Inherited chromosomally integrated HHV-6 demonstrates tissue-specific RNA expression in vivo that correlates with increased antibody immune response

Human herpesvirus-6A and 6B (HHV-6A, HHV-6B) are human viruses capable of chromosomal integration. Approximately 1% of the human population carry one copy of HHV-6A/B integrated into every cell in their body, referred to as inherited chromosomally integrated human herpesvirus 6A/B (iciHHV-6A/B). Whether iciHHV-6A/B is transcriptionally active in vivo and how it shapes the immunological response is still unclear. Here, we screened DNA-Seq and RNA-Seq data for 650 individuals available through the Genotype-Tissue Expression (GTEx) project and identified 2 iciHHV-6A and 4 iciHHV-6B positive candidates. When corresponding tissue-specific gene expression signatures were analyzed, low levels HHV-6A/B gene expression was found across multiple tissues, with the highest levels of gene expression in the brain (specifically for iciHHV-6A), testis, esophagus, and adrenal gland. U90 and U100 were the most highly expressed HHV-6 genes in both iciHHV-6A and iciHHV-6B individuals. To assess whether tissue-specific gene expression from iciHHV-6A/B influences the immune response, a cohort of 15,498 subjects was screened and 85 iciHHV-6A/B+ subjects were identified. Plasma samples from iciHHV-6A/B+ and age and sex matched controls were analyzed for antibodies to control antigens (CMV, EBV, FLU) or HHV-6A/B antigens. Our results indicate that iciHHV-6A/B+ subjects have significantly more antibodies against the U90 gene product (IE1) relative to non-iciHHV-6 individuals. Antibody responses against EBV and FLU antigens or HHV-6A/B gene products either not expressed or expressed at low levels, such as U47, U57 or U72, were identical between controls and iciHHV-6A/B+ subjects. CMV seropositive individuals with iciHHV-6A/B+ have more antibodies against CMV pp150, relative to CMV seropositive controls. These results argue that spontaneous gene expression from integrated HHV-6A/B leads to an increase in antigenic burden that translates into a more robust HHV-6A/B specific antibody response. Importance HHV-6A/B are human herpesviruses that have the unique property of being able to integrate into the subtelomeric regions of human chromosomes. Approximately 1% of the world’s population carries integrated HHV-6A/B genome in every cell of their body. Whether viral genes are transcriptionally active in these individuals is unclear. By taking advantage of a unique tissue-specific gene expression dataset, we show the majority of tissues from iciHHV-6 individuals do not show HHV-6 gene expression. Brain and testes showed the highest tissue-specific expression of HHV-6 genes in two separate datasets. Two HHV-6 genes, U90 (immediate early 1 protein) and U100 (glycoproteins Q1 and Q2), were found to be selectively and consistently expressed across several human tissues. Expression of U90 translates into an increase in antigen-specific antibody response in iciHHV-6A/B+ subjects relative to controls. Future studies will be needed to determine the mechanism of gene expression, the effects of these genes on human gene transcription networks and the pathophysiological impact of having increased viral protein expression in tissue in conjunction with increased antigen-specific antibody production.


78
Primary HHV-6B infection occurs in 90% of children within their first two years of life and causes 79 Roseola, also known as sixth disease, and has been strongly associated with febrile (1). HHV-80 6B reactivation has been observed in 56% of post hematopoietic stem cell transplant recipients.

81
Those with post-transplant HHV-6B reactivation have also been observed to have a higher 82 chance of human cytomegalovirus reactivation (2).

83
As with all herpesviruses, HHV-6A/B establish lifelong latency, though it is unique in that   was associated with pathogenesis and was successfully treated with antivirals. At the molecular 103 level, the left direct repeat region of the genome is fused to the subtelomeric region (8-10). At 104 the other side of the genome, the right direct repeat ends with telomeric DNA repeat extensions 105 whose lengths are often shorter than those of the other chromosomes (9). In Europe and 106 America, integration in chromosome 17p is overrepresented while overrepresentation of 107 integration in chromosome 22q is observed in Asia (11, 12). Such overrepresentation in 108 chromosome 17p and 22q are unlikely due to multiple independent integration events but likely 109 originates in ancestral integration events that were propagated over time (11,12).  RNA-seq data for various tissues within that donor. Here, we report the results of the RNAseq 121 screen, and tissue-based iciHHV-6A/B activity from two unique gene expression datasets.

122
Furthermore, we studied whether the HHV-6A/B gene expression detected was correlated with 123 antigen specific antibody responses. Our hypothesis was that iciHHV-6A/B + subjects may be 124 routinely exposed to a higher antigenic burden than iciHHV-6subjects and this would translate 125 into a more robust anti-HHV-6A/B immune response.

6
GTEx data analysis 129 GTEx genotype data was downloaded from dbgap (June 1 st , 2018) using prefetch. 650 130 DNA sequence SRA files were clipped, decompressed, and extracted using fastq-dump with the 131 following flags: -W (remove tag sequences from dataset), -I (uniquely labels paired-end reads), 132 and -split-files (splits paired-end reads into separate files). The fastq-dump output was piped to 133 bowtie2 (14) for alignment using the --no-unal (unaligned reads not saved in resulting SAM file)

171
FASTQ files were then combined and aligned to HHV-6A and HHV-6B reference genomes as 172 described above. Twenty randomly selected samples that were negative for HHV-6 DNA were 173 also screened for HHV-6 RNA. All reads were confirmed as HHV-6 using BLASTn with an  The Luciferase ImmunoPrecipitation assay System has been described previously (19,20).. In sequencing. The following constructs were previously described and generously provided by 205 Peter Burbelo (NIH): pREN2-p18 EBV, pREN2-pp150-d1 CMV and pREN2-HA2 influenza (21).

295
RNA-seq reads and were noticeably different to the high insert sizes achieved for the human 296 WGS data ( Figure 4A). We also detected reads consistent with splicing in the U90 and U100 297 genes for three of the four iciHHV-6B individuals ( Figure 4B). Unique sequence polymorphisms 298 in the HHV-6 RNA-seq data could be found that segregated the four iciHHV-6B sequences and 299 these polymorphisms matched the respective iciHHV-6B DNA sequence for that individual 300 ( Figure 4C).

302
Screening and identification of iciHHV-6 + subjects from the MHI biobank 303 DNA samples from 15498 of the MHI biobank were screen by qPCR to identify iciHHV-6A/B + 304 subjects. In total, 85 iciHHV-6A/B + individuals (40 females) were identified indicating a 305 prevalence of 0.55% (Table 1)    showed > 2 log 10 reactivity against the FLU antigen with no significant differences between the 325 groups ( Figure 6A). Similar results were observed with the EBV antigen with the exception that 326 iciHHV-6A + and iciHHV-6B + subjects had slightly higher (1.7X) antibody levels relative to control 327 subjects. Such difference did not however reach statistical significance. Results for the CMV 328 pp150 antigen indicate that the cohort contains both CMV seronegative and seropositive 329 subjects ( Figure 6C). The proportion of seropositive subjects varied between 25-30% and was 330 no different between groups. Intriguingly, among CMV seropositive subjects, iciHHV-6A + and 331 iciHHV-6B + subjects displayed >5x higher antibody reactivity against the CMV antigen than 332 14 control subjects with such a difference reaching statistical significance for the iciHHV-6A + group 333 (p<0.01).

336
The GTEx results indicate that certain HHV-6A/B genes such as U90 (IE1) and U100 (gQ) are 337 preferentially expressed relative to others such as U47 (gO), U57 (MCP) or U72 (gM). Whether 338 this would translate into a differential antibody response was examined next. Constructs

344
Sera reactivity against gO, gM and MCP were detected in most subjects with no significant 345 differences observed between iciHHV-6A + , iciHHV-6B + and control subjects (Figures 7A-C).

346
Considering that IE1 is the most divergent protein between HHV-6A and HHV-6B (62% identity), 347 antibody reactivity against both proteins were measured. The mean reactivity of sera from 348 iciHHV-6A + and iciHHV-6B + subjects against the IE1A antigen (U90) was significantly higher 349 (p<0.01) that that of sera from control subjects ( Figure 7D). Of interest, two distinctive patterns 350 of reactivity were observed in sera of iciHHV-6A + and iciHHV-6B + subjects. Approximately two-351 thirds of the iciHHV-6 + subjects had antibody reactivity against IE1A that was similar to that of 352 the control group. The remaining third iciHHV-6 + sera had antibodies levels against IE1A that 353 were >25X that of controls. Analysis of antibody response against IE1B also indicate that 354 iciHHV-6B + subjects has significantly more antibodies that control subjects or iciHHV-6A + 355 subjects ( Figure 7E). Mean antibody response against IE1B of iciHHV-6B + individuals was 13X 356 greater than that of control subjects (p<0.05).

368
We also found that coverage from WES data could prove to be a reliable metric for 369 detecting iciHHV-6A/B in WES sequence. In the GTEx dataset we first analyzed WGS data to