Zika virus alters osteogenic lineage progression of human mesenchymal stromal cells

Arboviruses target bone forming osteoblasts and perturb bone remodeling via paracrine factors. We previously reported that Zika virus (ZIKV) infection of early‐stage human mesenchymal stromal cells (MSCs) inhibited the osteogenic lineage commitment of MSCs. To understand the physiological interplay between bone development and ZIKV pathogenesis, we employed a primary in vitro model to examine the biological responses of MSCs to ZIKV infection at different stages of osteogenesis. Precommitted MSCs were infected at the late stage of osteogenic stimulation (Day 7) with ZIKV (multiplicity of infection of 5). We observe that MSCs infected at the late stage of differentiation are highly susceptible to ZIKV infection similar to previous observations with early stage infected MSCs (Day 0). However, in contrast to ZIKV infection at the early stage of differentiation, infection at a later stage significantly elevates the key osteogenic markers and calcium content. Comparative RNA sequencing (RNA‐seq) of early and late stage infected MSCs reveals that ZIKV infection alters the mRNA transcriptome during osteogenic induction of MSCs (1251 genes). ZIKV infection provokes a robust antiviral response at both stages of osteogenic differentiation as reflected by the upregulation of interferon responsive genes (n > 140). ZIKV infection enhances the expression of immune‐related genes in early stage MSCs while increasing cell cycle genes in late stage MSCs. Remarkably, ZIKA infection in early stage MSCs also activates lipid metabolism‐related pathways. In conclusion, ZIKV infection has differentiation stage‐dependent effects on MSCs and this mechanistic understanding may permit the development of new therapeutic or preventative measures for bone‐related effects of ZIKV infection.


| INTRODUCTION
Zika virus (ZIKV) was first isolated in 1947 from a rhesus monkey in Uganda (Brasil et al., 2016). From the 1960s to the 1980s, ZIKV cases of human infections were rarely found across Asia and Africa (Bordi et al., 2017). ZIKV has become more of a global threat because of its emergence and spread around the globe over the last decade, especially after the first major outbreak of ZIKV occurred in 2007 in the Yap Islands of Micronesia (Lanciotti et al., 2008), followed by another large outbreak in 2013 in French Polynesia (Sikka et al., 2016).
More recently, the ZIKV spread into and beyond Brazil in 2015−2016, with the further geographic expansion of ZIKV in the United States (Marcondes & Ximenes, 2016;White et al., 2016).
The clinical presentation of acute symptomatic ZIKV infections in adults is generally mild with symptoms similar to that of other arboviruses, such as Dengue virus (DENV), Chikungunya virus (CHIKV), and Ross River virus (RRV) (Alshammari et al., 2018;Haddow et al., 2012), and includes osteoarticular complications like articular joint pain (arthralgia) (Brasil et al., 2016;Colombo et al., 2017;Wimalasiri-Yapa et al., 2020). The bone-related complaints may present in patients because ZIKV and other arboviruses can infect the cell types, such as osteoblasts and osteoclasts, which play an essential role in bone remodeling (Borgherini et al., 2008;W. Chen et al., 2014;Huang et al., 2016;Mumtaz et al., 2018). Intriguingly, microcephaly observed in ZIKV-infected pregnant women has been linked to infection of cranial neural crest cells (CNCCs), which give rise to cranial bones and influence the developing brain (Bayless et al., 2016;Del Campo et al., 2017;Chung et al., 2009). Furthermore, observations with a fetus from a ZIKV-infected pregnant woman revealed that ZIKV exhibits tropism to mesenchymal stromal cells (MSCs), which represent fibroblastic precursors of bone-forming osteoblasts (van der Eijk et al., 2016). This osteo-tropism of ZIKV suggests a direct possible link between ZIKV infection and bone-related clinical outcomes.
In our previously published study, we reported that ZIKV infected osteoprogenitor cells and affected the differentiation and mineralization (Mumtaz et al., 2018). Previously, we infected the osteoblast precursors at early stage of differentiation, and did not determine if the effect of infection is stage dependent. Differentiation of MSCs into osteoblasts is a multistep process and several key regulators of differentiation and maturation tightly regulate distinct stages with characteristic biomarkers (Stephens et al., 2011). Therefore, in this follow-up study, we determined the susceptibility of late stage MSCs, to ZIKV infection and monitored the replication kinetics and effects of ZIKV infection on osteogenic differentiation. Since we found phenotypic differences between early and late stage MSC, we performed comparative transcriptome profiling of ZIKV infected cells at different stages of osteogenic differentiation (early and late stage) to gain initial insights into regulatory pathways linked to osteoarticular complications. Our results identified key pathways associated with ZIKV infection that depend on the stage of osteogenic differentiation in MSCs.
Briefly, differentiation into calcium depositing osteoblasts was initiated on Day 3 post-seeding (Day 0-osteogenic stimulation; Figure 1) and alpha minimum essential medium containing 50 mM ascorbic acid was supplemented with 100 mM dexamethasone (dex) and 10 mM β-glycerophosphate (osteogenic medium) (Bruedigam et al., 2011). Vero F I G U R E 1 Schematic overview of the ZIKV infection regimen relative to osteogenic differentiation of human mesenchymal stromal cells (hMSCs). Bone marrow-derived hMSCs undergo osteogenic differentiation through preosteoblasts into mature, mineralizing osteoblasts. (a) Our previous data demonstrated that ZIKV infection of MSCs during early differentiation stages (Day 0) reduced their differentiation and maturation potential compared to mock-infected controls. (b) Our present studies aim to assess the susceptibility of MSCs to ZIKV infection during later stages of osteogenic differentiation (Day 7 poststimulation). ZIKV, Zika virus. cells (African green monkey kidney epithelial cells, ATCC CCL-81) were maintained to perform ZIKV replication kinetics. Cells were cultured in Dulbecco's modified Eagle's medium (Lonza) supplemented with 10% heat-inactivated fetal bovine serum (Greiner Bio-One), 2 mM L-glutamine, 1% sodium bicarbonate, 1% HEPES, 100 U/ml penicillin, and 100 µg/ml streptomycin (all from Lonza) at 37°C and 5% CO 2 in a humidified atmosphere.

| Virus
ZIKV strain Suriname ZIKVNL00013 (ZIKVAS-Sur16; EVAg no. 011V-01621) was grown in Vero cells and passage number 3 was used for the current study. Virus titers in the supernatants were determined as described previously (Mumtaz et al., 2018).

| Replication kinetics of ZIKV in late infected MSCs
ZIKV replication in late infected MSCs was determined as described previously (Mumtaz et al., 2018). Briefly, MSCs were subjected to osteogenic induction for 7 days and then infected at a multiplicity of infection (MOI) of 5 with ZIKV for 1 h at 37°C in 5% CO 2 . After incubation, the supernatant was removed and cells were cultured in an osteogenic medium. Cells were refreshed twice a week for approximately 2 weeks. Mock-infected controls were cultured in parallel. To determine the titers produced by ZIKV infection, cell supernatants were collected at different time points postinfection. The supernatant was stored at −80°C for titration. Experiments were performed in triplicate.

| Immunofluorescence assay (IFA)
Infected cells from the replication growth kinetics assay were fixed with 4% PFA at Day 4 postinfection, permeabilized with 70% ethanol, and stained for ZIKV virus with mouse monoclonal antibody anti-flavivirus group antigen (MAB 10216), clone D1-4G2-4-15 (Millipore), using an IFA as described previously (Mumtaz et al., 2018). 2.5 | ALP, mineralization, and protein assays ALP and calcium measurements were performed as described previously (Bruedigam et al., 2011;Granchi et al., 2010). Briefly, ALP activity was determined by measuring its conversion of paranitro phenyl phosphate (pNPP) (Sigma) to paranitrophenol for 10 min at 37°C and measured at 405 nm, using a Victor2 plate reader. ALP results were adjusted for protein content of the cell lysates as described before (Mumtaz et al., 2018). Calcium measurements were performed after overnight incubation of cell lysates with 0.24 M HCl at 4°C. Calcium content was determined calorimetrically using a calcium assay reagent prepared by combining 1 M ethanolamine buffer (pH 10.6) with 0.35 mM O-cresolphthalein complex one in a ratio of 1:1. All measurements were done at 595 nm using a Victor2 plate reader.

| RNA preparation
To perform host transcriptomic analysis, MSCs (Donor# 1) were subjected to osteogenic stimulation and were infected at different stages of differentiation; at Day 0 (early) and Day 7 (late) post-osteogenic stimulation with ZIKV at a MOI of 5 for 1 h at 37°C in 5% CO 2 . For both ZIKV-infected treatment groups and the corresponding mock-infected control, total RNA was extracted on Day 15 post-initial osteogenic induction of MSCs using TRIzol reagent (Thermo Fisher Scientific). RNA was purified with the miRNAeasy mini kit (Qiagen). The quantity and integrity of RNA were assessed by ultraviolet absorbance using a NanoDrop device (Thermo Fisher Scientific) and RNA integrity number (RIN score) (Agilent Technologies). High throughput RNA sequencing (RNA-seq) was carried out using cDNA samples that were indexed using TruSeq Kits. The library size distribution was examined using an Agilent Bioanalyzer DNA 1000 chip and Qubit fluorometry (Invitrogen) followed by the generation of paired-end sequencing reads on an Illumina HiSeq.

| Pathway and functional enrichment analysis
Functional enrichment and pathway analyses were carried out using the web-based tools DAVID v 6.8 (Dennis et al., 2003) and Ingenuity pathway analysis software (IPA, http://www.ingenuity.com/index. html). Gene ontology (GO) terms and pathways were considered enriched when their Benjamini−Hochberg-corrected enrichment p-value was below 0.05.

| Quantification of mRNA expression
For gene expression validation by qPCR, RNA from ZIKV infected or mock infected MSCs were used for cDNA synthesis and PCR reactions as described previously (Bruedigam et al., 2011).
Oligonucleotide primer pairs were designed to be either on exon boundaries or spanning at least one intron. Data are presented as relative mRNA levels calculated by the formula: All primer sequences used are summarized in Table 1.

| Statistical analysis
Statistical analyses of quantitative values obtained by biochemical analysis including ALP activity, calcium deposition, and qPCR validation of mRNAs, were performed using Graph Pad Prism 9 software. All results are expressed as means with a standard error of the mean. Mann−Whitney U test was used for the comparison between two groups (infected vs. uninfected). p ≤ 0.05 was considered significant.
In addition, the invariability of AXL and TYRO3 expression in both ZIKV-infected and mock-treated MSCs indicates that ZIKV does not directly control mRNA expression of proteins that support its infectious cycle via host cell entry (assuming this indeed is the entry pathway). with a developmental disorder called craniosynostosis (Katsianou et al., 2016). While ZIKV infection may increase the incidence of microcephaly by precipitating a neurological disorder (Brasil et al., 2016;Moore et al., 2017), our results suggest that ZIKV may perhaps promote premature fusion of cranial sutures by affecting CNCCs that are required for cranio-facial development (Bayless et al., 2016;Chung et al., 2009).

|
Host transcriptomic analysis reflects the opposing phenotypic changes following ZIKV infection in early versus later stage MSCs.
Using functional enrichment analysis for ZIKV-infected early stage MSCs, we mainly found an enrichment of genes linked to immunerelated responses. It might explain the underlying factors responsible for reduced differentiation and maturation in ZIKV-infected early stage MSCs, as it was previously reported that IFN-mediated immune response could affect early stages of differentiation of MSCs and lineage-commitment toward the osteoblast phenotype (Abukawa et al., 2006;Hatzfeld et al., 2007).  Angelo et al., 2021;Goede et al., 2006;Yu & Song, 2020).
Thus, intrinsic susceptibility of MSCs to ZIKV, which is functionally linked to robust expression of the AXL receptor which mediates viral entry, triggers an IFN response that may suffice to produce cranial bone abnormalities.
Several other lines of evidence support the interpretation that perturbation of immunoregulation may alter osteoblast differentiation and bone formation. For example, endogenous type-I IFN based feedback inhibition mechanisms control bone remodeling via IFN stimulated genes and their upstream immune regulators (Deng et al., 2020). Hence, any disturbance in immune regulation can perturb bone remodeling either by affecting bone formation or bone resorption. Interferonopathies along with an enhanced proinflammatory cytokine profile are evident during viral infection as has been observed upon CHIKV infection (Chirathaworn et al., 2010). CHIKV infected patients exhibit upregulation of type 1 IFN and inflammatory cytokine profiles that are associated with arthritis, imbalanced bone remodeling, and excessive bone loss (Amdekar et al., 2017;Kelvin et al., 2011;Mori et al., 2011;Rulli et al., 2007). The cyclin-dependent kinases (CDKs) had also been investigated for their effect on osteoblast differentiation (Drissi et al., 1999;Farhat et al., 2021). In general, differentiation of osteoblasts is