The capsid and envelope of a virus protect the genome and enable efficient transmission of the virus from cell to cell and to new prospective hosts. Most common viral infections are spread by direct contact, by ingestion of contaminated water or food, or by inhalation of aerosolized particles. In all these situations, infection begins on an epithelial or mucosal surface and spreads along the mucosa and into deeper tissues. Infection may spread to cells that can enter blood vessels, lymphatics, or neural circuits. HBV, HCV, HTLV, and HIV are dependent on transmission by parenteral inoculation. Some viruses are transmitted only between humans. The dependence of smallpox virus and poliovirus infections on interhuman transmission makes it feasible to eliminate these viruses from human circulation by mass vaccination. Herpesviruses also survive by interhuman transmission but may be more difficult to eliminate because they establish persistent latent infection in humans and continuously reactivate to infect new and naïve generations.
Animals are also important reservoirs and vectors for transmission of viruses causing human disease. Insect vectors can mediate parenteral transfer of viruses that reach high titers in animal or human hosts. Arboviruses are parenterally transmitted from mammalian species to humans by mosquito vectors. Herpes B, monkeypox, rabies, and viral hemorrhagic fevers are other examples of zoonotic infections caused by direct contact with animals, animal tissues, or arthropod vectors.
Initial viral infections usually last for several days or weeks. During this period, the concentration of virus at sites of infection rises and then falls, usually to unmeasurable levels. The rise and fall of viral replication at a given site depend on local innate immune responses and the access of systemic antibody and cell immune effectors to the virus. Typically, primary infections with enteroviruses, mumps virus, measles virus, rubella virus, rotavirus, influenza virus, AAV, adenovirus, HSV, and VZV are cleared from almost all sites within 3–4 weeks. Some viruses are especially proficient in altering or evading innate and acquired immune responses. Primary infection with AAV, EBV, or CMV can last for several months. Characteristically, primary infections due to HBV, HCV, hepatitis D virus (HDV), HIV, HPV, and molluscum contagiosum virus (MCV) extend beyond several weeks. For some of these viruses (e.g., HPV, HBV, HCV, HDV, and MCV), the manifestations of primary infection are almost indistinguishable from the persistent phase.
Disease manifestations usually arise as a consequence of viral replication, infected-cell injury or death, and local inflammatory and innate immune responses. Disease severity may not necessarily correlate with the level of viral replication alone. For example, the clinical manifestations of intense primary infection with poliovirus, enterovirus, rabies virus, measles virus, mumps virus, or HSV at mucosal surfaces may be inapparent or relatively mild, whereas limited replication in neural cells can have dramatic consequences. Similarly, rubella virus or CMV infections in utero or neonatal HSV infections may have much more devastating effects than infections in adults.
Primary infections are cleared by nonspecific innate and specific adaptive immune responses. Thereafter, an immunocompetent host is usually immune to the disease manifestations of reinfection by the same virus. Immunity frequently does not prevent transient surface colonization on reexposure, persistent colonization, or even limited deeper infection.
PERSISTENT AND LATENT INFECTIONS
Relatively few viruses cause persistent or latent infections. HBV, HCV, rabies virus, measles virus, HIV, HTLV, HPV, HHVs, and MCV are notable exceptions. The mechanisms for persistent infection vary. HCV RNA polymerase and HIV reverse transcriptase are error-prone and generate variant genomes. Genome variation can be sufficient to permit evasion of host immune responses, thereby allowing persistent infection. HIV is also directly immunosuppressive, depleting CD4+ T lymphocytes and compromising CD8+ cytotoxic T cell immune responsiveness. Moreover, HIV encodes the Nef protein, which downmodulates MHC class I expression, rendering HIV-infected cells partially resistant to immune CD8+ T cell lysis.
DNA viruses have low mutation rates. Their persistence in human populations usually depends on their ability to establish latent infection in some cells, to reactivate from latency, and then to replicate at epithelial surfaces. Latency is defined as a state of infection in which virus is not replicating, viral genes associated with lytic infection are not expressed, and infectious virus is not made. The complete viral genome is present and may be replicated by cellular DNA polymerase in conjunction with replication of the cell’s genome. HPVs establish latent infection in basal epithelial cells. The latently infected basal cell replicates, along with the HPV episome, by using cellular DNA polymerase. Some of the progeny cells provide new latently infected basal cells, whereas others go on to squamous differentiation. Infected cells that differentiate to squamous cells become permissive for lytic viral infection. Herpesviruses establish latent infection in nonreplicating neural cells (HSV and VZV) or in replicating cells of hematopoietic lineages (EBV, CMV, HHV-6, HHV-7, and Kaposi’s sarcoma–associated herpesvirus [KSHV, also known as HHV-8]). In their latent stage, HPV and herpesvirus genomes are largely hidden from the normal immune response. Reactivated HPV and herpesvirus infections escape immediate and effective immune responses in highly immune hosts by inhibiting host innate immune and inflammatory responses. In addition, HPV, HSV, and VZV are somewhat protected because they replicate in the middle and upper layers of the squamous epithelium—sites not routinely visited by cells that mediate or amplify immune and inflammatory responses. HSV and CMV are also known to encode proteins that downregulate MHC class I expression and antigenic peptide presentation, enabling infected cells to escape recognition by and cytotoxic effects of CD8+ T lymphocytes.
Like other poxviruses, MCV cannot establish latent infection. This virus causes persistent infection in hypertrophic skin lesions that last for months or years. MCV encodes a chemokine homologue that probably blocks inflammatory responses, an MHC class I analogue that blocks cytotoxic T lymphocyte attack, and inhibitors of cell death that prolong infected-cell viability.
PERSISTENT VIRAL INFECTIONS AND CANCER
Persistent viral infection is estimated to be the root cause of as many as 20% of human malignancies. Cancer is an accidental and highly unusual or long-term effect of oncogenic human viral infection. With most “oncogenic viruses,” infection is a critical and ultimately determinative early step in carcinogenesis. Latent HPV infection can block cell death and cause cervical cells to proliferate. A virus-infected cell with an integrated HPV genome overexpressing E6 and E7 undergoes subsequent cellular genetic changes that enhance autonomous malignant cell growth.
Most hepatocellular carcinoma is believed to be caused by chronic inflammatory, immune, and regenerative responses to HBV or HCV infection. Epidemiologic data firmly link HBV and HCV infections to hepatocellular carcinoma. These infections elicit repetitive cycles of virus-induced liver injury followed by tissue repair and regeneration. Over decades, chronic viral infection, repetitive tissue regeneration, and acquired chromosomal changes can result in proliferative nodules. Further chromosomal mutations can lead to the degeneration of cells in a proliferating nodule into hepatocellular carcinoma. In rare instances, HBV DNA integrates into cellular DNA, promoting overexpression of a cell gene that can also contribute to oncogenesis.
Most cervical carcinoma is caused by persistent infection with “high-risk” HPV type 16 or 18. In contrast to HBV and HCV infections, which stimulate cell growth as a consequence of virus-induced cell death, HPV type 16 or 18 proteins E6 and E7 destroy p53 and pRB, respectively. Elimination of these key tumor-suppressive cell proteins increases cell growth, cell survival, and cell genome instability. However, like HBV and HCV infections, HPV infection alone is not sufficient for carcinogenesis. Cervical carcinoma is inevitably associated with persistent HPV infection and integration of the HPV genome into chromosomal DNA. Integrations that result in overexpression of E6 and E7 from HPV type 16 or 18 cause more profound changes in cell growth and survival and permit subsequent chromosomal changes that result in cervical carcinoma.
EBV is the most unusual oncogenic virus in that normal B cell infection results in latency with expression of viral proteins that can cause endless B lymphocyte growth. In almost all humans, strong CD4+ and CD8+ T cell immune responses to the antigenic EBV latent-infection nuclear proteins prevent uncontrolled B cell lymphoproliferation. However, when humans are severely immunosuppressed by transplantation-associated medication, HIV infection, or genetic immune deficiencies, EBV-induced B cell malignancies can emerge.
EBV infection also has a role in the long-term development of B lymphocyte and epithelial cell malignancies. Persistent EBV infection with expression of an EBV latency-associated integral membrane protein (LMP1) in latently infected epithelial cells appears to be a critical early step in the evolution of anaplastic nasopharyngeal carcinoma, a common malignancy in populations in southern China and northern Africa. Genomic instability and chromosomal abnormalities also contribute to the development of EBV-associated nasopharyngeal carcinoma. EBV is an important cause of Hodgkin’s lymphoma. High-level expression of LMP1 or LMP2 in Reed-Sternberg cells is a hallmark in up to 50% of Hodgkin’s lymphoma cases. LMP1-induced nuclear factor-κB (NF-κB) activity may prolong the survival of defective B cells that are normally eliminated by apoptosis, thereby allowing other genetic changes leading to the development of malignant Reed-Sternberg cells.
The HTLV-1 Tax and Rex proteins are critical to the initiation of cutaneous adult T cell lymphoma/leukemias that occur long after primary HTLV-1 infection. Tax-induced NF-κB activation may contribute to cytokine production, infected-cell survival, and eventual outgrowth of malignant cells.
Molecular data confirm the presence of KSHV DNA in all Kaposi’s tumors, including those associated with HIV infection, transplantation, and familial transmission. KSHV infection is also etiologically implicated in pleural-effusion lymphomas and multicentric Castleman’s disease, which are more common among HIV-infected than among HIV-uninfected people. KSHV also has a virus-encoded cyclin, an IFN regulatory factor, and a latency-associated nuclear antigen that are implicated in increased cell proliferation and survival.
Evidence supporting a causal role for viral infection in all of these malignancies includes (1) epidemiologic data, (2) the presence of viral DNA in all tumor cells, (3) the ability of the viruses to transform human cells in culture, (4) the results of in vitro cell culture–based assays that reveal transforming effects of specific viral genes on cell growth or survival, (5) pathologic data indicating the expression of transforming viral genes in premalignant or malignant cells in vivo, (6) the demonstration in animal models that these viral genes can cause malignant cell growth, and (7) the ability of virus-specific vaccines to reduce the incidence of virus-associated malignancy.
Virus-related malignancies provide an opportunity to expand our understanding of the biologic mechanisms important in the development of cancer. They also offer unique opportunities to develop diagnostics, vaccines, or therapeutics that could prevent or specifically treat cancers associated with viral infection. Widespread immunization against hepatitis B has resulted in a decreased prevalence of HBV-associated hepatitis and will probably prevent most HBV-related liver cancers. Current HPV vaccines can reduce rates of colonization with high-risk HPV strains and thereby decrease the risk of cervical cancer. The successful use of in vitro–expanded EBV-specific T cell populations to treat or prevent EBV-associated posttransplantation lymphoproliferative disease demonstrates the potential of immunoprevention or immunotherapy against virus-associated cancers.
RESISTANCE TO VIRAL INFECTIONS
Resistance to viral infections is initially provided by factors that are not virus-specific. Physical protection is afforded by the cornified layers of the skin and by mucous secretions that continuously sweep over mucosal surfaces. Once the first cell is infected, IFNs are induced and confer resistance to RNA virus replication. Viral infection may also trigger the release of other cytokines from infected cells. These cytokines may be chemotactic to inflammatory and immune cells. Viral protein epitopes expressed on the cell surface in the context of MHC class I and II proteins can stimulate the expansion of T cell populations with receptors that can recognize virus-encoded peptides presented on the cell surface by MHC class I proteins. Cytokines and antigens released by virus-induced cell death further attract inflammatory cells, dendritic cells, granulocytes, natural killer (NK) cells, and B lymphocytes to sites of infection and to draining lymph nodes. IFNs and NK cells are particularly important in containing viral infection for the first several days. Granulocytes and macrophages are also important in the phagocytosis and degradation of viruses, especially after an initial antibody response.
By 7–10 days after infection, virus-specific antibody responses, virus-specific human leukocyte antigen (HLA) class II–restricted CD4+ helper T lymphocyte responses, and virus-specific HLA class I–restricted CD8+ cytotoxic T lymphocyte responses develop. These responses, whose magnitude typically increases over the second and third weeks of infection, are important for rapid recovery. Also between the second and third weeks, the antibody type usually changes from IgM to IgG; IgG or IgA antibody can then be detected at infected mucosal surfaces. Antibody may directly neutralize virus by binding to its surface and preventing cell attachment or penetration. Complement can significantly enhance antibody-mediated virus neutralization. Antibody and complement can also lyse virus-infected cells that express viral membrane proteins on the cell surface. Cells infected with a replicating enveloped virus usually express the virus-envelope glycoproteins on the cell plasma membrane. Specific antibodies can bind to the glycoproteins, fix complement, and lyse the infected cell.
Antibody and CD4+/CD8+ T lymphocyte responses to viral infection can remain at high levels for several months after primary infection but usually wane over time. Low-level persistence of antibody-producing B lymphocytes and CD4+ or CD8+ T lymphocyte responses as memory cells can provide a rapid response to a second infection or an early barrier to reinfection with the same virus. Redevelopment of T cell immunity may take longer than secondary antibody responses, particularly when many years have elapsed between primary infection and reexposure. However, persistent infections or frequent reactivations from latency can result in sustained high-level T cell responses. EBV and CMV typically induce high-level CD4+ and CD8+ T cell responses that are maintained for decades after primary infection.
Some viruses have genes that alter innate and acquired host defenses. Adenoviruses encode small RNAs that inhibit IFN-induced, protein kinase R (PKR)–mediated shutoff of infected-cell protein synthesis. Adenovirus E1A can also directly inhibit IFN-mediated changes in cell gene transcription. Moreover, adenovirus E3 proteins prevent tumor necrosis factor (TNF)–induced cytolysis and block HLA class I synthesis by the infected cell. HSV ICP47 and CMV US11 also block class I antigen presentation. EBV encodes an interleukin (IL) 10 homologue that inhibits NK and T cell responses. Vaccinia virus encodes a soluble receptor for IFN-α and binding proteins for IFN-γ, IL-1, IL-18, and TNF, which inhibit host innate and adaptive immune responses. Vaccinia virus also encodes a caspase inhibitor that inhibits the ability of CD8+ cytotoxic T cells to kill virus-infected cells. Some poxviruses and herpesviruses encode chemokine-binding proteins that inhibit cell inflammatory responses. The adoption of these strategies by viruses highlights the importance of the corresponding host resistance factors in containing viral infection and the importance of redundancy in host resistance.
The host inflammatory and immune responses to viral infection do not come without a price. These responses contribute to the symptoms, signs, and other pathophysiologic manifestations of viral infection. Inflammation at sites of viral infection can subvert an effective immune response and induce tissue death and dysfunction. Moreover, immune responses to viral infection could, in principle, result in immune attack upon cross-reactive epitopes on normal cells, with consequent autoimmunity.
All human cells can synthesize IFN-α or IFN-β in response to viral infection. These IFN responses are usually induced by the presence of double-strand viral RNA, which can be made by both RNA and DNA viruses and sensed by double-strand RNA binding proteins (e.g., PKR and RIG-I) in the cell cytoplasm. IFN-γ is not closely related to IFN-α or IFN-β and is produced mainly by NK cells and by immune T lymphocytes responding to IL-12. IFN-α and -β bind to the IFN-α receptor, whereas IFN-γ binds to a different but related receptor. Both receptors signal through receptor-associated JAK kinases and other cytoplasmic proteins, including “STAT” proteins, which are tyrosine-phosphorylated by JAK kinases, translocate to the nucleus, and activate promoters for specific cell genes. Three types of antiviral effects are induced by IFN at the transcriptional level. The first effect is attributable to the induction of 2′-5′ oligo(A) synthetases, which require double-strand RNA for their activation. Activated synthetase polymerizes oligo(A) and thereby activates RNAse L, which in turn degrades single-strand RNA. A second effect results from the induction of PKR, a serine and threonine kinase that is also activated by double-strand RNA. PKR phosphorylates and negatively regulates the translational initiation factor eIF2α, shutting down protein synthesis in the infected cell. A third effect is initiated through the induction of Mx proteins, a family of GTPases that is particularly important in inhibiting the replication of influenza virus and vesicular stomatitis virus. These IFN effects are mostly directed against the infected cell, causing virus and cell dysfunction and thereby limiting viral replication.
A wide variety of methods are used to diagnose viral infection. Serology and virus isolation in tissue culture remain important standards. Acute- and convalescent-phase sera with rising titers of antibody to virus-specific antigens and a shift from IgM to IgG antibodies are generally accepted as diagnostic of acute viral infection. Serologic diagnosis is based on a more than fourfold rise in IgG antibody concentration when acute- and convalescent-phase serum samples are analyzed at the same time.
Immunofluorescence, hemadsorption, and hemagglutination assays for antiviral antibodies are labor-intensive and have been replaced by enzyme-linked immunosorbent assays (ELISAs), which generally use the specific viral proteins most frequently targeted by the antibody response. The proteins are purified from virus-infected cells or produced by recombinant DNA technology and are attached to a solid phase, where they can be incubated with serum, washed to eliminate nonspecific antibodies, and allowed to react with an enzyme-linked reagent to detect human IgG or IgM antibody specifically adhering to the viral antigen. The amount of antibody can then be quantitated by the intensity of a color reaction mediated by the linked enzyme. ELISAs can be sensitive and automated. Western blots can simultaneously confirm the presence of antibody to multiple specific viral proteins. The proteins are separated by size and transferred to an inert membrane, where they are incubated with serum antibodies. Western blots have an internal specificity control because the level of reactivity for viral proteins can be compared with that for cellular proteins in the same sample. Western blots require individual evaluation and are inherently difficult to quantitate or automate.
Isolation of virus in tissue culture depends on infection and replication in susceptible cells. Growth of virus in cell cultures can frequently be identified by effects on cell morphology under light microscopy. For example, HSV produces a typical cytopathic effect in rabbit kidney cells within 3 days. Other viral cytopathic effects may not be as diagnostically distinctive. Identification usually requires confirmation by staining with virus-specific monoclonal antibodies. The efficiency and speed of virus identification can be enhanced by combining short-term culture with immune detection. In assays with “shell vials” of tissue culture cells growing on a coverslip, viral infection can be detected by staining with a monoclonal antibody to a specific viral protein expressed early in viral replication. Thus, virus-infected cells can be detected within hours or days of inoculation, whereas several rounds of infection would be required to produce visible cytopathic effects.
Isolation of virus in tissue culture also depends on the collection of specimens from appropriate sites and the rapid transport of these specimens in appropriate medium to the virology laboratory (Chap. 7). Rapid transport maintains viral viability and limits bacterial and fungal overgrowth. Enveloped viruses are generally more sensitive to freezing and thawing than nonenveloped viruses. The most appropriate site for culture depends on the pathogenesis of the virus in question. Nasopharyngeal, tracheal, or endobronchial aspirates are most appropriate for the identification of respiratory viruses. Sputum cultures generally are less appropriate because bacterial contamination and viscosity threaten tissue-culture cell viability. Aspirates of vesicular fluid are useful for isolation of HSV and VZV. Nasopharyngeal aspirates and stool specimens may be useful when the patient has fever and a rash and an enteroviral infection is suspected. Adenoviruses can be cultured from the urine of patients with hemorrhagic cystitis. CMV can frequently be isolated from cultures of urine or buffy coat. Biopsy material can be effectively cultured when viruses infect major organs, as in HSV encephalitis or adenovirus pneumonia.
The isolation of a virus does not necessarily establish disease causality. Viruses can persistently or intermittently colonize normal human mucosal surfaces. Saliva can be positive for herpesviruses, and normal urine samples can be positive for CMV. Isolations from blood, cerebrospinal fluid (CSF), or tissue are more often diagnostic of significant viral infection.
Another method aimed at increasing the speed of viral diagnosis is direct testing for antigen or cytopathic effects. Virus-infected cells from the patient may be detected by staining with virus-specific monoclonal antibodies. For example, epithelial cells obtained by nasopharyngeal aspiration can be stained with a variety of specific monoclonal antibodies to identify the specific infecting respiratory virus. Antigen and serologic assays can be multiplexed to detect multiple analytes simultaneously by coupling of reagents to color-coded beads for each analyte and detection by flow cytometry.
Nucleic acid amplification techniques bring speed, sensitivity, and specificity to diagnostic virology. The ability to directly amplify minute amounts of viral nucleic acids in specimens means that detection no longer depends on viable virus and its replication. For example, amplification and detection of HSV nucleic acids in the CSF of patients with HSV encephalitis is a more sensitive detection method than culture of virus from CSF. The extreme sensitivity of these tests can be a problem, because subclinical infection or contamination can lead to false-positive results. Detection of viral nucleic acids does not necessarily indicate virus-induced disease.
Measurement of the amount of viral RNA or DNA in peripheral blood is an important means for determining whether a patient is at increased risk for virus-induced disease and for evaluating clinical responses to antiviral chemotherapy. Nucleic acid technologies for RNA quantification are routinely used in AIDS patients to evaluate responses to antiviral agents and to detect viral resistance or noncompliance with therapy. Virus-load measurements are also useful for evaluating the treatment of patients with HBV and HCV infections. Nucleic acid testing or direct staining with CMV-specific monoclonal antibodies to quantitate virus-infected cells in the peripheral blood (CMV antigenemia) is useful for identifying immunosuppressed patients who may be at risk for CMV-induced disease.
DRUG TREATMENT FOR VIRAL INFECTIONS
Multiple steps in the life cycles of viruses can be effectively targeted by antiviral drugs (Chap. 87). Nucleoside and nonnucleoside reverse transcriptase inhibitors prevent HIV provirus synthesis, whereas protease inhibitors block maturation of the HIV and HCV polyprotein after infection of the cell. Enfuvirtide is a small peptide derived from HIV gp41 that acts before cell infection by preventing a conformational change required for initial fusion of the virus with the cell membrane. Raltegravir is an integrase inhibitor that is approved for use with other anti-HIV drugs. Amantadine and rimantadine inhibit the influenza M2 protein, preventing release of viral RNA early during infection, whereas zanamivir and oseltamivir inhibit the influenza neuraminidase, which is necessary for the efficient release of mature virions from infected cells.
Viral genomes can evolve resistance to drugs by mutation and selection, by recombination with a drug-resistant virus, or (in the case of influenza virus and other segmented RNA viral genomes) by reassortment. The emergence of drug-resistant strains can limit the efficacy of antiviral therapy. As in antibacterial therapy, excessive and inappropriate use of antiviral therapy can select for the emergence of drug-resistant strains. HIV genotyping is a rapid method for identifying drug-resistant viruses. Resistance to reverse transcriptase or protease inhibitors has been associated with specific mutations in the reverse transcriptase or protease genes. Identification of these mutations by polymerase chain reaction amplification and nucleic acid sequencing can be clinically useful for determining which antiviral agents may still be effective. Drug resistance also can arise in herpesviruses but is a less common clinical problem.
IMMUNIZATION FOR THE PREVENTION OF VIRAL INFECTIONS
Viral vaccines are among the outstanding accomplishments of medical science. Smallpox has been eradicated except as a potential weapon of biological warfare or bioterrorism (Chap. 10). Poliovirus eradication may soon follow. Measles can be contained or eliminated. Excess mortality due to influenza virus epidemics can be prevented, and the threat of influenza pandemics can be decreased by contemporary killed or live attenuated influenza vaccines. Mumps, rubella, and chickenpox are well controlled by childhood vaccination in the developed world. Reimmunization of mature adults can be used to control herpes zoster. New rotavirus vaccines can have a major impact on this leading cause of gastroenteritis and prominent cause of childhood death worldwide. Widespread HBV vaccination has dramatically lowered the frequency of acute and chronic hepatitis and is expected to lead to a dramatic decrease in the incidence of hepatocellular carcinoma. The HPV vaccine was the first vaccine specifically licensed to prevent virus-induced cancer. Use of purified proteins, genetically engineered live-virus vaccines, and recombinant DNA–based strategies will make it possible to immunize against severe infections with other viruses. The development of effective HIV and HCV vaccines is complicated by the high mutation rate of viral RNA polymerase and reverse transcriptase, the population-based and individual divergence of HIV or HCV genomes, and repeated high-level exposure in some populations. Concerns about the use of smallpox and other viruses as weapons necessitate maintenance of immunity to agents that are not encountered naturally.
VIRUSES AS NOVEL THERAPEUTIC TOOLS OR AGENTS
Viruses are being used experimentally to deliver biotherapeutic agents or novel vaccines. Foreign genes can be inserted into viral nucleic acids, and the recombinant virus vectors can be used to infect the patient or the patient’s cells ex vivo. Retrovirus integration into the cell genome has been used to functionally replace the abnormal gene in T cells of patients with severe combined immunodeficiency, thereby restoring immune function. Recombinant adenovirus, AAV, and retroviruses are being explored for use in diseases due to single-gene defects, such as cystic fibrosis and hemophilia. AAV carrying a lipoprotein lipase gene is now being used in Europe to treat a rare lipid-processing disease and is the first gene therapy approved for clinical use. Recombinant poxviruses, adenoviruses, and influenza viruses are also being used experimentally as vaccine vectors. Viral vectors are being tested experimentally for the expression of cytokines that can enhance immunity against tumor cells or for the expression of proteins that can increase the sensitivity of tumor cells to chemotherapy. HSV deficient for replication in resting cells is being used to selectively kill proliferating glioblastoma cells after injections into CNS tumors. For improved safety, nonreplicating viruses are frequently used in clinical trials. Potential adverse events associated with virus-mediated gene transfer include the induction of inflammatory and antiviral immune responses. Instances of retrovirus-induced human malignances have raised concerns about the safety of retroviral gene therapy vectors.