual viruses. The qualitative and quantitative diversity of viral species discovered in our study validates ViroFind as a suitable method for detection of DNA and RNA viruses of all sizes in human samples. Indeed, ViroFind can detect large viral species such as EBV with a genome of ~172 Kbp, that is potentially integrated in the human genome, as well as small and non-integrated polyomaviruses with genomes of ~10 Kbp. Conversely, studies limited to the screening of human genomic databases for viruses may be biased towards DNA viruses only, which may be lost if the sequencing library preparation is tailored for exome research. Furthermore, RNA or DNA viruses may be lost with poly-A selection that is biased for mRNA, which is commonly used in RNAseq transcriptome analysis [20].
Association of HIV-infection and the brain virome
Overall, brain samples from HIV+ individuals harbored a greater number of viral taxa than those of HIV- people, and HIV-infection was frequently associated with an increased total viral burden. It is possible that the immunosuppression associated with HIV was responsible for this increase in viral load in the brain. In addition, viruses found more frequently in HIV+ individuals included Adenovirus and EBV, which have been associated with brain diseases [21,22]. Torque teno virus (TTV) had a higher viral burden in HIV+ people. Of interest, TTV has an increased viremia in HIV+ patients [23].
Association of SUD and the brain virome
While brain samples from SUD+ people did not harbor a greater number of viral taxa compared to SUD- individuals, HCV was found more frequently present in brain tissue from SUD+ people. Although this virus has been associated with SUD [24], its presence in the brain is tantalizing. Indeed, HCV-associated neurologic disorder most frequently involves the peripheral nervous system, causing peripheral neuropathy because of proliferation of B lymphocytes and production of cryoglobulins [25]. However, central nervous system manifestations independent from hepatic encephalopathy secondary to liver cirrhosis have been described. Potential mechanisms include infection of astrocytes and microglia [26,27]. Of note, another hepatitis virus that was not expected to be found in the brain is HBV. Although HBsAg and HBV DNA have been detected in the CSF of HBV-infected patients with neuro-psychiatric manifestations, whether the virus itself or immune reaction to the virus is responsible for these symptoms remains unclear [28]. Limited available serologic and virologic data during the study subjects’ lifetime showed an association with ViroFind results for HCV but not HBV in the brain. Future studies will be necessary to understand the full range of the compartmentalization of those viruses in the body.
Combined association of HIV and SUD and the brain virome
Subgroup analyses allowed us to further explore the combined association of HIV and SUD in the brain virome. The association of HIV in SUD+ individuals was an increase in the total number of viruses and frequency of adenovirus and EBV as well as an increase in the TTV viral burden compared to HIV-/SUD+ group. In addition, the SUD phenotype in HIV+ individuals was associated with an increased frequency of HCV compared to the HIV+/SUD- group. Altogether, the variety of the brain virome, and increased frequency of some viral species and viral burden associated with HIV and SUD is enticing. However, confirmation of these findings and whether viral species detected by ViroFind are latent or replicating, and the characterization of their precise location and cellular localization in the brain parenchyma will require further investigations.
Potential role of virus infection of the brain in SUD
While we have identified brain virome associations with both HIV and SUD status in this study, it is conceivable that viruses may also be associated with the etiology of SUD. This study provides a first exploration into which viruses are present in the brain parenchyma of a matched cohort of people with and without HIV and SUD. Our data shows that many viruses can reach the CNS and be detected in the prefrontal cortex. While the cellular localization of those viruses requires further study, it is possible that they remain latent, thereby escaping detection from the immune system. It is therefore possible that recurrent reactivations in the setting of immunosuppression or inflammation may trigger an immune, metabolic, regulatory, or chemical imbalance that may contribute to drug use in certain individuals. This has been observed to occur through alterations of dopamine metabolism, as demonstrated with Japanese Encephalitis Virus [29], Borna Disease Virus [30], and HIV-Tat expression [31,32]. Activation of Extracellular Regulated Kinases (ERK) has been implicated in SUD [33] and HIV reactivation [34] and may be induced by other viral infections. Indeed, many viruses have been shown to affect the ERK pathway [35,36]. Certain viruses have also been shown to activate Phospho-Lipase C (PLC), an enzyme that increases drug seeking behavior when it is upregulated in the ventral tegmental area of the midbrain [29,37]. Alternatively, viral integration into the human genome may act as modulator of gene expression in diverse ways such as dopamine regulation, even in absence of active replication. This is the case for integration of HERV-K LTR in RASGRF2 gene, which is more frequent in IV drug users than controls [38]. The frequency and location of viral integration in the human genome will require further study. The data generated herein provide viral targets for future study to determine whether any of the identified taxa produce similar effects in CNS resident cells which might contribute to SUD.
Limitations
This study, as with all currently used viral detection methods, has limitations. Unlike bacterial species, which have conserved genetic regions for targeted 16S rRNA amplification, viral genomes are not only too diverse for a similar amplification strategy but also harbor genomes encoded on DNA and RNA. Therefore, designing whole virome identification strategies depends on NGS from both DNA and RNA samples mixed with host nucleic acids which remain far more abundant than viral nucleic acids due to the lack of conserved amplification targets on viral sequences. NGS strategies are rapidly expanding to increased sequencing density which allows a greater read depth per sample; however these new chemistries have shown increased misattribution of reads from sequencing errors and through index hopping [39–43]. Limited misattribution of reads between samples does not qualitatively affect genomic and transcriptomic analyses but can dramatically affect the observations made in metagenomics studies like whole virome sequencing. Thresholding by read number or by regions covered as well as dual-index sequencing can be used to improve the confidence in the qualitative results from previous work. We have elected to threshold on the basis of read numbers, though other strategies should continue to be developed.
We have analyzed the resulting data both qualitatively and semi-quantitatively, however we acknowledge that genome length and secondary structure, PCR efficiency, GC content, and host homology may all confound the number of reads obtained for each individual viral taxon. Consequently, establishing a viral titer from this data remains challenging. Nonetheless, our findings indicate that an increase in the prevalence of viral nucleic acids and the number of viral reads correlates with a higher starting viral load in the sample. In other words, samples with more reads likely had a greater initial viral burden compared to those with fewer reads from the same viral species. Moreover, sequencing based viromics, such as NGS, ViroFind, or viral amplification, is limited in it’s ability to define the status of infection. The presence of viral nucleic acids may be an indication of an active or latent infection but may also be caused by partial sequences/inactivated viral components contained in exosomes or maintained episomally rather than as an indication of localized infection, These and other sequencing based viromics surveys can inform viral targets for further tropism and culturing experiments and are therefore an important first step.
We examined one single brain region, the prefrontal cortex, and it is possible that the composition of the virome differs in other cortical areas and in the white matter, or in the cerebellum, brain stem and spinal cord. Of specific relevance in SUD, brain samples from the ventral tegmentum including the nucleus accumbens, were not available for study. Serologic and virologic data for HCV and HBV was not available from all study subjects, and therefore, correlations with ViroFind data for those two viruses are limited. While ViroFind may allow for detection of viral variants or potentially, novel species bearing some homology to known viruses, the genome coverage of each virus may vary depending on the viral load, from a few reads up to close to the entire genome. Nevertheless, even detection of minute amounts of any virus by ViroFind can inform targeted detection using species-specific PCR primers of various part of its genome, followed by further analysis using Sanger sequencing.
Conclusions
ViroFind allows for unbiased identification of DNA and RNA viruses in all human clinical samples. In addition to characterizing the entire human virome in various conditions and diseases, ViroFind could also be used as a surveillance tool for emergence of viral variants and, potentially, novel viruses in various populations. As the world slowly emerges from the COVID-19 pandemic, ViroFind could become a valuable tool for monitoring viral dynamics in various compartments, monitoring outbreaks, and informing vaccine development.
Supporting information
S1 Fig. Coverage plot of sequences with greater than 90% coverage.
Integrated Genome Viewer (IGV) coverage traces for viral species identified with coverage of greater than 90% of the genome length. Study ID number including group designation (G1: HIV+/SUD+; G2: HIV+/SUD-; G3: HIV-/SUD+; and G4: HIV-/SUD-), coverage breadth in percentage of reference genome covered, and depth range for nucleotide coverage are shown. A) JCV sequences aligned to reference NC_001699.1 from three Group 2 (HIV+/SUD-) subjects. B) HBV reads aligned to reference NC_003977.2 from two Group 1 (HIV+/SUD+) subjects. C) HIV reads aligned to reference NC_00182.1 from one Group 2 subject. D) BKV reads aligned to reference NC_001538.1 from one Group 2 subject. E) Sphinx 1.76 episomal DNA reads aligned to reference number LK931492.1 from one Group 1 and one Group 4 (HIV-/SUD-) subject. F) AAV-2 reads aligned to reference NC_001401.2 from a Group 3 (HIV-/SUD+) subject. Colored lines in coverage trace show nucleotide changes from the indicated reference sequence (A: Green; C: Blue; T: Red; and G: Orange).
https://doi.org/10.1371/journal.pone.0299891.s001
(TIF)
S1 Table. ViroFind probe coverage.
Table showing the NCBI accession number for each reference sequence, the viral name, total genome length, genome length which is covered by ViroFind probes, and the calculated percent of the genome which is covered by the probes.
https://doi.org/10.1371/journal.pone.0299891.s002
(XLSX)
Acknowledgments
This publication was made possible by sample and data sets obtained from Manhattan HIV Brain Bank (MHBB), Texas NeuroAIDS Research Center (TNRC) at UTMB, National Neurological AIDS Bank (NNAB), California NeuroAIDS Tissue Network (CNTN), and Data Coordinating Center (DCC).
References
- 1. Koay WLA, Siems LV, Persaud D. The microbiome and HIV persistence: implications for viral remission and cure. Curr Opin HIV AIDS. 2018;13(1):61–8. pmid:29135490
- 2. Taylor BS, Hammer SM. The challenge of HIV-1 subtype diversity. N Engl J Med. 2008;359(18):1965–6.
- 3. Li L, Deng X, Linsuwanon P, Bangsberg D, Bwana MB, Hunt P, et al. AIDS alters the commensal plasma virome. J Virol. 2013;87(19):10912–5. pmid:23903845
- 4. Handley SA, Desai C, Zhao G, Droit L, Monaco CL, Schroeder AC, et al. SIV Infection-Mediated Changes in Gastrointestinal Bacterial Microbiome and Virome Are Associated with Immunodeficiency and Prevented by Vaccination. Cell Host Microbe. 2016;19(3):323–35. pmid:26962943
- 5. Chalkias S, Gorham JM, Mazaika E, Parfenov M, Dang X, DePalma S, et al. ViroFind: A novel target-enrichment deep-sequencing platform reveals a complex JC virus population in the brain of PML patients. PLoS One. 2018;13(1):e0186945. pmid:29360822
- 6. Johnson DB, Balko JM, Compton ML, Chalkias S, Gorham J, Xu Y, et al. Fulminant Myocarditis with Combination Immune Checkpoint Blockade. N Engl J Med. 2016;375(18):1749–55. pmid:27806233
- 7. Morgello S, Gelman BB, Kozlowski PB, Vinters HV, Masliah E, Cornford M, et al. The National NeuroAIDS Tissue Consortium: a new paradigm in brain banking with an emphasis on infectious disease. Neuropathol Appl Neurobiol. 2001;27(4):326–35. pmid:11532163
- 8. Parfenov M, Seidman JG. Finding Pathogenic Nucleic Acid Sequences in Next Generation Sequencing Data. Curr Protoc Hum Genet. 2015;86:18 9 1- 9 0. pmid:26132004
- 9. Jiang H, Lei R, Ding SW, Zhu S. Skewer: a fast and accurate adapter trimmer for next-generation sequencing paired-end reads. BMC Bioinformatics. 2014;15:182. pmid:24925680
- 10. Schmieder R, Edwards R. Quality control and preprocessing of metagenomic datasets. Bioinformatics. 2011;27(6):863–4. pmid:21278185
- 11. Li H, Durbin R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics. 2010;26(5):589–95. pmid:20080505
- 12. Neph S, Kuehn MS, Reynolds AP, Haugen E, Thurman RE, Johnson AK, et al. BEDOPS: high-performance genomic feature operations. Bioinformatics. 2012;28(14):1919–20. pmid:22576172
- 13.
Picard. Available at: http://broadinstitute.github.io/picard/.
- 14. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol. 2012;19(5):455–77. pmid:22506599
- 15. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009;25(16):2078–9. pmid:19505943
- 16. Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics. 2016;32(18):2847–9. pmid:27207943
- 17. Wook Kim K, Allen DW, Briese T, Couper JJ, Barry SC, Colman PG, et al. Distinct Gut Virome Profile of Pregnant Women With Type 1 Diabetes in the ENDIA Study. Open Forum Infect Dis. 2019;6(2):ofz025. pmid:30815502
- 18. Schep AN, Kummerfeld SK. iheatmapr: interactive complex heatmaps in R. The Journal of Open Source Software. 2017; 2(16), 359, available at: https://joss.theoj.org/papers/10.21105/joss.00359.
- 19. Strickler HD, Rosenberg PS, Devesa SS, Hertel J, Fraumeni JF Jr., Goedert JJ. Contamination of poliovirus vaccines with simian virus 40 (1955-1963) and subsequent cancer rates. JAMA. 1998;279(4):292–5. pmid:9450713
- 20. Viscardi MJ, Arribere JA. Poly(a) selection introduces bias and undue noise in direct RNA-sequencing. BMC Genomics. 2022;23(1):530.
- 21. Sakrani N, Almazrouei S, Mohan S, Ramsi M. Adenovirus as a rare cause of acute necrotising encephalitis. BMJ Case Rep. 2019;12(12). pmid:31801782
- 22. Zhang N, Zuo Y, Jiang L, Peng Y, Huang X, Zuo L. Epstein-Barr Virus and Neurological Diseases. Front Mol Biosci. 2021;8:816098. pmid:35083281
- 23. Lapa D, Del Porto P, Minosse C, D’Offizi G, Antinori A, Capobianchi MR, et al. Clinical Relevance of Torque Teno Virus (TTV) in HIV/HCV Coinfected and HCV Monoinfected Patients Treated with Direct-Acting Antiviral Therapy. J Clin Med. 2021;10(10). pmid:34068071
- 24. Parikh N, Nonnemacher MR, Pirrone V, Block T, Mehta A, Wigdahl B. Substance abuse, HIV-1 and hepatitis. Curr HIV Res. 2012;10(7):557–71. pmid:22973853
- 25. Mathew S, Faheem M, Ibrahim SM, Iqbal W, Rauff B, Fatima K, et al. Hepatitis C virus and neurological damage. World J Hepatol. 2016;8(12):545–56. pmid:27134702
- 26. Faccioli J, Nardelli S, Gioia S, Riggio O, Ridola L. Neurological and psychiatric effects of hepatitis C virus infection. World J Gastroenterol. 2021;27(29):4846–61. pmid:34447230
- 27. Wilkinson J, Radkowski M, Laskus T. Hepatitis C virus neuroinvasion: identification of infected cells. J Virol. 2009;83(3):1312–9. pmid:19019968
- 28. Kappus MR, Sterling RK. Extrahepatic manifestations of acute hepatitis B virus infection. Gastroenterol Hepatol (N Y). 2013;9(2):123–6. pmid:23983659
- 29. Simanjuntak Y, Liang JJ, Lee YL, Lin YL. Japanese Encephalitis Virus Exploits Dopamine D2 Receptor-phospholipase C to Target Dopaminergic Human Neuronal Cells. Front Microbiol. 2017;8:651. pmid:28443089
- 30. Solbrig MV, Koob GF, Parsons LH, Kadota T, Horscroft N, Briese T, et al. Neurotrophic factor expression after CNS viral injury produces enhanced sensitivity to psychostimulants: potential mechanism for addiction vulnerability. J Neurosci. 2000;20(21):RC104. pmid:11050146
- 31. Gaskill PJ, Miller DR, Gamble-George J, Yano H, Khoshbouei H. HIV, Tat and dopamine transmission. Neurobiol Dis. 2017;105:51–73. pmid:28457951
- 32. Kesby JP, Najera JA, Romoli B, Fang Y, Basova L, Birmingham A, et al. HIV-1 TAT protein enhances sensitization to methamphetamine by affecting dopaminergic function. Brain Behav Immun. 2017;65:210–21. pmid:28495611
- 33. Sun WL, Quizon PM, Zhu J. Molecular Mechanism: ERK Signaling, Drug Addiction, and Behavioral Effects. Prog Mol Biol Transl Sci. 2016;137:1–40. pmid:26809997
- 34. Kim YK, Mbonye U, Hokello J, Karn J. T-cell receptor signaling enhances transcriptional elongation from latent HIV proviruses by activating P-TEFb through an ERK-dependent pathway. J Mol Biol. 2011;410(5):896–916. pmid:21763495
- 35. Bonjardim CA. Viral exploitation of the MEK/ERK pathway - A tale of vaccinia virus and other viruses. Virology. 2017;507:267–75. pmid:28526201
- 36. Hans A, Syan S, Crosio C, Sassone-Corsi P, Brahic M, Gonzalez-Dunia D. Borna disease virus persistent infection activates mitogen-activated protein kinase and blocks neuronal differentiation of PC12 cells. J Biol Chem. 2001;276(10):7258–65. pmid:11073944
- 37. Bolanos CA, Perrotti LI, Edwards S, Eisch AJ, Barrot M, Olson VG, et al. Phospholipase Cgamma in distinct regions of the ventral tegmental area differentially modulates mood-related behaviors. J Neurosci. 2003;23(20):7569–76. pmid:12930795
- 38. Karamitros T, Hurst T, Marchi E, Karamichali E, Georgopoulou U, Mentis A, et al. Human Endogenous Retrovirus-K HML-2 integration within RASGRF2 is associated with intravenous drug abuse and modulates transcription in a cell-line model. Proc Natl Acad Sci U S A. 2018;115(41):10434–9. pmid:30249655
- 39. Pyoria L, Pratas D, Toppinen M, Hedman K, Sajantila A, Perdomo MF. Unmasking the tissue-resident eukaryotic DNA virome in humans. Nucleic Acids Res. 2023;51(7):3223–39. pmid:36951096
- 40. Moustafa A, Xie C, Kirkness E, Biggs W, Wong E, Turpaz Y, et al. The blood DNA virome in 8,000 humans. PLoS Pathog. 2017;13(3):e1006292.
- 41. Guenay-Greunke Y, Bohan DA, Traugott M, Wallinger C. Handling of targeted amplicon sequencing data focusing on index hopping and demultiplexing using a nested metabarcoding approach in ecology. Sci Rep. 2021;11(1):19510. pmid:34593851
- 42. Farouni R, Djambazian H, Ferri LE, Ragoussis J, Najafabadi HS. Model-based analysis of sample index hopping reveals its widespread artifacts in multiplexed single-cell RNA-sequencing. Nat Commun. 2020;11(1):2704. pmid:32483174
- 43. Jurasz H, Pawlowski T, Perlejewski K. Contamination Issue in Viral Metagenomics: Problems, Solutions, and Clinical Perspectives. Front Microbiol. 2021;12:745076. pmid:34745046