Varicella zoster virus differentially alters morphology and suppresses proinflammatory cytokines in primary human spinal cord and hippocampal astrocytes

The VZV Gilden strain, isolated from the zoster vesicle of a 75-year-old male, was propagated in human fetal lung fibroblasts (ATCC, Manassas, VA, USA) and cryopreserved at passage 3. Whole-genome-based genotyping identified the strain as a Clade 3 isolate (GenBank accession number MH379685).

Primary human spinal cord astrocytes (HA-sps; ScienCell, Carlsbad, CA, USA) and hippocampal astrocytes (HA-hps; ScienCell) were used at passage 3. Cell type was confirmed by immunofluorescence antibody assay (IFA) using an anti-GFAP antibody and 4′,6-diamidino-2-phenylindole (DAPI) nuclear DNA stain then quantitated by Fiji image processing software (https://​fiji.​sc/).

Cell-associated infections were used in all experiments to more accurately represent infection in vivo as described [15]. HA-sps and HA-hps were seeded at 5000 cells/cm² in basal astrocyte medium containing 2% fetal bovine serum (FBS), 1% astrocyte growth supplement, and 1% 100× penicillin-streptomycin (ScienCell). After 24 h, medium was changed to basal astrocyte medium containing 0.1% FBS and 1% 100× penicillin-streptomycin and replenished every 72 h for 7 days, establishing quiescence. Quiescent HA-sps (qHA-sps) were co-cultivated with stocks of VZV-infected HA-sps (0.008 multiplicity of infection (MOI)) or mock-infected HA-sps and analyzed at 3 days post-infection (DPI) as described [15]; parallel experiments using quiescent HA-hps (qHA-hps) were also completed at the same MOI. This same protocol was followed for culturing and infecting primary human perineurial cells (HPNCs; ScienCell), which comprise the nerve-extrafascicular barrier.

While VZV is highly cell-associated and cell-free virus has not been reproducibly obtained from any primary human cell type in vitro, conditioned supernatants from each cell type were applied to uninfected HPNCs for 3 days to test for live virus. VZV DNA was not detected by qRT-PCR in any sample (data not shown) indicating that cell-free, infectious viral particles were not present in the conditioned supernatant.

Frozen human PBMCs (Stemcell Technologies, Vancouver, BC) were thawed and cultured in RPMI (ThermoFisher, Waltham, MA) supplemented with 10% human serum (Gemini Bio-Products, West Sacramento, CA) and 1% PS (Sciencell) and immediately used for migration assays.

HA-sps, HA-hps, and HPNCs were plated in 24-well clear bottom plates (Ibidi, Martinsried, Germany) and cultured and infected as above. At 3 DPI, cells were washed once with 1× phosphate-buffered saline (PBS), fixed for 20 min in 4% paraformaldehyde, permeabilized with 0.3% Triton-X for 20 min, and blocked in 10% normal donkey serum. Cells were incubated with mouse anti-human VZV glycoprotein B (VZV gB; 1:500 dilution; Abcam, Cambridge, MA, USA) and chicken anti-GFAP (1:500 dilution; Abcam) antibodies overnight at 4 °C and probed with donkey anti-mouse Alexa Fluor 594 IgG (1:500 dilution; LifeTechnologies; Grand Island, NY, USA) and donkey anti-chicken Alexa Fluor 488 (1:500 dilution; Jackson ImmunoResearch Laboratories Inc.; West Grove, PA, USA). Cell nuclei were stained with 2 μg/mL DAPI. Cells were visualized by confocal microscopy (3I Marianas inverted spinning disk on Zeiss Axio observer Z1; Oberkochen, Germany) and analyzed using 3I Slidebook 6 software.

Conditioned supernatant from mock- and VZV-infected qHA-sps, qHA-hps, and quiescent HPNCs (qHPNCs) at 3 DPI were analyzed for levels of proinflammatory cytokines using the Meso Scale Discovery Proinflammatory Panel 1 enzyme-linked immunosorbent assay (ELISA) (Meso Scale Discovery, Rockville, MD, USA) for detection of interleukin (IL)-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, interferon (IFN)-γ, and tumor necrosis factor (TNF)-α. Cytokine concentrations were calculated by reference to a standard curve for each molecule derived using various concentrations of the standards assayed in the same manner as supernatant samples. The lower and upper limits of detection were calculated based on the concentration of signal equal to 2.5 SDs above the zero calibrator and below the upper plateau of the standards curve, respectively. All samples were analyzed in duplicate.

At 3 DPI, the ability of conditioned supernatant from VZV-infected qHA-sps, qHA-hps, and qHPNCs to recruit immune cells was determined. PBMC migration towards mock- or VZV-infected supernatant from each of the three cell lines was quantified using a chemotaxis 96-well plate with a 5-μm pore size filter (Neuroprobe, Gaithersburg, MD, USA). PBMCs were incubated with 10 μM fluorescently labeled 647-nm CellTracker (Thermo Fisher Scientific, Waltham, MA, USA) for 30 min. Sample supernatant was loaded in the wells of the 96-well chemotaxis plate according to the manufacturer’s instructions and the filter was placed on top. Labeled immune cells were pipetted onto the filter over the sample supernatant (50,000 cells/well) and incubated at 37 °C for 4 h, when the filter was removed and immune cells that migrated through the filter to the supernatant were visualized by fluorescence microscopy and quantified using Fiji 3D Objects Counter.

To determine reactivity of HA-sps and HA-hps, quiescent cells were treated with 10 ng/mL of recombinant human IL-1β (R&D Systems, Minneapolis, MN, USA) or vehicle PBS containing 0.01% bovine serum albumin then IL-6 was quantified as previously described [17]. After 24 h, supernatant was collected, flash-frozen, and analyzed for IL-6 concentrations using the Human IL-6 ELISA MAX Deluxe set as per the manufacturer’s instructions (Biolegend, San Diego, CA, USA). In addition, cells were fixed and analyzed at 24 h post-treatment by IFA for GFAP expression and morphological changes.

RNA was extracted using the Direct-zol RNA MiniPrep kit (Zymo Research, Irvine) and reverse transcribed using the Transcriptor High Fidelity cDNA Synthesis Kit (Roche, Basel, Switzerland). Complementary DNA were analyzed with qPCR primers corresponding to each cytokine (IL-1β, Hs.PT.58.1518186; IL-2, Hs.PT.58.1142676; IL-4, Hs.PT.58.46539563g; IL-6, Hs.PT.58.40226675; IL-8, Hs.PT.58.39926886g; IL-10, Hs.PT.58.2807216; IL-12p70, Hs.PT.58.1687020; IL-13, Hs.PT.58.40619905.g; IFN-γ, Hs.PT.58.3781960; and TNF-α, Hs.PT.58.45380900; Integrated DNA Technologies, Coralville, IA) and glyceraldehyde-3-phosphate-dehydrogenase (GAPdH) as previously described [15]. Data were normalized to GAPdH and analyzed using the ∆∆ threshold cycle method.

Statistical analysis was performed using GraphPad Prism (GraphPad, San Diego, CA, USA). Individual unpaired t tests were used to determine differences between IL-6 concentrations following IL-1β treatment. Differences in cytokine concentrations among mock- and VZV-infected qHA-sps, qHA-hps, and qHPNCs were determined using multiple unpaired t tests with a false discovery rate (q value ≤ 0.05) and the two-stage step-up method of Benjamini et al. [18]. Significant differences in number of migrated PBMCs were determined using one-way ANOVA with Tukey’s multiple comparisons test. Alpha was set at 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001).

GFAP was expressed in all DAPI-positive cells in mock-infected qHA-sp cultures (1208 counted) and in 99% of cells in qHA-hps (953 counted) cultures, indicating predominantly pure astrocyte cultures (Fig. 1a, e, respectively, green); in addition, qHA-sps were larger than qHA-hps (Fig. 1a, e, respectively). VZV-infected qHA-sps expressed GFAP (Fig. 1b, d, green), as well as VZV gB (Fig. 1c, d, red); note the VZV gB-positive, filipodia-like projections (arrows) sprouting from infected qHA-sps. VZV-infected qHA-hps also expressed GFAP (Fig. 1f, h) and VZV gB (Fig. 1g, h). Unlike VZV-infected qHA-sps, VZV-infected qHA-hps did not have an extensive interconnected cellular network and no VZV gB-positive, filopodia-like projections were seen. Compared to the fibroblast-like morphology of mock-infected qHA-hps (Fig. 1e), VZV-infected qHA-hps appeared swollen (Fig. 1g, h).

Enhanced contrast imaging of VZV-infected qHA-sps (Fig. 2a, b) revealed extensive morphological changes, with abundant VZV gB-positive, filipodia-like projection sprouting along cellular processes (arrows); surface plot imaging illustrated the net-like VZV gB-positive (red) projections that link neighboring cells (Fig. 2c). The same analysis of VZV-infected qHA-hps showed distinct morphological differences, with swollen, globular cells (Fig. 2d, e) compared to the fibroblast-like structure of mock-infected qHA-hps (Fig. 1e) and without the cytoplasmic extensions seen in VZV-infected qHA-sps (Fig. 2a, b). Surface plot imaging further demonstrated the island of globular VZV-infected qHA-hps (Fig. 2f).

At 24 h after treatment of qHA-sps and qHA-hps with 10 ng/mL of IL-1β, both cell types exhibited morphological changes, with cell process extensions seen on IFA using anti-GFAP antibodies, as compared to vehicle-treated cells (Fig. 3a). IL-1β-treated qHA-sps (n = 3) and qHA-hps (n = 3) had significantly increased concentrations of IL-6 (416.77 ± 69.6 versus 780.98 ± 8.48, p < 0.001; 42.21 ± 20.24 versus 761.73 ± 5.87, p < 0.001; respectively; mean pg/mL ± SD), i.e., a fold change of 1.87 ± 0.02 and 18.04 ± 0.14, respectively (mean fold change ± SD), as compared to the supernatant from vehicle-treated cells (Fig. 3b). The morphological changes in both astrocyte types, as well as increased secretion of IL-6 in response to IL-1β stimuli, provide evidence of the capability of these cells to respond to exogenous factors and increase cytokine production/release.

At 3 DPI, compared to conditioned supernatant from mock-infected qHA-sps (n = 4), supernatant from VZV-infected qHA-sps (n = 5) had significantly reduced concentrations of IL-2 (p = 0.04; q = 0.042), IL-4 (p = 0.014; q = 0.03), IL-6 (p = 0.03; q = 0.04), IL-12p70 (p = 0.016; q = 0.03), and IL-13 (p = 0.01; q = 0.03) released into the supernatant, translating to a significant fold change of VZV/mock-infected qHA-sps for IL-2 (0.74 ± 0.17), IL-4 (0.70 ± 0.1), IL-6 (0.60 ± 0.17), IL-12p70 (0.70 ± 0.1), and IL-13 (0.73 ± 0.11) (mean ± SD) (Fig. 4a, black bars); no differences in IL-1β, IL-8, IL-10, IFN-γ, or TNF-α were detected. Compared to supernatant from mock-infected qHA-hps (n = 5), VZV-infected qHA-hps (n = 4) had only significantly reduced concentrations of IL-8 (p = 0.002; q = 0.024), with a fold change of 0.49 ± 0.12; (mean ± SD) (Fig. 4a, light gray bars). RNA transcripts were not detected for any cytokines corroborating the reduced or unchanged secreted protein levels measured in both mock- or VZV-infected qHA-sps and qHA-hps.

In contrast, qHPNCs infected with the same VZV Gilden strain displayed significantly elevated levels of all secreted cytokines measured by ELISA compared to mock-infected qHPNCs (n = 3): IL-1β (p = 0.0002; q = 0.0005), IL-2 (p = 0.022; q = 0.026), IL-4 (p = 0.041; q = 0.043), IL-6 (p = 0.0002; q = 0.0005), IL-8 (p = 0.0001; q = 0.0005), IL-10 (p = 0.007; q = 0.014), IL-12p70 (p = 0.018, q = 0.025), IL-13 (p < 0.0001; q = 0.0005), IFN-γ (p = 0.016; q = 0.025), and TNF-α (p = 0.005; q = 0.01), translating to a significant fold change of VZV/mock-infected qHPNCs for IL-1β (5.31 ± 0.56), IL-2 (2.55 ± 0.74), IL-4 (2.09 ± 0.62), IL-6 (3.09 ± 0.27), IL-8 (6.11 ± 0.59), IL-10 (4.59 ± 1.21), IL-12p70 (2.1 ± 0.48), IL-13 (2.9 ± 0.18), IFN-γ (1.63 ± 0.27), and TNF-α (4.71 ± 1.17) (mean ± SD; Fig. 4a, dark gray bars). The ability of the VZV Gilden strain to induce secretion of proinflammatory cytokines in HPNCs is consistent with results using the VZV Ellen strain in a prior report [16]. Transcripts for IL-10 were not detected in either mock- or VZV-infected qHPNCs. RNA transcripts for IL-2 and IFN-γ were not detected in mock-infected qHPNCs but were detected in all VZV-infected qHPNCs (17.5 ± 1.16 and 16.78 ± 0.71, respectively; n = 3, delta CT ± SD relative to GAPDH). Compared to mock-infected qHPNCs, transcripts in VZV-infected qHPNCs showed an upward trend for IL-4 (1.56 ± 0.29), IL-8 (6.7 ± 5.42), IL-13 (19.7 ± 18.21), and TNF-α (53.68 ± 18.31) but a downward trend for IL-1β (0.19 ± 0.11) and IL-12p70 (0.05 ± 0.03) (mean fold change ± SD; n = 3). IL-6 transcripts remained unchanged (0.98 ± 0.05) (mean fold change ± SD; n = 3). The discrepancy between secreted cytokine increases in VZV-infected cells and the undetectable, unchanged, or decreased transcripts for IL-10, IL-1β, IL-12p70, and IL-6 may be due to a negative feedback loop, transcripts below the level of quantitation on qPCR, or the fact that secreted cytokines represented cumulative cytokine production over the 3 days of infection, whereas transcripts only represented a single point.

To determine the functional significance of cytokines detected, the ability of conditioned supernatant to attract PBMCs was measured in a chemotaxis assay. Consistent with the reduced or unchanged cytokine concentrations, the number of migrating PBMCs towards supernatant from either VZV-infected qHA-sps or qHA-hps did not differ from that in their respective mock groups (0.82 ± 0.41 and 0.92 ± 0.32, respectively; mean fold change ± SD; p > 0.05, n = 7) (Fig. 4b, black and light gray bars, respectively). In contrast, conditioned supernatant from VZV-infected qHPNCs, which had elevated levels of multiple proinflammatory cytokines, attracted significantly more PBMCs compared to mock-infected qHPNCs (5.24 ± 2.61; mean fold change ± SD; p < 0.001, n = 4) (Fig. 4b, dark gray bars, respectively). Chemokine (C-C motif) ligand 2 (CCL-2), used as a positive control (20 pg/mL, diluted in quiescent astrocyte media), attracted a significant number of PBMCs (4.88 ± 1.22, mean fold change ± SD relative to medium only, p < 0.01, n = 2); quiescent astrocyte medium (0.56 ± 0.19, n = 3; mean fold change ± SD) was used as a negative control (Fig. 4b).