VIRUSES

INTRODUCTION

Technical Field

The field of this invention is the screening of drugs, treatment for virus infection and evaluation of attenuated virus strains.

Background

The field of medicine relies heavily on animal models. These models have relied on pathogens with a broad enough host range to include both human and animal hosts, or pathogens which were capable of causing an animal disease analogous to a human disease. There are many shortcomings with these techniques. Since the diseased cells and tissue are not human tissue, it is uncertain whether the animal cells can provide a reasonable prediction for the human correlate. With many diseases, the same lesion in animals and humans will result in different clinical symptoms. Furthermore, a number of pathogens are known to infect only humans. Thes study of these agents, and the development of therapies, are seriously hampered by the lack of a suitable animal model.

Primary infection with varicella-zoster virus (VZV) causes chicken pox (varicella). About 3.7 million Americans, mostly children, get chickenpox each year. 9,600 of these are hospitalized, and as many as 100 die from the

disease. Cell-associated viremia occurs during the incubation period of varicella and leads to the eruption of the widespread vesiculo-papular exanthem which characterizes this disease and distinguishes it from infection with the related virus, herpes simplex virus (HSV). The most serious complications occur when the virus invades the brain or lungs, causing encephalitis or pneumonia, or in children whose immune systems are compromised. After primary infection, the virus establishes latency in dorsal root ganglia and can reactivate years later to cause shingles (herpes zoster). Recently the FDA approved a live attenuated vaccine for Varicella. The vaccine is similar to one developed in Japan in the early 1970s.

The lack of a small animal model which mimics human disease is a major obstacle to the study of events in VZV pathogenesis. Guinea pigs infected with varicella-zoster virus (VZV) have been used as an animal model for chicken pox. However, the guinea pig model has a number of drawbacks. Not every animal becomes infected, there is little or no evidence of disease, i.e. vesicular lesions, and long-term latency does not seem to be established. There are no systems available for investigating VZV replication in epidermal cells. In view of current efforts being made to vaccinate children worldwide against VZV, an improved small animal model for infection is highly desirable. Having viable human tissue in an animal model provides numerous advantages. One can investigate the effect of agents on the tissue at various stages in the development of the disease. The interactions of cells, secreted agents and tissue can also be analyzed. Mutations introduced into the viral genome can be used to map the functions of viral proteins, and to determine which domains are responsible for various aspects of the infection, i.e. in establishing latency, transforming cells, viral replication, etc. A xenogeneic animal model further provides a means of testing the effect of infection on cells which are difficult to maintain in culture. Short-lived lymphocyte subsets,

neural cells, complex tissues, etc. which cannot easily be grown in culture for extended periods of time may be infected.

It is therefore of substantial interest to develop and provide animal models comprising human tissue that remains viable for extended periods of time and which is susceptible to infection by VZV. Such an animal model would permit investigation of the changes in the tissue, the etiology of disease and the effect of agents on pathogens.

Relevant Literature A description of the SCID-hu mouse may be found in J.M. McCune et al.

(1988) Science 241 :1632-1639: R. Namikawa et al. (1990) J. EXD. Med.

172:1055-1063 and J.M. McCune et al. (1991 ) Ann. Rev. Immunol. 9:395-429.

Immunocompromised mouse strains are described in S. Nonoyama et al.

(1993) J. Immunol 150:3817-3824: I. Gerling et al. (1994) Diabetes 43:433- 440; Bosma, et al. (1983) Nature 301 :52; and P. Mombaerts et al. (1992) Cell

68:869-877.

Infection of guinea pigs with varicella-zoster virus is discussed in P.

Lowry et al. (1993) J. Infect. Dis. 167:78-83; Y. Matsunaga et al. (1982) Infect.

Immun. 37:407-412; and Myers et al. (1979) J. Inf. Pis. 140:229-233. Varicella- zoster infection of human mononuclear cells is disclosed in Gilden et al.

(1992) Viral Res. 7:117-129. Investigation of varicella-zoster virus infection of human lymphocytes is disclosed in Koropchak et al. (1989) J. Virol. 63:2392-

2395. Lymphocyte associated viremia in varicella is discussed in Ozaki et al.

(1986) J. Med. Virol. 19:249-253. Vonsover et al. (1987) J. Med. Virol. 21:57-66 reports the detection of VZV in lymphocytes by DNA hybridization.

The generation of infectious varicella-zoster virus (VZV) and viral mutants from cosmid DNAs is described in Cohen and Seidel (1993) P.N.A.S.

90:7376-80. The viral genome was cut into 4 roughly equal, overlapping pieces with restriction enzymes, and cloned individually into cosmid vectors.

To regerate VZV, the cosmids are digested with restriction enzymes to linearize the VZV DNA fragments, and transfected together into a suitable cell line. The segments of the viral genome recombine in the cell and initiate the viral life cycle. After 7-10 days, viral plaques of clonally pure recombinant virus are seen. The serine protein kinase associated with varicella- zoster virus

ORF 47 is described in Ng and Grose (1992) Virology 191 :9-18. The complete DNA sequence of VZV may be found in Davison and Scott (1986) J. Gen. Virol. 67:1759-1816.

SUMMARY OF THE INVENTION

Methods and compositions are provided for investigating the response of human cells engrafted in a chimeric immunocompromised mammal to virus infection. This model system allows determination of tropism and infectivity for the different cell types that constitute specific human tissues in vivo and a comparison of different strains for their pathogenic potential. The infection of human lymphocytes and skin is of particular interest. The virus may be used directly, or regions of interest in the viral genome may be molecularly cloned into an assay vector. The screening is useful in the determination of peptides and protein for immunization, evaluation of attenuated strains, and for screening drugs that may block particular protein determinants. The sequences identified as conferring in vivo infectivity are used in drug screening and vaccine development.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a graph showing data from VZV replication in Thy/Liv implants. Figure 1A is a scattergraph showing the total number of infectious foci for each Thy/Liv implant infected with wild type (□) or Oka (•) strains of VZV. Figure 1 B shows the geometric means of the scatter data, the bars represent the standard error of the mean. Only positive samples were used to

calculate the data shown for day 14. Dotted lines represent the level of detection for the assay.

Figures 2A, 2B and 2C show a fluorescence activated cell sorter analysis of VZV-infected T-cells stained with antibodies specific for CD4, CD8 and VZV proteins. The cells in 2A are uninfected; 2B are infected with wild type-virus; and 2C are infected with Oka-strain virus.

Figure 3 summarizes the staining data from FACS analysis. The percentage of VZV-positive T-cells of each subpopulation are shown for implants infected with either wild type or Oka VZV strains.

DESCRIPTION OF THE SPECIFIC EMBODIMENTS Methods are provided for the investigation of human tropic virus infectivity and pathogenicity in an in vivo model system. Human cells that are capable of being infected by a virus are engrafted in an immunocompromised heterologous mammalian host, particularly a mouse, for extended periods of time. The method comprises engrafting one or more of normal human tissue, e.g. skin, non-dispersed fetal liver tissue and thymus tissue in juxtaposition, ganglia, or other suitable tissue in an appropriate site in an immunocompromised host. The xenografted human tissue is infected with virus, which then replicates in vivo.

The subject model provides a valuable diagnostic tool for evaluating strains of alphaherpesviruses, e.g. VZV, HSV-1 and HSV-2; and gammaherpesviruses, e.g. HHV-8 and human EBV; including vaccine strains, for pathogenicity and mechanisms of infection, therapeutic drugs, as a model for latency and the progression of infection with respect to different tissues.

Other herpesvirus of interest include HHV-7 and HHV-8. The alphaherpesvirus are of particular interest. These viruses spread widely to produce infection of the skin, mucus membranes and nervous system, and establish a life-long latent infection of sensory nerve ganglia, from which later

reactivation may occur. The system allows experimental manipulation and treatment of infection, so that human subjects need not be tested before a vaccine or therapy has been proven effective.

The subject animal model also allows evaluation of in vivo infectivity of non-retroviral virus isolates, where the in vivo infectivity of the virus is determined by its ability to grow in the xenografts. The in vivo virus replication may be compared to the in vitro replication of the virus. Systems for growth of viruses in in vitro cell culture are known in the art, and readily available. Conveniently, a cell line permissive for replication of the virus is infected the the test strain, and quantitation of virus replication is made from culture supernatants, cell lysates, etc.

Of interest are human tropic viruses, which infect or cause disease in human cells, many of which cannot easily be studied in conventional animal models. In general, human tropic viruses primarily cause productive infections, i.e. infection which results in viral replication and release of new infectious particles, in human cells. The virus may infect cells of closely related primate species, but will usually not cause the disease symptoms seen in humans, e.g. measles virus. The reason for such a tropism to human cells may be due to specific binding of the virus to a particular cell surface antigen required for entry of the virus into the cell.

Human tropic, non-retroviral viruses that may be evaluated in a comparison of in vitro, i.e. cell culture, and in vivo infectivity of interest include enteric viruses, e.g. coxsackievirus, echovirus, reovirus, hepatitis A virus; respiratory viruses, e. g. rhinovirus, adenovirus, coronavirus, parainfluenzavirus, influenzavirus; picomavirus; rhabdovirus; rubeola; poxvirus; herpesviruses; paramyxovirus; morbillivirus (measles), hepatitis B, C and D viruses. Viruses of particular interest for comparison of in vitro and in vivo infectivity include the alphaherpesviruses and gammaherpesviruses.

In one embodiment of the invention, cosmid vectors encoding portions of the VZV genome are used for screening particular regions of the VZV genome as to their contribution towards tropism and infectivity in vivo. The screening is useful in the determination of specific peptide and protein sequences useful for immunization, and for screening drugs that block specific regions of the viral proteins. Additionally, sequences may be subjected to directed or random mutation to generate libraries of altered sequences for screening.

The genome of VZV contains 71 open reading frames. The protein products include regulatory proteins (ORF62, ORF4, ORF63, ORF61 , and

ORF10), glycoproteins (gpA, gpB, gpC, gpD and gpE), ribonucleotide synthase, ribonucletide reductase, serine protein kinase, thymidylate synthetase, origin binding protein, etc. It has been reported that several of these proteins are not required for virus growth in vitro, including gpE, serine protein kinase and thymidylate synthetase. Evaluation of the role of these genes and proteins in in vivo pathogenicity is of interest for the understanding of the virus, and for development of vaccines and theraputics.

The virus may be wild type, e.g. clinical isolates, conventional strains, ere ; attenuated strains; or may be genetically engineered to enhance or reduce infectivity, pathogenicity, etc. Such modifications in the viral genome may include deletion of virulence genes, mutations in viral coat proteins that alter the host range, changes in viral nucleic acid polymerases, changes in transactivator genes, deletion of the serine protein kinase, etc. Mutations introduced into the viral genome are useful to map the functions of viral proteins, and to determine which domains are responsible for various aspects of the infection, e.g. in establishing or breaking latency, transforming cells, viral replication, etc.

The infectivity assay uses the human xenografts, e.g. thy/liv, skin, ganglia, ere. as a substrate for growth of virus. The method comprises

inoculating xenografts with an infectious level of a test virus and then determining the growth of the virus as a function of time. The ability of the virus to replicate in the host tissue is indicative of the potential for in vivo infectivity. The virus will usually be allowed to grow for an effective period of time, e.g. at least about 3 days, more usually about 1 week, and may grow for as long as four weeks. Data may be obtained as to the immune response of human cells to the virus; products that are secreted by infected or involved cells in response to infection, e.g. cytokines, interferons, antibodies, etc.; the viability and growth of the human epidermal cells, T cells (T lymphopoiesis), myeloid cells, and stromal cells that are present either in the implant or in the host circulation; and virus replication, e.g. production of new infectious particles.

Modes of administration are well-known to those of skill in the art. Certain viruses, e.g. VZV, may be administered by parenteral injection of infected cells suspended in a physiologically acceptable medium, where the injected cells will produce infectious virus over time. The cells will deliver a dose of virus of at least about 10 ² infectious units, preferably from about 10 ³ to 10 ⁵ infectious units of virus. Alternatively, infection may be achieved by direct injection of the virus suspended in a physiologically acceptable medium, either parenteral or injected into the xenograft. Usually, the injection will involve at least about 10 ² infectious units. The test virus may be an uncloned clinical isolate, a molecularly-cloned clinical isolate, a molecularly-cloned isolate with specific mutations, or a recombinant made by cloning a portion of DNA from one isolate into another. By "isolate" is intended a virus population having a substantially homogeneous DNA sequence, usually having fewer than about 10% variant sequences. The virus isolate may be isolated from biological fluids such as blood, cerebrospinal fluid, tears, saliva, lymph, organ or tissue culture derived fluids, and the like.

In one embodiment of the invention, a portion of the test virus genome will be used to replace the corresponding portion of the genome of an assay

vector. For example, the assay vector may comprise portions of the VZV genome cloned into cosmid vectors (see Cohen and Seidel (1993), supra.) The set of cosmid clones containing the complete genome are then used to transfect suitable host cells. The assay vector provides a uniform genetic background for testing the infectivity of different mutations. The assay vector may be cleaved with a restriction endonuclease to provide an insertion site for the replacement sequence.

The polymerase chain reaction is conveniently used to generate specific sequence changes and to amplify regions of naturally occurring mutations. The amplification primers may engineered to provide asymmetric restriction sites, and to amplify the sequence of interest and flanking sequences. The amplified DNA is then cleaved with the appropriate restriction endonucleases, and cloned into the assay vector. Directed or random mutations are made at specific sites by incorporation of changes into the primers used for amplification. Alternatively, a number of methods are known in the art for site directed mutagenesis.

The subject animal model may be used to study the effectiveness of known or potential anti-viral drugs or therapies, e.g. gene therapy. The administration of a drug may begin prior to, substantially concomitant with, or subsequent to the administration of the infectious dose of virus. Usually, administration of the drug will begin not earlier than 7 days prior to infection, and usually not later than about 7 days after infection, more usually not later than about 1 day after infection. However, after initial screening, different periods of time may be of interest in establishing the effectiveness of the drug in suppressing maintenance of virus, as may be determined by virus replication or cell depletion. Those compounds which are found to be effective in the assay are defined as those which prevent, either quantitatively or qualitatively, loss of viability of human cells, or which counteract virus-induced

suppression of hematopoiesis. Such compounds may then be used in the treatment of virus-infected individuals to enhance cell survival after infection.

The amount of drug which is administered will vary with the nature of the drug. The determination of how large a dosage to be used may be determined using the small animal model and relating the dosage based on pharmacokinetics, e.g. with equations predictive of interspecies scaling. Usually, the lowest effective dose will be used. The drug may be administered by any convenient route, orally or parenterally, e.g. intravascularly, or by inhalation, depot or the like. The effect of the drug may be monitored for any convenient time, usually at least 1 week from the initiation of administration of the drug, more usually at least 2 weeks, and at times for periods as long as 6 weeks or more.

Preferably, determinations will be made in the period from about 2-6 weeks.

Various measurements can be made as to the effectiveness of the drug in suppressing virus-induced tissue damage, cell death, thymocyte depletion and in maintaining T lymphopoiesis. Viral replication may be determined by plaque titration on Vero cell monolayers after coculture with cells from the sampled tissue, e.g. thymus, skin, ganglia, etc. By employing flow cytometry (fluorescence-activated cell scanning flow cytometry), one can analyze the cell surface expression of human markers in the skin graft, peripheral blood, neural tissue, the cell populations in a cell dispersion prepared from the thymic implant, or other human fetal tissue which is present, as appropriate. Infection may be correlated with the phenotype of human cells by analysis and staining for viral antigens and histological or flow-cytometric inspection of the human tissues for expression of human proteins, particularly such markers as CD4, CD8 and CD3 for T cells. One may also monitor for the presence of virus, by monitoring the level of viral proteins in the peripheral blood or the implant using an ELISA assay, or by monitoring the level of virus RNA or portion thereof, or virus DNA, using the polymerase chain reaction. In addition,

one may use histological analysis, employing immunochemistry, for detecting the presence of human specific markers or proteins of virus, and in situ hybridization for detecting the presence of viral genes that are present in the thymic implant. The subject animal model provides a convenient means of performing comparison studies on wild type and attenuated strains of a virus. Live attenuated strains are commonly used for vaccination, e.g. Oka-Merck strain of varicella-zoster virus, etc. Successful vaccination often requires a balance between an infection which allows long-lasting host immune protection and an infection which causes significant morbidity and/or mortality. After infection of a suitable number of animals with wild type and attenuated strains, determinations are made as to the effects of infection on human cells, as described above, e.g. histology, viability of various cell subsets, response to drugs and/or immunotherapy, viral replication, etc. The comparative studies are useful for determining how much damage is caused by attenuated strains, using the wild type strain as a reference point. This information further allows ranking of attenuated strains as to their virulence.

Comparisons may also be made of pathogenicity, infectivity, tropism, attenuation, etc. between related viruses, for example between different herpesviruses, particularly between the alphaherpesviruses, e.g. HSV-1 and

VZV.

The thymus and liver "Thy/Liv" implants are prepared by engrafting non-dispersed human fetal liver tissue and thymus tissue in juxtaposition, in an appropriate site in an immunocompromised host. The implants are found to include numerous subsets of cells, including immature hematopoietic progenitor cells, cells of the myelomonocytic lineage including polymorphonuclear granulocytes and eosinophilic granulocytes, cells of the megakaryocytic and lymphoid lineages, immature adipocytes, epithelial cells and stromal cells. The thymic microenvironment allows maturation of

thymocytes and phenotypically normal human T cells. The implants remain functional for periods up to about 18 months, usually at least about 6 months or more. Thus, the Thy/Liv tissue supports the presence and differentiation of the human pluripotential hematopoietic stem cell. The skin implants are prepared by engrafting non-dispersed human skin, usually human fetal skin, subcutaneously in an immunocompromised host. Ganglia implants are prepared from neural tissue, and may be implanted sub-cutaneously, or in the eye, e.g. the anterior chamber, sub- retinal space, etc. The human fetal tissue may be fresh tissue, obtained from an aborted donor within about 48 hours of death, or freshly frozen tissue within about 12 hours of death and maintained at below about -10°C, usually at about liquid nitrogen temperature (-150°C) indefinitely. The frozen tissue may be stabilized with suitable preservation agents and DMSO may be added. The tissue will generally be non-malignant, non-transformed tissue. The tissue may be from an organ implanted in a chimeric host, where the tissue may be removed from 2-4 months after implantation, or longer. In this manner, the tissue originally obtained from the host source may be greatly expanded, substantially increasing the total number of chimeric hosts which may be obtained. The tissue obtained from the chimeric host may be treated analogously to the tissue obtained from the human source.

The thymus and liver tissue are provided as solid tissue pieces of whole organs, and will include such stromal and epithelial cells as are normally present. To generate Thy/Liv, the two tissues are placed substantially contiguously. The hematopoietic component of the liver tissue is found to proliferate and differentiate within discrete subanatomic locations of the growing human fetal thymus, as evidenced by collections of cells of various hematopoietic lineages and stromal cells. The type, quantity and

spatial organization of these hematopoietic and stromal cells is similar to that found in normal human bone marrow.

For Thy/Liv, the tissue will generally be slices of a size in the range of about 0.5-6 mm, more usually 4-6 mm, with a thickness in the range of about 1 -2 mm for implantation with a 15- to 20-gauge trocar. Skin tissue will generally be full-thickness including epidermis and dermis, usually from about 1 to 5 cm ² in surface area. Generally, the fetal tissue will be of an age in the range of at least about 9 gestational weeks (g.w), generally about 9-24 g.w. Liver tissue will generally be of an age from about 10-24 g.w., while the age of the thymus tissue will generally be from about 9-24 g.w., more usually between about 18-20 g.w. Skin tissue may be from any fetal age of greater than about 9 g.w. While any vascularized convenient site for Thy/Liv implantation may be employed, of particular interest is the renal capsule, which provides a sanctuary for the tissue. Other sites include the splenic capsule, ear pinna, various subcutaneous locations and the intraperitoneal cavity. Methods of inserting tissue into the renal capsule have been described in the literature and are substantially described in EPA 88.312222.8, filed December 22, 1988.

Immunocompromised mammalian hosts suitable for implantation and having the desired immune incapacity exist or can be created. The significant factor is that the immunocompromised host is incapable naturally, or in conjunction with the introduced organs, of mounting an immune response against the xenogeneic tissue or cells. Therefore it is not sufficient that a host be immunocompromised, but that the host may not be able to mount an immune response after grafting, as evidenced by the inability to produce functional syngeneic host B cells, particularly plasma cells, and/or T cells, particularly CD4 ⁺ and/or CD8 ⁺ T cells after implantation. Of particular interest are small mammals, e.g. rabbits, gerbils, hamsters, guinea pigs, etc. , particularly murines, e.g. mouse and rat, which are immunocompromised due

to a genetic defect which results in an inability to undergo germline DNA rearrangement at the loci encoding immunoglobulins and T-cell antigen receptors.

Presently available hosts with this characteristic include mice that have been genetically engineered by transgenic disruption to lack the recombinase function associated with RAG-1 and/or RAG-2 (e.g. commercially available TIM™ RAG-2 transgenic), to lack Class I and/or Class II MHC antigens (e.g. the commercially available C1 D and C2D transgenic strains), or to lack expression of the Bcl-2 proto-oncogene. Of particular interest are mice that have a homozygous mutation at the scid locus, causing a severe combined immunodeficiency which is manifested by a lack of functionally recombined immunoglobulin and T cell receptor genes. The scid/scid mutation is available or may be bred into a number of different genetic backgrounds, e.g. C.B-17, ICR (outbred), C3H, BALB/c, C57BI/6, AKR, BA, B10, 129, etc. Other mice which are useful as recipients are NOD scid/scid; SGB scid/scid, bh/bh;

C.B-17 scid/hr; NIH-3 bg/nu/xid and META nu/nu. Transgenic mice, rats and pigs are available that lack functional T cells due a homozygous disruption in the CD38 gene. Immunocompromised rats include HsdHan÷RNU-rnu;

HsdHan:RNU- 7?_'/+; HsdHan:NZNU-rw; HsdHan:NZNU-m_ +; LEW/HanHsd- rnu; LEW/HanHsd-mi//+; WAG/HanHsd-/77_- and WAG/HanHsd-mϋ/+.

The host will usually be of an age less than about 25% of the normal lifetime of an immunocompetent host, usually about 1 to 20% of the normal lifetime. Generally, the host will be at least about six weeks old and large enough to manipulate for introduction of the donor human cells at the desired site. For example, mice are usually used at about 6 to 10 weeks of age.

Growth of the tissue within the host will vary with the organ.

The mammalian host will be grown in conventional ways. Depending on the degree of immunocompromised status of the mammalian host, it may be protected to varying degrees from infection. In some instances a sterile

- H -

environment or prophylactic antibiosis may be indicated. Prophylactic antibiosis may be achieved for scid/scid mice with 25-75 mg trimethoprim and 100-300 mg sulfamethoxazole in 5 ml of suspension, 0.125 ml of suspension per 4 ml of drinking water per mouse per day. Alternatively, it may be satisfactory to isolate the potential xenogeneic hosts from other animals in germ-free environments after cesarean derivation. The feeding and maintenance of the chimeric host will for the most part follow conventional techniques.

By employing varied DNA sequences from naturally occurring isolates or from sequences that have been modified in vitro, a library can be produced and screened for infectivity. In this manner one or a few consensus sequences for a particular viral protein of interest can be established that are shown to have high infectivity, as well as establishing those sequences that have low infectivity. The consensus sequences that have high infectivity can be used as antagonists to inhibit infectivity of VZV. In addition, the proteins associated with high infectivity may be used to screen antagonists, both peptidic and non-peptidic in the subject assay. The high infectivity sequences are useful in comparisons of drug or receptor binding to VZV proteins, to determine whether a particular binding agent will affect infectious virus. The structure of the high infectivity sequences may be determined, and used to model the interaction with potential therapeutic agents.

The peptides produced by the subject methods may be used as therapeutics or in the development of therapeutics. As therapeutics, the peptides will usually be at least about 8 amino acids and not more than 50, more usually not more than 36 amino acids. To enhance stability, the peptide may be terminally acetylated or amidated or the like. Alternatively, one or more of the amino acids may be substituted with the unnatural D-amino acid. Each of these compounds may be screened in accordance with the subject assay

Administration of peptides to a mammalian host is a well established procedure and does not require elaboration here.

The peptides may be used in competitive assays to evaluate drugs in their ability to inhibit binding to the cellular target. By employing helper T cells susceptible to infection with an infectious strain of VZV, a subject peptide and the candidate drug, one can evaluate the binding activity of the drug. For those candidates which effectively compete for a peptide identified in accordance with the subject invention, the candidates may be screened in the subject assay. The compounds having the desired pharmacological activity may be administered in a physiologically acceptable carrier to a host for treatment of infection. The inhibitory agents may be administered in a variety of ways, orally, topically, parenterally e.g. subcutaneously, intraperitoneally, intravascularly, etc. Depending upon the manner of introduction, the compounds may be formulated in a variety of ways. The concentration of therapeutically active compound in the formulation may vary from about 0.1-100 wt.%.

The following examples are offered by way of illustration and not by way of limitation.

EXPERIMENTAL

Example 1 Infection of Human Xenografts with VZV Varicella-zoster virus (VZV) is highly restricted in its host range, causing disease only in humans. Guinea pigs are useful to evaluate VZV immunity and some aspects of pathogenesis, but VZV replication is limited in this model. In order to facilitate the analysis of pathogenicity of wild type and genetically modified VZV, VZV replication was investigated in C.B-17 scid/scid

mice implanted with human fetal thymus and human fetal liver (Thy/Liv) under the kidney capsule, or with human fetal skin implanted subcutaneously.

SCID-hu mice. Male homozygous C.B-17 scid/scid mice were bred and maintained at SyStemix, Inc. At 8 weeks of age, co-implants of human fetal thymus and liver tissue from 18-23 week fetuses were introduced under the kidney capsule as a conjoint implant using an 18-gauge trocar. Skin was introduced subcutaneously as full-thickness dermal grafts. Human fetal tissues were obtained with informed consent according to federal and state regulations and were screened for HIV. Viral strains and culture conditions. A fresh clinical VZV isolate, designated strain Schenke, and VZV Oka strain (ATCC) were passed twice in human foreskin fibroblasts and twice in MRC-5 human lung fibroblasts and stored at -70°C in tissue culture media (MEM (Mediatech, Washington, D.C.) supplemented with 2mM L-glutamine (Gibco, Gaithersburg, MD), 50 I.U. penicillin, 50 μg streptomycin (Pen/Strep, ICN Biomedicals, Inc. Costa Mesa,

CA), 0.5 μg fungizone (Flow Laboratories, McLean, VA) with 10% DMSO. Animal inoculations were done with VZV-infected MRC-5 cells that were maintained in tissue culture media (TCM) with 10% fetal calf serum (FCS, Tissue Culture Biologicals, Tulare, CA), and used at 3-4+ cytopathic effects (CPE). The VZV-monolayer was trypsinized, the cells were counted, centrifuged, resuspended in TCM, and briefly stored on ice until they were injected into the SCID-hu mouse Thy/Liv implants. Mock-infected implants were injected with an equal number of uninfected MRC-5 cells prepared in the same manner. The virus inoculum was determined by plaque titration on Vero cell monolayers for each experiment. Vero cells were passed twice weekly and maintained in TCM with 5% FCS at 37°C in humidified 5% C0 ₂ . Unless otherwise stated, all chemicals used were obtained from Sigma Chemical Co., St. Louis, MO.

Infection of thymus/liver (Thy/Liv) implants. Experiments in SCID-hu mice with Thy/Liv implants were done 1 to 6 months after implantation. The mice were anesthetized with a solution of 5% ketamine (Aveco Co., Inc., Fort Dodge, IA) and 2.5% xylazine (LyphoMed, Inc., Rosemont, IL) (wt/vol in PBS, i.p.) and the left kidney was exposed surgically. The implant, which appeared as 100 mm ³ pale tissue, was inoculated with approximately 20-50 μl of a suspension of VZV-infected cells using a 27-gauge needle. The peritoneal incision was sutured, the skin was stapled closed, and the animals were observed daily for abnormal behavior or illness. At 2, 7, 14, and 21 days post inoculation the implants were removed. Half of each tissue block was fixed in

10% formalin; the other half was compressed and disrupted between ground glass slides. The cell suspension was filtered through a sterile nylon mesh to remove large debris and the released cells were counted, washed, and used for FACS analysis and virus titration. Infection of skin implants. Three to 5 weeks after implantation of human fetal skin, the mice were anesthetized, an 8 mm dorsal, mid-saginal incision was made, and the bilateral skin implants were exposed for inoculation with wild type VZV (clinical strain Schenke). One graft was inoculated by scraping the surface with a scalpel and covering it with 1 drop of the inoculum from a 27-gauge needle; 1-10 μl of the inoculum were injected into the other graft using a 27-gauge needle. Control implants were inoculated with an equal number of uninfected MRC-5 cells. The skin was then stapled closed to cover each implant. At 7, 14, and 21 days post inoculation the implants were dissected from the murine skin and divided, with one half being fixed in 10% formalin and the other half frozen in PBS (140 mM NaC1 , 2.7 mM KCI, 15 nM Na2HPθ4, 1.5 mM KH2PO4, pH 7.6) at -20°C.

Infectious focus assay. Cells from Thy/Liv implants or small pieces of skin implants were serially diluted 10-fold in TCM with 5% FCS. A 0.1 ml cell suspension, mixed with 1.5x10 ⁵ Vero cells in 0.9 ml TCM, was added to 24-

well dishes in triplicate. The dishes were incubated for 6 days at 37°C in 5% C 0 ₂ ; 1.0 ml fresh TCM with 5% FCS was added on Day 3. Following aspiration of the supernatant, the wells were flooded with crystal violet stain (5% ethanol, 5% formaldehyde, 0.13% crystal violet in PBS) for 2-5 min. The stain was aspirated, the wells were air dried, and the plaques were counted using an inverted light microscope (magnification x 40). The number of infectious foci per implant and the titer of the inoculum were calculated. The difference in virus titer was analyzed for statistical significance using the Student t test. A modification of the infectious focus assay was performed with 3.0 μm pore transwells (Costar Corp., Cambridge, MA). Cell suspensions of infected and uninfected Thy/Liv implants were obtained 7 days after inoculation and added to the transwell chamber at concentrations of 10 ⁵ cells in 0.1 ml TCM were placed in the transwell chamber. The transwells were placed in 24-well tissue culture plates seeded with Vero cell monolayers in 1.0 ml TCM and incubated for 6 days; the monolayers were stained with crystal violet and examined for characteristic VZV plaques by light microscopy.

FACS analysis and cell separation for infectious focus assay. Aliquots of approximately 10 ⁶ lymphocytes obtained from infected and uninfected Thy/Liv implants were washed in PBS with 2% FCS, resuspended in 100 μl PBS with

2% FCS in a 96-well V-bottom microtiter dish, and incubated on ice with VZV- immune or non-immune polyclonal IgG diluted 1 :40 in PBS with 2% FCS for 45 min. The IgG fractions of these antibodies were isolated from the sera of immune and non-immune human donors. The cell samples were washed by layering 100 μl FCS under the cells, centrifuging the plates and aspirating the supernatant. The cells were resuspended in 100 μl PBS with 2% FCS and incubated on ice with mouse monoclonal anti-CD4-phycoerythrin (PE) and anti-CD8-tricolor (TRI) conjugates (both diluted 1 :50), and goat anti-human- fluorescein isothiocyanate (FITC) conjugated F(ab')2 fragments (diluted 1 :100;

CalTag Laboratories, So. San Francisco, CA) for 45 min. The cells were washed again and resuspended in PBS with 2% FCS and 1 % paraformaldehyde. Cell suspensions were analyzed with a Becton/Dickinsoπ FacScan apparatus. To separate T cell subpopulations for the infectious focus assay, approximately 10 ⁶ lymphocytes from Thy/Liv implants were treated with anti- CD8-FITC and anti-CD4-PE antibody conjugates (both diluted 1 :50) on ice for 45 min. The cells were washed, resuspended in 0.5 ml PBS with 2% FCS, and sorted into CD4+, CD4+/CD8+, and CD8+ subpopulations with a Becton/Dickinson fluorescent activated cell sorter. Each sample was centrifuged, resuspended in 0.5 ml PBS with 2% FCS and titered on Vero cells using the infectious focus assay.

Histology and immunohistochemistry. Thy/Liv and skin implants were fixed in 10% formalin overnight, embedded in paraffin, cut into 3 μm sections and stained with hematoxylin and eosin (H&E). Unstained sections were deparaffinated in xylene for 3 minutes and rehydrated in graded ethanols before use. For immunohistochemistry, tissue sections were treated with a polyclonal human anti-VZV serum for 30 min; the polyclonal serum was obtained from a patient with herpes zoster infection and diluted 1 :500 in PBS with 2% FCS and 1% NaNβ. Sections were then rinsed twice with PBS and a secondary biotinylated goat anti-human antibody was added for 40 min. (diluted 1 :1000 in PBS with 2% FCS, 1%, NaN3), and rinsed again with PBS followed by TRIS buffered saline (TBS, 125 mM NaCI, 15mM TRIS, pH 8.0). Sections were then incubated in a streptavidin-alkaline phosphatase conjugate for 40 min, diluted 1 :400 in TBS with 2% FCS and 1 % NaNβ

(Jackson ImmunoResearch Labs, Inc., West Grove, PA) and rinsed twice with TBS. Red color was produced by flooding the tissue sections with a Fast Red substrate mix (2% dimethylformamide, 0.1% Fast Red, 0.02% Naphtol AS-MX phosphate, 100 mM TRIS, pH 8.2) for 20-40 min. The slides were rinsed with

water to stop the reaction, counterstained for 15 seconds with hematoxylin, air dried and covered with Crystal/Mount (Biomeda Corp., Foster City, Ca). After drying, the slides were coverslipped with Permount (Fisher Scientific, Fair Lawn, NJ) and examined by light microscopy. In situ hybridization. The VZV probe consisted of a 12.9 kb biotinylated plasmid, pVZV-C, that is a pBR322 vector carrying the Hindlll fragment C of VZV genomic DNA (Koropchak et al. (1991 ) J. Virol. 63:2392-2395). The probe was biotinylated using the BioNick kit (BRL, Gaithersburg, MD) according to the manufacturer's instructions and purified by ethanol precipitation. A negative control probe consisting of pBR322 vector alone was prepared in the same manner and used at the same concentration as the VZV-specific probe. The sectioned and deparaffinated tissue was treated with 3 μg/ml proteinase K for 30 min in a humidified chamber at 37"C and then rinsed with water. The hybridization solution (5X SSC, 1X Denhardt's solution, 100 μg/ml salmon sperm DNA, 80 ng of VZV probe DNA, and 40% formamide) was placed on the tissue section. A coverslip was placed on the tissue section and the DNA was denatured at 82°C for 10 min.

Following hybridization overnight in a humidified chamber at 37°C, the coverslip was removed with Triton buffer (50 mM TRIS base, 0.5 mM NaCI, 0.1 % Triton X-100, pH 7.5), the slide was rinsed with TBS, and the tissue was treated with streptavidin-alkaline phosphatase for 40 min, rinsed with TBS, then TRIS/saline (100 mM TRIS base, 100 mM NaCI, 50 mM MgCI _2l pH 9.5). Blue color was produced by adding an NBT/BCIP reagent for 20-40 min (0.0035% nitro blue tetrazolium chloride (NBT), 0.0017% 5-bromo-4-chloro-3- indoyl phosphate p-toluidine salt (BCIP, Amresco, Solon, OH), in TRIS/saline), then rinsing the tissue with water. The tissue sections were counterstained with hematoxylin, coated with Crystal/Mount, and coverslipped with Permount.

Preparation of VZV Inoculum MRC-5 cells were infected by adsorbing a frozen stock of VZV onto subconfluent cells in a T25 flask for 1 hour in

approx. 3 ml of medium (DMEM, 10% FCS). Virus suspension was drawn off and replaced with fresh media. The virus was then allowed to grow for 3-5 days until plaques and cytopathic effects are visible. The virus strain used was a fresh clinical isolate, (Schenke strain) or the Oka/Merck strain. The infected MRC-5 cells are then trypsinized, the contents of the flask added to a T175 flask containing a subconfluent layer of MRC-5 cells. The virus is adsorbed for one hour as described in the first step. Virus is grown until large plaques form, usually about 3 days. The infected cells in the T175 flask are trypsinized, and the entire contents placed in a new T175 flask. This disperses the infected cells and spreads the infection to all cells. The culture is grown overnight. The infected cells are trypsinized, and added to about 10 ml of fresh medium. The cells are centrifuged at 1200 rpm for 5 mins. and then resuspended in 1 ml of fresh medium. This cell suspension is used to inoculate the Thy/Liv implant. Each implant is injected with 20-50 μl of cell suspension.

Combined Immunohistochemical Staining and in situ Hybridization Tissue specimens were prepared by standard methods. Formalin-fixed and paraffin-embedded tissue sections were deparaffinated before use by baking overnight at 60-80° C, then dipping in xylene for 3 minutes, followed by dipping in 100; 95; 80; and 70% ethanol for three minutes each, then rinsed in water.

Results:

Thy/Liv tissue supports the growth of VZV and shows areas of tissue destruction that co-localize with viral antigen. MRC-5 fetal lung fibroblasts infected with VZV (low passage clinical isolate) were used as the inoculum because cell-free VZV preparations contain very low virus titers; approximately 1.5x10 ³ VZV-infected cells were injected into the implants. The implants were surgically removed 2, 7 and 14 days post-infection (p.i.).

On day 7 p.L, the implants were gently homogenized to release lymphoid cells from stromal components; these lymphoid cells were adsorbed onto Vero cell monolayers to detect VZV-infected cells by infectious focus assay. The titer of VZV-infected cells ranged from 4.8x10 ³ to 1.1x10 ⁴ PFU/implant. By day 14 p.L, the titer of VZV in the implants ranged from 9.5x10 ³ to 1.1x10 ⁴ PFU/implant. The results are shown in Table 1.

Table 1

Growth of Varicella-Zoster Virus in Thy/Liv Implants

Exp. no. 2 days post- 7 days post- 14 days post- infection infection infection

PFU/implant

1 6.0 x 10 ³ 9.5 x 10 ³

4.8 x 10 ³ 1.0 x 104

1.1 X 10 ⁴ 1.1 x 104

2 3.9 x lO ⁴ 2.5 x lO ⁴

5.1 x 104 3.5 x 104

3 1.7 X 10 ⁴ 1.5 x 104

2.3 X 10 ⁴

Histological analysis of Thy/Liv implants infected with VZV. The cytopathic effect (CPE) and the distribution of viral protein expression and viral DNA in infected implants were demonstrated by histological and immunohistochemical analysis. Hematoxylin and eosin (H&E) stained sections of VZV-infected Thy/Liv implants revealed progressive cytopathology over the 21 -day course of viral replication. On Day 2 post infection, very little evidence of CPE was seen in the lymphocytes or thymic epithelium. Macrophages had infiltrated the implants and engulfed cells which expressed VZV proteins as shown by immunohistochemical staining. On Day 7, the

lymphoid lobes showed areas of fibrosis, interstitial hemorrhage, and lymphocyte depletion that contrasted with the completely normal histologic appearance of uninfected implant tissue stained with H&E and the absence of any viral DNA signal or protein expression in these specimens. Necrosis was pronounced in infected implants by Day 14. In situ hybridization and immunohistochemical staining revealed extensive viral DNA and protein expression in the tissue sections. Viral protein expression was apparent in an even distribution throughout the lymphoid lobes of the implant but no VZV protein synthesis was detected in cells composing the thymic stroma or capsule structures. The in situ hybridization signal for viral DNA was strong but present only in areas where lymphocytes remained intact; clear regions without a VZV DNA signal were observed in the areas with the most severe necrosis. At Day 21 , the infected implants were involuted and shrunken; residual evidence of viral protein expression and viral DNA was noted in the least damaged areas. These histologic results parallel the observation that viral replication continued in Thy/Liv implants to Day 14, and decreased significantly by Day 21 post infection. Controls using non-immune serum for immunohistochemical stains and using the vector alone as a probe for in situ hybridization were negative when tested with infected implant tissue.

In order to determine the phenotype of cells infected by VZV, antibody staining and FACS analysis were performed on lymphocytes from the infected graft. The animals were infected as described above and five days later the graft was removed. The tissue was teased apart and the non-adherent cells were collected. Thymocytes were washed and counted, and the cells stained with (a) non-immune polyclonal human serum to determine background or (b)

VZV-immune polyclonal human IgG, as described in the materials and methods. The cells were also stained for the T cell markers CD4 and CD8, as described above.

The number of cells that stained positively for each antibody was determined by flow cytometry. The net % VZV infected cells was calculated as the difference between the number of cells staining with antibody (b) VZV- immune serum and the number of cells staining with (a) non-immune human serum. The data are presented in Table 2.

Table 2.

FACS Analysis of VZV-infected Thymocytes

Cell Type Net % VZV-infected

Uninfected MRC5 (fibroblast line) 0

VZV-infected MRC5 9.86

Uninfected Thy/Liv Lymphocytes

CD4 ⁺ CD8- 0

CD4 ⁺ CD8 ⁺ .49

CD4-CD8 ⁺ .87

VZV Infected Lymphocytes, mouse #1

CD4 ⁺ CD8- 5.02

CD4 ⁺ CD8 ⁺ 4.26

CD4-CD8 ⁺ 4.02

VZV Infected Lymphocytes, mouse

• ^• #2

CD4 ⁺ CD8- 14.57

CD4 ⁺ CD8 ⁺ 6 19

CD4-CD8 ⁺ 8.33

VZV Infected Lymphocytes, mouse #4

CD4 ⁺ CD8- 10.99

CD4 ⁺ CD8 ⁺ 9.06

CD4-CD8 ⁺ 10.05

These results demonstrate that lymphocytes in a Thy/Liv implant are infected with VZV, and that the CD4 ⁺ CD8-, CD4 ⁺ CD8 ⁺ , and C D4-C D8 ⁺ populations are all equally susceptible to infection.

Replication of wild type VZV and Oka strain in Thy/Liv implants. The replication of wild type VZV and Oka strain in Thy/Liv implants was measured over 21 days using the infectious focus assay. Lymphocytes were recovered from infected and uninfected implants; the number of infected cells was determined by plaque titration on Vero cell monolayers. The number of VZV- infected MRC-5 cells in the inoculum was measured in the same assay. In Figure 1A, the total number of infectious foci is shown for each Thy/Liv implant; at least 3 implants were analyzed at each time point. Although there was some variation between implants, the replication of wild type and Oka strains was comparable except on Day 14 post infection. Comparison of the geometric means of the titers (Figure 1 B) showed that wild type VZV replication peaked at Day 7 post infection (to a peak of 5.0 x 10 ⁴ ) and decreased at Day 14; no infectious virus was detectable by Day 21. Oka strain grew somewhat more slowly, reaching a peak titer of 8.0 x 10 ⁴ on Day 14, which was significantly higher than the wild type strain titer of 1.1 x 10 ⁴ on Day

14 (Student t-test, p=0.04). The failure to recover infectious virus by Day 21 from implants inoculated with wild type or Oka strains correlated with an extensive depletion of thymic CD4+ and CD8+ lymphocytes, as determined by histological and FACS analysis. No VZV was isolated from the uninfected control implants.

VZV release from infected Thy/Liv cells. The distribution of VZV in infected Thy/Liv implants observed by immunohistochemistry and by in situ hybridization indicated that the virus spreads by conventional cell fusion and also probably by transport of single particles to areas distant from the site of injection. Evidence for the release of free virus particles was obtained in the transwell assay, using VZV-infected Thy/Liv cells obtained at 7 days post infection; implant cells were separated from a Vero cell monolayer by a 3.0 μm pore membrane. VZV plaques were identified in three of three samples of 10§

lymphocytes from infected Thy/Liv implants. Uninfected lymphocytes and tissue culture media alone yielded no plaques when tested in the transwell assay.

FACS analysis of T cells from Thy/Liv implants. In order to identify the target cell populations that supported VZV replication in the Thy/Liv implants, cells from four implants infected with wild type or Oka strains and from uninfected controls were obtained at Day 7 post inoculation and analyzed by FACS using polyclonal VZV-immune serum. FACS analysis showed that viral protein expression was present in all T cell subpopulations, including CD4 ⁺ ,

CD8 ⁺ , and CD4 ⁺ CD8 ⁺ cells. The percentage of infected T cells within each subpopulation was comparable for wild type and Oka VZV strains. VZV protein expression reached a peak of 10-30% in each T cell subpopulation by Day 7 post infection and declined thereafter. Uninfected control cells incubated with VZV-immune IgG antibodies had similar fluorescence to the average background levels of 2.4-4.5% seen when infected cells were incubated with non-immune polyclonal human IgG.

The total number of lymphocytes in infected Thy/Liv implants decreased more than 10-fold by 7 days after inoculation, although the relative proportion of each T cell subpopulation remained unchanged compared to control tissues, as shown by the data in Table 3. Representative FACS plots from Thy/Liv implants infected with wild type or Oka VZV strains show that the ratios of each T cell subpopulation varied somewhat between implants yet were not significantly different from those of uninfected implants. The percentage CD4 ⁺ CD8 ⁺ T cells expressing VZV protein was 7.6% for wild type-infected cells, 28% for Oka-infected cells, and 1.7% for uninfected cells. While there was some variability between implants, the differences were not significant. The stability of the T cell distribution in conjunction with the FACS data

indicated that all T cell subpopulations were infected to a similar extent and that no particular subpopulation was a preferred target of VZV replication.

The frequency of VZV-infected cells among the sorted T cell subpopulations was 1-3 infected cells per 10,000 (Table 3). With such an infection rate, an implant containing 4.4 x 10 ⁴ lymphoid cells (the average number calculated at Day 7 post inoculation) would contain from 4.4 x 10 ⁴ to 1.3 x 10 ⁵ pfu of infectious virus. Thus titer corresponds to the average of 5.0 x 10 ⁴ infectious foci per implant when unsorted VZV-infected cells were cultured on Day 7 post infection.

Table 3.

Proportion of T cell subpopulations in Thy/Liv implants and the frequency of

^a The values do not add to 100% because CD4-/CD8- double-negative lymphocytes constitute a small fraction of the cells.

Histological analysis of VZV-infected skin implants. By Day 7 post infection, areas of the epidermis in VZV-infected skin implants showed thickening, the presence of multinucleated cells, and a mononuclear cell infiltrate. The uninfected implants retained their normal structure and appearance throughout the experiment. The severity and number of lesions was equivalent in the skin implants whether the virus was inoculated by injection or scarification. More advanced lesions were associated with disruption of the keratinized outer layer of the skin and with spread of the virus deeper into the dermis. In situ hybridization of these lesions revealed VZV

DNA within cells where tissue damage was the most obvious, as well as in the glandular cells and fibroblasts of the dermis. These cutaneous lesions showed a marked increase in size by Day 14 post infection; the infection penetrated into deeper layers of the dermis, balloon cells were prevalent, and acellular material from degenerated cells was enclosed by a keratinized surface layer.

These experiments demonstrate that the SCID-hu mouse model provides a useful system for investigating VZV interactions with cell types that are critical for VZV pathogenesis in the human host. Primary VZV infection is characterized by the occurrence of cell-associated viremia, followed by the appearance of widely distributed, cutaneous vesicles. In an immunocompromised host, failure to terminate viremia results in fatal varicella. Cell-associated virus can be documented by the recovery of infectious VZV and by in situ hybridization, Southern blot or polymerase chain reaction (PCR) to detect viral gene sequences. VZV tropism for human lymphocytes has been suspected based on isolation of infectious virus from the non-adherent fraction of peripheral blood mononuclear cells (PBMC) as well as by the morphology of PBMC that harbored viral DNA using in situ hybridization, but a more precise identification of the subpopulation of PBMC involved in VZV viremia has not been possible because of technical difficulties. In healthy individuals with varicella, VZV infects only about 1 in 100,000 PBMC. While PCR is very sensitive for demonstrating viremia, it cannot be used to define the cell type that is infected. The data from infected Thy/Liv mice show that the virus replicates very efficiently in both CD4+ and CD8+ T cells as well as immature CD4+/CD8+ T- cells in vivo. Experiments in the subject model system can also be used to explain the recovery of VZV from the adherent monocyte/macrophage fraction of PBMC. The subject data show that macrophages that had ingested one or

more VZV-infected lymphocytes were common throughout the thymic tissue sections. VZV replication to titers well above the inoculum occurred in T cells within the environment of the thymus over a shorter time interval than would be expected following infection of permissive human fibroblast cell lines in vitro. The frequency of productively infected cells is very low when primary cultures of PBMC or activated T lymphocytes are inoculated with VZV in vitro. The capacity to investigate VZV interactions with T lymphocytes in the subject animal model will be essential for experiments to identify VZV gene functions that account for this important tropism of the virus. The model will also permit useful comparisons of VZV T cell tropism with the pathogenicity of other human herpesviruses for this target cell population.

VZV is well known for its completely cell-associated replication pattern in vitro regardless of the cell line used for its propagation; infectious virus is not released into the supernatant even when cytopathic effects are extensive. Release by infected lymphocytes may be responsible for accelerating the spread of VZV through lymphoid and reticuloendothelial tissues. Transfer of VZV to epidermal cells may be enhanced by release of virus from infected T cells trafficking through skin capillaries.

Example 2

In Vivo Infectivity of VZV The varicella-zoster virus (VZV) open reading frame 47 (ORF47) protein kinase was previously reported to be dispensable for viral replication in vitro (Heineman and Cohen (1995) J. Virol. 69:7367-70). A VZV mutant was generated in which two contiguous stop codons were introduced into ORF47, thus eliminating expression of the ORF47 kinase (herein 47S). ORF47 kinase was not essential for the growth of VZV in cultured cells, and the growth rate of the VZV mutant lacking ORF47 protein was indistinguishable from that of parental VZV (ROka). Immunoprecipitation of ³² P-labeled proteins from cells

infected with parental virus and those infected with ORF47 mutant virus yielded similar amounts of the VZV phosphoproteins encoded by ORF4, ORF62, ORF63, and ORF68 (VZV gE), and the electrophoretic migration of these proteins was not affected by the lack of ORF47 kinase.

Replication of ROka and ORF 47s in thy/liv implants. In order to determine if the ORF 47 kinase was required for growth in T-cells, thy/liv implants were inoculated with a suspension of VZV-infected MRC-5 cells; the implant received either the parental VZV ROka strain or the VZV ORF 47 mutant strain 47S. The amount of input virus injected into the implants was calculated on Day 0 by titering the inoculum in a Vero cell plaque assay. On Days 7, 14, and 21 post inoculation, the implants were surgically removed and the T-cells were dispersed. Infected T-cells were titered in the Vero cell plaque assay and the number of infected T-cells per implant was calculated. The data is shown in Table 4. No mutant virus, strain 47S, was detected until Day 21 when one implant had a titer of 144 pfu. The ROka strain reached a peak on Day 14 and had kinetics typical of this strain. This experiment shows the ability of the SCID-hu thy/liv implants to reveal differences between VZV strains that are identical in tissue culture.

Flow cytometry: ROka- and 47s-infected T-cells. The percentages of CD4+ and CD8+ T-cells expressing VZV proteins from implants infected with either ROka or 47S were measured by flow cytometry. T-cells from infected implants harvested on Day 14 were prepared for flow cytometry using the following antibodies: polyclonal human anti-VZV IgG and anti-human FITC conjugate, anti-CD4 phycoerythrin (PE) conjugate, and anti-CD8 tricolor conjugate. The labeled T-cells were analyzed on a Becton-Dickinson FacScan flow cytometer and the percentages of VZV+, CD4+, and CD8+ T-cells were determined. In a representative sample of ROka-infected T-cells, 17.5% of the CD4+ cells and 13.4% of the CD8+ cells expressed VZV proteins. Only 1.5% of the CD4+ cells and 1.3% of the CD8+ cells infected with 47S expressed VZV proteins, which is only slightly above the 0.5% background. These results showed that the mutant VZV 47S strain was not able to infect T-cells in thy/liv implants and produce viral proteins.

Infectivity of a gC variant in vivo. Varicella and zoster are diseases characterized by a distinctive vesicular exanthem caused by VZV replication in the skin. Experiments described in Example 1 showed that wild type VZV replicated and produced vesicles in subcutaneous skin implants of the SCID-hu mouse that were indistinguishable from human biopsy material. It was not known whether the live vaccine strain, VZV Oka, or other laboratory passaged and mutant strains would also be able to replicate in skin implants.

A first experiment compared a wild type clinical isolate of VZV to the Oka vaccine strain. Skin implants were inoculated by directly injecting a suspension of VZV-infected MRC-5 cells; the implant received either wild type

VZV or Oka strain. The concentration of input virus was calculated on Day 0 by titering the inoculum in a Vero cell plaque assay. On Days 7, 14, and 21 post inoculation, the implants were surgically removed and cell lysate were prepared. The infected skin implants were minced to a paste, then sonicated

in a detergent lysis buffer to release cellular and viral proteins. Skin cell lysate and tissue culture grown virus as positive controls were separated by SDS-PAGE, transferred to nylon membranes, and viral proteins were detected by Western Blot using a VZV immune human polyclonal antiserum. The results showed that high molecular weight (more than 97 kD) viral proteins, an indicator of virion synthesis, were detected in implants infected with wild type VZV at low levels in 2 of 3 implants on Day 7. Increasing amounts of viral protein were detected in all wild type VZV-infected implants on Days 14 and 21. In contrast, Oka-infected skin implants had a low level of viral protein in only 1 of 3 implants on Day 14 and a large amount of viral protein in 1 of 2 implants on Day 21. The Oka vaccine strain was impaired in its ability to infect skin tissue compared to wild type VZV.

To determine if glycoprotein C of VZV was important for infectivity in skin, a naturally occurring gC-variant of VZV was tested in the SCID-hu model. This VZV gC- variant, L-N strain, does not express gC and was isolated by repeated plaque purification from its parent, VZV Ellen strain. HSV-1 gC is involved in adherence to polarized cells and therefore VZV gC may be necessary for replication in the highly polarized cell layers found in skin tissue.

Skin implants were inoculated by directly injecting a suspension of VZV-infected MRC-5 cells; the implant received either wildtype VZV, Oka, Ellen, or L-N strains. The concentration of input virus was calculated on Day 0 by titering the inoculum in a Vero cell plaque assay. On Days 7, 14, and 21 post inoculation, the implants were surgically removed and cell lysate were prepared. The infected skin implants were minced to a paste, then sonicated in a detergent lysis buffer to release cellular and viral proteins. Skin cell lysate and tissue culture grown virus as positive controls were separated by SDS- PAGE, transferred to nylon membranes, and viral proteins were detected by Western Blot using a VZV-immune human polyclonal antiserum. The results showed that high molecular weight (more than 91 kD) viral proteins, an

indicator of virion synthesis, were detected at high levels in implants infected with wild type VZY in 2 of 2 implants on Day 21 and at low levels in 1 of 2 implants infected with Oka. The Ellen strain, a laboratory strain with a long passage history in tissue culture, and the gC- L-N strain were unable to replicate in skin implants after 21 days. This experiment confirms the ability of the SCID-hu skin implant model to reveal differences in the pathogenicity of VZV strains that are identical in tissue culture.

Example 3 Comparison of HSV-1 and VZV Infectivity

To study differences in the cell tropism of VZV and HSV-1 , two closely related members of the alphaherpesviruses, thy/liv implants were inoculated with a suspension of HSV-1 infected MRC5 cells. The amount of input virus injected into the implants on Day 0 was calculated by titering the inoculum in a Vero cell plaque assay. On Days 2, 4, 6, and 7 post inoculation, implants were surgically removed and the lymphoid cells were dispersed. Infected cells and free virus were titered in the Vero cell plaque assay and the number of pfu per implant was calculated. Viral replication reached peak titer on Day 4 and slowly declined on Day 6 and 7, shown in Table 5. In situ hybridization using a biotinylated fragment of HSV-1 genome DNA as probe showed that virus was concentrated in the cortical epithelial cells. This is in contrast with the tropism of VZV for T-cells in thy/liv implants.

Table 5

Days post infection KOS

0 1.6x103

2 8.1x10 ⁶ 1.1x10 ⁸ 1.1x10 ⁷

4 6.5x10 ⁷ 1.2x10 ⁸ 2.3x10 ⁸

6 4.7x10? 8.9x10 ⁷ 4.1x10 ⁷

7 3.7x10 ⁷ 4.5x10? 1.2x10 ⁷ 4.4x10 ⁶

To determine if glycoprotein C of HSV-1 was necessary for infectivity in thy/liv implants, a gC deletion mutant was tested in the SCID-hu mouse model. Thy/liv implants were inoculated with a suspension of HSV-1 infected MRC5 cells; the implants received either the gC deletion strain, the gC+ revertant, or the parental KOS strain. The amount of input virus injected into the implants on Day 0 was calculated by titering the inoculum in a Vero cell plaque assay. On Day 4 post inoculation, implants were surgically removed and the lymphoid cells were dispersed. Infected cells and free virus were titered in the Vero cell plaque assay and the number of pfu per implant was calculated. There was no significant difference in the ability of the gC mutant to grow in thy/liv implants compared to the revertant or parent strain, indicating that gC was not important for HSV-1 replication in thymus tissue.

Table 5

Strain Inoculum pfu/implant Day 4

KOS 6.0x102 2.8x10$

Revertant 1.7x103 4.0x1 θ6 gC deletion 2.0x103 5.8x10$ 9.4x10^ 7.6x1 Oδ

It is evident from the above results that the subject invention provides for a method to evaluate viral infectivity. Attenuated viral isolates can be pre- evaluated in SCID-hu mice to confirm their replication capacity and pathogenicity. Identification of the attenuated gene functions will continue to help understand the disease process and rational design of live vaccines.

Direct investigation of VZV tropism for human epidermal cell subpopulations is possible in the subject animal model, as shown by the typical cutaneous lesions produced by VZV inoculation of skin implants. Both experimentally induced and naturally occurring VZV vesicles contain balloon cells with intranuclear inclusions, multinucleated giant cells, and a fibrinopurulent exudate covered by a keratinized roof.

Experiments in the subject model demonstrate that the Oka strain of VZV, which is used to prepare live varicella vaccine, is not altered in its tropism for CD4 ⁺ and CD8 ⁺ T cells. The initial replication of Oka strain was somewhat slower than the wild type strain but the infectious virus titers and cytopathic effects were comparable at later time points. The opportunity now exists to evaluate the replication of naturally occurring variants of VZV and genetically engineered recombinant strains of VZV in differentiated human cells in vivo.

All publications and patent applications cited in this specification are herein incorporated by reference as if each individual publication or patent application were specifically and individually indicated to be incorporated by reference. Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be readily apparent to those of ordinary skill in the art in light of the teachings of this invention that certain changes and modifications may be made thereto without departing from the spirit or scope of the appended claims.