Stem cell culture establishment and characterisationAdipose tissue collection, cAD-MSCs extraction, and propagation

This study obtained adipose tissue samples from 12 clinically healthy dogs (Canis lupus familiaris), 11 females and one male who underwent elective surgery. The collection of adipose tissue, extraction of cAD-MSCs and propagation were performed according to previously established protocols [8, 31]. To accomplish the objectives of this investigation, we used cells from cAD-MSC donors 6/21, 9/21, 13/21, 14/21, 1/22, 2/22, 3/22, 6/22, and 7/22, which have been described in a prior publication [8]. In addition to these samples, cells from three novel donors (7/21, 8/21, and 8/22) were used following the same procedure. Nonetheless, this research provides a distinct objective, experimental framework, and conclusions by contrasting the baseline data from uninfected cAD-MSCs with new findings following CHV infection. Table 1 contains information on the age, breed, health status, adipose tissue collection site and mass of the donors. Sterility-tested cAD-MSCs for aerobic and anaerobic bacteria, fungi and mycoplasma were used for all experiments following a previously established protocol [31]. All donor cells were cryobanked in liquid nitrogen via the standard cryobanking procedure with 10% dimethyl sulfoxide (Sigma‒Aldrich, St. Louis, MO, USA, Cat. No. D2650-100ML) at passage 2 (P2) or P3 for future experiments.

Table 1 Canine adipose tissue donor informationImmunophenotyping and multipotency testing of cAD-MSCs

As previously described [8], the immunophenotyping and multipotency testing of the cAD-MSCs were performed at P3. FACSVerse (BD, Franklin Lakes, NJ, USA) flow cytometry was used to confirm the immunophenotype, while adipogenic, osteogenic and chondrogenic in vitro differentiation was performed to verify multipotency, following the criteria of the International Society for Cellular Therapy [32].

Testing of established cAD-MSCs for CHV

All donors were tested for CHV to gain insight into the possible latent infection of cAD-MSCs extracted from adipose tissue. A cryobanked batch of cells per donor at P2/P3 was first transferred at -20 °C to induce lysis of the cell membranes. After 24 h, the cell lysate was thawed at room temperature for 30 min, vortexed and subjected to nucleic acid extraction using a MagMAX CORE nucleic acid purification kit (Thermo Fisher Scientific, Waltham, MA, USA, Cat. No. A32702) on a KingFisher Flex Purification System (Thermo Fisher Scientific) according to the manufacturer’s instructions. Quantitative real-time PCR (qPCR) was applied for the detection of the CHV glycoprotein B gene according to a previously published protocol [33] using a QuantiFast Pathogen PCR + IC Kit (Qiagen, Hilden, Germany, Cat. No. 211352) on a Rotor-Gene Q (Qiagen) instrument. Beta-actin served as an endogenous control, employing the same reagents, instrument, and a previously established protocol [34]. The reaction mixture setup and thermal cycling conditions were performed as recommended by the manufacturer. The reaction mixtures’ final primer and probe concentrations were adjusted to 1,000 nmol/L for CHV-For (5’-ACAGAGTTGATTGATAGAAGAGGTATG-3’) and CHV-Rev (5’-CTGGTGTATTAAACTTTGAAGGCTTTA-3’) and 500 nmol/L for CHV-Pb (5’-6-FAM-TCTCTGGGGTCTTCATCCTTATCAAATGCG-BHQ1-3’). For beta-actin, primers were adjusted to 83.3 nmol/L for ACT2-1030-F (5’-AGCGCAAGTACTCCGTGTG-3’) and ACT-1135-R (5’-CGGACTCATCGTACTCCTGCTT-3’) and 41.7 nmol/L for ACT-1081-HEX (5’-HEX-TCGCTGTCCACCTTCCAGCAGATGT-BHQ1-3’).

Isolation and characterisation of autochthonous wild-type CHVCHV recovery from clinical specimen and virus stock production

The autochthonous wild-type CHV strain 29107 was obtained from the organs (liver, spleen, and lungs) of a 6-day-old golden retriever undergoing routine CHV diagnostics at the Croatian Veterinary Institute. The organ samples (1 × 1 cm each) were combined and homogenised with a cold mortar and pestle containing sterile sand and 10 mL of DMEM Low Glucose. The homogenate was freeze-thawed, centrifuged at 2,100 × g for 10 min, filtered using a Millex-HP syringe filter unit 0.45 µm (Merck, Darmstadt, Germany) and stored at − 80 °C. For in vitro propagation, the Madin-Darby Canine Kidney (MDCK) cell line (ATCC, Manassas, VA, USA; Cat. No. CCL-34), which is known to be susceptible to CHV [35], was used. Before inoculation, the MDCK cell line was confirmed to be CHV contamination-free. A 90% confluent MDCK (P34) monolayer in a T25 flask (Thermo Fisher Scientific) was infected with 1 mL of stock supernatant. Following two hours of adsorption at 37 °C (5% CO2, 80% humidity), 10 mL of basal medium (79% DMEM Low Glucose (Thermo Fisher Scientific, Cat. No. 31885049), 20% fetal bovine serum (FBS) (Thermo Fisher Scientific, Cat. No. 1027010), and 1% penicillin/streptomycin (Sigma‒Aldrich, Cat. No. P4333-100ML)) was added. Upon full CPE development (monitored using a Lux2 live imaging platform, Axion BioSystems, Atlanta, GA, USA) or 96 h postinfection (p.i.), the infected cell culture flask underwent a single freeze‒thaw-centrifugation cycle. The final CHV stock was generated after the third viral passage on MDCK cells in T75 flasks and stored at -80 °C.

CHV virus stock titration was conducted in triplicate using a confluent MDCK monolayer (P35) seeded in a 96-well microplate (Thermo Fisher Scientific). Eight separate tenfold dilutions of stock supernatant (100 µL per well) were added to the cells. After a two-hour adsorption period, 180 µL of the basal medium was added to the inoculum, and the plates were incubated at 37 °C with 5% CO2 and 80% humidity for 72 h. The virus titre (TCID50) was calculated using the Spearman–Kärber method.

Verification of the autochthonous wild-type CHV strain by NGS

To verify the autochthonous wild-type CHV strain 29107 and generate a whole-genome sequence, we performed next-generation sequencing (NGS). Specifically, viral DNA was extracted from 200 µL of CHV organ suspension homogenate using a DNA Blood and Tissue kit (Qiagen, Cat. No. 69506) according to the manufacturer’s instructions. Sequencing libraries were prepared using the Nextera XT DNA Library Preparation Kit (Illumina Inc., San Diego, USA, Cat. No. 15032354 and No. 15032355) with Nextera DNA UD Indexes (Illumina Inc., Cat. No. 20026934) and sequenced on a NextSeq 550 sequencer (Illumina Inc., Cat. No. SY-415-1002) loaded with a NextSeq 500/550 High Output Kit v 2.5 (300 cycles) (Illumina Inc., Cat. No. 20024908) following the manufacturer’s instructions. Library fragment size control and quantification were performed using a 2100 Bioanalyzer instrument with an Agilent High Sensitivity DNA Kit (Agilent Technologies, Santa Clara, CA, USA, Cat. No. 5067-4626) and a Qubit™ 4 Fluorometer with a Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific, Cat. No. Q32854), respectively.

The sequence reads were assembled into contigs using the Spades software v3.15 [36]. The contigs were compared to known complete CHV genome sequences in NCBI GenBank. Sequence reads were mapped to the GenBank sequence MW353136 using Bwa mem v0.7.17 [37] and SAMtools v1.19 [38] and then used to generate a consensus assembly with Ivar v1.0 [39]. MW353136 was used as a reference because it was the most thoroughly covered by sequencing reads among the complete genome sequences of CHV (taxid:170325) in GenBank. The final novel sequence scaffold of the autochthonous wild-type CHV strain was scaffolded, curated and annotated manually. PROKKA software [40] was used to generate initial functional annotation. The autochthonous wild-type CHV strain 29107 genome sequence was deposited in GenBank under accession number PP349830.

Furthermore, phylogenetic analysis of the novel autochthonous wild-type CHV complete genome sequence was constructed from complete genome alignment of a total of 23 CHV sequences (22 reference sequences from the GenBank database) using IQTree2 software [41] and substitution model HKY + F + I. An optimal substitution model was found using ModelFinder [42]. Multiple sequence alignment was prepared using MAFFT software [43]. The phylogenetic tree was visualised with Python scripting with the help of the module Toytree with UFBoot node support [44] values shown. Calculation of the similarity plot was aided by the Python module Numpy, and visualisation was performed using the Toyplot module. Recombinations were analysed by RDP5 [45].

In vitro cAD-MSCs infection with wild-type CHVCHV serial passages on cAD-MSCs

To demonstrate successful CHV infection in cultured cAD-MSCs, a cohort of six donors (9/21, 13/21, 14/21, 2/22, 3/22, and 7/22) was randomly chosen for five consecutive viral passage experiments. In contrast to prior experiments involving freshly utilized cAD-MSCs, cryopreserved cAD-MSCs at P2 or P3 were used for these specific infections. After thawing and expansion, cells from each donor were distributed as six replicates into 24-well plates (Thermo Fisher Scientific) at a density of 105 cells/well in 1 mL of basal medium and maintained at 37 °C and 5% CO2 (80% humidity) until they reached 90% confluence. Subsequently, the basal medium was removed, and three wells per donor were inoculated with CHV virus stock at a multiplicity of infection (MOI) of 0.5. Following a two-hour incubation period to allow virus adsorption, basal medium was added to the inoculum to a volume of 1 mL. The CHV infection experiments were conducted until the CPE reached 80% or for a maximum of 120 h if the CPE was minimal or absent. Upon meeting the criteria, the plates were frozen at − 80 °C. Thawed cell lysate suspensions from each donor triplicate were individually mixed, transferred to sterile 5 mL tubes (Eppendorf, Hamburg, Germany), and subjected to a second freezing cycle. After the second thaw, the cell lysate suspensions were centrifuged at 2665 × g for 10 min, and the supernatant was filtered through a 0.45 µm filter and stored in 2 mL cryovials (Cryoking, Newcastle, Australia). Subsequent passages of the CHV virus were initiated by preparing new 24-well plates as described previously, with 500 µL/well of the preceding viral passage used as inoculum for virus absorption. Five passages of the CHV virus were conducted for each cAD-MSC donor and control cell line (MDCK). The progression of CPE was documented using a microscopic camera (Axiocam ER/105/208/HD, Axio Observer D1, Zeiss, Jena, Germany) and a Lux 2 live imaging platform. Each passage included three wells of negative controls. Supernatants from each passage (CHV-infected and uninfected) were stored at -80 °C for further qPCR experiments.

Quantification of CHV DNA with qPCR

To confirm the success of the infection, i.e., the presence of viral DNA in the supernatant of infected cAD-MSCs, we quantified the number of CHV genome copies. Total DNA was extracted from 200 µL of filtered supernatant from each viral passage after two freeze–thaw cycles using a DNA Blood and Tissue Kit (Qiagen, Cat. No. 69506) according to the manufacturer’s cell extraction protocol. DNA was quantified using a Qubit 1X dsDNA High Sensitivity (HS) kit (Thermo Fisher Scientific, Cat. No. Q33230) on a Qubit 4 Fluorometer (Thermo Fisher Scientific). CHV detection was performed by qPCR as previously described.

For quantification, a triple 5-point standard curve was generated with quantitative genomic DNA from CHV strain D 004 (ATCC VR-552DQ, lot: 70054940). The following values of the standard curve were obtained: R2 = 0.99945, slope = − 3.57263, Y-intercept = 35.24627, and reaction efficiency = 91%. The limit of detection (LOD, ≥ 95% detection in 20 replicates) was 3.31 genomic copies (gc)/reaction. Theoretically, this assay provides an LOD of 6.44 × 102 gc/mL of cell lysate supernatant. The limit of quantification (LOQ, coefficient of variability ≤ 35% in 20 replicates) was set at 136.94 gc/reaction, theoretically providing a CHV LOQ of 2.66 × 104 gc/mL for the cell lysate supernatant. The results are presented as the mean ± SEM unless otherwise stated.

Gene expression profiling of CHV-infected cAD-MSCs

Gene expression analysis via RT‒qPCR array was conducted on P3 of CHV-infected and uninfected cAD-MSCs from twelve donors. Two T75 flasks (Thermo Fisher Scientific) were seeded with ≈106 cAD-MSCs per flask in basal medium and incubated until they reached 90% confluence (24–48 h). One flask was inoculated with CHV stock at an MOI of 0.5, allowing for a two-hour adsorption period, while the second flask served as a negative control. Twenty-four hours p.i., gene expression profiling was conducted following an established procedure [8]. In brief, RNA extraction was performed with an RNeasy Mini kit (Qiagen, Cat. No. 74106) following the manufacturer’s instructions. The quality of the extracted RNA was verified using an RNA QC kit (Qiagen, Cat. No. 50–727-743). Finally, gene expression profiling was conducted using the RT2 Profiler™ PCR Array for Dog Mesenchymal Stem Cells (PAFD-082ZR, Qiagen). The RNA QC and raw gene expression data from nine uninfected cAD-MSC donors (6/21, 9/21, 13/21, 14/21, 1/22, 2/22, 3/22, 6/22, and 7/22), which were previously published [8], were analysed together with new data from three additional uninfected cAD-MSC donors (7/21, 8/21, and 8/22). This way, a comparative analysis was performed with new results from all 12 CHV-infected cAD-MSC donors.

After data acquisition, the specialised RT2 Profiler PCR Array Data Analysis Software, accessible online at https://dataanalysis2.qiagen.com/pcr (accessed 19 March 2024), enabled normalisation and comprehensive analysis. The gene expression analysis results, researched gene names, symbols, and NCBI sequences are listed in Additional file 1. Statistical significance was determined via Student’s t-test applied to replicated 2^ (-Delta CT) values within both the control and treatment groups with p < 0.05. The software automatically established a fold change cut-off value of 2.0, corresponding to a log2fold change ± 1.0. Gene expression profile data were publicly deposited in the NCBI Gene Expression Omnibus database under accession number GSE267402. The data were visualised with GraphPad Prism 10.2.2.

Proteomic analysis of the CHV-infected cAD-MSCs secretome

The alterations in the proteomic composition of the secretome of cAD-MSCs were also analysed in P3 under two conditions, uninfected and CHV-infected cAD-MSCs, in six randomly selected donors (6/21, 9/21, 14/21, 1/22, 6/22, and 7/22). The cells were seeded in six replicates at 105 cells/mL density in 24-well plates (Thermo Fisher Scientific) and conditioned in the basal medium at 37 °C, 5% CO2 and 80% humidity until they reached 90% confluence. The culture medium was then aspirated, and three wells of uninfected cells were rinsed with 2 × 2 mL DMEM Low Glucose before being incubated in 2 mL of the same medium. On the other hand, three wells of CHV-infected cAD-MSCs were inoculated with a MOI 0.5 of CHV viral stock, the virus was allowed to absorb for 2 h, and then the CHV-infected cells were rinsed with 2 × 2 mL of DMEM Low Glucose and incubated in 2 mL of the same medium. Forty-eight hours later, the secretome of the cAD-MSCs was collected as previously described [8].

Following a previously published protocol [8], the samples were prepared for liquid chromatography with tandem mass spectrometry (LC–MS/MS) analysis. In brief, the secretome proteins were reduced and extracted from the culture medium. The protein concentrations were adjusted via the Bradford assay, and enzymatic digestion followed, with peptide separation conducted using the nanoLC EASY-nLC 1200 system (Thermo Fisher Scientific). The mass spectra were recorded using a Q Exactive Plus Hybrid Quadrupole-Orbitrap tandem mass spectrometer (Thermo Fisher Scientific).

Raw data analysis utilised Scaffold Quant Q + S 5.3.0, employing protein sequence data from the Canis lupus familiaris reference proteome (UniProt Proteome ID UP000805418, accessed on 30 October 2023, with a total of 20,991 entries). Scaffold Quant version 5.0.3 was utilised for subsequent analysis, employing untargeted label-free quantification and statistical analysis based on spectral counting. Statistical significance, verified via t-tests, was defined as p < 0.05, with proteins filtered to include only those with at least two identified peptide sequences. A cut-off value of 1.3 was applied, corresponding to a log2fold change ± 0.3785. The mass spectrometry proteomics data were deposited with the ProteomeXchange consortium via the PRIDE [46] partner repository with the dataset identifiers PXD052289 and https://doi.org/10.6019/PXD052289. In this study, we incorporated previously published raw proteomic data from six uninfected cAD-MSC donors (6/21, 9/21, 14/21, 1/22, 6/22, and 7/22) [8] to facilitate a comparative analysis with new secretome proteome data from six cAD-MSCs following CHV infection.

Bioinformatics analysis of the detected proteins was performed with Gene Ontology (GO) Panther 18.0 to analyse cellular components, protein classes, molecular functions, and biological processes. GO enrichment analysis was used to determine affected protein pathways using Fisher’s exact test and false discovery rate (FDR) correction, with data presented as raw p values < 0.05 and FDR < 0.05. Additionally, a protein–protein interaction network analysis was conducted using STRING (v12.0) [47], employing a high confidence interaction score of 0.700, an FDR < 0.05, and a strength score > 0.75. To elucidate protein pathways and interactions lost due to CHV infection, the proteins secreted distinctively in uninfected samples were grouped with the biologically significant downregulated proteins (uninfected group). In contrast, distinct proteins secreted in CHV-infected samples were grouped with biologically significantly upregulated proteins (CHV-infected group) to elucidate the protein pathways activated after CHV infection. The data visualisation was performed with GraphPad Prism 10.2.2.