INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Herpes Simplex Virus-1 (HSV-1) is a neurotropic virus that normally infects dorsal root ganglion cells (DRGs) in host animals but can infect a broad range of other neurons. Its broad selectivity, ability to infect postmitotic neurons, and capacity to accommodate large fragments of foreign DNA make it an attractive vehicle for the delivery of foreign genes into neurons for therapeutic or experimental purposes (see Fink et al. 1996; Kaplitt and Makimura 1997). The cytotoxicity of the wild-type virus, however, requires that some, or all, HSV-1 genes be removed from vectors derived from it to allow their use for gene transfer. Minimal vectors, which package bacterial plasmids, or "amplicons," into an HSV-1 coat, represent a particularly elegant solution to this problem (reviewed by Ho 1994).

HSV-1 amplicons are plasmids that have been genetically modified to include viral control elements but that completely lack HSV-1 genes. Amplicons contain signals for DNA replication and viral packaging and an HSV-I immediate early gene promoter to drive expression of any gene(s) of interest inserted downstream. HSV-1 amplicons are replicated and packaged into viral capsids in permissive cell lines with the aid of replication-defective HSV-1 helper viruses. In addition to amplicon-containing virus particles, this process also produces more helper virus, which remains as part of the final preparation. Procedures for making amplicon preparations free of helper virus have been developed but lower yields have limited their utility (see Robbins and Ghivizzani 1998; but also Wang et al. 2000).

Although helper viruses are a common component of amplicon preparations, little has been reported on the frequency with which they co-infect amplicon-transduced neurons under typical experimental conditions or on their cytopathological effects on co-infected neurons. While it is clear that the cytotoxicity of helper viruses is significantly attenuated relative to that of wild-type HSV-1 (Lim et al. 1996), sublethal effects on cellular physiology have also not generally been investigated. The successful use of amplicon preparations to express, or overexpress, genes of interest in neurons both in vitro and in vivo (Meier et al. 1997; Neve et al. 1997; Phillips et al. 1999; Song et al. 1998) has no doubt blunted interest in these issues, particularly as the effects of helper virus can be controlled for by amplicon preparations lacking the gene of interest. The results we report here suggest, however, that the physiological effects of helper virus on amplicon-transduced neurons call for increased caution in matching the titers of helper virus in control and experimental preparations to ensure that results are comparable.

It has long been known that HSV-1 infection results in the rapid loss of neuronal excitability. Early evidence suggesting that this loss resulted from the downregulation of voltage-sensitive Na⁺ currents (Fukuda and Kurata 1981; Mayer 1986; Oakes et al. 1981) has been confirmed by more recent work (Howard et al. 1998). In the course of using amplicons to overexpress K⁺ channels in cultured rat cortical neurons, we observed widespread suppression of inward current in amplicon-transduced neurons, coupled with an almost complete loss of excitability. Surprisingly, we obtained the same result with an amplicon preparation expressing only Green Fluorescent Protein (GFP), suggesting that the effect was not due to increased K⁺ conductance. Because the suppression of excitability by wild-type HSV-1 is known to depend on viral gene expression, infection by amplicon-containing virus alone is not expected to exert this effect. Instead, we have traced the suppression to the 5dl1.2 helper virus used in amplicon packaging. We demonstrate that 5dl1.2, like wild-type HSV-1, inhibits voltage-sensitive Na⁺ currents in infected neurons. Consistent with our physiological findings, we directly demonstrate a high-frequency of co-infection of amplicon-transduced neurons by helper virus, even at low multiplicities of infection (MOI = 0.2). Using anti-HSV antibodies to detect helper-virus-infected neurons treated with the GFP-expressing amplicon preparation, we find that helper virus co-infects 80% of amplicon-transduced cortical neurons under our standard conditions. Because the loss of electrical activity in amplicon-transduced neurons may under some circumstances be mistaken for the effects of an introduced gene, our results imply that 5dl1.2 titers in experimental and control amplicon preparations should be matched to produce equal degrees of co-infection. In addition, we describe an immunofluorescence assay that should prove more useful than the standard plaque assay in determining accurate titers of helper virus on the neurons of interest.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell culture

Whole cerebral cortices from newborn (P0) Sprague Dawley rats were dissected free of meninges, washed in Hank's balanced salt solution on ice, then digested with 0.05% trypsin/0.53 mM EDTA at 37°C for 20 min. (Unless otherwise noted all reagents here and in the following text were from Gibco, Grand Island, NY.) The cell suspension was then diluted 1:1 into Dulbecco's modified Eagle's medium (DMEM) and treated briefly with DNAse I to disperse tissue clumps prior to trituration with a fire-polished Pasteur pipette and passage through a Falcon cell strainer (Applied Scientific, South San Francisco, CA). Dissociated cells were pelleted by centrifugation at 1,000 g for 5 min and resuspended in DMEM/5% fetal bovine serum (FBS) after removal of the supernatant. This procedure was repeated twice, and the cell pellet was finally resuspended in Neurobasal medium supplemented with B27, 5% FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, penicillin (100 U/ml), and streptomycin (100 µg/ml). The cell suspension was plated at a density of 100,000 cells/well in 24-well plates containing polyornithine (0.01 mg/ml)- and laminin (2 µg/ml)-coated glass coverslips. Long-term cultures were fed every third day by exchanging half the media with fresh media.

Adult rat dorsal root ganglion (DRG) neurons were cultured as previously described (Caffrey et al. 1992). Briefly, the L₄ and L₅ DRG ganglia were harvested from adult male Sprague-Dawley rats. The DRG were treated with collagenase (1 mg/ml) for 25 min and collagenase (1 mg/ml) plus papain (30 U/ml) for 25 min, dissociated in DMEM and Ham's F12 medium supplemented with 10% FBS and plated on glass coverslips. The cultures were maintained at 37°C in a humidified 95% air-5% CO₂ incubator.

Virus preparation and titering

Amplicon preparations were prepared as described by Lim and Neve (1999). Briefly, amplicon DNA containing the gene for Green Fluorescent Protein (GFP) was transfected into cultured 2-2 Vero cells using 2 µg maxiprep DNA and 12 µl lipofectamine in 1.0 ml of serum-free media; serum containing media was added after 5 h. Prior to transfection, 2-2 cells were grown in 30-mm culture dishes until 80% confluent, at 10% CO₂, in DMEM media (Hyclone Laboratories, Logan, UT) containing 10% fetal calf serum (FCS), and 1% Pen/Strep. Twenty hours after transfection, 20 µl (1 × 10⁶ pfu) of helper virus was added, and 36 h later cells were lysed, and this lysate was added onto fresh 2-2 cells. This amplification process was repeated twice, resulting in a final 60 ml of virus-containing cell lysate. The lysate was purified over a sucrose step gradient (10%/30%/60%) by centrifugation at 125,000 g for 1 h. The virus bands at the 30%/60% interface were removed and diluted 1:10 into PBS. Virus was then concentrated by ultracentrifugation at 125,000 g for 1 h. The resulting pellet was resuspended in 200 µl PBS/10% sucrose and stored at 80°C in 30-µl aliquots. Unpurified helper virus preparations were made by infecting untransfected 2-2 cells with 5dl1.2 helper virus as described in the preceding text and isolating the virus-containing lysate. Control lysates were prepared from uninfected 2-2 cells.

Plaque assays to determine the titer of 5dl1.2 helper virus in the viral preparations were carried out as follows. A 100 µl aliquot of viral preparation was added to a 6-cm dish containing 1 × 10⁶ 2-2 cells in 2 ml DMEM/10% FCS. After allowing the virus to adsorb for 2 h, the media was replaced with 3 ml 1% Seaplaque agarose (BioWhittaker Molecular Applications, Rockland, ME) at 40°C in DMEM/5% FCS. Agarose was allowed to solidify before returning the culture dishes to a 37°C incubator for 24 h. DMEM/5% FCS (2 ml) was added after 24 h and exchanged again after a further 24 h, and 1 day later cells were fixed with 3 ml 5% methanol/10% acetic acid for 30 min. Plaques were visualized by staining with crystal violet and counted. Amplicon titers were determined by infecting PC12 cells, plated in the wells of a 24-well culture dish at a density of 500,000 cells/well, with up to 4 µl of HSV-GFP for 24 h. One day after infection, cells were fixed with 4% paraformaldehyde/0.1 M PBS for 25 min then washed with PBS. Amplicon-transduced cells were identified by the presence of GFP fluorescence using an Olympus BX-60 fluorescence microscope and counted.

Viral transduction/infection

TRANSDUCTION. Gradient-purified GFP-HSV amplicon preparations were stored at 80°C and thawed at 37°C before dilution into media and application to neuronal cultures in a volume of 500 µl. The estimated ratio of virus particles per neuron, or MOI, used for transduction of the cortical neuron cultures was 0.2 but was higher for DRG neurons, which were plated at lower (and variable) densities. Neuronal cultures were incubated for 6 h with the amplicon preparation prior to washout, and electrophysiological recordings were carried out 2 days after transduction except where otherwise noted. Cortical neurons were typically transduced 6-7 days after plating. DRG neurons were transduced 2 days after plating.

Whole cell patch-clamp recording of neurons

CORTICAL NEURONS. Whole cell patch-clamp recordings were conducted at room temperature using an Axopatch 1D amplifier and pCLAMP 6 data-acquisition software (both from Axon Instruments, Foster City, CA). Patch electrodes were pulled using a vertical Narishige glass microelectrode puller and had resistances of 3-4 M. The offset potential was zeroed before patching the cells. Voltage errors were minimized using 80-90% series resistance compensation. The capacitance artifact after breakthrough was cancelled by capacitance compensation and the magnitude of compensation applied was used to estimate cell membrane capacitances. Linear leak subtraction, based on resistance estimates from four to five hyperpolarizing pulses applied before the depolarizing test potential, was used for all voltage-clamp recordings. The pipette solution contained (in mM) 97.5 KGluconate, 32.5 KCl, 5 EGTA, and 10 HEPES, pH 7.2, and the bath solution contained (in mM) 140 NaCl, 1.3 CaCl2, 5.4 KCl, 25 HEPES, and 33 glucose, pH 7.2. Voltage-clamp data were sampled at a rate of 8 kHz and filtered at 2 kHz. Amplicon-expressing neurons, identified as GFP-positive by fluorescence microscopy, were recorded 2 days after transduction except where noted, and control cells were handled in parallel but were not incubated with the amplicon preparation. To identify the inward currents, voltage-clamp recordings were made in 10 control cells before and after bath application of 250 nM tetrodotoxin (TTX). Neurons from 5dl1.2-helper-virus-infected cultures were picked at random and recorded 1 day after helper-virus treatment. This was prior to the onset of extensive neurite degeneration and cell death, which was evident by 2 days. Neurons from parallel cultures treated with a cell lysate from 2-2 helper cells uninfected with 5dl1.2 helper virus were used as controls.

DRG NEURONS. For whole cell patch-clamp recordings of DRG neurons, data were acquired on a Windows-based Pentium-III computer using an EPC-9 amplifier and the Pulse program (v 8.1, HEKA Electronic, Germany). Fire-polished electrodes (0.8-1.5 M) were fabricated from 1.7-mm capillary glass using a Sutter P-97 puller (Novato, CA). Small DRG neurons with a soma diameter of 18-30 µm were selected for recording. Cells were not considered for analysis if the initial seal resistance was less than 2 G, if they had high leakage currents (holding current more than 0.6 nA for DRG neurons at 80 mV), or if they had an access resistance more than 4 M. The pipette solution was designed for the isolation of Na⁺ currents and contained (in mM) 140 CsF, 1 EGTA, 10 NaCl, and 10 HEPES (pH 7.3). The bathing solution was (in mM) 140 NaCl, 3 KCl, 1 MgCl2, 1 CaCl2, and 10 HEPES (pH 7.3). The osmolarity of all solutions was adjusted to 310 mosM (Wescor 5500 osmometer, Logan, UT). All P values reported in the text were derived from two-tailed t-tests.

Immunocytochemistry

Cover-slip plated neurons were fixed for 25 min in 4% paraformaldehyde/phosphate-buffered saline (PBS), then permeabilized for 5 min with 0.3% Triton-X 100 (Sigma, St. Louis, MO). After blocking with 10% normal goat serum in PBS, coverslips were incubated overnight at 4°C with a rabbit anti-HSV Type 1 and 2 polyclonal antibody (Chemicon International, Temecula, CA, Catalog No. AB1125), used at a dilution of 1:500 in 10% goat serum in PBS. Following washes with PBS, coverslips were incubated for 3 h at room temperature with a fluorescein- or Texas Red-conjugated goat-anti-rabbit IgG secondary antibody (Jackson ImmunoResearch) at a dilution of 1:100 in 10% goat serum/PBS. After further washes with PBS, coverslips were mounted on glass slides with Vectashield mountant (Vector Laboratories, Burlingame, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Suppression of neuronal Na+ currents by an amplicon preparation

To investigate the physiological effects of HSV-1 amplicon preparations on neurons, we treated cultured rat cortical neurons with an amplicon preparation containing the gene for GFP and recorded from amplicon-transduced neurons identified by the presence of green fluorescence (Fig. 1A). Initial current-clamp recordings from neurons 2 days after transduction revealed a significant difference in the excitability of the amplicon-transduced neurons compared with that of controls (Fig. 1, B and C). Whereas 90% of control cells fired multiple action potentials in response to depolarizing current injections of 300 nA (average peak firing frequency: 23 ± 4 Hz), 88% of the transduced neurons were completely unexcitable and the remaining 12% fired a single spike. Examination of whole cell currents in the amplicon-transduced neurons strongly suggested that the suppression of excitability derived from a profound inhibition of the inward currents (Fig. 2A). Peak inward currents in amplicon-transduced neurons declined sevenfold relative to those of uninfected control cells. In contrast, steady-state outward currents declined less than twofold, a difference that was not statistically significant (Fig. 2C). We identified the inward currents in these cells as Na⁺ currents, using the specific Na⁺ channel blocker tetrodotoxin (TTX; Fig. 2B). Application of 250 nM TTX to control cells selectively inhibited 97% of the inward current (I_inward = 5.7 ± 0.8 nA before TTX vs. 0.13 ± 0.05 nA after TTX; n = 10).

View larger version (34K): [in this window] [in a new window]   Fig. 1. A Herpes Simplex Virus (HSV) amplicon preparation expressing Green Fluorescent Protein (GFP) suppresses the excitability of embryonic rat cortical neurons. Rat cortical neurons were incubated with a GFP-expressing amplicon at a multiplicity of infection (MOI) of 0.2 for 6 h, 1 wk after plating. Amplicon-transduced cells were identified by GFP fluorescence after 2 days (A), and the excitability of the cells was assessed by whole cell current-clamp recording. B: representative responses to current injections of 10, 50, and 100 pA for control (top) and amplicon-transduced (bottom) neurons are shown (holding potential 60 mV). C: the maximum firing frequencies in response to current injections of 10-310 pA were averaged for all amplicon-treated (n = 8) and control (n = 10) neurons (P < 103 by 2-tailed t-test).

View larger version (22K): [in this window] [in a new window]   Fig. 2. The loss of excitability on HSV amplicon transduction results from suppression of Na+ currents in embryonic rat cortical neurons. A: representative whole cell currents from voltage-clamped cortical neurons. Shown are currents from control (left) and amplicon-treated (right) cells from parallel cultures. Cells were held at 70 mV and stepped to +80 mV in 10-mV increments. B: currents from a control cell measured before (left) and after (right) application of 250 nM TTX indicate that the inward currents inhibited by the amplicon preparation are Na+ currents. Scale bar same as in A. C: peak Na+ (left) and outward (right) currents were normalized to cell capacitance and averaged for control and amplicon-transduced cells. (For Na+ currents, P < 104 by 2-tailed t-test.)

Suppression of Na⁺ currents by wild-type HSV-1 is well established in adult DRG neurons (Howard et al. 1998; Mayer et al. 1986), the natural host cell of this virus, and although live virus was absent from our amplicon preparation, the effect we observed with cortical neurons appeared similar. To investigate whether the amplicon preparation, like live HSV-1, suppressed Na⁺ currents in DRG neurons, we transduced cultured DRG neurons from adult rat and 2 days later conducted whole cell recordings under conditions designed to isolate Na⁺ currents. Consistent with the results obtained in cortical neurons, we observed a 78% inhibition of the Na⁺ current in transduced cells versus untransduced controls (Fig. 3, A and B).

View larger version (15K): [in this window] [in a new window]   Fig. 3. Na+ currents are suppressed in amplicon-transduced dorsal root ganglion cells (DRGs). A: representative Na+ currents from amplicon-transduced (bottom) and parallel control (top) DRGs. L4 and L5 lumbar DRGs were cultured from adult rats and treated with amplicon as described in METHODS. Na+ currents were recorded by whole cell voltage clamp from small-diameter neurons (18-25 µm) in the presence of 140 mM CsCl and 100 µM cadmium to block K+ and Ca2+ currents, respectively. Cells were held at 100 mV and stepped from 80 mV to +40 mV in 10-mV increments. B: average Na+ currents from control (n = 22) and amplicon-transduced (n = 22) cells (P < 5 × 105 by 2-tailed t-test).

To determine whether transduced neurons recovered their inward currents over time, we compared the normalized Na⁺ currents in transduced cortical neurons 1-2 wk posttransduction with those observed 1-2 days posttransduction. As indicated in Fig. 4, there is, at later times (Fig. 4C, bottom), a profound reduction in the percentage of neurons with small current densities (normalized values of 0.0-0.4) when compared with neurons assayed 1-2 days posttransduction (43 vs. 70%, Fig. 4C, top). As is also indicated in Fig. 4, the loss of transduced neurons with small inward current densities is paralleled by a rapid decline in the number of GFP-labeled neurons (Fig. 4A). We found that within 4 days of transduction the number of green neurons declined by more than 50%. By 1 wk after transduction, less than 20% of the original number of labeled neurons remained. Some of the loss of label is likely due to downregulation of GFP expression in transduced neurons, a common feature of amplicon-expressed genes (Lim et al. 1996), but morphological characteristics of many of the labeled neurons suggested that they were dying as many of them showed signs of rounding and neurite degeneration (Fig. 4B). The electrophysiological results further suggest that the amplicon-transduced neurons fall into two classes: a class with very low Na⁺-current densities that degenerate and perhaps die, and another more viable population with Na⁺-current densities that are less severely depressed.

View larger version (55K): [in this window] [in a new window]   Fig. 4. Most amplicon-transduced cortical neurons appear to degenerate and die within a week. A: average number of GFP-labeled neurons per field as a function of time after transduction. The number of GFP-positive cortical neurons in seven randomly chosen microscope fields from 3 coverslips were counted at each indicated day and averaged. Each field contained approximately 900 cells. B: typical amplicon-transduced neuron (green) at day 4 posttransduction showing signs of neurite degeneration. Note the absence of neurites extending from this cell, while most unlabeled cells show healthy processes. C: the distribution of neurons with suppressed Na+ currents also differs at short and long intervals after transduction. Na+ currents in amplicon-transduced (black) and control (gray) neurons 1-2 wk (bottom) or 1-2 days (top) after transduction. Because the average Na+ current amplitudes increase with time in culture, the inward current densities for cells of each age group were normalized to the average current density of the control cells for that age. The values on the abscissa are therefore ratios and have no units.

Immunodetection of helper-virus infection in amplicon-treated neurons

Although wild-type virus was absent from our preparation, the 5dl1.2 helper virus used to prepare it was present at a titer similar to that of amplicon-containing virus as determined by the standard plaque and fluorescence assays described in METHODS (1.2 × 10⁸ pfu/ml vs. 1.0 × 10⁸ infectious units ifu/ml, respectively). While ratios of amplicon-containing virus to helper virus of 10:1 have been reported in enriched amplicon preparations (Kwong and Frenkel 1995), 1:1 ratios are typical (Ho 1994; Smith et al. 1995). Because the 5dl1.2 helper virus has previously been reported to reduce neuronal viability with a similar time course to our observed loss of GFP fluorescence and because it expresses viral genes, which is known to be required for the suppression of Na⁺ currents (Mayer et al. 1986), it seemed possible that both the observed reductions in viability and excitability derived from co-infection of the transduced neurons with 5dl1.2 helper virus. To test this possibility, we developed an immunofluorescence technique using anti-HSV antibodies to simultaneously measure infection by both amplicon and helper virus.

Because infection of neurons by helper virus results in the expression of viral genes, whereas infection by amplicon-containing virus does not, cells infected with helper virus should be selectively immunoreactive to antibodies against HSV secondary proteins. To test the ability of anti-HSV antibodies to detect helper virus infection and to distinguish it from virion attachment, which will also occur with amplicon-containing virus, we incubated cortical neuron cultures with 5dl1.2 helper virus for 5 h in the presence or absence of cycloheximide (CHX, at 10 µg/ml) to block viral protein synthesis. Cells were then fixed and immunolabeled with anti-HSV antibodies and a fluorescein-labeled secondary antibody. In the absence of CHX, 9% of the cells showed clear anti-HSV immunofluorescence with the cell bodies typically filled and neurites more lightly labeled (Fig. 5A). In contrast, none of cells in the CHX-treated cultures showed this pattern of staining (Fig. 5B), although some were lightly "decorated" with dots, which may represent staining of the structural proteins of attached virions. Both patterns of staining were enhanced by extending the incubation with helper virus to 24 h (Fig. 5, C and D), a condition that led to the labeling of 37% of cells in the absence of CHX. A small number of cells (0.8%) in the 24-h CHX-treated cultures showed dim, somatic immunofluorescence, but this labeling was easily distinguishable from that seen in cells untreated with CHX and is likely to be due to residual protein synthesis in the prolonged presence of helper virus. Confident that we could discriminate between infected cells and those having only attached viral particles, we used the anti-HSV antibodies to selectively identify helper virus-infected neurons in our amplicon-treated cultures.

View larger version (13K): [in this window] [in a new window]   Fig. 5. Neurons infected with 5dl1.2 helper virus can be identified by immunofluorescence using anti-HSV antibodies. Cortical neuron cultures were incubated with helper virus for 5 (A) and 24 h (C) before fixation and staining with anti-HSV antibodies using a fluorescein-conjugated secondary antibody. Immunoreactivity is evident within 5 h of infection and very strong within 24 h with green anti-HSV immunofluorescence clearly filling cell bodies of infected cells within 5 h and neurites within 24 h. Parallel cultures co-incubated with 10 µg/ml cycloheximide during the 5 (B) or 24 h (D) helper-virus incubation do not exhibit the same robust cell-filling pattern of staining and are instead sometimes "decorated" with dots, which may represent staining of the structural proteins of attached virions.

To examine the degree of helper-virus infection of neurons treated with our amplicon preparation, we incubated cortical neuron cultures at MOIs (0.2) previously found to suppress Na⁺ currents in approximately 70% of transduced cells (Fig. 4C, top). Incubation with amplicon was carried out for 6 h prior to washout, and neurons were allowed to grow for a further 48 h prior to examination by immunofluorescence. Amplicon-transduced neurons were visualized by green fluorescence (Fig. 6, A and D). As described in the preceding text, helper-virus-infected neurons were identified by immunofluorescence using anti-HSV antibodies but this time with a Texas-red-labeled secondary antibody (Fig. 6, B and E). Coincident labeling of cells could readily be evaluated by overlapping the signals in the two fluorescence channels with double-labeled neurons appearing yellow (Fig. 6, C and F).

View larger version (9K): [in this window] [in a new window]   Fig. 6. Most amplicon-infected neurons are also infected with helper virus. Amplicon-treated (A-C) and untreated control (D-F) cortical neurons were assayed for amplicon-transduction by GFP fluorescence (A and D) and for helper virus infection by immunostaining using anti-HSV antibodies and a Texas Red-conjugated secondary antibody (B and E). Double labeling was monitored by double exposure (C and F). Most of the GFP-positive cells can be seen to also be immunopositive for HSV proteins and appear some shade of yellow in the double exposure, though some neurons remain solid green (arrows) and are thus uninfected by helper virus. The efficiencies of amplicon and helper virus infection given in the text were estimated from 8 randomly photographed fields containing 908 neurons, and the frequency of co-infection was estimated from examination of 258 GFP-positive neurons. The criterion for co-infection was the coincidence of green GFP fluorescence and red anti-HSV immunofluorescence in the same neurons, confirmed by the presence of yellow as seen in C. The absence of signal in the controls (D-F) indicates that all labeling observed is specific.

While only 3.3% of the neurons had been transduced by amplicon, we found that 80% of these GFP-positive neurons were also anti-HSV immunopositive (Fig. 6C). This was nearly an order of magnitude higher than the co-infection rate predicted from the helper-virus infection rate (10.2%) if all infections were to occur randomly and independently. Consistent with our physiological data, these results strongly indicate that viral particles containing amplicon and helper virus DNA preferentially co-infect the same neurons, even at the low MOI we have employed.

Interestingly, the relative titer of helper virus to amplicon-containing virus, when measured by the rates of infection of cortical neurons, was approximately threefold higher than that estimated from the titers derived from the standard assays. The rate of cortical neuron infection by helper virus (10.2%) corresponds to a titer almost sixfold more than that estimated by "plaque assay" on 2-2 Vero cells (6 × 10⁸ ifu/ml vs. 1.2 × 10⁸ pfu/ml), while the rate of amplicon transduction (3.3%), corresponds to an infectious titer of 2 × 10⁸ ifu/ml, or twice that measured by GFP fluorescence on amplicon-treated PC12 cells (1.0 × 10⁸ ifu/ml). This suggests that the standard assays may differentially underestimate helper virus titers, which will also lead to higher than expected rates of co-infection of amplicon-transduced neurons by helper virus. The discrepancy also highlights one of the advantages of the immunofluorescence assay employed here, namely, that it allows the titers of both amplicon-containing virus and helper virus to be determined on the same population of cells by similar techniques.

Suppression of neuronal Na+ currents by helper virus alone

The high frequency of co-infection of amplicon-transduced neurons by helper virus provides a rational explanation for the widespread suppression of Na⁺ currents in these cells if the 5dl1.2 helper virus, like wild-type HSV-1, downregulates these currents. To directly test the ability of 5dl1.2 to inhibit Na⁺ currents, we carried out patch-clamp recordings from cultured rat cortical neurons infected with helper virus alone. We incubated neuronal cultures with titers of 5dl1.2 helper virus sufficient to infect 67% of the cells as determined by immunostaining with anti-HSV antibodies. Sampling inward currents in randomly selected cells by whole cell patch-clamp techniques 1 day later, we found that two-thirds of the cells recorded from (12/15) had Na⁺ current densities of less than or equal to 0.2 nA/pF (Fig. 7, A and B). Only one-fifth of control cells (3/13) had current densities this small, with the average Na⁺ current in the controls having a value of 0.53 nA/pF (Fig. 7, A and B). One-third of the cells treated with helper virus (3/15) had large inward currents (0.93 nA/pF) and were presumably uninfected. The strong inhibition of Na⁺ currents in helper virus treated cells at a frequency similar to the frequency of infection, is strong evidence that 5dl1.2, like wild-type HSV-1, suppresses neuronal Na⁺ currents. We conclude that the 5dl1.2 helper virus is the cause of Na⁺ current suppression observed in amplicon transduced neurons.

View larger version (29K): [in this window] [in a new window]   Fig. 7. The 5dl1.2 helper virus profoundly suppresses inward current in cortical neurons. A: representative whole cell currents from voltage-clamped cortical neurons 1 day after incubation with helper virus (bottom) or control lysate (top). Cells were incubated for 6 h with a volume of helper virus sufficient to infect 67% of the neurons as assayed by immunostaining. Control cells were incubated with an equal volume of 2-2 cell lysate. Neurons were held at 70 mV and stepped to +80 mV in 10-mV increments. B: distribution of inward currents (normalized to cell capacitance) for helper virus-treated () and control () cells. Consistent with the efficiency of helper virus infection, two-thirds of the cells (12/15) have strongly suppressed Na+ currents relative to controls. Inset: average inward current densities for control and helper-virus-infected cells (P = 0.03 by 2-tailed t-test).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We report here several findings that, taken together, strongly support the conclusion that helper virus present in HSV-1 amplicon preparations can lead to the loss of Na⁺ currents and neuronal excitability. We find that within 2 days, 70% of the neurons transduced with an amplicon preparation packaged using the 5dl1.2 helper virus exhibit substantially suppressed voltage-sensitive Na⁺ currents. This effect is similar to the inhibition of Na⁺ currents reported for wild-type HSV-1 and is likely to derive from infection by either the amplicon-containing virions or the helper virus in the preparation. Downregulation of Na⁺ currents by native HSV-1 is known to require viral gene expression and is absent in strains in which specific genes are mutated or deleted (Howard et al. 1998; Mayer et al. 1986; Storey et al. 1996). The fact that the 5dl1.2 helper virus, but not the amplicon, carries viral genes therefore suggests that helper virus is responsible for the suppression of Na⁺ current in amplicon-infected neurons. This conclusion is confirmed by our observations that the helper virus has the capacity of the wild-type virus to suppress Na⁺ currents and that it broadly co-infects amplicon-transduced neurons under our experimental conditions.

We have not investigated the mechanism for the widespread co-infection of neurons by helper virus and amplicon-containing virus that occurs at an eightfold higher rate than would be predicted for random and independent infection events. Because amplicon DNA and helper virus DNA are thought to be packaged into independent, but otherwise identical, viral particles, however, the preferential co-infection of cells by amplicon-containing virus and helper virus must reflect the different susceptibilities of individual neurons to viral infection, perhaps due to differential expression of the viral receptor.

Differing susceptibilities to infection of different cell types may also underlie the conflicting estimates of viral titer we obtained using different assay types. The immunofluorescence technique introduced in this paper gave a helper virus titer sixfold higher on cortical neurons than that determined by the standard plaque assay performed on African Green Monkey Kidney cells (i.e., 2-2 Vero cells). Similarly, amplicon titers determined directly on cortical neurons were twice those found by the standard assay using pheochromocytoma (PC12) cells. As a consequence the relative titer of amplicon to helper virus measured on cortical neurons was approximately three times that estimated by the usual methods. These observations highlight the advantages of the immunofluorescence technique for estimating viral titers introduced here. It not only permits titering the levels of contaminating helper virus directly on the cell type of interest but also allows the amplicon-containing virus to be titered on the same cell type by the same methodology as long as the expressed gene product is labeled (e.g., by GFP) or if antibodies to it are available. A more accurate estimate of the extent of helper virus contamination can then be obtained.

The mechanism of HSV suppression of Na⁺ currents is unknown. One possible mechanism involves the rapid internalization of Na⁺ channels, but using a variety of anti-Na⁺ channel antibodies we were not able to demonstrate any overt redistribution of immunoreactivity in amplicon-transduced neurons (data not shown). Na⁺ channel function, rather than surface expression, may thus be affected by HSV infection though further work will be necessary to address this possibility. Interestingly, not all channel types may be targets for suppression as we failed to observe suppression of Na⁺ currents by the amplicon preparation in a human embryonic kidney (HEK) cell line stably transfected with the muscle-type Na⁺ channel (data not shown). HEK cells may also lack regulatory factors required for mediating channel inhibition, but it remains an interesting possibility that only specific Na⁺ channel types are targeted by the virus. Investigation of this point may also provide clues as to which viral genes mediate Na⁺ current suppression.

Our data indicate that Na⁺ current suppression is a common feature of infection in neurons. Previous studies of Na⁺ current inhibition by wild-type HSV have typically focused on DRG neurons as these are the normal host for herpes simplex virus infections. Our observation that Na⁺ currents were similarly inhibited in both rat embryonic cortical neurons and adult DRG neurons indicates not only that the suppression of Na⁺ current occurs generally but that the co-infection of transduced neurons by helper virus is also a general characteristic of treatment with the amplicon preparation. Also, we observed similar suppression of inward current with amplicon preparations containing transgenes other than GFP, demonstrating that the inhibition was not a property of the GFP-containing amplicon preparation reported on here. Indeed, we first observed the profound suppression of Na⁺ currents during experiments to test the effects of expressing mutant K⁺ channels on cortical neuron physiology (data not shown).

Overall, our results indicate that some care should be taken in interpreting the results of experiments using amplicon preparations because even substantial removal of contaminating helper virus and working at low MOIs do not guarantee elimination of effects of the helper virus. In particular, it seems critical that control preparations have titers of helper virus similar to those of the experimental amplicon preparation to ensure like frequencies of co-infection by helper virus. Otherwise, effects of co-infection seen with the experimental preparation may be erroneously attributed to the transgene. The immunofluorescence method introduced here to determine viral titers may also be helpful in establishing accurate relative titers of helper and amplicon-containing virus on the cell type to be infected.

What implications our results have for transduction experiments carried out in animals or for gene therapy is not yet clear. Indeed, amplicon preparations made with the 5dl1.2 helper virus have been successfully used in numerous gene transfer experiments in vivo without apparent deleterious consequences (Antonawich et al. 1999; Carlezon et al. 1997, 1998; Chen et al. 2001). Likewise, preliminary experiments on slice preparations from amplicon-expressing neurons in the locus coeruleus of rats have failed to show noticeable suppression of excitability in transduced neurons after 1 wk (G. Aghajanian, personal communication). Similarly, experiments involving the injection of an amplicon preparation into rat dentate gyrus have been reported to leave the population responses of both dentate gyrus and hippocampal neurons unchanged (Dumas et al. 1999). The helper virus used for these last experiments, however, differed genetically from 5dl1.2 and has not, to our knowledge, been tested for its ability to suppress Na⁺ currents. It is, however, also possible that the loss of some transduced neurons due to co-infection can be tolerated in vivo or that neurons in vivo tolerate helper virus infection better than those in culture. Given the profound effects we observe in culture, however, possible effects of helper virus co-infection cannot necessarily be ignored.