Regulation of Expression of the Neuronal POU Protein Oct-2 by Nerve Growth Factor*

POU proteins are a class of homeobox-containing transcription factors that regulate tissue-specific gene expression and influence cell differentiation and func- tion. We have investigated the possible role of such factors in mediating the actions of nerve growth factor (NGF) on sensory neurons. NGF has been found to have differential effects on the levels of three POU protein transcription factors that are expressed in adult rat sensory neurons. A sensory neuron octamer-binding protein with the properties of the transcription factor Oct-2 is up-regulated 3-4-fold by NGF, as measured by mobility shift assays using nuclear extracts from adult rat dorsal root ganglion neurons grown in the presence or absence of NGF. Quantitation of Oct-2 mRNA by polymerase chain reaction amplification of RNA from such cells shows a parallel increase in Oct-2 mRNA levels. In contrast, the levels of mRNA encoding the ubiquitous POU protein Oct-1 or the neuron- specific POU protein Brn-3, also present in sensory neurons, are unaffected by NGF. These observations suggest a role for Oct-2 in mediating transcriptional effects induced by NGF. In particular, as Oct-2 is known to inhibit herpes simplex virus immediate-early gene expression in neuronal cells, these findings pro- vide a mechanism for the known action of NGF in the maintenance of latent herpes virus infections in sen- sory neurons.

Nerve growth factor (NGF)' is one of a family of neurotrophic factors that, among other actions, promote the survival of developing peripheral sensory and sympathetic neurons. Neuronal survival in the developing peripheral nervous system has been demonstrated to depend in part upon the presence of limiting amounts of NGF (Barde, 1989). A family of related factors (BDNF, NT3, NT4/5) with remarkably conserved protein sequences between species that show subtly different profiles of cellular specificity as survival factors in vitro have recently been identified by molecular cloning (Maisonpierre et al., 1990;Berkemeier et al., 1991;Hallbook et al., * This work was supported in part by grants from Action Research (to D. S. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. Tel.: . 'The abbreviations used are: NGF, nerve growth factor; DRG, dorsal root ganglion; SDS, sodium dodecyl sulfate; Hepes, 4-(2-hy-droxyethy1)-1-piperazineethanesulfonic acid PCR, polymerase chain reaction; HSV, herpes simplex virus. 1991). A high affinity receptor for NGF has also been identified and been demonstrated to be a transmembrane protein with intrinsic tyrosine kinase activity (TrkA) (Kaplan et al., 1991;Klein et al., 1991). Structurally related transmembrane tyrosine kinases (TrkB, TrkC) are good candidates as cognate receptors for the other neurotrophins so far identified (Squint0 et al., Lamballe et al., 1991). Despite these advances, and the cataloguing of NGF-induced second messenger changes and effects on gene expression in susceptible cells (Chao, 1992), the physiological role of neurotrophic factors in the mature nervous system and the mechanism of their action are poorly understood. Artificial elevation of NGF levels has been shown to alter the phenotype of nociceptive sensory neurons that play a role in the neuronal component of inflammatory responses, and effects on the expression of various neuropeptides and ion channels at the mRNA level in these cells have been described (Lindsay and Harmer, 1988). It therefore seems likely that many actions of NGF are mediated through effects on transcriptional regulation. Consistent with this, a number of NGF-regulated transcription factors have been identified in the pheochromocytoma-derived PC12 cell line by differential screening methods (Milbrandt, 1987;Oppenheim, 1991).
We have used primary cultures of peripheral neurons as an in uitro system to investigate the action of NGF on the POU protein class of transcription factors that appears to have particular importance in tissue-specific gene regulation (Herr et al., 1988;He et al., 1989). Adult rat sensory neurons are known to express a number of such proteins, including Oct-1, Oct-2, Brn-3, and a number of Brn-3-related factors, the partial sequences of which have been identified by PCR (Latchman et al., 1992). Two such factors, Brn-3 and Oct-2, are of particular interest because of their restricted cellular distribution. The octamer-binding protein Oct-2, a factor first identified as a B cell immunoglobulin-specific transcription factor has subsequently been found in a number of types of neurons including sensory neurons (He et al., 1989;Lillycrop et al., 1991). The brain-derived putative transcription factor Brn-3 is additionally interesting because of its close structural similarity to a protein that determines the developmental fate of some sensory neurons in Caenorhabditis elegans, first identified in the Unc-86 mutant (Herr et al., 1988). We therefore analyzed the effects of NGF on the levels of mRNA encoding the neuron-specific class IV POU protein Brn-3 and the tissue-specific octamer binding protein Oct-2 as well as Oct-1, a ubiquitously expressed POU protein, that is known to play a critical role in regulating the expression of a number of cellular and viral genes by binding to the octameric sequence ATGCAAAT in their promoters (Falkner et al., 1986). Adult rat sensory neurons, unlike their neonatal counterparts, can survive in culture without NGF (Lindsay et al., 1989;Winter et al., 1988). We were therefore able to examine the effect of depleting NGF on the expression of transcription factors that have been identified in adult rat sensory neurons without compromising the viability of the cultures. We report here that studies using quantitative polymerase chain reaction amplification of sensory neuron RNA demonstrate a specific up-regulation of Oct-2 mRNA levels by NGF in adult rat sensory neurons, together with an increase in Oct-2 protein levels measured in mobility shift assays. This effect is specific for Oct-2, as the POU proteins Brn-3 and Oct-1 as well as other non-POU transcription factors such as TFIIIC are unaffected by alterations in the levels of NGF.

MATERIALS AND METHODS
Cell Culture-Dorsal root ganglia from all spinal levels of adult male Sprague-Dawley rats were dissected aseptically and collected in Ham's F14 medium supplemented with 1.176 g/liter sodium bicarbonate, 1 mM glutamine, 100 pg/ml penicillin, and 100 units/ml streptomycin. Ganglia were digested with 0.125% collagenase (Boehringer) and mechanically dissociated through a fire polished Pasteur pipette. Cells were then preplated overnight on polyornithine-coated Petri dishes. After 24 h, lightly adherent neurons were removed from the dishes in a stream of medium and replated on 13-mm glass coverslips previously coated with polyornithine and laminin (Bethesda Research Laboratories) at a density of 20,000 neurons per coverslip. Cultures were then supplemented with 0.1 pg/ml2.5 S NGF prepared from mouse salivary glands (Suda et al., 1978) or grown in the absence of NGF and presence of neutralizing sheep anti-NGF antiserum at concentrations capable of neutralizing 0.2 pg/ml NGF to block any endogenously synthesized NGF produced by nonneuronal cells in the cultures. An equivalent concentration of normal sheep serum was added to cultures grown in the presence of NGF. Cytosine arabinoside (10 p~) was included in the medium for the first 2 days to kill dividing cells, then removed, and 5 days after plating either total RNA was extracted from the cells by a guanidine chloride/phenol extraction procedure (Chomczynski and Sacchi, 1987) or nuclear extracts were prepared for mobility shift assays.
cDNA Probe Synthesis and Northern Blot Analysis-A neonatal rat dorsal root ganglion (DRG) cDNA library constructed in X zap 11 (the kind gift of Dr. J. Boulter) was plated at 1-2 X lo6 plaqueforming units per 132-mm-square plate and plaques grown for 7 h at 42 "C before binding to Hybond N membranes (Amersham). The filters were denatured and neutralized and UV-irradiated. The filters were prehybridized for 5 h in 4 X SSC, 5 X Denhardt's solution 100 pg/ml salmon sperm DNA, 5 mM EDTA, and 0.2% SDS at 68 "C and hybridized in the same solution containing lo6 cpm/ml of a denatured oligo-labeled Brn-3 probe comprising the POU and homeobox domain. After 16-h hybridization, the filters were washed at room temperature for 20 min in 0.2 X SSC and 3 X 30 min in 0.2 X SSC/ 0.2% SDS at 68 "C. Plaque-purified positive clones were sequenced using Sequenase I1 (United States Biochemical). Neonatal rat DRG total RNA (20 pg) or poly(A+) RNA (2 pg) was fractionated on 1% agarose/formaldehyde gels and blotted on to Hybond N filters. After blotting, the RNA was cross-linked to the membrane by exposing the moist blot to 254-nm light at 100 kJ/cm for 3 min, followed by baking a t 80 "C for 1 h. Baked filters were prehybridized for 4 h in 50% formamide 5 X SSC, 5 x Denhardt's, 100 pg/ml salmon sperm DNA, 50 pg/ml yeast RNA, 0.1% SDS and hybridized for 40 h in the same solution containing 3 X lo6 cpm/ml of a denatured oligo-labeled probe at 42 "C. Filters were washed at room temperature for 20 min in 0.2 x SSC and for 3 X 20 min in 0.2% X SSC/O.2% SDS at 65 'C. Washed filters were exposed to Kodax XAR-5 film at -70 'C for 12 h.
DNA Mobility Shift Assays-Oligonucleotides were labeled after annealing by phosphorylation with [-y-32P]ATP (Amersham) and T4 polynucleotide kinase (Boehringer). Nuclear extracts were made from pools of approximately lo5 DRG neurons. The washed cell pellet was first resuspended in 5 volumes 10 mM Hepes, pH 7.9, 10 mM KCl, 1.5 mM MgC12, 0.5 mM DTT cells for 10 min at 4 "C. After centrifugation at 1000 X g for 10 mins, the cell pellet was resuspended in 3 volumes of the above buffer to which Nonidet P-40 (0.05%) was added. The cells were disrupted in a Potter homogenizer (20 strokes), and nuclei isolated by centrifugation at 1000 X g for 10 mins. The nuclear pellet was then resuspended in 100 pl of buffer C (25 pl of 2 M Hepes, pH 7.9, 1.25 ml of 50% glycerol, 3.75 pl of MgClZ, 1.25 pl of 1 M DTT, 12.5 p1 of 100 mM phenylmethylsulfonyl fluoride in a total volume of 2.5 ml), and 4 M NaCl added to a final concentration of 300 mM. After 30 min at 4 "C, the nuclear extracts were microfuged for 15 min at 4 "C, and supernatants were stored in 20-pl aliquots. 10 fmol of labeled probe was added to 1 p1 of nuclear extract in a total volume of 20 pl of Hepes, pH 7.9, 5 mM MgC12, 50 mM KC1, 0.5 mM DTT, 4% Ficoll, 2 pg of poly(d1. dC) (Pharmacia LKB Biotechnology Inc.) at 4 "C for 40 min, after which DNA protein complexes were separated by electrophoresis on 4% polyacrylamide gels run for 2-3 h at 150 V at 4 "C in 2.5 mM Tris, 2.5 mM boric acid, 0.5 mM EDTA, pH 8.3.
Quantitative PCR-In preliminary experiments reverse transcription and PCR amplification were carried out using dilutions of RNA from 0.1 to 2 pg with 20-30 cycles of amplification to identify conditions in which the PCR product signal was quantitatively related to input RNA, and controls using RNA samples without reverse transcription were used to demonstrate that contaminating DNA was absent. Polymerase chain reaction amplification of a ribosomal protein (L27) 130-bp cDNA was used as a measure of the amount of input RNA (LeBeau et al., 1991). Amplifications were carried out for Oct-1 and Oct-2 using 25 cycles of amplification as previously described (Lillycrop et al., 1991). Polymerase chain reaction of the Oct-2 mRNA used a control internal template prepared by the transcription of a human Oct-2 cDNA (Clerc et al., 1988) using T7 polymerase. A 10-ng aliquot of this RNA was added to mRNA samples from DRG cells cultured in the presence (+) or absence (-) of NGF. Following amplification with Oct-2-specific primers, the product from the control human Oct-2 RNA and from the rat Oct-2 mRNA in the sample was distinguished by restriction enzyme digestion with Bgll which cuts only the human Oct-2 PCR product to produce fragments of 142 and 108 base pairs, while not affecting the rat Oct-2 product. PCR amplifications from the 3' region of Brn-3 (30 s at 95 "C, 45 s at 57 "C, 1 min at 72 "C) used 50 pmol of primers 5'CAAATAGGTCT GCACTTATCCG3' and 5'TTGGATTATTAGTATGAGATACC3' in a 50-pl volume containing 0.4 units of Taq polymerase (Promega), 1 mM dNTPs, 50 mM Tris-HC1, pH 8.5, 50 mM NaC1, 5 mM MgC12, 2 mM DTT, and cDNA synthesized from 0.1 pg of DRG RNA reversetranscribed with superscript (BRL) and primed with random seprimers were 5'AACTACAACCACCTCATGCC3' and 5'ATCGCT quence hexanucleotides (Promega). Rat ribosomal protein L27 CCTCAAACTTGACCB'. The PCR reactions were analyzed by Southern blotting or spiked with 1 pCi of [ C~-~~P ]~A T P and reaction products separated on a 6% polyacrylamide gel, dried, and autoradiographed.

RESULTS AND DISCUSSION
Actions of NGF on POU Protein mRNA Levels-RNA was extracted from adult rat sensory neurons grown in the presence or absence of NGF, with anti-NGF antiserum added to NGF-free cultures to neutralize any NGF released from nonneuronal cells. Because of the limited amounts of material available from the DRG cultures and the low abundance of transcription factor mRNAs, the mRNAs were quantitated using PCR with oligonucleotide primers specific to each of the POU domain sequences of the Oct-1, Oct-2, or Brn-3 mRNAs. Because a variety of Brn-3-like clones with similar but not identical POU domain sequences have been identified in neuronal cells by PCR amplification (Latchman et al., 1992), we screened a rat DRG library using the originally described Brn-3 POU domain as a probe (He et al., 1989) and isolated and sequenced overlapping clones, in order to identify PCR primers uniquely directed at Brn-3 itself. Those clones that contained partial sequence with 100% homology to the original clone were used to probe Northern blots of DRG RNA. Both the original clone and a 3' overlapping clone hybridized to a 4-kb mRNA (Fig. 1). We therefore used additional primers across the 3"untranslated region of Brn-3 sequence derived from the new clone, as well as primers for the POU domain of Brn-3 to be certain that we were measuring Brn-3 mRNA levels. As an external control, primers specific for a constitutively expressed control mRNA encoding the ribosomal protein L27 were also used (Lebeau et al., 1991). In each experiment, the identity of the PCR product was confirmed both by digestion with appropriate restriction enzymes and by hybridization with Oct-1, Oct-2, or Brn-3specific cDNA clones. To ensure that the PCR was quantitative, preliminary experiments were carried out by varying the amounts of mRNA and cycle numbers to identify conditions in which the signal obtained was linearly related to the amount of input RNA. In addition, in some experiments we included equal amounts of a control human Oct-2 RNA template prepared by the transcription of a human Oct-2 plasmid whose PCR product could be distinguished from that of rat Oct-2 through sequence differences encoding distinct restriction enzyme recognition sites.
We found that NGF did not alter the levels of Oct-1 or Brn-3 mRNA, or that of the control ribosomal protein mRNA. In contrast, Oct-2 mRNA levels were elevated 2.7-3.5-fold in RNA isolated from three independent NGF-treated cultures when compared with RNA isolated from parallel cultures grown in the absence of NGF. The levels of amplification of the internal Oct-2 standard were identical, however, confirming that differences in Oct-2 mRNA levels obtained in the NGF-treated and untreated samples was not due to artefactual differences in amplification efficiency (Fig. 2).

NGF Effects on POU Protein
Leuekr-In order to test whether the alteration in mRNA levels was reflected in changes in the levels of Oct-2 DNA binding, nuclear extracts were isolated from adult sensory neurons grown in the presence or absence of NGF, and DNA mobility shift assays using the radiolabeled octamer ATGCTAATGATAT, which is a high-affinity binding site for both Oct-1 and Oct-2 , were used to quantitate the levels of octamer binding proteins. The DNA binding specificity of full length clones of Brn-3 has yet to be determined, precluding the measurement of this protein by DNA mobility shift assays. In these experiments the levels of a complex with a mobility identical to that produced by the Oct-2 protein also present in both B cells and a neuronal cell line were 3-4-fold greater in NGF-treated cells than in control cells cultured without NGF (Fig. 3A). The protein producing this complex was identified as the neuronal form of Oct-2 on the basis of its  2. a, PCR amplification of cDNA prepared from total RNA isolated from adult rat dorsal root ganglion cells cultured in the presence (+) or absence (-1 of NGF separated on a 1.8% agarose gel stained with 1 pg/ml ethidium bromide. Amplifications were carried out as previously described using oligonucleotides specific for the POU domain of either the Oct-1 ( I ) or Oct-2 ( 2 ) mRNAs (Lillycrop et al., 1991). Arrows indicate the positions of DNA size markers of the indicated sizes. The expected sizes of the Oct-1 and Oct-2 PCR products is 250 base pairs. b, PCR amplification of Oct-2 mRNA using a control internal template prepared by the transcription of a human Oct-2 cDNA, separated on a 1.8% agarose gel. The product from the control human Oct-2 RNA and from the rat Oct-2 mRNA in the sample was distinguished by restriction enzyme digestion with Bgll which cuts only the human Oct-2 PCR product to produce fragments of 142 and 108 base pairs, while not affecting the rat Oct-2 product. The 250-base pair product of the rat Oct-2 mRNA is indicated by arrows. c, PCR amplification of both Brn-3 170-bp cDNA fragment, and a ribosomal protein L27 130-bp cDNA used as a measure of the amount of input RNA (LeBeau et al., 1991). Amplifications were carried out using cDNA derived from 0.1 pg of RNA from DRG cultures grown in the presence (+) or absence (-1 of NGF for 22 ( C ) , 26 ( B ) , and 30 ( A ) cycles using PCR primers across the 3"untranslated region of Brn-3 (arrow B ) or from rat L27 primers (arrow L). PCR reactions were spiked with 1 pCi of [a-"PJdATP, and reaction products were separated on a 6% polyacrylamide gel, dried, and autoradiographed.

I"
FIG. 3. DNA mobility shift assay  using the octamer binding site ATGCTAATGATAT ( A ) or a binding site (5'CAAGAGTTCAAGACCAAC3' ( B ) ) for the RNA polymerase 111 transcription factor TFIIIC ( B ) and extracts prepared from DRG cells cultured in the presence (+) or absence (-) of NGF. Note the specific increase in the Oct-2 complex in the presence of NGF. The arrows indicate the positions of Oct-2 (02) and of an octamer-binding complex unaffected by NGF treatment. C, characterization of the NGF-induced octamer binding protein by DNA mobility shift assay using the octamer binding site ATGCTAATGATAT and extracts from Oct-2 containing rat B cells (lane I ) , from the Oct-2 containing neuronal cell line ND7 (lane 21, and from DRG cells cultured in the presence of NGF (lanes 3-6). Incubation with the DRG extracts were carried out in the absence of unlabeled competitor (lane 3) or in the presence of a 100-fold excess of the homologous octamer oligonucleotide competitor (lane 4 ) , an octamer oligonucleotide with a single base change (T-G a t position 11) which reduces Oct-2 binding (lane 5 ) , or the unrelated oligonucleotide containing the binding site for the SP1 transcription factor (lane 6). The arrows indicate the position of Oct-1 and Oct-2. sequence specificity for different octamer oligonucleotides which we have previously shown distinguish neuronal Oct-2 from the B cell form of Oct-2 and other octamer binding proteins . Thus in competition experiments the protein bound strongly to the octamer oligonucleotide ATGCTAATGATAT but less strongly to an oligonucleotide ATGCTAATGAGAT containing a single base change which reduces the binding of neuronal Oct-2 while not affecting binding of B cell Oct-2 or Oct-1 (Fig. 3C). The up-regulation of neuronal Oct-2 in the NGF-treated cells was specific to this factor, as the multiple complexes produced by association of the ubiquitous polymerase 111-associated transcription factor TFIIIC with its binding site were not diminished by removal of NGF (Fig. 3B). Similarly, no increase was observed in NGF-treated cells of the levels of a high mobility octamerbinding complex (arrowed in Fig. 3) which was specifically competed only by octamer oligonucleotide (Fig. 3C) and is therefore likely to be formed by one of the low molecular weight octamer-binding proteins identified by others in neuronal cell lines (Scholer et al., 1989). Moreover, on longer autoradiographic exposure, a low mobility octamer-binding complex identical in size to that produced by binding of Oct-1 was observed at equal levels in both the NGF-treated and untreated cells (data not shown). These results therefore demonstrate that the levels of Oct-2 protein are regulated by NGF at the mRNA level, while the related transcription factor Oct-1 remains unaffected.
NGF is known to up-regulate the mRNA for a number of transcription factors in PC12 cells, such as the NGFI-A zincfinger protein that is also induced in a variety of neuronal and nonneuronal cell types through activation of SRE-related sequences (Changelian etal., 1989). Brn-3, found at high levels in developing and adult DRG neurons was first isolated from rat brain by PCR and found to have structural similarities to the Unc-86 protein that determines neuronal cell fate in C. elegans. As sensory neurons are known to depend upon NGF for survival during development, it was of particular interest t o see if Brn-3 mRNA levels were up-regulated by NGF. The studies on adult neurons in culture here provide no evidence that this is the case, but the possibility that NGF regulates Brn-3 levels during development, which is not amenable to analysis in this experimental system remains an open question. DRG neurons also contain a number of POU proteins that are related to Brn-3 (Latchman et al., 1992). In this study we have used PCR primers derived from the POU domain of Brn-3, as well as primers derived from an overlapping 3' clone that hybridize to a 4-kb Brn-3 transcript in DRG RNA. There remains the possibility either that related Brn-3-like proteins are transcriptionally regulated by NGF, or that posttranscriptional regulation occurs; further sequence information about this family of proteins is required to assess these possibilities. The recent demonstration of differentially spliced homologous class IV POU proteins from Drosophila (I-POU and It-POU) with opposing effects on transcription (Treacy et al., 1992) suggest that an analogous situation could exist for Brn-3 like-proteins.
The role played by the Oct-2 protein in B cells in positively regulating the expression of immunoglobulin genes (Scheideret, 1987) suggests that neuronal Oct-2 may play a similarly significant role in gene regulation in sensory neurons and their response to NGF. Although the cellular genes regulated by Oct-2 in sensory neurons remain to be identified, one obvious candidate is the gene encoding the neuropeptide CGRP which contains two octamer motifs in its promoter (Broad et al., 1989) and whose mRNA increases in abundance in adult sensory neurons treated with NGF (Lindsay and Harmer, 1989). It is probable however, that Oct-2 may act primarily to inhibit rather than activate gene expression in neuronal cells. A number of differently spliced variants of Oct-2 have been identified (Wirth et al., 1991), and there is evidence that the forms expressed in neuronal cells may differ in their activity from those expressed in B cells . Thus unlike B cell Oct-2, neuronal Oct-2 has been shown to be ineffective at activating reporter constructs containing an octamer motif and can interfere with the activation of such constructs by Oct-1 , suggesting that Oct-2 acts as an inhibitor of octamer-mediated gene regulation. This potential inhibitory role of neuronal Oct-2 is of particular interest with respect to the mechanism of infection of these cells by herpes simplex virus (HSV). Sensory neurons support a latent form of asymptomatic HSV infection in an NGF-dependent manner (Wilcox and Johnson, 1988). Depleting the supply of nerve growth factor to the cells results in the re-expression of virus which lytically infects cells innervated by the infected neuron. Evidence for the existence of an NGF-regulated repressor of viral reactivation has been obtained in rat, monkey, and human sensory neurons (Wilcox et al., 1990). The failure of the HSV lytic cycle in neuronal cells with the consequent production of a latent infection has been shown to be dependent on the weak activity of HSV immediate-early promoters in sensory neurons, due to the presence of an inhibitory factor that binds to the viral regulatory octamer-related motif TAATGARAT. This inhibitory factor has been identified as neuronal Oct-2 by mobility shift assays, and elevation of the levels of Oct-2 in neuronal cell lines has been shown to exert an inhibitory action on immediate-early gene expression using reporter gene constructs (Kemp et al., 1990;Lillycrop et al., 1991). This inhibitory action reflects the competition for the viral octamer binding site by Oct-1 which forms a productive trans-activating transcriptional complex with the HSV virion protein Vmw65 and Oct-2 which does not (Gerster and Roeder, 1988). The demonstration that NGF up-regulates Oct-2 without altering the levels of Oct-1 thus provides a mechanism for the repression of HSV immediate-early gene expression and maintenance of latent infection by NGF and is consistent with the view that Oct-2 is the inhibitor of herpes virus expression. Failure of retrograde transport of NGF caused by tissue damage would lead to a fall in Oct-2 levels with a consequent activation of HSV immediate-early gene transcription and production of virus. It will therefore be important to determine the precise molecular structure of sensory neuron Oct-2 and to test the actions on Oct-2 mRNA levels of other neurotrophic factors such as NT3 and BDNF as well as mediators such as glucocorticoids that play important homeostatic roles and may also influence HSV expression.