Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Sep 1.
Published in final edited form as: Virology. 2013 Jun 12;443(2):285–293. doi: 10.1016/j.virol.2013.05.021

Productive vs non-productive infection by cell-free Varicella zoster virus of human neurons derived from embryonic stem cells is dependent upon infectious viral dose

Anna Sloutskin 1, Paul R Kinchington 2, Ronald S Goldstein 1,*
PMCID: PMC3733239  NIHMSID: NIHMS483092  PMID: 23769240

Abstract

Varicella Zoster virus (VZV) productively infects humans causing varicella upon primary infection and herpes zoster upon reactivation from latency in neurons. In-vitro studies using cell-associated VZV infection have demonstrated productive VZV-infection, while a few recent studies of human neurons derived from stem cells incubated with cell-free, vaccine-derived VZV did not result in generation of infectious virus.

In the present study, 90%-pure human embryonic stem cell-derived neurons were incubated with recombinant cell-free pOka-derived made with an improved method or with VZV vaccine. We found that cell-free pOka and vOka at higher multiplicities of infection elicited productive infection in neurons followed by spread of infection, cytopathic effect and release of infectious virus into the medium. These results further validate the use of this unlimited source of human neurons for studying unexplored aspects of VZV interaction with neurons such as entry, latency and reactivation.

Keywords: varicella zoster virus, neuronal infection, human embryonic stem cells

Introduction

The infection of human sensory neurons by varicella zoster virus (VZV) is of clinical importance. the widespread and painful reactivation disease herpes zoster. The virus gains access to neurons by 1) either infection of sensory axons by virus produced in the epidermis in varicella vesicles, and from there is transported retrogradely to the ganglia or 2) by infection of ganglionic cells via lymphocytes (reviewed in (Zerboni and Arvin, 2008)). Following primary infection, VZV enters a latent state in peripheral ganglia throughout the body which can reactivate decades later to produce zoster with painful sequellae, via a mechanism that is poorly undersood.. In most clinical cases the reactivation of VZV in the ganglia leads to a productive infection where viruses are transported anterogradely where they infect fibroblasts and keratinocytes to produce characteristic zoster lesions in the skin.

Unlike its close relative HSV, VZV usually does not replicate in non-human neuronal models (i.e. (Bourdon-Wouters et al., 1990; Gershon et al., 2008; Zerboni and Arvin, 2011)). This has limited progress in understanding the interactions of VZV with neurons, especially the vital pathological aspect of latency. Only in one animal model, guinea pig enteric ganglia (Gershon et al., 2008), has a productive infection of neurons been reported (see below). A second issue that has caused some difficulties in studies of VZV, is the comparative difficulty of generating cell-free virus (Harper et al., 1998). Although the initial infection of neurons may be partially via immune cells (Gershon et al., 2010), the infection of nerve endings in the skin may be from free virus produced in varicella vesicles, and since neurons release cell-free infectious virus (Gowrishankar et al., 2007) (see below). It is also possible that some spread of virus within the ganglia may occur in this manner (satellite cells in direct contact with ganlglionic neurons have been strongly implicated in spread of VZV within ganglia (Reichelt et al., 2008). Until recently, there have been relatively few studies of VZV neuronal infection, leading to the development of several models including dorsal root ganglia from 8–20 week human fetuses either in-vitro or transplanted to SCID-mice (reviewed in (Zerboni et al., 2010)). Over the past two years, several new sources of cells for studying VZV infection in-vitro have been proposed. Our group showed that infection of human embryonic stem cell-derived (hESC) neurons by VZV elicits a productive infection with release of infectious virions (Markus et al., 2011), although neurons do not die by apoptosis as a result of infection (as also shown for fetal ganglia by (Hood et al., 2003)). The model was used more recently by us to visualize and measure kinetic parameters of axonal transport of VZV in living axons (Grigoryan et al., 2012). Productive infection by cell-associated VZV was also shown for cultured fetal DRG (Gowrishankar et al., 2007) for neuronally-differentiated human neuroblastoma cells (Christensen et al., 2011) and neuronally-differentiated human iPS cells (Lee et al, 2013). By contrast, two recent reports examining infection of neurons derived from either 1) human neuronal stem cells (Pugazhenthi et al., 2011) or 2) human iPS cells (Yu et al., 2012) by VZV, have not observed productive infection. An important difference between studies obtaining productive neuronal infection and these studies was the mode of infection used. In studies using cell-associated VZV (infected fibroblasts/MeWo cells), productive infection was observed, whereas the studies obtaining non-productive infection used cell-free virus preparation to infect neurons. Indeed, parallel observations were obtained using the guinea pig enteric ganglion model: cell-associated virus in that model elicited a productive, and cell-free virus, a non-productive “latent-like” infection (Gershon et al., 2008).

It was shown that VZV released from cells used to propagate the virus in culture includes many defective particles unable to infect permissive cells (Carpenter et al., 2009). The present study was therefore undertaken to examine whether human embryonic stem cells-derived neurons incubated with cell-free virus can elicit a productive infection.

Results

Generation of cell free VZV

Retinal pigment epithelium (RPE) cells have reported to be infected by VZV, and can be used to generate higher titers of cell-free virus than fibroblasts and MeWo cells, the most common cells for propagating the virus (Schmidt-Chanasit et al., 2008). ARPE-19, a retinal pigmented epithelia (RPE) cell line, was therefore used to generate cell-free virus. We also used these cells as a “sensor” for productive VZV infection because of their clear cytopathic effect (CPE) and rapid infection. Cell-free VZV was prepared using a modification of published methods (Grose et al., 1979; Schmidt and Lennette, 1976) and detailed in the Methods section (Fig. 1A). We then concentrated the virus to comparatively high PFU levels with either PEG6000 (Kutner et al., 2009) or the commercial lentivirus concentration reagent Lenti-X.

Figure 1. Infection of hESC-derived neurons with cell-free VZV.

Figure 1

Open in a new tab

(A) Schematic representation of the methods used to infect hESC-derived neurons with cell-free VZV. hESC (brown) were propagated on mitoticaly inactivated human foreskin fibroblasts (light grey) (a) and induced to differentiate neurally by co-culture with PA6 cells (dark grey) (b) and plated on laminin for terminal differentiation to neurons (yellow) (c). In parallel, ARPE19 cells infected with GFP-expressing VZV (d) were sonicated to release cell free virus (e) and clarified by centrifugation and concentrated. The medium containing virus (green) was incubated with neurons (f), resulting in VZV infected neurons (g). (B) Effect upon cell-free virus infection by heparin inhibition. ARPE19 cells were infected by cell-free VZV in the presence (heparin) or absence (control) of 100µg/ml heparin, and the PFU was determined. The mean of 10 samples from 3 different experiments is shown; error bars represent the SE values. The difference was significant as determined by Student’s t-test (p<0.01).

The cell-free nature of the VZV preparations was tested using several methods. The first was an infectious foci assay in the presence or absence of heparin, a known inhibitor of cell-free VZV entry into susceptible cells (Li et al., 2006). When heparin was present during the attachment of the virus to the cell monolayer, it inhibited plaque formation by more than 70% (Fig. 1B; 1.8×103 PFU/ml with heparin vs. 6.4×103 without, p <0.01). Additional evidence was that when both supernatant and pellet fractions after 15 minutes centrifugation at 3000g were plated in medium, no live cells were observed after several days of culture (not shown). Finally, a cell-free preparation passed through a 0.45 micron filter infected hESC-derived neurons and ARPE cells (not shown). It should be noted that both of reagents used to concentrate the virus would likely have killed any remaining cells.

Differentiation of human embryonic stem cells (hESC) into neurons

Initial induction of neuronal differentiation of hESC was carried out using mouse PA6 stromal cells as described previously (Pomp et al., 2008). After co-culture, neurally-induced hESC colonies were manually removed from the stromal cells and plated directly on poly-lysine/laminin coated coverslips for terminal differentiation (Grigoryan et al., 2012). This "direct plating" procedure results in cultures consisting of more than 90% neurons (Fig. 2A,B and Reis et al, ms in preparation) immunopositive for neurofilament proteins (Fig. 2C,D). Presence of neurofilament protein indicates that neurons had matured past the stage where only early neuronal markers such as β-III-tubulin are expressed. Some of the neurons expressed peripherin, a marker of peripheral nervous system neurons (not shown).

Figure 2. hESC-derived neurons differentiated without a suspension culture step.

Figure 2

Open in a new tab

Low (A) and high (B) magnification of a neurosphere (NSP) induced by co-culture of hESC with PA6 stromal cells, seeded on a laminin substrate and allowed to differentiate for 2 weeks. (A) Large numbers of neuronal processes radiate from the neurosphere, and many cells have migrated away from the central mass. (B) Cells that migrated away from the neurospheres display neuronal morphology and sit on a plexus of neurites. (C–D) Immunostaining for neurofilament-M protein (red) reveals that the vast majority of cells in the culture are neurons. (C) Low magnification of neurons outside the neurospheres stained for neurofilament-M, the neurospheres mass was also overwhelming immunopositive for neurofilament protein (not shown). (D) High magnification of a field in (C). The arrow points to a rare nucleus which is not surrounded by neurofilament positive cytoplasm. Such rare cells are likely to be neural precursors, immature neurons or glial cells/precursors. Scale bars = 50µm.

Infection of neurons with cell-free alpha-herpesviruses

In order to assess the ability of cell-free α-herpesviruses to infect our neurons, we first used HSV-1, which unlike VZV infects efficiently in a cell-independent manner. Neurons were infected with HSV1 expressing GFP under the glycoprotein C promoter (HSV1 gC-GFP) at ~103 PFU/well (Fig. 3A–C). 2 days post infection (PI), the large majority of neurons in the cultures were infected GFP+, and by 7 days PI, virtually all of the neurons had died.

Figure 3. Infection of hESC-derived neurons with cell-free alpha-herpesviruses.

Figure 3

Open in a new tab

Panels A–I are photomicrographs of living cultures. (A–C) Medium containing HSV-1 expressing GFP readily infects virtually all hESC-derived neurons in cultures by 3 days PI. (A) GFP diffusely fills the entire neurons, including the axons. (B) Phase-contrast image of the same field, showing neuronal morphology. (C) Merged phase and fluorescence image.

(D–F) Cell-free VZV expressing GFP (VZV-66GFP) infects hESC-derived neurons. (D) GFP expression is mostly detected in the neuronal cell bodies, and not in axons. (E) Phase-contrast image of the same field, showing neuronal morphology of the cells. (F) Merged phase and fluorescence image.

(G–I) Cell-free VZV expressing GFP as a fusion protein to the viral ORF23 capsid protein (VZV-23GFP). (G) GFP expression is mostly restricted to the nuclei of the cells, where the production of capsids takes place. (H) Phase-contrast image of the same field. The cells in this culture are at a later stage of infection (6 days vs. 2 days PI) than in the culture with VZV-66GFP, and the cells display a strong cytopathic affect, rounding up and detaching from the substrate. (I) Merged phase and fluorescence image.

(J–M) Immunofluorescence confirmation of cell-free VZV infection of neurons. The sections are stained with 4 fluorescent stains, 2 of which are shown in J–L, and 3 in M. (J) GFP-fluorescence indicating infection by VZV-66GFP combined with blue staining for nuclei with Hoechst. (K) Immunofluorescent staining for the late gE membrane-specific protein (red), outlining cells infected by VZV with nuclear counterstain. (L) Staining for neurofilament-M (NF, white), demonstrating the neuronal phenotype of the infected cells. (M) Merged image of the native GFP (ORF66, green), gE immunostaining (red) and blue staining of the nuclei (blue). All the GFP+ cells are positive for gE stain. Scale bars=50µm.

After demonstrating that hESC-derived neurons could be infected in a cell-free manner by alpha-herpesviruses, we used cell-free VZV expressing GFP as fusion protein to the ORF66 kinase gene (VZV-66GFP, (Eisfeld et al., 2007)) to infect neurons in a 24 well plate (~250PFU/well). ARPE19 cells were infected with this virus at the same concentration used to infect the neurons as a positive control. 2 days after introduction of the cell-free VZV, infection of neurons was detected by GFP fluorescence (Fig. 3D), followed by spreading of the infection and neuronal death within 1 week post infection. A similar result was obtained for VZV expressing GFP as fusion protein to the capsid protein ORF23 (VZV-23GFP, (Markus et al., 2011)) (Fig. 3G–I). VZV infection of the neurons was easily detected by the presence of GFP fluorescence. In addition to being fluorescent, the appearance of the neurons was indicative of a cytopathic process (i.e Fig. 3H), as we had previously observed following cell-associated infection (Markus et al., 2011). Cell-free VZV ORF66-GFP infected neurons were also immunostained for the late VZV gE protein (Fig. 3J–M). The neurons exhibited clear gE and neurofilament immunostaining (Fig. 3K and L, respectively). The expression of the late ORF23 capsid protein and gE glycoprotein, as well as the spread of the infection and the death of the neurons, together strongly suggest that incubation of cell-free VZV preparations with hESC-derived neurons results in a productive infection.

Neurons infected with cell-free VZV generate and release infectious virions

After demonstrating that cell-free viruses prepared from pOka-derived recombinant VZV displayed a CPE and expressed gE in neurons, we obtained direct evidence that the infection was productive by testing whether infected neurons could infect recipient cells. In order to assess the infectivity of virus produced by neurons infected with cell-free virus, we used two approaches. In the first approach, neurons were infected with free-virus prepared from VZV-GFP (Zhang et al., 2008a). Infection by this virus diffusely fills infected cells with GFP, which allows ease of visualization in living cultures (Fig. 4A–C). After most of the cultures showed GFP-expression, the neurospheres and neurons were detached from coverslips and triturated using the tip of a 200µl pipettor. The generated suspension was transferred to an uninfected monolayer of human foreskin fibroblasts (HFF). No living neurons or portions of neurospheres were observed after the transfer to the fibroblasts, which are grown in a different medium. The fibroblasts became infected with VZV after 3 days, readily detected by GFP fluorescence and CPE (Fig. 4D,E). The infection of the HFF by a suspension of neurons that had been infected with cell free-VZV strongly suggests that these neurons produced VZV that readily infect susceptible cells. It has been shown by ourselves and others that cultured neurons infected with cell-associated VZV release virions into the medium (Gowrishankar et al., 2007), (Markus et al., 2011), (Lee et al., 2012). In a second approach, we tested the ability of hESC-derived neurons infected by cell-free VZV to not only produce but release infectious virions. Media were collected from 3 wells of neurons infected with cell-free VZV-GFP66 after extensive infection was observed, 5 days PI (Fig. 4G–I). The media were clarified by centrifugation and 10-fold concentrated and then transferred to a monolayer of uninfected ARPE19 cells. Four days after addition of the medium that had bathed the cell free VZV-infected neurons, GFP-positive centers of infection were observed ARPE19 (Fig. 4J–L). The same results were obtained with neurons infected by cell-free VZV-GFP23 (not shown). By contrast, medium bathing MeWo cells infected by cell-free VZV-GFP66 transferred to ARPE19 cells and processed in parallel with the medium from the neuronal cultures, failed to infect the ARPE19 cells. These results from both of these approaches demonstrate the cell-free preparations of VZV can infect neurons in a manner that leads to production of infective virus.

Figure 4. Infection of hESC-derived neurons by cell-free VZV is productive.

Figure 4

Open in a new tab

(A–F): Secondary cell-associated infection by neurons that were infected with cell-free virus. (A) hESC-derived neurons infected with cell free VZV expressing GFP under a constitutive SV40 promoter (VZV-GFP). (B) Phase contrast image of the same field. (C) Merged phase and fluorescence image. (D–F). Neurons from the culture shown in panels A–C were homogenized by trituration and transferred to uninfected HFF. The fibroblasts were infected by VZV after 3 days, as determined both by GFP fluorescence (D) and by cytopathic effect, by their rounding up and detaching from the culture dish. (E) Phase-contrast image of the same field. (F) is a merge of the micrographs in D and E. (G–L): Neurons infected with cell-free virus release infection virus into the medium. hESC-derived neurons were infected with cell-free VZV-66GFP and medium was collected from the cultures after most neurons were infected as indicated by strong GFP fluorescence (G) and death of many neurons (H). (I) Merge of panels G and H. (J–L) Show ARPE19 cells 4 days after the transfer of clarified and concentrated medium that had bathed cultures such as in G–I. Infection of the ARPE19 cells was detected by both GFP fluorescence (J) and cytopathic effect (rounding up of the flat ARPE19 cells (K)). (L) is a merge of panels J and K. Scale bar = 50µm.

Productive infection of neurons is determined by the MOI used

One possible explanation for our observation of productive infection (in contrast to results of Pugazhenthi et al. and Yu et al) is that the outcome of cell free infection may depend on the viral dose. An estimate of the number of the cells per coverslip in our system is between 2,000–20,000 (depending on the number of spheres seeded and their size). This would result in our infecting with preparations of VZV at MOI between 1.25×10−1 and 1.25×10−2. By contrast, Pugazhenthi et al. used 2,500PFU for 1 million cells, giving an MOI of 2.5×10−3, at least one order of magnitude lower. To test whether the type of infection obtained is dependent upon viral dose, we performed a dilution series of cell-free virus infections, starting from the concentration we used in the experiments described above (MOI of 10−1–10−2), and 1/10 and 1/100 dilutions.

Extensive fluorescence was present 7 days post infection when approximate MOIs of 10−1–10−2 or 10−2–10−3 were used (Fig. 5A). By contrast, no fluorescence at all was detected one week PI when an MOI of 10−3–10−4 (Fig. 5C) was used. Essentially the same results were obtained using several batches of cell-free virus prepared separately. Therefore, it appears that the difference in MOI used to infect neurons is responsible for the difference in the outcome of infection observed.

Figure 5. Productive infection of neurons depends on the MOI used.

Figure 5

Open in a new tab

hESC-derived neurons were incubated with cell-free VZV66GFP at a MOI of 10−1–10−2 (A, B) or 10−3–10−4 (C,D) and monitored for infection. (A) 7 days after addition of virus, prominent infection is observed for the higher MOI, as detected by GFP fluorescence, while no fluorescence is seen for the lower MOI (C). (B,D) Phase-contrast images of the fields shown in (A,C).

Another possible explanation for why Pugazhenthi et al did not obtain productive infection was that they used the attenuated vOka strain (Horien and Grose, 2012; Somekh and Levin, 1993) from the Zostavax vaccine, while we (and others obtaining productive infection of neurons in-vitro (i.e. Hood et al., 2003) use recombinant VZV based on the parent Oka strain. In order to test this possibility, we infected hESC-derived neurons with either vOka or cell-free pOka-derived virus at MOIs of 10−1–10−2 or 10−3–10−4. 5 days after the initial infection the cultures were fixed and stained for VZV glycoprotein E and neurofilament-H protein to identify neurons. VZV given to cultures at the higher MOI used (10−1–10−2) resulted a distinct cytopathic effect by three days PI and increasing at 5 days PI indicating productive infection by both attenuated (Figure 6 A–B) and parental (not shown) strains. At low MOI, vOka did not generate a CPE in neurons 5 day PI (Figure 6C), as was observed for the pOka-derived recombinant VZV (see above). Neurons infected with both strains at high MOI strongly expressed gE (Figure 6E–J). Although it is possible that the lower neurovirulence of vOka contributed to observation of stem-cell derived neurons not being productively infected, it appears that the amount of input virus used plays the determining role in whether cultured stem-cell derived neurons replicate VZV after infection.

Figure 6. The VZV vaccine strain vOka productively infects hESC-derived neurons at higher MOI (A–D).

Figure 6

Open in a new tab

Neurons were incubated with vOka at a MOI of 0.1 (A&B) and 0.001 (C) and monitored for 5 days. At the higher MOI, 3 days post infection neurons appear healthy (A). However, by 5 days the neurons exhibit a distinct CPE (B) indicated by swelling of somata (arrows, also see (Markus et al., 2011)) and detachment, as compared to the normal morphology of neurons in a parallel culture not receiving VZV (D). When neurons were bathed in vOka at the lower MOI, no CPE was observed after 5 days infection (C). (E–J) Immunocytochemical evidence for productive infection of hESC-derived neurons by pOka and vOka. Neurons were infected at “high” MOI (10–1–10−2) with either vOka (E–G) or pOka (H–J) and stained for the late gE membrane-specific protein (shown in green (E,H)) and neurofilament-200 (shown in red (F,I)). (G&J) Merged images of gE immunostaining (green), NF-200 (red) and blue staining of the nuclei (blue). Scale bars = 50µm.

Discussion

We demonstrate here using several methods that hESC-derived neurons infected by cell-free VZV replicate and release infectious virus. Thus, the reason for not obtaining productive infection by VZV in two recent studies, is not the use of cell-free versus cell-associated virus (as observed in guinea pig neurons, see Introduction). It is important to note that cell-free VZV produced in a different laboratory (in China) has also recently been shown to generate productive infection (as shown by spread of infection, for example) of hESC-derived neurons and neuroblastoma-derived neurons (Selariu et al., 2012). Even more recently, incubation of human iPS-derived neurons with cell-free VZV was shown to result in release of infectious virus into the medium in (Lee et al., 2012).

The cell free nature of the virus used in our study was demonstrated by reduction of the number of the foci formed in the presence of heparin, the absence of intact cells at the end of the production process, and the ability to infect neurons after being filtered through a 0.45 micron filter. The highest titers were obtained by generating virus in ARPE19 cells (as compared to HFF and MeWo cells) and concentration with a commercial reagent marketed for the concentration of lentivirus, Lenti-X (4-fold higher concentration than using PEG (Grose et al., 1979)). It should be noted that the concentration step also likely kills any cells that could have survived the sonication procedure. The achieved titers were as high as 105 PFU/ml, as detected by infectious foci assay. This improved method should prove useful for other studies of VZV infection requiring lack of contamination by “input” cells or for analyses of kinetics of viral gene/protein expression.

There are several potential explanations for the discrepancy between studies obtaining productive infection results and those obtaining non-productive infection by cell-free VZV. First, there is the possibility that neurons derived from hESC and those derived from other sources are somehow different in their ability to support a productive infection. The finding that iPS derived neurons ((Lee et al., 2012)) and even cells that are not bonafide neurons such as undifferentiated neuroblastoma cells (Selariu et al., 2012) and unpublished observations) support productive infection by cell-free VZV suggest that the source of neuronal cells is not the determining factor. Another possible explanation is the use of different viral strains. We routinely use recombinant VZV derived from pOka((Eisfeld et al., 2007; Markus et al., 2011; Zhang et al., 2008b)), while the studies not obtaining productive infection ((Pugazhenthi et al., 2011) and (Yu et al., 2012) used the attenuated vaccine strain vOka (Quinlivan et al., 2011). The molecular basis of the vaccine attenuation is being elucidated (Peters et al., 2012), but the vaccine has already shown to be less neurovirulent (Horien and Grose, 2012). Our observation that vaccine VZV also productively infects hESC neurons suggests that the VZV strain is not the reason for the discrepancy. Consistent with our results, Lee et al (2012) also reported productive infection of iPS-derived neurons with the vOka strain, including release of infectious virus. It should be noted that while a few gE positive neurons out of the thousands in the culture were observed when we infected with recombinant pOka at lowest MOI, no gE positive cells were observed when we used vOka at the same concentration.

It was suggested in Yu et al that the reason that we had observed productive neuronal infection using cell-associated VZV in Markus et al, was because our cultures were not sufficiently pure, and the productive neuronal infection was via non-neuronal cells. The formation of syncitia between satellite cells and dorsal root ganglion neurons in SCID-hu mice is strong evidence that neurons in-vivo are infected by neighboring cells (Reichelt et al., 2008). However, we believe that their explanation for our in-vitro observations in Markus et al and in the present study is not correct. First, the difference in purity of neurons used in the two studies was only 5%, their cultures contained 95% neurons and those in our current study 90%-pure neurons. This means there were non-neuronal cells present in their experiments and thus potential for indirect infection, just as in ours. Second, infection of 95% pure hESC-derived neurons by cell-free virus in Selariu et al (generated using the more time-consuming differentiation method of Pomp et al, 2008), also resulted in spread of viral infection within the culture. Indeed, the ability of the infection to spread between neurons in-vitro was used in order to obtain information about the function of VZV protein ORF7.

Therefore, we believe that the reason that we (as well as Lee et al) obtained, and others did not obtain productive infection, is due to the amount of virus used to infect. When VZV is given to neurons at high MOI a productive infection ensues, while at low MOI, either a non-productive infection, or no infection is obtained. In order to show this directly, we performed a dilution series experiment and found at low MOI we did not obtain GFP expression and spread (Figure 5A–D). It is possible that when using lower input concentration of VZV the infection itself takes more time to become evident. The cultures in the experiment of Figure 5 were monitored for 14 days, with no signs of GFP fluorescence in the low MOI treated wells. The lack of GFP fluorescence we observed at low MOI may be either due to non-productive infection or no infection at all. Performing PCR for VZV genomes would not allow distinction between these possibilities because the reaction would detect DNA from input virus that had adhered or even entered the cells. This MOI-dependence of the type of infection obtained might reflect a situation in vivo following primary infection. It is possible that the levels of infectious virus delivered to any individual neuron in primary infection would be relatively low when transported from the skin, since this is probably not a very efficient process. Lower levels of initial infection of neurons are therefore more likely in vivo, perhaps promoting the establishment of latency and not a productive, potentially lytic infection.

Our demonstration that VZV can generate productive infection of stem-cell derived neurons in-vitro has wider implications than resolving a difference between published studies. Zoster requires reactivation of VZV into a productive infection, usually resulting in virus delivery to the periphery via axons and zoster lesions. Even in herpes zoster sine herpete, with its potential serious clinical manifestations, virus is generated in infected sensory neurons. The present results demonstrate that our accessible in-vitro system recreates this important aspect of neuronal infection by VZV, and, with further development, the hESC-derived neuron model should be useful for studying reactivation and anterograde axonal transport, which could lead to improved treatment and prevention of zoster.

Methods

Cells and viruses

The H9 (US National Stem Cell Bank (WA09)) human embryonic stem cell line was maintained as described previously (Markus et al., 2011). Human neonatal foreskin fibroblasts (HFF), PA6, MeWo (human melanoma cell line), ARPE19 (human retinal pigment epithelium(ATCC #CRL-2302)), and Vero (green African monkey kidney epithelia) cells were maintained in DMEM containing 10% FCS, 2mM glutamine, 50U/ml penicillin and 50µg/ml streptomycin.

Parent Oka-based VZV expressing GFP as fusion proteins to ORF66 and ORF23 were described elsewhere ((Eisfeld et al., 2007), (Markus et al., 2011)). VZV expressing GFP from a constitutive SV40 promoter were described previously (Zhang et al., 2008b). All viruses were maintained in MeWo or ARPE19 cells, as described previously (Markus et al., 2011). HSV-1 expressing GFP under the promoter of the late glycoprotein C gene was described previously (Decman et al., 2005) and was propagated and titrated on Vero cells.

Neuronal differentiation from hESC

Induction of neuronal precursors was performed by co-culture of hESC on PA6 mouse stromal cellsas described previously (Pomp et al., 2008). Subsequently, colonies were terminally differentiated for two weeks (Grigoryan et al., 2012) until extensive neurite outgrowth was observed. The vast majority of cells in the cultures stained positive for the intermediate neurofilament proteins but the cultures probably contained some neural precursors and glial cells at the time of infection.

Generation of cell free VZV and infection

ARPE19 cells were infected with VZV at a ratio of 1 infected to 10 uninfected cells. After an extensive cytopathic effect was observed (3–4 days) the cells were scraped in PSGC buffer (Harper et al., 1998), freeze thawed, sonicated for 3×2min (PC3 ultrasonic bath sonicator 22W, 50KHz) and centrifuged 15 min at 3000g. After scraping, all suspensions were kept either at 4°C or on ice. The supernatant containing cell-free VZV (confirmed by plating the supernatant and microscopic examination) was concentrated 50-fold using PEG6000 as described in (Kutner et al., 2009), according to a previously published procedure (Grose et al., 1979), or Lenti-X (Clontech, USA, cat. #631231). The concentrated virus was aliquoted and stored in PSGC buffer in liquid nitrogen. All experiments were performed with concentrated virus that had been frozen and then thawed. To infect cells, aliquots were thawed, diluted no more than 1:5 in DMEM:F12 and allowed to adsorb to a cell monolayer at 37°c in a volume of 150µl/well of 24-well-plate. After 2 hours the innoculum was replaced with growth medium. Infection was observed after 2 days, and was detected by fluorescence or cytopathic effect. Infection of neurons by HSV1 was performed as described above for VZV.

Infectious foci assays

All assays were performed on ARPE19 cells in triplicates. 10-fold dilutions of the thawed cell-free VZV were prepared, and allowed to adsorb to the cell monolayer for 2 hours, agitating occasionally, after which the innoculum was removed. The cells were monitored for fluorescent focus formation. At 5 days post infection the foci were counted, and virus concentration (as PFU/ml) was determined. The infection was performed in the presence or absence of 100µg/ml heparin in order to further confirm the cell-free nature of the virus (Li et al., 2006).

Demonstration of productive neuronal infection

For testing for cell-associated infection, neurons were infected with cell-free VZV expressing GFP under the constitutive SV40 promoter. After 3 days, the neurons were heavily infected as detected by GFP fluorescence and cytopathic effect (CPE). At this point the medium was changed and neurons were detached by pipetation and plated on “sensor” HFF cells. After allowing 2 hours at 37°C for infection, the innoculum was replaced with fresh medium. The “sensor” cells were observed daily for infection by GFP fluorescence.

In order to test for “cell-free” infection, neurons or MeWo cells were infected with cell-free VZV ORF66-GFP. After 5 days, extensive infection was observed, and the culture medium bathing 3 wells of each type of cell was collected and clarified at 10,000g for 10min. 10-fold concentration (v/v) of the virus was achieved by Lenti-X concentrator (Clontech, USA, cat. #631231), and the concentrated virus was added to the “sensor” cell monolayer (ARPE19) and allowed to adsorb to the cells for 2 hours. The sensor cells were observed daily for infection.

Immunofluorescence staining

Coverslips were stained as described in (Pomp et al., 2008). Primary ntibodies used (at the indicated dilutions) were rabbit anti-neurofilament-H (200KD subunit) (N4142, 1:200; Sigma-Aldrich), goat anti-neurofilament-intermediate subunit (NF-M) (1:200; Santa Cruz), mouse anti-NF-M (2H3, 1:15; DSHB) and mouse monoclonal anti-gpI (now named gE) (1:2000; Chemicon). Secondary antibodies coupled to Alexa Fluor 594 were then applied for 30 min, and counterstained with Hoechst 33258 to visualize nuclei. In one experiment where four distinct fluorescent wavelengths were required (GFP, immunostaining for neurofilament and VZV-gE as well as nuclei, Figure 3J–M), Alexa Fluor 647 (near-infrared) secondary antibodies were used for detecting the 2nd antigen. Negative controls were performed by omitting the primary antibodies and by staining of uninfected cultures.

Microscopy

Preparations were viewed with Olympus IX70 (live cultures) and BX51 (immunostainings) microscopes, photographed using digital cameras (Scion and Jenoptiks Coolscan monchrome) and processed using ImageJ software. Images were enhanced using ImageJ and Paint Shop Pro software with all changes in the images (i.e., contrast, brightness, gamma, and sharpening) made evenly across the entire field, and no features were removed or added digitally.

Highlights.

Human embryonic stem cell-(hESC) derived neurons are productively infected by cell-free pOka VZV.

Cell-free vaccine VZV (vOka) also generates a productive infection in hESC-derived neurons.

Productive infection of hESC-derived neurons by cell-free pOka and vOka is dependent on MOI.

An improved protocol for generating relatively high-titer cell-free VZV is described.

Acknowledgements

This study was supported by Israel Science Foundation grant #238/11 (RSG): NIH grants NS064022, NS082662, EY08098, and funds from the Research to Prevent Blindness Inc. and The Eye & Ear Foundation of Pittsburgh (PRK). Our thanks to Hua Zhu for VZV-GFP. Thanks as always to Chaya Morgenstern for expert technical and logistic support.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Disclosure of potential conflicts of interest: No conflicts of interest are declared.

References

  1. Bourdon-Wouters C, Merville-Louis MP, Sadzot-Delvaux C, Marc P, Piette J, Delrée P, Moonen G, Rentier B. Acute and persistent varicella-zoster virus infection of human and murine neuroblastoma cell lines. J. Neurosci. Res. 1990;26:90–97. doi: 10.1002/jnr.490260111. [DOI] [PubMed] [Google Scholar]
  2. Carpenter JE, Henderson EP, Grose C. Enumeration of an extremely high particle-to-PFU ratio for Varicella-zoster virus. J. Virol. 2009;83:6917–6921. doi: 10.1128/JVI.00081-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Christensen J, Steain M, Slobedman B, Abendroth A. Differentiated neuroblastoma cells provide a highly efficient model for studies of productive varicella-zoster virus infection of neuronal cells. J Virol. 2011 doi: 10.1128/JVI.00515-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Decman V, Kinchington PR, Harvey SAK, Hendricks RL. Gamma interferon can block herpes simplex virus type 1 reactivation from latency, even in the presence of late gene expression. J. Virol. 2005;79:10339–10347. doi: 10.1128/JVI.79.16.10339-10347.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Eisfeld AJ, Yee MB, Erazo A, Abendroth A, Kinchington PR. Downregulation of class I major histocompatibility complex surface expression by varicella-zoster virus involves open reading frame 66 protein kinase-dependent and -independent mechanisms. J. Virol. 2007;81:9034–9049. doi: 10.1128/JVI.00711-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Gershon AA, Chen J, Gershon MD. A model of lytic, latent, and reactivating varicella-zoster virus infections in isolated enteric neurons. J. Infect. Dis. 2008;197(Suppl 2):S61–S65. doi: 10.1086/522149. [DOI] [PubMed] [Google Scholar]
  7. Gershon AA, Gershon MD, Breuer J, Levin MJ, Oaklander AL, Griffiths PD. Advances in the understanding of the pathogenesis and epidemiology of herpes zoster. J. Clin. Virol. 2010;48(Suppl 1):S2–S7. doi: 10.1016/S1386-6532(10)70002-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Gowrishankar K, Slobedman B, Cunningham AL, Miranda-Saksena M, Boadle RA, Abendroth A. Productive varicella-zoster virus infection of cultured intact human ganglia. J. Virol. 2007;81:6752–6756. doi: 10.1128/JVI.02793-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Grigoryan S, Kinchington PR, Yang IH, Selariu A, Zhu H, Yee M, Goldstein RS. Retrograde axonal transport of VZV: kinetic studies in hESC-derived neurons. J. Neurovirol. 2012;18:462–470. doi: 10.1007/s13365-012-0124-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Grose C, Perrotta DM, Brunell PA, Smith GC. Cell-free varicella-zoster virus in cultured human melanoma cells. J. Gen. Virol. 1979;43:15–27. doi: 10.1099/0022-1317-43-1-15. [DOI] [PubMed] [Google Scholar]
  11. Harper DR, Mathieu N, Mullarkey J. High-titre, cryostable cell-free varicella zoster virus. Arch. Virol. 1998;143:1163–1170. doi: 10.1007/s007050050364. [DOI] [PubMed] [Google Scholar]
  12. Hood C, Cunningham AL, Slobedman B, Boadle RA, Abendroth A. Varicella-zoster virus-infected human sensory neurons are resistant to apoptosis, yet human foreskin fibroblasts are susceptible: evidence for a cell-type-specific apoptotic response. J. Virol. 2003;77:12852–12864. doi: 10.1128/JVI.77.23.12852-12864.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Horien C, Grose C. Neurovirulence of varicella and the live attenuated varicella vaccine virus. Semin Pediatr Neurol. 2012;19:124–129. doi: 10.1016/j.spen.2012.02.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kutner RH, Zhang X-Y, Reiser J. Production, concentration and titration of pseudotyped HIV-1-based lentiviral vectors. Nat Protoc. 2009;4:495–505. doi: 10.1038/nprot.2009.22. [DOI] [PubMed] [Google Scholar]
  15. Lee KS, Zhou W, Scott-McKean JJ, Emmerling KL, Cai G-Y, Krah DL, Costa AC, Freed CR, Levin MJ. Human sensory neurons derived from induced pluripotent stem cells support varicella-zoster virus infection. PLoS ONE. 2012;7:e53010. doi: 10.1371/journal.pone.0053010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Li Q, Ali MA, Cohen JI. Insulin Degrading Enzyme Is a Cellular Receptor Mediating Varicella-Zoster Virus Infection and Cell-to-Cell Spread. Cell. 2006;127:305–316. doi: 10.1016/j.cell.2006.08.046. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Markus A, Grigoryan S, Sloutskin A, Yee MB, Zhu H, Yang IH, Thakor NV, Sarid R, Kinchington PR, Goldstein RS. Varicella-Zoster Virus (VZV) Infection of Neurons Derived from Human Embryonic Stem Cells: Direct Demonstration of Axonal Infection, Transport of VZV, and Productive Neuronal Infection. J. Virol. 2011;85:6220–6233. doi: 10.1128/JVI.02396-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Peters GA, Tyler SD, Carpenter JE, Jackson W, Mori Y, Arvin AM, Grose C. The attenuated genotype of varicella-zoster virus includes an ORF0 transitional stop codon mutation. J. Virol. 2012;86:10695–10703. doi: 10.1128/JVI.01067-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Pomp O, Brokhman I, Ziegler L, Almog M, Korngreen A, Tavian M, Goldstein RS. PA6-induced human embryonic stem cell-derived neurospheres: a new source of human peripheral sensory neurons and neural crest cells. Brain Res. 2008;1230:50–60. doi: 10.1016/j.brainres.2008.07.029. [DOI] [PubMed] [Google Scholar]
  20. Pugazhenthi S, Nair S, Velmurugan K, Liang Q, Mahalingam R, Cohrs RJ, Nagel MA, Gilden D. Varicella-zoster virus infection of differentiated human neural stem cells. J. Virol. 2011;85:6678–6686. doi: 10.1128/JVI.00445-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Quinlivan M, Breuer J, Schmid DS. Molecular studies of the Oka varicella vaccine. Expert Review of Vaccines. 2011;10:1321–1336. doi: 10.1586/erv.11.93. [DOI] [PubMed] [Google Scholar]
  22. Reichelt M, Zerboni L, Arvin AM. Mechanisms of varicella-zoster virus neuropathogenesis in human dorsal root ganglia. J. Virol. 2008;82:3971–3983. doi: 10.1128/JVI.02592-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Richter ER, Dias JK, Gilbert JE, 2nd, Atherton SS. Distribution of herpes simplex virus type 1 and varicella zoster virus in ganglia of the human head and neck. J. Infect. Dis. 2009;200:1901–1906. doi: 10.1086/648474. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Schmidt NJ, Lennette EH. Improved yields of cell-free varicella-zoster virus. Infect. Immun. 1976;14:709–715. doi: 10.1128/iai.14.3.709-715.1976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Schmidt-Chanasit J, Bleymehl K, Rabenau HF, Ulrich RG, Cinatl J, Jr, Doerr HW. In vitro replication of varicella-zoster virus in human retinal pigment epithelial cells. J. Clin. Microbiol. 2008;46:2122–2124. doi: 10.1128/JCM.00122-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Selariu A, Cheng T, Tang Q, Silver B, Yang L, Liu C, Ye X, Markus A, Goldstein RS, Cruz Cosme RS, Lin Y, Wen L, Qian H, Han J, Dulal K, Huang Y, Li Y, Xia N, Zhu H. ORF7 of Varicella Zoster Virus is a Neurotropic Factor. Journal of virology. 2012;86:8614–8624. doi: 10.1128/JVI.00128-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Somekh E, Levin MJ. Infection of human fetal dorsal root neurons with wild type varicella virus and the Oka strain varicella vaccine. J. Med. Virol. 1993;40:241–243. doi: 10.1002/jmv.1890400313. [DOI] [PubMed] [Google Scholar]
  28. Yu X, Seitz S, Pointon T, Bowlin JL, Cohrs RJ, Jonjić S, Haas J, Wellish M, Gilden D. Varicella zoster virus infection of highly pure terminally differentiated human neurons. J. Neurovirol. 2012 doi: 10.1007/s13365-012-0142-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Zerboni L, Arvin A. Investigation of varicella-zoster virus neurotropism and neurovirulence using SCID mouse-human DRG xenografts. J. Neurovirol. 2011;17:570–577. doi: 10.1007/s13365-011-0066-x. [DOI] [PubMed] [Google Scholar]
  30. Zerboni L, Arvin AM. Neurotropic Viral Infections. Cambridge University Press; 2008. The pathogenesis of varicella-zoster virus neurotropism and infection; pp. 225–250. [Google Scholar]
  31. Zerboni L, Reichelt M, Arvin A. Varicella-zoster virus neurotropism in SCID mouse-human dorsal root ganglia xenografts. Curr. Top. Microbiol. Immunol. 2010;342:255–276. doi: 10.1007/82_2009_8. [DOI] [PubMed] [Google Scholar]
  32. Zhang Z, Huang Y, Zhu H. A highly efficient protocol of generating and analyzing VZV ORF deletion mutants based on a newly developed luciferase VZV BAC system. J. Virol. Methods. 2008a;148:197–204. doi: 10.1016/j.jviromet.2007.11.012. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES