Identification of Equid herpesvirus 2 in tissue-engineered equine tendon

Background: Incidental findings of virus-like particles were identified following electron microscopy of tissue-engineered tendon constructs (TETC) derived from equine tenocytes. We set out to determine the nature of these particles, as there are few studies which identify virus in tendons per se, and their presence could have implications for tissue-engineering using allogenic grafts. Methods: Virus particles were identified in electron microscopy of TETCs. Virion morphology was used to initially hypothesise the virus identity. Next generation sequencing was implemented to identify the virus. A pan herpesvirus PCR was used to validate the RNASeq findings using an independent platform. Histological analysis and biochemical analysis was undertaken on the TETCs. Results: Morphological features suggested the virus to be either a retrovirus or herpesvirus. Subsequent next generation sequencing mapped reads to Equid herpesvirus 2 (EHV2). Histological examination and biochemical testing for collagen content revealed no significant differences between virally affected TETCs and non-affected TETCs. An independent set of equine superficial digital flexor tendon tissue (n=10) examined using designed primers for specific EHV2 contigs identified at sequencing were negative. These data suggest that EHV is resident in some equine tendon. Conclusions: EHV2 was demonstrated in equine tenocytes for the first time; likely from in vivo infection. The presence of EHV2 could have implications to both tissue-engineering and tendinopathy.


Amendments from Version 1 Introduction
Tendons transfer force from muscle to bone, in addition to playing a key role in positioning and reducing locomotion expenditure, such as in the equine superficial digital flexor tendon (SDFT). Their structure is hierarchical 1 , with the predominating component being a collagen-dominated extracellular matrix (ECM) 2 . Sparsely distributed fibroblasts (tenocytes) produce ECM. Tenocytes are utilised in the production of tissue-engineered tendon constructs (TETCs), which are used within tendon research as a model for in vivo tendons, and have been explored as a potential therapeutic modality for tendinopathy 3 .
There are few causes of musculoskeletal disease which have been attributed to viral infection, with no previous literature citing viral causes of tendinopathy in the horse. Previous viruses implicated in tendinopathy of other species include adenovirus and reovirus in tenosynovitis of broiler hens 4 . This study in broiler hens suggests that a primary viral infection coupled with subsequent bacterial infection leads to catastrophic musculoskeletal infection and death.
Next generation sequencing (NGS) is frequently used as a research tool to characterise the transcriptome of eukaryotes, providing superior profiling to previously used methods such as microarrays. Wang et al. 5 describe NGS as a high throughput, highly sensitive method for transcriptome analysis. We have previously used it to characterise transcriptome-wide gene expression in numerous studies, including tendon ageing and disease 6 . Additionally, it has infrequently been used to identify viral isolates within tissues 7 with a high sensitivity.
In this study equine TETCs were produced to report changing protein profiles with ageing. Following electron microscopy, two samples were found to contain virus-like structures. The aim of this study was to identify these particles using NGS and investigate their functional consequences on TETCs.

Materials and methods
All reagents were supplied by Sigma unless otherwise stated.

Sample collection
Equine SDFT was collected from young (n=7; age: 5 years ± 1.095 SD) and old (n=6; age: 18.5 years ± 2.429SD) donors from the equine hospital and abattoir as a by-product of the agricultural industry. The Animal (Scientific Procedures) Act 1986, Schedule 2 does not define abattoir collection as a scientific procedure and hence ethical approval was not required. Samples collected from an equine hospital were subject to the University of Liverpool ethical approval and consent (VREC462).
Tissue-engineered tendon production Tenocytes were digested from equine SDFT using standard collagenase protocol 8 . Constructs were engineered as previously described by Kharaz et al. 9 . In brief, tenocytes were seeded at 600,000 cells per well then scored every other day until full contraction of the construct. Constructs were harvested at 28 days and either snap frozen, collected for standard histology or transmission electron microscopy (TEM).
Transmission electron microscopy TEM of tendon constructs was performed following fixation in 2.5% glutaraldehyde in 0.1M sodium cacodylate buffer for 8 hours, followed by buffer washing procedures and second fixation and contrast stain with 0.1% osmium tetroxide for 90 minutes. Samples were stained with 8% uranyl acetate in 0.69% maleic acid for 90 minutes, dehydrated in ascending ethanol concentrations and embedded in epoxy resin. Ultrathin cross-sections (60-90 nm) were cut with a Reichert-Jung Ultracut on an ultramicrotome (Leica Microsystems, Wetzlar, Germany) using a diamond knife. Sections were then mounted on 200 mesh copper grids and stained with 'Reynold's Lead citrate' stain for 4 minutes. Images were viewed in Philips EM208S Transmission Electron Microscope (Philips UK Limited, Guildford, UK) at 80K.

Virion investigation
On examination of TEM images, two TETCs were identified as containing virus-like particles (V: Y1, Y6). These particles were morphologically assessed for virion identification. Particle diameter was estimated from electron micrograph images. Particle morphological characteristics were observed for virion diameter, presence of a viral envelope, nucleocapsid shape and surface projection presence. The number of full capsids, nucleocapsids and empty capsids were counted in 22-26 TEM images from each of the infected donors. For each sample, resin sections were mounted onto copper grids and viewed in the TEM at X 44000 magnification. 25-30 successive grid squares were viewed and the first area in each grid square which was found to contain virus was photographed. The images were loaded into ImageJ (version 1.51n) 10 and counts made of virus with: i) envelope + nucleocapsid + DNA (full) ii) envelope + nucleocapsid -DNA (nucleocapsid only), iii) envelope only (empty).
TEM images were compared to current literature in order to provide a morphological reference (Supplementary File 1).

Histological analysis
TETCs were fixed in 4% paraformaldehyde and paraffin embedded. Sections were cut at 4µm onto polylysine slides and subsequently stained with haematoxylin and eosin (H&E) and Masson's Trichrome. Histology was assessed using a scoring system developed in an unpublished report by Charters 11 , and shown in Supplementary File 2.

NGS analysis
Two samples of TETCs were submitted for RNASeq, one virally affected (V: Y1) and one control (NV). RNA was extracted from constructs as previously described, and NGS was performed in accordance with the method described by Peffers et al. 6 . Analysis was undertaken by the Centre for Genomic Research, University of Liverpool. One µg of total RNA was ribosome depleted with the RiboZero Magnetic kit (Illumina, San Diego, California, United States). NGS libraries were prepared using the ScriptSeq v2 NGS Library Preparation Kit (Illumina, San Diego, California, United States). All of the enriched material was used as input material and following 15 cycles of amplification, libraries were purified using AMPure XP beads. Each library was quantified using Qubit and the size distribution assessed using the Agilent 2100 Bioanalyser (Agilent, Santa Clara, CA, USA). Libraries were pooled in equimolar amounts and quantity and quality of each pool assessed by using Bioanalyzer (Agilent, Santa Clara, CA, USA) and subsequently by qPCR using the Illumina Library Quantification Kit from Kapa (KK4854) on a Roche Light Cycler LC480II according to manufacturer's instructions. The template DNA was loaded at 300 pM. The sequencing was carried out on one lane of an Illumina HiSeq4000 at 2×150 bp paired-end sequencing with v1 chemistry.
Initial processing and quality assessment was undertaken as previously described 12 . Coverage plots were created using R. Initially a reference sequence file was created using known herpesvirus genomes. Reads were then aligned to the reference sequence files using Bowtie2 global alignment. PCR and optical duplicate reads were subsequently removed using Picard. Resulting alignment files were then used as input to Bedtools, (version 2.16.2) which was used to calculate coverage across the genomes. The coverage data output of this was used as input for a custom R script. Once the reads were mapped the number of reads mapped to each transcript was undertaken. These counts were calculated using HTSeq-count (version 0.6.1p1) 13 , accepting only hits of quality 10+, and excluding ambiguous hits and hits on the opposite strand.  15 ) search of the assembled transcripts against the NT database from NCBI was carried out and full taxa information (species, order, phylum, superkingdom) was derived based on the best hit, using the NCBI taxonomy database (ftp://ftp.ncbi.nlm.nih.gov/ pub/taxonomy/taxdump.tar.gz). The search was carried out with an e-value cut off of 1e-5. The results were filtered to identify all hits to the virus superkingdom.
All raw read data produced in this study has been submitted to the EBI ENA, primary accession PRJEB20552, secondary accession ERP022713.

Hydroxyproline assay
Freeze-dried samples (V (n=2) and NV (n=5) of TETC derived from the young donors were hydrolysed overnight at 60°C using papain reagent. Digests were stored at -20°C and subsequently assayed for collagen content by hydroxyproline assay 16 .

Statistical analysis
Data for the hydroxyproline assay were normality tested prior to statistical analysis. Data was considered statistically significant at P≤0.05. All normality testing and statistical analysis was undertaken using Excel (2010, Microsoft, Redmond, WA, USA) and GraphPad prism (2016, version 7, GraphPad Software Inc, CA, USA).

Pan herpesvirus PCR
As RNAseq is not definitive, we used a pan-herpes PCR to confirm that the virus in question was indeed a herpes virus and there were no other herpesviruses present.
In order to validate the RNASeq findings with a different platform, a pan-herpesvirus PCR was undertaken. Samples V; Y1 and Y6 (virally affected) and a negative control (O3) were tested using a pan-herpesvirus PCR using a modified methodology described by Ehlers et al. 17 . This protocol uses dI-substituted primers that offer improved sensitivity and specificity than previous protocols 18 .
Modifications of the components included using 12.5ul of 2x Qiagen multiplex PCR master mix containing HotStar Taq DNA polymerase (Qiagen, Crawley, UK). The list of primers used in pan herpesvirus PCR is available in Supplementary File 3.
Specific EHV-2 PCR assay of equine SDFT In order to determine the presence of EHV-2 in SDFT of a larger equine population, PCR primers were designed to amplify a region within a 37kb contig demonstrating an excellent depth of coverage (mean 4919x coverage) and 99% homology to EHV-2. The primer sequence was; V1 forward GGGCGGAGAATGTAGAGACG, V1 reverse GGTGGACTTTAACGGGGAGG (product size 443). DNA was extracted from 10 (mean ±SD age: 12.8±7.1) grossly normal equine SDFT collected from an abattoir. A QIAmp DNA extraction kit (Qiagen, Crawley, UK) following the manufacturer's protocol was used. A sample of EHV-2 positive DNA (kindly gifted by Microbiology Diagnostic Laboratory, Institute of Veterinary Science, Liverpool) was used as a positive control.
PCR was performed in a final volume of 50ul consisting of 12.5ul ThermoPrime 2x ReddyMix PCR Master Mix, 1.5ul forward primer, 1.5ul reverse primer, 24.5ul water and 10ul extracted DNA template. Following an initial denaturation at 94°C for 9 min, products were amplified by 5 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 1 min and elongation at 72°C for 1 minute. This was immediately followed by 30 cycles of denaturation at 94°C for 1 min, annealing at 55°C for one min and elongation at 72°C for one minute. Amplification was followed by a final extension at 72°C for 7 min. 15µl of product was electrophoresed on a 1.5% 1× Tris-acetate-EDTA agarose gel containing 10ul of PeqGreen. TrackIT 1Kb Plus DNA ladder was used as the standard.

Phylogeny
All phylogenetic analysis was carried out within MEGA (version 5). DNA sequences were aligned using Clustal W (EMBL-EBI, Hinxton, UK). Neighbour-joining trees constructed. Bootstrap analysis (1,000 replicates) was used to provide support for individual nodes.

Results
Transmission electron microscopy Two out of 13 samples contained virus-like particles. The virus-like particles identified were approximately 100-150nm in diameter, and had an indistinct outer border with surface protrusions and a cylindrical nucleocapsid ( Figure 1A and 1B). Viral particles were identified both intracellularly (T) (including intranuclearly) and within the extracellular matrix ( Figure 1A). A further characteristic present in several of the images observed were 'empty capsids' within cells ( Figure 1C and 1D). Viral families have distinct morphological characteristics relating to size, nucleocapsid shape, and envelope present, which were used to tentatively identity the particles. The virus particles identified within the TETCs exhibited a spherical shape, with a visible envelope, contained within an icosahedral capsid shape. Virion diameter was calculated and was suggestive of a retrovirus or herpesvirus. The apparent spiked edge of the viral particles was suggestive of a retrovirus. In order to tentatively assess the level of infection within the TETCs, the number of tenocytes affected with virus particles was counted by examining the TEM images of all the donors. Whilst no virus particles were evident in images of the ten non-infected donors, 65.4% and 100% of the tenocytes of the two infected samples contained virus particles. The number of empty, full and nucleocapsids was determined for a set of images from both of the virus-infected samples. Results are shown in Table 1.

Histological scoring
There were no significant differences in histological score between V and NV based upon the characteristics measured ( Figure 2).

Hydroxyproline assay
Contrasting V and NV TETCs showed that there was no significant difference in collagen content ( Figure 3).

NGS analysis NGS analysis: Transcript assembly and removal of host genomic reads.
Illumina sequence reads were adapter-and quality-trimmed using Cutadapt version 1.2.1 (Martin, 2011) and Sickle version 1.200. A summary of raw and trimmed sequence data is shown in Supplementary File 4, following host sequence removal, and subsequent filtering of mapped reads. Resulting paired-end reads were assembled using Trinity version r2013_08_14 19 for both samples. V resulted in 913,443 transcripts (including isoforms) and NV resulted in 788,076 transcripts (including isoforms).    (Figure 4a) and V (Y6) (Figure 4b), it is clear that the virus isolated from TETC Y6 is likely EHV2. There does not appear to be any EHV2 in the NV sample. Occasional coverage spikes were due to mapping low complexity reads which would map well against many genomes.
The coverage plot shows that the genomic coverage of the TETC virus of EHV2 strain G9/92 was 86.26%. The virus was mapped to 3,875.605 contigs to EHV-2, compared with 2-339 from the rest of the herpesvirus panel (EHV1, 3, 4, 5).

Pan herpesvirus PCR.
Pan herpesvirus PCR was found to be positive for EHV2 for the two V samples and negative for NV samples ( Figure 5). Sequencing of PCR products was used to confirm EHV2 presence within the sample.

Specific PCR.
Of the ten samples assessed, none were found to contain EHV2 (Figure 6).

Phylogenic analysis.
A phylogenetic tree was produced in order to characterise the relationship between the TETC virus and currently identified EHVs (Figure 7). Phylogenetic analysis of the TETC isolated virus glycoprotein B gene shows branching with Equid herpesvirus 2 strain 275, with further close relationships with strains G9/92, 86/87 and 86. The phylogenetic tree produced clearly demonstrates that the strain isolated from the TETCs is an EHV2, although not a strain which currently has been genome sequenced in the NCBI database.

Discussion
This study is the first to identify EHV2 within cells derived from equine tendons. Findings of virus particles in samples of TETCs in which TEM was undertaken were incidental. At the time in our laboratory, equine tenocytes were the only cell type being cultured. We were interested in the nature of the particles and thus set out to investigate further. As this was a post-experiment attendant result identified a number of weeks after the end of the study, we were unable to undertake culture of the virus-infected tenocytes or determine what the titre of the virus was and if this was infectious.
We realise this is a limitation of our work. However, as the presence of EHV2 was previously unreported in tendons, and due to the potential use of TETCs as allogenic transplant agents, we believe the finding of EHV2 in tenocytes could have implications for future tissue-engineering studies. Additionally, it should serve as a warning that viral contamination of musculoskeletal tissues may not be appreciated if screening is not undertaken. Whilst all culturing was undertaken in sterile conditions the identified virions could have originated from a culture contaminant. Likely causes include cross-contamination between TETCs and other tissue within the laboratory or use of contaminated reagents during production. Previous studies have found viral contamination in commercially available foetal calf serum such as that used in TETC production here. However, as the virus was limited to a subset of cultured TETC this is unlikely. Viral isolates include BVDV, bovine parainfluenza virus-3 and bovine herpesvirus-1 (BoHV1) 31,32 . Membrane filtration removes many contaminants, however these may be an ineffective against viruses due to their small size 33 .
It is important to consider host-range and tissue specificity of viruses, which could be present as a contaminant of tissue culture reagents. Foetal calf serum was used in the tissue-engineered culture system. As a result, we compared the viral reads obtained to a variety of bovine virus' with suggestive morphology. One read to BoHV1 was identified in virally affected TETC. As an alphaherpesvirus BoHV1 has previously shown limited scope for crossing species-barriers 34 . The single read to BoHV1 when compared with EHV2 (129) suggest that BoHV-1 was not the virus isolated here. Moreover, the BoHV1 read could be a consequence of cross mapping. One read was found to EIA. This is a notifiable disease within the UK. The single read was of only 52bp hence it was suspected due to cross-mapping and further investigation was not pursued.
One additional potential explanation for the presence of EHV-2 in the samples is the ubiquitous nature of EHV-2 in both foals to adult horses 35 . One might predict that it is not surprising to identify EHV-2 in tendon but perhaps more surprising that it was identified in only two of 13 samples. A number of studies have described equine gammaherpesviruses such as EHV-2 in peripheral blood mononuclear cells (PBMCs) 36 . EHV-2 is latent in B lymphocytes and direct cell to cell contact between epithelial cells and B lymphocytes is required for the production of infection 37 . In a study of Icelandic horses of EHV-2 in PBMCs and a selection of tissues using a co-cultivation technique 80% of horses examined demonstrated infectious replicating virus 38 . These studies suggest that EHV-2 or cells containing the viral genes could have been present in a high proportion of tendon taken. Whilst every effort was made to remove tendon from the lower limbs during dissection in a aseptic manner cross contamination by other infected or latently-infected cells cannot be ruled out. Thus it is possible the tendon derived EHV-2 came from cell to cell contact of either PBMC or other donor cells and that this was the route of infection. However of the two infected samples (both from young donors) one was from the abattoir collection and one from the equine hospital. It would have been expected that more than one abattoir sample would be affected if this was the case as these were harvested, handled and processed at the same time.
Whilst the PCR for EHV-2 on tendon did not find any viral DNA this test cannot confirm lack of presence of latency associated transcript RNA. We did not identify the presence of latency associated transcripts in the RNAseq data. When the two sequences in NCBI that result from a search for 'latency associated transcript and herpes' were used neither resulted in any significant hits. Thus whilst latency associated transcripts cannot be ruled out with data currently available this is unlikely. However in order to confirm this a more in-depth analysis than this would require further research in latency associated transcripts linked to herpes in other animals which is beyond the scope of this study.
Unfortunately the EHV status of the donor horses was unknown.
The clinical history for one of the TETC donors containing virus was available. The donor was presented at the hospital with right hind lameness and proprioceptive deficits. The horse was euthanased on humane grounds. EHV myeloencephalopathy was not suspected on presentation due to the absence of cauda equina signs 39 . Moreover, the equid herpesvirus most associated with myeloencephalopathy is EHV1 40 . NGS data in this study revealed minimal coverage of the EHV1 genome when compared with EHV-2. There was no clinical history of the TETC donor as this was derived from an abattoir and hence further conclusions cannot be drawn. Since a small population of other horses were assessed for the presence of EHV2 RJW 248419 it would appear that the presence of this virus in equine tendon is uncommon.
Finally, an interesting point to consider is the potential role of an equid herpes virus in tendinopathy. EHV2 has not previously been isolated in cases of musculoskeletal disease in the horse. Whether this is because it has not been investigated or because it has no significant role in disease has not been established. Due to the unknown status of horse from the abattoir it is difficult to draw conclusions from the single case in which hind limb pathology was described. Whilst histological examination and limited biochemical tests within this study implies that EHV2 has no significant affect upon the structure of TETCs the mechanical properties of the TETCs were not assessed. Further work is required to determine if EHV-2 has a role in tendinopathy.

Conclusions
This study identifies EHV2 in equine tendons for the first time, and describes NGS as a useful tool for virus identification. The implications of the presence of EHV2 in tendon to both tissueengineering and tendinopathy requires further work. However, there are potential implications for the use of TETCs as allogenic transplant agents, as the presence of virus could result in transplant failure.

Data availability
All raw read data produced in this study has been submitted to the EBI ENA, primary accession PRJEB20552, secondary accession ERP022713. The data underlying this work has been uploaded to the Open Science Framework Database, and can be accessed via DOI, 10.17605/OSF.IO/WYPKQ 41 .

5.
"This study suggests that a primary viral infection coupled with subsequent bacterial infection leads to catastrophic musculoskeletal infection and death" Please clarify that this is the study on broiler hens and not the current work.
It is not clear from the current work what statistical tests were carried out in the programmes described in the materials and methods.
Were any of the RNA seq hits for the latency associated transcript RNA?
The specific PCR for EHV2 on tendon tissue did not find any viral DNA however can this test confirm lack of presence of latency associated transcript RNA? Was there any RNA from the virus infected tenocytes left over for analysis? Did the authors collect these tenocytes from tissue that was known to be septic antemortem? I wonder if this statement is a jump too far?
Is the work clearly and accurately presented and does it cite the current literature? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Partly
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Partly
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Author Response 13 Oct 2017 , University of Liverpool

Reviewer
This study suggests that a primary viral infection coupled with subsequent bacterial infection leads to catastrophic musculoskeletal infection and death" Please clarify that this is the study on broiler hens and not the current work.

Response
We have amended the manuscript to clarify that this was indeed in broiler hens.

Reviewer
It is not clear from the current work what statistical tests were carried out in the programmes described in the materials and methods.

Response
We have amended the manuscript to clarify that these statistical tests relate to the hydroxyproline assay.

Reviewer
Was there any RNA from the virus infected tenocytes left over for analysis?

Response
All the RNA samples were used in the studies pertaining to this manuscript.

Reviewer
Did the authors collect these tenocytes from tissue that was known to be septic antemortem? I wonder if this statement is a jump too far.

Response
All samples were grossly normal on collection. Hospital derived samples showed no signs of sepsis.

Reviewer
Were any of the RNA seq hits for the latency associated transcript RNA?
Response Two different types of mapping were undertaken. Firstly a megablast run against the entire nt database to get taxonomic information, and secondly a blastx run against a database created of all know ERV genes in animals from NCBI. Therefore the' latency associated transcripts' (only appear to be sequences for these in humans in NCBI) would not have been included. We assessed the two sequences in NCBI that result from a search for 'latency associated transcript and herpes' and ran a blastx for both samples against these. Neither resulted in any significant hits. Thus whilst latency associated transcripts cannot be ruled out with data currently available this is unlikely. A more in-depth analysis than this would require further research in latency associated transcripts linked to herpes in other animals and is beyond the scope of this paper. We have amended the manuscript accordingly.

Reviewer
The specific PCR for EHV-2 on tendon tissue did not find any viral DNA however can this test confirm lack of presence of latency associated transcript RNA?

Response
The reviewer is correct in their assumption. We have amended the manuscript.
No competing interests were disclosed. Competing Interests: 29 August 2017 Referee Report