Identification of virus resistant tumor cell subpopulations in three-dimensional uveal melanoma cultures.

To better understand melanoma resistance to herpes simplex virus type 1 (HSV-1)-mediated oncolysis, traditional two-dimensional (2D) cultures and extracellular matrix (ECM) containing three-dimensional (3D) cultures of OCM1 and C918 uveal melanoma cells were infected with an HSV-1 strain that expresses the green fluorescent protein (GFP) marker during replication. Although 2D cultures were completely destroyed within a few days of HSV-1 inoculation, viable GFP-negative tumor cells remained detectable in 3D cultures for several weeks. Tumor cells with increased resistance to HSV-1 included cells that formed vasculogenic mimicry patterns and multicellular spheroids and cells that invaded Matrigel individually. Mechanisms of tumor resistance against HSV-1 in the 3D environment included impaired virus spread in the ECM and ECM-mediated inhibition of viral replication after viral entry into tumor cells. Observations also suggested that HSV-1 established quiescent infection in some tumor cells present in multicellular spheroids and that this could revert to productive viral infection when the tumor growth pattern changed. These findings indicate that 3D tumor cell cultures can be used to identify distinct tumor cell populations with increased resistance to HSV-1 and to explore mechanisms of ECM-mediated tumor resistance to oncolytic virotherapy.

Double phakomatosis; neurofibromatosis type-1 and tuberous sclerosis.

Neurocutaneous syndromes represent some of the most common inherited disorders of the nervous system. Neurofibromatosis type-1 (NF-1) and tuberous sclerosis are well described. Yet, the presentation of both syndromes in the same patient is quite rare. We performed a thorough review of the literature of such double phakomatosis including pattern of inheritance. Eleven cases were reported in the literature. In addition we report a young patient who presented with clinical picture suggestive of both NF-1 and tuberous sclerosis, and present a radiographic and histopathological description of the case.

Expression of entry receptor nectin-1 of herpes simplex virus 1 and/or herpes simplex virus 2 in normal and neoplastic human nervous system tissues.

Herpes simplex virus 1 and/or Herpes simplex virus 2 (HSV) are important pathogens of human nervous system (NS) and genetically modified HSV strains have been proposed as vectors for gene therapy targeting the brain and brain tumors. Nectin-1 is an immunoglobulin-like adhesion molecule that participates in the formation of synapses and serves as an entry receptor for HSV. The expression pattern of nectin-1 in normal human NS and brain tumors is not well understood. To better understand the nectin-1 expression in normal and neoplastic human NS, immunohistochemistry was used to detect the nectin-1 expression in sections of normal human brain, spinal cord and trigeminal and dorsal root ganglia (n=10) and in sections of primary NS neoplasms (n=22). In normal human NS, nectin-1 was detected in the soma and processes of central and peripheral neurons, in ependymal cells, choroid plexus epithelial cells, vascular endothelial cells and meningothelial cells. Oligodendrocytes, astrocytes, vascular smooth muscle cells, and Schwann cells showed variable immunoreactivity. Among tumors, schwannoma, fibrous meningioma, and medulloblastoma were nectin-1 negative. Oligodendroglioma, ependymoma, pilocytic astrocytoma, pleomorphic xanthoastrocytoma, diffuse astrocytoma, anaplastic astrocytoma, glioblastoma multiforme and meningothelial meningioma showed weak focal nectin-1-positivity. Ganglion cells of ganglioglioma were strongly positive. These studies provide novel information about the expression of nectin-1 in normal and neoplastic NS, and thus may lead to a better understanding of cell targeting by HSV during HSV-induced neurological disease and during a HSV-based gene therapy.

Role of oxidative damage in the pathogenesis of viral infections of the nervous system.

Oxidative stress, primarily due to increased generation of reactive oxygen species (ROS) and reactive nitrogen species (RNS), is a feature of many viral infections. ROS and RNS modulate the permissiveness of cells to viral replication, regulate host inflammatory and immune responses, and cause oxidative damage to both host tissue and progeny virus. The lipid-rich nervous system is particularly susceptible to lipid peroxidation, an autocatalytic process that damages lipid-containing structures and yields reactive by-products, which can covalently modify and damage cellular macromolecules. Oxidative injury is a component of acute encephalitis caused by herpes simplex virus type 1 and reovirus, neurodegenerative disease caused by human immunodeficiency virus and murine leukemia virus, and subacute sclerosing panencephalitis caused by measles virus. The extent to which oxidative damage plays a beneficial role for the host by limiting viral replication is largely unknown. An enhanced understanding of the role of oxidative damage in viral infections of the nervous system may lead to therapeutic strategies to reduce tissue damage during viral infection without impeding the host antiviral response.

Molecular genetic and proteomic analysis of synchronous malignant gliomas.

Described is a patient with concurrent discrete gliomas: a pleomorphic xanthoastrocytoma with anaplastic features and an anaplastic oligoastrocytoma. The distinct and morphologically dissimilar tumors demonstrated similar genetic abnormalities by loss of heterozygosity and comparative genome hybridization. Clonality and proteomic analyses highlighted an independent origin for the two tumors. Proteomic methods may prove useful in cases where the differential diagnosis and pathogenetic origin of tumors are uncertain, as well as more globally for its ability to provide insight into specific expression of proteins that may serve as unique markers of tumorigenesis or as novel targets of therapy.

Chlamydia pneumoniae exacerbates aortic inflammatory foci caused by murine cytomegalovirus infection in normocholesterolemic mice.

Inflammatory foci induced by murine cytomegalovirus infection in normocholesterolemic mice were present temporarily in the aortic wall, but some of these foci developed into advanced lesions that persisted late after infection. The early foci induced by virus infection were significantly exacerbated following a single inoculation with Chlamydia pneumoniae.

Herpes simplex virus type 1 latency in the murine nervous system is associated with oxidative damage to neurons.

The pathological consequences of herpes simplex virus type 1 (HSV-1) latency in the nervous system are not well understood. To determine whether acute and latent HSV-1 infections of the nervous system are associated with oxidative damage, mice were inoculated with HSV-1 by the corneal route, and the extent of viral infection and oxidative damage in trigeminal ganglia and brain was assessed at 7, 90, and 220 days after inoculation. Abundant HSV-1 protein expression in the nervous system was observed in neurons and non-neuronal cells at 7 days after inoculation, consistent with viral replication and spread through the trigeminal and olfactory systems. Acute HSV-1 infection was associated with focal, neuronal and non-neuronal 4-hydroxy-2-nonenal- and 8-hydroxyguanosine-specific immunoreactivity, indicating oxidative damage. Rare HSV-1 antigen-positive cells were observed at 90 and 220 days after inoculation; however, widespread HSV-1 latency-associated transcript expression was detected, consistent with latent HSV-1 infection in the nervous system. HSV-1 latency was detected predominantly in the trigeminal ganglia, brainstem, olfactory bulbs, and temporal cortex. Latent HSV-1 infection was associated with focal chronic inflammation and consistently detectable evidence of oxidative damage involving primarily neurons. These results indicate that both acute and latent HSV-1 infections in the murine nervous system are associated with oxidative damage.

Roles of interleukin-12 and gamma interferon in murine Chlamydia pneumoniae infection.

BALB/c and strain 129 mice infected intranasally with Chlamydia pneumoniae displayed a moderate-to-severe inflammation in the lungs and produced interleukin-12 (IL-12), gamma interferon (IFN-gamma), tumor necrosis factor alpha (TNF-alpha), and IL-10, with peak levels on days 1 to 3 postinfection (p.i.), returning to basal levels by day 16 p.i. Anti-IL-12 treatment resulted in less-severe pathological changes but higher bacterial titers on days 3 and 7 p.i. By day 16 p.i., the inflammatory responses of control antibody-treated mice subsided. The bacterial titers of both anti-IL-12- and control antibody-treated mice decreased within 3 weeks to marginally detectable levels. Anti-IL-12 treatment significantly reduced lung IFN-gamma production and in vitro spleen cell IFN-gamma production in response to either C. pneumoniae or concanavalin A. In gamma-irradiated infected mice, cytokine production was delayed, and this delay correlated with high bacterial titers in the lungs. Following C. pneumoniae infection, 129 mice lacking the IFN-gamma receptor alpha chain gene (G129 mice) produced similar IL-12 levels and exhibited similarly severe pathological changes but had higher bacterial titers than 129 mice. However, by day 45 p.i., bacterial titers became undetectable in both wild-type 129 and G129 mice. Thus, during C. pneumoniae lung infection, IL-12, more than IFN-gamma, plays a role in pulmonary-cell infiltration. IFN-gamma and IL-12, acting mostly through its induction of IFN-gamma and Th1 responses, play an important role in controlling acute C. pneumoniae infection in the lungs, but eventually all mice control the infection to undetectable levels by IL-12- and IFN-gamma-independent mechanisms.

Divergence of brain prostaglandin H synthase activity and oxidative damage in mice with encephalitis.

Several lines of evidence point to inflammation and increased oxidant injury in brain regions of patients with Alzheimer disease (AD). Prostaglandin H synthase (PGHS) catalyzes the limiting step in prostaglandin synthesis and generates a potent oxidizing agent as by-product. One form of PGHS, PGHS-2, is induced by pro-inflammatory signals; thus leading to the 2-step hypothesis that pro-inflammatory signals in AD brain induce PGHS-2 that in turn contributes to brain oxidant injury. Here we have tested directly this 2-step hypothesis in a murine reovirus type 3 encephalitis model by measuring cerebral PGHS activity and quantifying oxidant injury. Our results showed a robust chronic inflammatory infiltrate and a 2-fold increase in PGHS activity in encephalitic mice compared with controls. Despite these changes, there was no significant increase in F2-isoprostanes or F4-neuroprostanes, accurate in vivo biomarkers of oxidant injury, and only minimal accumulation of protein adducts from the lipid peroxidation product 4-hydroxy-2-nonenal in the most intensely inflamed brain regions. These results challenge the proposal of others that pro-inflammatory induction of PGHS activity significantly contributes to oxidant injury in brain.

beta-Gal enzyme activity driven by the HSV LAT promoter does not correspond to beta-gal RNA levels in mouse trigeminal ganglia.

beta-Galactosidase enzyme expression can be detected in only a small percentage of trigeminal ganglia (TG) neurons acutely and latently infected with herpes simplex virus (HSV), in which the lacZ reporter gene was placed down-stream of the latency associated transcript (LAT) promoter at the LAT locus. However, DNA quantification suggests that a larger percentage of cells is infected than is expressing beta-galactosidase enzyme. To investigate the mechanism involved in regulation of genes expressed from the LAT promoter in trigeminal ganglia, in situ hybridization and histochemical staining assays were employed to determine on a cell-by-cell basis beta-gal gene expression both at the RNA and protein level. Using a LAT promoter-driven beta-gal construct in HSV-1 strain HFEM, it was found that there were 89-fold more cells positive for beta-gal transcript than cells positive for beta-gal enzyme in acutely infected trigeminal ganglia and a 10-fold difference in latently infected trigeminal ganglia. Thus, there is a discordance between beta-gal mRNA and beta-gal enzyme levels in HFEM/LAT-lacZ infected cells during acute and latent infection, and the beta-gal reporter gene activity does not faithfully compare the LAT promoter activity between acute and latently infected tissue. In contrast, in situ hybridization and histochemical staining assays were performed in mice acutely infected with a virus in which 140 bp of the LAT promoter sequences flanking the TATA element were replaced by 1.8 kbp of the neurofilament promoter (HSV-1 HFEM/NF-lacZ). This construct showed a correlation between beta-gal mRNA and enzyme expression in trigeminal ganglia in acute and latent infections. These findings suggest that sequences at the 5' end of the beta-gal transcript influence translation of the beta-gal message.

Oral inoculation of SCID mice with an attenuated herpes simplex virus-1 strain causes persistent enteric nervous system infection and gastric ulcers without direct mucosal infection.

BACKGROUND: After placement of herpes simplex virus type 1 (HSV-1) into the esophageal lumen of BALB/c mice, the virus replicated in enteric neurons within the esophagus and stomach and was transported to the sensory ganglia of the vagus nerve (nodose ganglia), where viral replication also occurs and where ultimately a long term latent infection is established. This described infection of immunocompetent mice primarily involved neuronal cells and associated satellite cells. EXPERIMENTAL DESIGN: Severe combined immunodeficient (SCID) mice were orally infected with an attenuated strain of HSV-1 to better identify sites of viral involvement in the gastrointestinal tract, particularly the mucosa. RESULTS: Three to five weeks after oral inoculation of SCID mice with HSV-1 strain in1814, a persistent viral infection of the gastrointestinal tract was established in most of the mice. Extensive viral replication was detected by immunohistochemistry throughout pathways of the vagus nerve and within the intrinsic enteric nervous system. Despite this ultimately fatal infection, viral replication in the gut occurred almost exclusively in enteric neurons and their processes; viral proteins were occasionally seen in smooth muscle cells immediately adjacent to heavily infected enteric ganglia. More than 50% of these persistently infected mice, when killed 18 to 31 days postinoculation, had gastric ulcers that were identified grossly and histologically. Only one of the 40 gastric ulcers was found to contain viral Ag. The remaining ulcers, although devoid of viral proteins, were found adjacent to virus-infected ganglia. CONCLUSIONS: HSV-1 can enter enteric neurons with minimal initial mucosal involvement, and once inside the nervous system, the virus is contained there despite the absence of a specific host immune response. Furthermore, chronically infected enteric neurons may provide an indirect mechanism for the pathogenesis of gastric ulcers in these immune-deficient mice.

Therapy of experimental human brain tumors using a neuroattenuated herpes simplex virus mutant.

BACKGROUND: Engineered herpes simplex virus (HSV) strains previously have been shown to offer a potential therapeutic alternative to conventional treatment modalities for brain tumors. Because HSV Type 1 strain 1716 has a deletion in the gamma 34.5 neurovirulence gene that renders it avirulent in the mouse central nervous system, we have assessed its potential to induce selective lysis of tumor cells versus neurons in vitro and in vivo. EXPERIMENTAL DESIGN: To do this, we studied parental HSV-1 strain 17+ and strain 1716 using human embryonal carcinoma cells (NT2 cells). These cells resemble neuronal progenitor cells and can be induced to differentiate into neurons (NT2N) with retinoic acid. Intracerebral grafts of NT2 cells into the brains of nude mice resulted in lethal brain tumors, and grafts of NT2N cells resulted in the integration of NT2N cells. RESULTS: In vitro studies showed that strain 1716 replicates in and spreads on monolayers of NT2 cells but not in NT2N cells. In vivo, strain 1716 replicated preferentially in NT2 tumors as evidenced by immunohistochemical staining for viral antigens, by in situ hybridization for HSV-specific transcripts, and by titration of virus from brains with tumor after intracranial injection of the virus into these mice. The temporal regression of NT2 tumors in mice treated with strain 1716 was demonstrated in vivo by magnetic resonance imaging. Electron microscopy and studies of DNA fragmentation suggested that regression of NT2 brain tumors in strain 1716-treated mice was mainly caused by a nonapoptotic, lytic mode of cell death. Finally, strain 1716-treated NT2 tumor-bearing mice survived more than twice as long as mock-treated tumor-bearing mice, and these differences in survival (25 vs. 9 weeks) were statistically significant (p < 0.03). CONCLUSIONS: We conclude from these studies that strain 1716 induces regression of human neural tumors established in the brains of nude mice, resulting in their prolonged survival.

An HSV-1 containing the rat beta-glucuronidase cDNA inserted within the LAT gene is less efficient than the parental strain at establishing a transcriptionally active state during latency in neurons.

The herpes simplex virus vector 17/LAT-RGUSB has previously been shown to express beta-glucuronidase enzyme activity stably in the trigeminal ganglia and brain stems of beta-glucuronidase-deficient mutant mice. However, the number of beta-glucuronidase expressing cells in trigeminal ganglia latently infected with 17/LAT-RGUSB was smaller than expected. Using normal mice for further characterization of 17/LAT-RGUSB latent infection, no appreciable differences were found between the vector and wild-type virus in: (1) their abilities to replicate in acutely infected ganglia; (2) their abilities to reactivate from latently infected ganglia: or (3) the quantities of viral DNA in tissues during the acute or the latent phases of infection. Using a minor LAT (mLAT)-specific probe to detect transcription by in situ hybridization, it was found that the intensity of the signal from individual cells latently-infected with 17/LAT-RGUSB or wild-type virus was similar. However, the vector-infected ganglia had only 20% as many positive cells as in wild-type infection. These data suggest that 17/LAT-RGUSB virus established latency similarly to wild-type virus, but that the LAT-promoter driven gene expression was compromised.

Replication, establishment of latent infection, expression of the latency-associated transcripts and explant reactivation of herpes simplex virus type 1 gamma 34.5 mutants in a mouse eye model.

The herpes simplex virus type 1 (HSV-1) gamma 34.5 gene is located within a region that is transcriptionally active during latent HSV-1 infection. To determine whether the gamma 34.5 gene deletion affects latency-associated transcript (LAT) gene expression or latent HSV-1 infection, a gamma 34.5 gene deletion mutant, 1716, and a stop codon insertion mutant, 1771, were studied in the mouse eye model. Although the gamma 34.5 gene is not essential, 1716 and 1771 replicated poorly in mouse eyes and trigeminal ganglia (TG). When mice were inoculated with 1716, infectious virus was detected in eyes only on the first day post-infection (p.i.), and was not detected at any time point in TG. Following inoculation with 1771, a small amount of virus was detected in the eyes on days 2 and 4 p.i., and in the TG of one animal on day 2 p.i. Reactivation of virus from mice latently infected with 1716 (0/30 TG) and 1771 (1/20 TG) was extremely low compared with the parental strain, 17+, and appropriate rescuants (80 to 100% reactivation), even though latent 1716 DNA was detected by PCR in 50% of TG. These results differ from those obtained following footpad inoculation; in the footpad there was limited 1716 replication and reactivatable latent infection was established in some dorsal root ganglia. The data support the hypothesis that the role of gamma 34.5 may be tissue and/or cell type specific. The synthesis, processing, and stability of the 2.0 kb LAT during 1716 and 1771 replication was not affected by these mutations in the gamma 34.5 gene. However, during latent infection of 1716 in mice the LATs were not detectable in TG by Northern blot, and were present in reduced amounts (approximately 10-fold less) during 1771 latency. The LATs from 1716 were barely detectable in a few neurons by in situ hybridization. Therefore, the gamma 34.5 gene might (i) affect replication in the eye, and reduce the amount of virus available to establish latent infection, be directly involved in (ii) establishment of latency, and/or (iii) the reactivation process.

In situ DNA PCR and RNA hybridization detection of herpes simplex virus sequences in trigeminal ganglia of latently infected mice.

The presence of herpes simplex virus (HSV-1) DNA in the trigeminal ganglia of latently infected mice was detected by an in situ DNA polymerase chain reaction (PCR), which includes a DNA:DNA hybridization step (indirect in situ PCR). These results were compared to the number of neurons possessing the HSV-1 latency associated transcript (LAT), as detected by in situ RNA hybridization with LAT probes. Sensitivity assays were shown to detect a single copy cellular gene in 48% of neuronal cell bodies. The results suggest that in situ PCR is an effective method to locate and detect HSV-1 within latently infected neurons. Moreover, the number of neurons found to be harboring HSV-1, by the method of in situ PCR, which does not depend upon virus gene expression, is within threefold of the number detected by in situ hybridization for LAT. Therefore, this report describes the first detection of HSV-1 DNA in latently infected murine trigeminal ganglia by the method of indirect in situ PCR, and compares the findings to the number of neurons expressing LAT, as assessed by conventional in situ hybridization.

Oral-oesophageal inoculation of mice with herpes simplex virus type 1 causes latent infection of the vagal sensory ganglia (nodose ganglia).

Herpes simplex virus type 1 (HSV-1) gingivostomatitis during childhood is known to result in a latent infection of the trigeminal ganglion neurons, which innervate the oral mucosa. During latency the viral genome is maintained in a non-infectious state. However, stimuli such as stress, fever or localized trauma can cause HSV-1 to reactivate in neurons and produce recrudescent disease in the peripheral tissues. Recently, HSV-1 proteins and nucleic acids have been detected in biopsies from human duodenal and gastric ulcers, raising the possibility that HSV-1 latency within the enteric nervous system is involved in this chronic recurrent gastrointestinal disorder. The studies in mice described here were done to determine whether HSV-1 latency could be established in neurons that innervate the murine gut. We found that after either intraperitoneal or oral-oesophageal inoculation of mice, HSV-1 establishes a latent infection in nodose ganglia of the vagus nerve, whose sensory neurons project to the gastrointestinal tract. This animal model of HSV-1 latency in the vagal sensory ganglia will be useful to examine the possible relationship between HSV-1 and recurrent gastrointestinal disease.

Long term herpes simplex virus type 1 infection of nerve growth factor-treated PC12 cells.

The behaviour of herpes simplex virus type 1 (HSV-1) strain 17 in tissue cultures of PC12 cells treated with nerve growth factor (NGF) was studied. PC12 cells respond to NGF by ceasing to proliferate and extending long neurites. After differentiation with NGF, cultures were infected with HSV-1 and maintained in the presence of the hormone for several weeks. These long-term infected cultures were tested for HSV DNA, transcripts and the ability to produce virus, before and after NGF removal. Before NGF removal, the cultures were characterized by little or no virus production and the presence of HSV-1 DNA in a predominantly endless form. In situ analysis of long-term infected cultures revealed latency-associated transcript expression in only a portion of the cells. However, as shown by an infectious centre assay, virus was present in almost all cells in the population. Moreover, removal of NGF from long-term cultures resulted in the appearance of significantly increased amounts of virus in the media. The degree to which this system resembles HSV latency in vivo is discussed.

Early region 3-replacement adenovirus recombinants are less pathogenic in cotton rats and mice than early region 3-deleted viruses.

BACKGROUND: Adenovirus type-5 (Ad5) recombinant viruses with replacement of the 1.9 kb XbaI fragment in the early region 3 (E3) by foreign genes have been constructed with the ultimate goal of inducing immune responses to the product of the inserted gene against a variety of virus infections. The pathogenicity of these recombinants, however, has not been studied. EXPERIMENTAL DESIGN: Histopathologic changes induced in cotton rat and mouse lung by E3-replacement-Ad5 recombinant or wild-type Ad (Wt-Ad) or E3-deleted mutant (Ad5-delta E3) viruses were compared. Expression of viral mRNA and replication of these viruses in cotton rat and mouse lungs, as well as in human tissue culture cells, were assayed. Expression of class I major histocompatibility complex antigens and the E3-14.7 kilodalton protein in virus-infected cells were also analyzed. RESULTS: An Ad5 recombinant, Ad-human cytomegalovirus glycoprotein B (Ad-HCMV.gB), in which the E3 region is replaced by the full-length gB gene of HCMV and with a genome size exceeding that of Wt-Ad, induced mild histopathologic responses in cotton rat and mouse lungs, comparable with those of Wt-Ad, but less severe than those of Ad5-delta E3. Analysis indicated that neither class I major histocompatibility complex expression on the cell surface nor differential expression of the protective E3-14.7 kilodalton protein underlies the pathologic differences observed in cells infected with Ad5-delta E3 or the Ad-HCMV.gB recombinant. In the mouse lung, another Ad-E3 replacement recombinant, Ad-herpes simplex glycoprotein B (HSV.gB), containing the complete HSV.gB gene and with a genome size larger than that of Wt-Ad, also induced a very mild inflammatory response. However, two recombinants with truncated forms of the HCMV.gB (Ad-HCMV.gB.155) or HSV.gB genes (Ad-HSV.gB.147) produced more severe histopathologic changes than the Wt-Ad or the recombinants with the full complement of HCMV.gB or HSV.gB genes. Ad5 and some of the recombinants replicated in mouse and cotton rat lung, and the extent of replication was inversely proportional to genome size, both in the lung and in human tissue culture cells. Infectious virus titers were, however, higher in cotton rat than in mouse lung. In situ hybridization analysis of cotton rat and mouse lung infected with Wt-Ad, Ad5-delta E3, or Ad-HCMV.gB virus revealed expression of Ad early/late mRNA predominantly in bronchial epithelial cells. CONCLUSIONS: These data not only confirm that E3-deleted viruses induce more severe pathologic changes in cotton rat lungs than Wt-Ad viruses (Ginsberg et al., Proc Natl Acad Sci USA 1989;86:3823-7) but led to the observation that some E3 replacement recombinants also lacking the expression of the 19 and 14.7 kilodalton proteins are significantly less pathogenic in cotton rats and mice than an E3-deleted virus. Pathogenicity and replication of the recombinant viruses inversely correlate with the genomic size.

The herpes simplex virus type 1 strain 17+ gamma 34.5 deletion mutant 1716 is avirulent in SCID mice.

Laboratory animal models are important tools for the identification of avirulent herpes simplex virus type 1 (HSV-1) strains which have potential for use in humans as vaccine strains or gene therapy vectors. We have studied an HSV-1 17+ variant, 1716, that has a deletion in the gamma 34.5 gene and which replicates poorly in the footpads of mice and is unable to grow in the mouse central nervous system or dorsal root ganglia (DRG) of the peripheral nervous system following peripheral inoculation. However, 1716 is known to be capable of establishing latent infections in the DRG of mice. Here we show that 1716 is avirulent after ocular infection and has low virulence after intracranial inoculation in SCID mice. Since SCID mice are much more sensitive to HSV-1 infection than immunocompetent mice, our results clearly demonstrate the drastically reduced virulence of the variant 1716 and provide additional support for the hypothesis that this variant would be avirulent in humans.