Figure 1.
VZV infection of primary keratinocytes.
Primary human keratinocyte were culture in a 24-well plate either in a low [0.6 mM] (−calcium) or high [1.2 mM] calcium (+calcium) containing media and infected with VZV at an mo.i. of 0.2. For the final condition, cells were infected with VZV as above and then at day 3 p.i. the calcium concentration was increased to 1.2 mM (calcium added Day 3). Cells were cultured for up to 5 days p.i. and fixed in 4% PFA. VZV infected cells were identified using IHC and images of the VZV colonies were captured and counted. (A) Representative VZV colonies at day 5 post infection in triplicate from keratinocytes grown in the low, high or switched at day 3 calcium media. (B) Representative images of VZV infected cells at day 5 p.i. grown in the three different conditions (scale bar = 500 µm). (C) The VZV foci number in primary keratinocytes was counted up to 5 days post infection and the result represented as ± standard deviation (n = 3). (D) VZV infected primary keratinocytes treated with trypsin and titred onto MeWo cells at days 4 and 5 p.i. VZV colonies were identified by IHC and the ffu/ml calculated and represented as ± standard deviation (n = 3) the statistical difference between the conditions was determines (p values less than 0.05 (*) are indicated. E) VZV infected primary human keratinocytes were processed for total RNA extraction following the calcium switch at day 3 post infection. The levels of VZV ORF29 cDNA were determined by qPCR and normalised to the housekeeping gene RN5S, experiment was carried out in triplicate, p-values of less than 0.05 (*) by Student's t-test are shown. F) Representation of conditions used in RNA-seq experiment, keratinocytes were infected with cell-free VZV (m.o.i of 0.2) at day 2 and either maintained in a low calcium media or changed to a high calcium media at day 5 (day 3 p.i.), RNA was harvested at day 7. Four conditions were analysed K = Keratinocytes; KV = Keratinocytes infected with VZV; KC = Keratinocytes+calcium induced differentiation; KCV = Keratinocytes infected with VZV and calcium differentiated.
Figure 2.
Overview of transcription data.
A) Summary of the number of significantly up and down regulated genes (FDR<0.01) observed across all six possible comparisons of the four sample types e.g. 463 genes are up regulated in KC/K whilst 53 genes are down regulated in KV/K. B) Venn diagram indicating the number of significant (FDR<0.01) differentially expressed genes across four key comparisons (KCV/KC, KC/K, KV/K and KCV/KC) and the overlap between each set of genes. C–F) Detailed analysis between the four comparisons illustrating the overlap between up and down regulated gene lists. Pairs of arrows in the intersection refer to the direction of fold change in the comparisons on the left and right hand sides respectively. Comparisons are shown for: C) KCV/KV and KCV/KV, D) KC/K and KCV/KC, E) KV/K and KCV/KC and F) KC/K and KCV/KV genes. For these four Venn diagrams (e.g comparison X vs. comparison Y), those genes that are up or down-regulated in X but not significantly altered in Y are shown on the left-hand side with single arrows denoting the direction of fold change. The converse is shown on the right-hand side (i.e. genes that are differentially expressed in Y but not in X), again with up or down arrows denoting the direction of fold change. The overlaps themselves show the 4 possible options: up in both X and Y, down in both X and Y, up in X and down in Y, or down in X and up in Y. These are denoted by pairs of arrows, with the left-hand arrow referring to the direction of fold change in X and the right-hand arrow denoting the direction of fold change in Y e.g. in KC/K vs. KCV/KC (D), 4 genes are up regulated in both comparisons, 2 are down regulated in both comparisons, 73 are up regulated in KC/K but down regulated in KCV/KC whilst 159 are down regulated in KC/K but up regulated in KCV/KC.
Figure 3.
Representative functional groups enriched in differentially expressed gene lists.
The genes which were differentially expressed in the (A) KC/K (B) KCV/KC and (C) KV/K comparisons were analysed using the functional annotation tool in DAVID. Enriched functional groups (Benjamini-Hochberg adjusted P-value<0.05) were identified for each comparison and groups representative of overall enrichment results for each comparison are shown.
Figure 4.
VZV infection dysregulates the expression of epidermal cytokeratins.
A) Average fold change for KCV/KC and KV/K comparisons for epithelial cytokeratins from the RNA-seq data, showing upregulation of basal (KRT15) and mucosal (KRT4/13) and downregulation of the suprabasal cytokeratins (KRT 1/10). VZV infected and mock-infected primary human keratinocytes were processed for total RNA and protein extraction. At day 5, 48 hrs after the addition of calcium, the levels of epidermal cytokeratins were determined by qPCR and normalised to the housekeeping gene RN5S (B). The fold change between the uninfected and VZV infected samples with and without the addition of calcium was determined and plotted ± stdev. A representative graph from 4 individual experiments is shown and the dataset is comparable to the RNA-seq analysis for all the KRT genes tested. The VZV inoculum was UV irradiated prior to infection and the qPCR was repeated for (C) KRT1 and (D) KRT10. Fold change was calculated for the VZV and UV-VZV relative to the mock control for both genes, p-values<0.05 are shown (*). VZV infection again downregulated both KRT1 and KRT10, but the downregulation was not observed in the UV-treated VZV for KRT1 and was partially restored for KRT10. Quantification of VZV ORF68 was used to confirm absence of viral transcripts in UV-VZV treated cDNA (data not shown). E) Protein extracts were analysed by immunoblotting for KRT10 and −15 at 24 and 48 hrs after the calcium switch. The VZV infected cells are denoted by the presence of the late viral protein gE and GAPDH was used as a loading control. Change in density of KRT10 (F) and KRT15 (H) expression after the addition of calcium relative to the no calcium control were calculated using imageJ from figure 4E after VZV infection (D) KRT1, (E) KRT10 and (F) KRT15. The change in relative density of KRT10 (G) and KRT15 (I) by VZV infection was calculated against the mock infected control at each timepoint. VZV infection reduced KRT10 expression and increased KRT15 expression after 48 hrs regardless of the addition of calcium. J–K) Immunofluorescent staining of calcium treated primary keratinocytes from uninfected and VZV samples. KRT10 (red) was downregulated and KRT15 (red) was upregulated in the presence of VZV expression as reported by GFP-ORF23 (green). DAPI is shown in blue. Scale bar 50 µm. The data are representative of three individual experiments.
Figure 5.
Confirming the downregulation of KRT10 by VZV in a keratinocyte cell line.
nTERTs were infected with an m.o.i of 0.2. and processed for RNA and protein extraction. A) KRT10 gene expression is reduced by VZV infection up to 72 hrs p.i. as measured by real time PCR and normalised to RN5S. KRT10 expression increases in the mock infected cells as they become more confluent. Experiment carried out in quadruplicates and p-values calculated by Student's t-test p<0.05 (*). B) KRT10 protein levels are reduced at 24 and 48 hrs in VZV infected cells which are shown by the presence of the ORF63 protein, GAPDH was used as a loading control. C) Pre-treatment of cells with PAA, which inhibits VZV DNA polymerase and viral replication prevents the reduction of KRT10 expression measured by western blot. Immunofluorescence showing downregulation of KRT10 in VZV infected nTERTs, KRT10 staining (red) is abundant in the no virus control (D) whereas in cells positive for IE62 (E) or gE (F) (both green), KRT10 protein expression is absent, DAPI is shown in blue, scale bar = 50 µm. G–H) Higher magnification of (E), show that early in VZV infection, when IE62 expression (green and indicated by white arrows) is confined to the nuclei, KRT10 expression is still present in the VZV infected cells, scale bar = 10 µm.
Figure 6.
VZV downregulated KRT1/10 expression in the epidermis.
A) H&E of a cross-section of VZV infected primary organotypic rafts to show swollen cells within the intact epidermis, typical of early VZV lesion formation which are not seen in the mock infected control, scale bar 200 µm. KRT10 (red) expression is suprabasal in the mock control (iii) but reduced in VZV infected pocket (vi) as indicated by VZV gE (green). Panels vii and viii represent the merged images with DAPI stained nuclei. Scale bars 25 µm. (B–G) Immunofluorescent staining of human varicella skin biopsies, showing KRT10 (red, top panels) and KRT1 (red, bottom panels) expression in the spinous layer. Dashed line represents the dermis/epidermis junction. The loss of both KRTs are seen in VZV positive areas (green), scale bars 50 µm. D and G represent the mean fluorescent intensity (MFI) of KRT10 (D) and KRT1 (G) from infected and uninfected cells from the spinous layer. Ten infected (green) and ten uninfected (not green) cells were taken from the spinous layer and the intensity of the KRT10 fluorescence (red) measured from each using imageJ. The mean intensity of KRT10 was then measured ± standard deviation and the associated p values calculated by a Student's t-test.
Figure 7.
VZV infection affects the integrity of the epidermis by disrupting the desmosomal junctions.
From the RNA-seq dataset the average fold change (KCV/KC) show that genes associated with the desmosomes (A) were significantly downregulated (*p<0.01) in differentiated keratinocytes after VZV infection. B) VZV infected and mock-infected primary human keratinocytes were processed for total RNA and protein extraction. At day 5 p.i., 48 hrs after the addition of calcium, the levels of the desmosomal genes were determined by qPCR and normalised to the housekeeping gene RN5S (B). The fold change between the uninfected and VZV infected samples was determined and plotted ± stdev. C) Immune blotting to confirm decreased DSG1 and DSC1 expression in VZV infected keratinocytes. Electron microscopy images of keratinocytes that were (D+G) uninfected, (E+H) early in VZV infection as shown by the presence of viral envelopes but not complete virions, and F+I) Late in VZV infection, where intact virions are easily detectable. White arrows denote the desmosomal junctions, which are absent in the last panels, scale bar D–F = 2 µm and G–I = 500 nm. Pre-treatment of keratinocytes with PAA, which inhibits VZV DNA polymerase and viral replication prevents the reduction of DSG1 (J) and partially restores DSC1 expression (not significant) (K) expression measured by qPCR, n = 3 *p<0.05 by Student's t-test.
Figure 8.
VZV infection increases KLK expression.
A) Heatmap analysis of transcriptome changes in the serine peptidases and non-peptidase homologues group (IPR001314). A clear upregulation (yellow) of the majority of the genes in this group was observed in all KCV lanes. B) The majority of the kallikreins were upregulated in the differentiated keratinocytes compared to the uninfected cells (KCV/KC). Upregulation of KLKs in transcriptome of VZV infected differentiated keratinocytes was confirmed by qPCR for KLK5 and KLK7 (C–D). The relative expression of both genes ± standard deviation, normalised to GAPDH is shown for the uninfected and VZV infected samples at day 5 p.i. after the addition of calcium at day3, p values less than 0.05 are indicated (*). Immune blotting for KLK5 and 7 from concentrated supernatants of uninfected and VZV infected keratinocytes at 1–3 days post-differentiation (E). GAPDH and gE from the cell lysates was used as a loading control and to show VZV infection respectively.
Figure 9.
Epidermal differentiation increases VZV transcription.
A) Coverage plots of viral RNA-seq reads mapped to the annotated VZV pOka genome. The top three panels represent KCV1–3 samples and the lower 5 samples are KV1–5. A schematic representation of the VZV open reading frames is shown below. Each coverage plot uses an identical log scale, illustrating the higher degree of coverage in the KCV samples. Regions with zero coverage are shaded grey. B) RPKM values for each of the VZV ORFs across the infected samples. The RPKM values are normalised relative to the sum of human and viral reads per sample. The pattern of expression is similar across all samples with RPKM values for each ORF being greater than 1 in every sample. RPKM values for the KCV samples are less variable than those for the KV samples with a median 9-fold upregulation per ORF (KCV/KV, range 4–15 fold upregulation, all ORFs significantly upregulated with pFDRs<0.01). RPKM values for KV1 are noticeably lower than all other samples. At day 5 p.i., 48 hrs after the addition of calcium, the levels of three VZV genes were determined by qPCR and normalised to the housekeeping gene RN5S, no amplification was seen in the mock infected controls (data not shown). The upregulation of KRT10 in the uninfected controls was used as a marker to ensure that the addition of calcium had caused keratinocyte differentiation (data not shown). The expression of (C) ORF63 (IE) ORF29 (early) and ORF14 (late) genes was determined (n = 3) and plotted. Expression of ORF29 (D) and ORF14 (E) was significantly increased after the addition of calcium, *p<0.05 by Student's t-test. F) The rate of vDNA replication, as measured by qPCR increases after the addition of calcium (at 48 hrs) in the differentiated cells (n = 3, ± stdev).
Figure 10.
Altered VZV gene expression following keratinocyte differentiation.
A) Heatmap illustrating VZV ORF expression profiles for the infected samples. RPKM values were calculated for each sample by normalising to viral reads alone and then median-centred for each ORF to highlight relative fold changes across all samples. Samples are hierarchically clustered by Pearson's correlation coefficient. ORFs are ordered by temporal gene expression (i.e. whether they are currently categorised as late, early, immediate-early genes). Log2(fold changes) are shown relative to the median RPKM for each ORF. B) Heatmap of 1463 human genes altered solely in the KC/K comparison. TMM-normalised CPM values are median-centred for each gene to highlight relative fold changes across all samples. Genes and samples are clustered by Pearson's correlation coefficient. Log2 (fold changes) are shown relative to the median CPM for each gene. Samples cluster primarily by experimental condition (i.e. addition of calcium) with the exception of KV2 and KV5 which cluster alongside the calcium-shifted KC samples suggesting these 2 samples had undergone spontaneous contact-induced differentiation. C) Primary keratinocytes were infected VZV at an mo.i. of 0.2 and treated with either DAPT [1 µM] or Jagged-1 [50 µM] and harvested for analysis by immunoblotting 48 hrs later. Jagged-1 treatment caused an increase and DAPT treatment showed a decrease in Notch 1 expression respectively as seen in the uninfected keratinocytes. Treatment of VZV infected cells with jagged-1 increased gE expression and DAPT treatment decreased gE expression relative to the untreated VZV infected keratinocytes, the changes in gE were quantified using ImageJ and the numbers below represent the gE expression relative to the untreated VZV infected keratinocytes. Primary keratinocytes were infected with VZV at day 0 with an m.o.i. of 0.2. At day 3 half of the samples were switched to a high calcium media. Duplicate samples were taken every 24 hrs to measure luciferase and/or renilla. Three different recombinant viruses were used D) VZVLUC under the control of an IE promoter (ORF4), E) ORF14Luc and F–G) ORF63LucORF9Renilla. The average relative luciferase/renilla units RLU/RRU for each virus is plotted ± stdev. H) Skin explants were injected into the dermis with cell associated VZV (1250 PFU), placed on grids, and harvested at 3 days post infection, fixed and embedded in paraffin. Sections were stained with antibodies against the immediate early viral marker, IE63, and the late marker glycoprotein E (gE). Staining was developed with the chromogen VIP. No staining was seen in the uninfected controls (data not shown). Although epidermal infection was evident, this model did not demonstrate epidermal blistering. VZV staining is present throughout the epidermis and the expression gE was more apparent in the higher layers of the epidermis.