Role of early genes in pathogenesis of adenovirus pneumonia

The role of human adenovirus early region 3 proteins (gp19K, 10.4K, 14.5K, and 14.7K) in a murine pneumonia model

Products of human adenovirus (Ad) early region 3 (E3) inhibit both specific (cytotoxic T lymphocytes [CTLs]) and innate (tumor necrosis factor alpha [TNF-alpha]) immune responses in vitro. The E3 gp19K protein prevents CTL recognition of Ad-infected fibroblasts by sequestering major histocompatibility complex class I proteins in the endoplasmic reticulum. E3 proteins 10.4K, 14.5K, and 14.7K function to protect infected cells from TNF-alpha cytolysis. To address the in vivo functions of these proteins, Ad mutants that lack the E3 genes encoding these proteins were inoculated intranasally into C57BL/10SnJ (H-2b) mice. Mutants that lack the gp19K gene failed to alter CTL generation or to affect Ad-induced pulmonary infiltrates. Since gamma interferon (IFN-gamma) is capable of overcoming gp19K suppression of CTL lysis in vitro, mice were depleted of IFN-gamma and inoculated with gp19K mutants. Even when IFN-gamma was depleted, gp19K was incapable of altering pulmonary lesions. These resuls are not in accord with the function of gp19K in vitro and suggest that gp19K does not affect immune recognition in vivo during an acute virus infection, yet they do not exclude the possibility that gp19K blocks immune recognition of Ad during a persistent infection. In contrast, when mice were inoculated with Ad mutants that lack the TNF resistance genes (14.7K and either 10.4K or 14.5K), there was a marked increase in alveolar infiltration and no change in the amounts of perivascular/peribronchiolar infiltration compared with wild-type-Ad-induced pathology. These findings demonstrate the importance of TNF susceptibility and TNF by-products for recruiting inflammatory cells into the lungs during Ad infections.

Safety of adenovirus-mediated transfer of the human cystic fibrosis transmembrane conductance regulator cDNA to the lungs of nonhuman primates

To define the toxicity of cystic fibrosis transmembrane conductance regulator gene (CFTR) gene therapy with a replication-deficient recombinant adenovirus (Av1Cf2) in a nonhuman primate model, 10(10) plaque forming units (pfu) were instilled directly through a bronchoscope into the right lung of 5 macaques, and a lower dose of 4 x 10(6) pfu was administered to the right lung of 1 macaque. One sham-treated control received phosphate-buffered saline (PBS). The macaques were evaluated sequentially by clinical examination, vital signs, weight, hematology, blood chemistry, chest radiography, pulse oximetry, and bronchoalveolar lavage (BAL) at baseline and 3-28 days post-treatment. After the period of observation, macaques were sacrificed for autopsy and histological examination. The macaques tolerated the experimental therapy clinically with no changes in body temperature, oxygen saturation, heart rate, body weight, or blood pressure. However, 1 macaque with visible evidence of aspiration at the time of initial bronchoscopy developed tachypnea with right lower lobe (RLL) pneumonia on chest radiograph and by histology. There were no changes in Hgb, Wbc, BUN, plasma electrolytes, bilirubin, or hepatic transaminases. In the macaques that received 10(10) pfu, there was a progressive increase in the number of CD8+ lymphocytes in BAL that was maximal at 28 days. Histological examination of the treated lungs of the high-dose macaques at 3 days showed marked peribronchial and perivascular cuffing by inflammatory cells and alveolar accumulation of neutrophils and macrophages. The alveolitis appeared to be resolving at 28 days, although the perivascular and peribronchial aggregates of mononuclear cells were still present. In the high-dose macaques, BAL interleukin-8 (IL-8) was increased at all time points (256-388 pg/ml versus 1-84 pg/ml at baseline and in control), whereas IL-1 beta was increased only at days 21 and 28 (341-852 pg/ml versus 30-92 pg/ml at baseline and in control). There were no increases in BAL cell counts, IL-1 beta or IL-8, and histological changes were mild in the macaque that received 4 x 10(6) pfu. Evaluation for Av1Cf2-derived human CFTR expression using RS-PCR demonstrated expression at 3, 10, and 21, but not 28 days in macaques treated with 10(10) pfu of Av1Cf2. In situ hybridization analysis demonstrated human CFTR mRNA in the alveolar regions of the lobes that received the vector at 10 and 21 days. There was no evidence of expression after treatment with 4 x 10(6) pfu. This study showed that high-dose adenoviral vector administration to the lung achieved CFTR gene transfer and expression but was associated with increased concentrations of cytokines in BAL and alveolar inflammation. A low dose, equivalent to the maximum clinical dose currently proposed for phase I trials in human subjects, was not associated with cellular or cytokine evidence of inflammation, and histological abnormalities were mild.

Gene therapy for cystic fibrosis: challenges and future directions

Images

FK506 immunosuppression to control the immune reactions triggered by first-generation adenovirus-mediated gene transfer

Despite good initial success in vivo, gene transfer using first-generation replication-defective adenovirus has been reported to lead to transient reporter gene expression and to trigger inflammatory reactions in various organs and animal models. To gain more knowledge on this phenomenon, immune reactions were investigated following in vivo transfection of adult immunocompetent mouse muscle using a delta E1/E3a adenoviral vector encoding a beta-galactosidase (beta-Gal) expression cassette. Cellular and humoral immune reactions, and rejection of beta-Gal-positive muscle fibers, occurred within 3 weeks. The muscles showed massive infiltration by macrophages, natural killer cells, and CD8+ leukocytes. The mRNA levels of granzyme B and interferon-gamma were increased 6 days after vector injection, indicating that the infiltrating lymphocytes were activated. Antibodies were formed against the adenovirus group antigen and the beta-Gal gene product 2 weeks after construct injection. The immunosuppressant FK506, however, blocked the cellular infiltration and the humoral response and allowed strong, stable transgene expression over 1 month. These data emphasize the immune problems related to the use of delta E1/E3a adenoviruses as vectors for gene therapy, and they underline the potential of FK506 as an immunosuppressant adjunct treatment for adenovirus-mediated gene transfer.

Adenovirus type 5 precursor terminal protein-expressing 293 and HeLa cell lines

HeLa and 293 cell lines that express biologically active adenovirus type 5 precursor terminal protein (pTP) have been made. The amount of pTP synthesized in these cell lines ranges from barely detectable to greater than that observed in cells infected with the wild-type virus. The pTP-expressing cell lines permit the growth of a temperature-sensitive terminal protein mutant virus sub100r at the nonpermissive temperature. A higher percentage of the stably transfected 293 cell lines expressed terminal protein, and generally at considerably higher levels, than did the HeLa cell lines. While 293 cells appeared to tolerate pTP better than did HeLa cells, high-level pTP expression in 293 cells led to a significantly reduced growth rate. The 293-pTP cell lines produce infectious virus after transfection with purified viral DNA and form plaques when overlaid with Noble agar after infection at low multiplicity. These cell lines offer promise for the production of adenoviruses lacking pTP expression and therefore completely defective for replication.

Expression of early and late adenoviral proteins in fatal adenovirus bronchopneumonia

The localization and distribution of three adenoviral proteins, hexon, E1A, and 55-kDa E1B, in 16 cases of fatal adenovirus bronchopneumonia in infants and children, are described. The proteins were immunohistochemically demonstrated in paraffin sections using monoclonal antibodies followed by the avidin-biotin-peroxidase method. The hexon antigen was present in inclusion-bearing bronchial, bronchiolar, and alveolar cells, mainly in the so-called rosette cells, as well as in necrotic debris in necrotizing areas. E1A antigen was also recognized in cells with nuclear inclusions where the reaction decorated the inclusion, nuclear chromatin, and cytoplasm but distributed mainly in alveolar cells and to a lesser extent in bronchial and bronchiolar cells. The 55-kDa E1B protein was extensively present in "activated," reactive-appearing, nuclei of bronchial, bronchiolar, and alveolar epithelial cells and in the cytoplasm of rare cells having nuclear inclusions. These activated nuclei did not stain for the other two antigens. "Smudge" cells reacted poorly or not at all with any of the antibodies. The reactivity found produced a sort of complementary pattern between the hexon-positive, inclusion-containing cells and the 55-kDa E1B-positive, inclusion-noncontaining cells. The relationships of present findings and virologic data are discussed.

Safety evaluation of Ad5CMV-p53 in vitro and in vivo

In preparation for a clinical trial of the recombinant p53 adenovirus Ad5CMV-p53 for the treatment of lung cancer, the potential adverse effects of Ad5CMV-p53 were assessed in vitro and in vivo. No infectious replication of Ad5CMV-p53 was detectable in HeLa cells infected with extracts from HeLa cells previously infected with Ad5CMV-p53. No Ad5CMV-p53 DNA replication was detected by 32Pi labeling in lung cancer cells infected with Ad5CMV-p53 at multiplicities of infection (moi) up to 1,000 pfu/cell (total of 5 x 10(9) pfu viruses). The infectivity and cytotoxicity of Ad5CMV-p53 were examined in vitro in normal human bronchial epithelial (NHBE) cells. At a moi of 50 pfu/cell, Ad5CMV-p53 infection and expression were detectable in 80% of the treated cells. The exogenous p53 protein was first detected by western blotting at 8 hr and peaked at 48 hr after infection. Growth of NHBE cells was not affected by Ad5CMV-p53 infection at a moi of 100 pfu/cell. The pathogenicity of Ad5CMV-p53 was assessed in BALB/c mice. The virus was given to four groups of mice by intratracheal injection at dosages from 10(7) to 10(10) pfu; a fifth group received phosphate-buffered saline alone. None of the viral injections proved to be lethal. Mild to moderate peribronchiolar and perivascular infiltration by mononuclear cells and lymphocytes, with patches of pneumonitis, was the most acute toxic effect detected by histologic analysis in the two high-dose groups. Immunohistochemical analysis of the same paraffin-embedded sections showed that infectivity and level of expression of p53 in lung tissue were dose-dependent. Our results demonstrate that Ad5CMV-p53 is a replication-defective virus that yields a relatively low degree of acute toxicity in mice; these data document a safety profile encouraging for clinical trials of Ad5CMV-p53 in the therapy of lung cancer.

Prolonged transgene expression in cotton rat lung with recombinant adenoviruses defective in E2a

Recombinant adenoviruses have tremendous potential for gene therapy of cystic fibrosis (CF) lung disease. First-generation recombinant viruses, rendered replication defective by deleting E1, have been associated with high-level recombinant gene expression in airway epithelial cells when administered directly to the lung. Experience in mice and non-human primates indicates that transgene expression is transient (i.e., lasting less than 21 days) and associated with the development of inflammation. We suggest an hypothesis to explain these findings that is based on expression of viral proteins in genetically modified cells that leads to destructive cellular immune responses and repopulation of lung epithelia with non-transgene-containing cells. This hypothesis has been evaluated in the current study using the cotton rat model. Instillation of the first-generation lacZ virus, H5.010CBlacZ, into cotton rat airway led to high-level gene expression in conducting and respiratory airway that was transient and associated with a substantial mononuclear, CD8-dominated, infiltrates. Treatment of the animals with cyclosporine blunted the inflammatory response and prolonged recombinant gene expression in both conducting and respiratory airways. Expression of viral early and late genes was detected in a subpopulation of lacZ-expressing epithelial cells of conducting airway and alveoli. Instillation of virus into cotton rat tracheal xenografts grown in athymic nu/nu mice led to efficient and stable transgene expression in the absence of pathology, underscoring the importance of T cell-mediated immunity. A recombinant adenovirus was constructed that is disabled in its capacity to replicate by the introduction of a temperature-sensitive mutation in the E2a gene as well deletion of E1 sequences. Instillation of this virus into cotton rat airway led to high-level transgene expression that was more stable than that achieved with the first-generation virus and was associated with less early and late gene expression as well as a diminished infiltration of CD8+ T cells in conducting airway epithelium. Interestingly, the introduction of the E2a mutation had no effect on the persistence of transgene expression, the pattern of late viral gene expression, nor the CD8+ T cell response within alveolar cells. These data suggest that cell-specific variation in the cell biology of recombinant adenoviruses exists in the lung. The present studies in cotton rats confirm the role of cellular immunity in the biology of adenovirus-mediated gene therapy to the lung and suggest that modifications in the design of recombinant adenoviruses to minimize or ablate transgene expression will be useful in improving the potential of this technology for gene therapy of CF.

Characterization of transgenic mice containing adenovirus early region 3 genomic DNA

Human adenoviruses (Ad) contain a complex transcription region (E3) which codes for proteins that interact with several arms of the immune system. However, E3 genes are not essential for replication in tissue culture. An E3-encoded 19,000-molecular-weight (19K) glycoprotein (gp19K) binds to the class I major histocompatibility complex (MHC) in the endoplasmic reticulum and prevents MHC transport to the cell surface. Three other E3 proteins are involved in the inhibition of apoptosis by tumor necrosis factor alpha. The entire E3 genomic DNA was utilized to produce transgenic mice to study the effect of the E3 proteins on pathogenesis of various infectious agents and to investigate the in vivo synthesis and processing of the multiple E3 mRNAs and proteins. There was basal expression of the E3 promoter in the thymus, kidneys, uterus, and testes and at all levels of the gastrointestinal tract. In addition, the E3 promoter of the transgene could be activated in some other organs, including the liver, by infection of these animals with an E3-deficient Ad (Ad7001) which contains a functional E1A region. Transactivation in vivo could also be demonstrated by infusion of bacterial lipopolysaccharide. There appeared to be differential ratios of expression between several of the E3 mRNAs in transgenic lung fibroblasts and primary kidney cells cultured from the transgenic animals. This observation suggested that there was differential mRNA splicing that was organ specific. These transgenic animals should provide a useful model for studying the effects of the E3 proteins on the immune system and on diseases affected either by control of MHC or by selected functions of tumor necrosis factor that are inhibitable by Ad E3 proteins.

Acute responses of non-human primates to airway delivery of an adenovirus vector containing the human cystic fibrosis transmembrane conductance regulator cDNA

Recombinant human adenovirus (Ad) vectors are leading candidates for human gene therapy for cystic fibrosis (CF) based on demonstration of efficient transfer of exogenous genes to rodent respiratory epithelium in vivo and human respiratory cells in vitro. The safety of Ad-mediated gene transfer to the respiratory epithelium and acute (up to 21 days) clinical responses to airway delivery of a replication-deficient recombinant, E1-, E3- Ad type 5-based vector containing the human cystic fibrosis transmembrane conductance regulator cDNA (AdCFTR) were evaluated in rhesus monkeys. Airway delivery of an Ad vector with the lacZ marker gene demonstrated beta-galactosidase expression in epithelial cells. Animals administered intratracheal AdCFTR demonstrated human CFTR cDNA expression in airway epithelial cells. Animals administered AdCFTR intranasal, and 24 hr later, intrabronchial [2 x 10(7) to 5 x 10(10) plaque-forming units (pfu), n = 12], in a fashion similar to a proposed human protocol, or only intrabronchial (10(11) pfu, n = 3), had no significant changes in clinical parameters compared to vehicle controls (n = 6). Microscopic analysis of the lung by necropsy or bronchoalveolar lavage demonstrated a dose-dependent increase in inflammatory cells, primarily lymphocytes, in the area where AdCFTR was delivered, which persisted for at least 2 months in some animals. Serum anti-Ad type 5 neutralizing antibody titers did not rise and shed Ad was not detected. The presence of AdCFTR DNA, analyzed by the polymerase chain reaction (PCR), was not detected in organs outside the lung. These data demonstrate that AdCFTR is well tolerated in non-human primates, although there is dose-dependent inflammation in the lung not clinically apparent.(ABSTRACT TRUNCATED AT 250 WORDS)

Inactivation of E2a in recombinant adenoviruses improves the prospect for gene therapy in cystic fibrosis

Although first generation recombinant adenoviruses, deleted of sequences spanning E1a and E1b, have been useful for in vivo applications of gene therapy, expression of the recombinant gene has been transient and often associated with the development of inflammation. We show that with first generation adenovirus-mediated gene transfer to the mouse lung, viral proteins are expressed leading to destructive cellular immune responses and repopulation of the lung with nontransgene containing cells. Second generation E1 deleted viruses further crippled by a temperature sensitive mutation in the E2a gene were associated with substantially longer recombinant gene expression and less inflammation. Stable expression of human CF transmembrane conductance regulator has been achieved in lungs of CF mice instilled with a second generation virus.

Adenovirus inhibition of cell translation facilitates release of virus particles and enhances degradation of the cytokeratin network

Infection of animal cells by a number of viruses generally results in an array of metabolic defects, including inhibition of host DNA, RNA, and protein synthesis, and morphological alterations known as cytopathic effects. For adenovirus infection there is a profound loss of cell structural integrity and a marked inhibition of host protein synthesis, the latter generally assumed necessary to enhance virus production. We examined the purpose of viral inhibition of cell translation and found that it was related in part to cytopathic wasting of infected cells. We show that viral shutoff of host translation promotes destruction of the intermediate filament network, particularly cytokeratins which are proteolysed at keratins K7 and K18 by the adenovirus late-acting L3 23-kDa proteinase. We found that if adenovirus is prevented from inhibiting cell translation, the intermediate filament network remains relatively intact, keratin proteins are still synthesized, and cells possess an almost normal morphological appearance and lyse poorly, reducing the release of nascent virus particles by several hundredfold. Remarkably, in tissue culture cells the accumulation of late viral structural proteins is only marginally reduced if host translation shutoff does not occur. Thus, a surprising major function for adenovirus inhibition of cellular protein synthesis is to enhance impairment of cellular structural integrity, facilitating cell lysis and release of progeny adenovirus particles.

Diversity of airway epithelial cell targets for in vivo recombinant adenovirus-mediated gene transfer

A variety of pulmonary disorders, including cystic fibrosis, are potentially amenable to treatment in which a therapeutic gene is directly transferred to the bronchial epithelium. This is difficult to accomplish because the majority of airway epithelial cells replicate slowly and/or are terminally differentiated. Adenovirus vectors may circumvent this problem, since they do not require target cell proliferation to express exogenous genes. To evaluate the diversity of airway epithelial cell targets for in vivo adenovirus-directed gene transfer, a replication deficient recombinant adenovirus containing the Escherichia coli lacZ (beta-galactosidase [beta-gal]) gene (Ad.RSV beta gal) was used to infect lungs of cotton rats. In contrast to uninfected animals, intratracheal Ad.RSV beta gal administration resulted in beta-gal activity in lung lysate and cytochemical staining in all cell types forming the airway epithelium. The expression of the exogenous gene was dose-dependent, and the distribution of the beta-gal positive airway epithelial cells in Ad.RSV beta gal-infected animals was similar to the normal cell differential of the control animals. Thus, a replication deficient recombinant adenovirus can transfer an exogenous gene to all major categories of airway epithelial cells in vivo, suggesting that adenovirus vectors may be an efficient strategy for in vivo gene transfer in airway disorders such as cystic fibrosis.

Pathogenesis of adenovirus type 5 pneumonia in cotton rats (Sigmodon hispidus)

Cotton rats (Sigmodon hispidus) were inoculated intranasally with 10(2.0) to 10(10.0) PFU of human adenovirus type 5. The virus replicated to a high titer in pulmonary tissues, with the peak titer being proportional to the input dose. The 50% lethal dose was 10(9.4) PFU. Histopathologic changes were proportional to the infecting inoculum and included the infiltration of interstitial and intra-alveolar areas, moderate damage to bronchiolar epithelium, and cellular infiltration of peribronchiolar and perivascular regions. These changes could be divided into two phases: an early phase (affecting alveoli, bronchiolar epithelium, and peribronchiolar regions) with an infiltrate consisting primarily of monocytes-macrophages and neutrophils, with occasional lymphocytes, and a later phase (affecting peribronchiolar and perivascular regions) with an infiltrate consisting almost exclusively of lymphocytes. In both phases, the predominant process was the response of the host to infection, rather than direct viral damage to infected cells. An infecting inoculum of 10(8.0) PFU or larger caused severe damage to type II alveolar cells, which were swollen, showed a loss of lamellar bodies, and were surrounded by polymorphonuclear leukocytes and macrophages. No evidence of complete viral replication was found in type II alveolar cells.

Murine adenovirus infection of SCID mice induces hepatic lesions that resemble human Reye syndrome

Murine adenovirus type 1 (MAV-1) infection of CB-17 SCID mice (which are homozygous for the severe combined immunodeficiency mutation) induces hepatic histopathologic and ultrastructural features that are strikingly similar to human Reye syndrome. Gross pathologic examination of MAV-1-infected mice revealed only pale yellow liver tissue. Histopathologic studies of tissue from MAV-1-infected mice revealed diffuse hepatic injury manifested by microvesicular fatty degenerative changes of hepatocytes and electron microscopic evidence of focal mitochondrial swelling with disruption of cristae and depletion of glycogen. Serum aminotransferase activities increased markedly in the infected animals; however, plasma ammonia levels were not elevated at the times assayed. Although all mice infected with MAV-1 died, neutralizing anti-MAV-1 monoclonal antibodies provided a dose-dependent delay in the appearance of clinical disease and hepatic histopathologic findings. Other findings included rare viral inclusions with only minimal inflammation in spleen, adrenal, and liver of infected mice. Our findings indicate that MAV-1 infection of SCID mice may provide important insights into the pathogenesis of the hepatic lesions of Reye syndrome.

A mouse model for investigating the molecular pathogenesis of adenovirus pneumonia

Intranasal inoculation of type 5 adenovirus (Ad5) produced pneumonia in mice even though the virus did not replicate. To induce the pneumonia, however, a large viral infectious dose was required--i.e., 10(10) plaque-forming units. Four strains of inbred mouse were studied (C57BL/6N, C57BL/10ScN, CBA/N, and C3H/N): all showed similar inflammatory responses, although the greatest infiltration occurred in the C57BL/6N mice. The pathological response to Ad5 infection resembled that previously described in cotton rats: it consisted of overlapping early and late phases, and the infiltration contained primarily lymphocytes and monocytes/macrophages with a scattering of polymorphonuclear leukocytes. The prominent early phase and the presence of polymorphonuclear leukocytes suggested that induction of cytokines may play an important role in the pathogenesis of this pneumonia. Assays showed the appearance of tumor necrosis factor alpha (TNF-alpha), interleukin 1 (IL-1), and IL-6 in the infected mouse lungs concomitant with the developing early-phase infiltration. Only IL-6 was found in the peripheral blood. IL-6 reached maximum titers 6-24 hr after infection, whereas maximum levels of TNF-alpha and IL-1 were attained 2-3 days after infection. Specific RNAs for each of these cytokines were demonstrated in the infected lungs. To test the hypothesis that a cytotoxic T-cell response was responsible for the second phase, which primarily consisted of a perivascular and peribronchial infiltration of lymphocytes, Ad5 was used to infect C57BL/10ScN Nu/Nu and parent mice. The nude mice showed a normal early-phase response, but essentially no peribronchial and only minimal perivascular infiltrations occurred.