Biology of adenovirus vectors with E1 and E4 deletions for liver-directed gene therapy.

Recombinant adenoviruses with E1 sequences deleted efficiently transfer genes into a wide variety of target cells. Antigen- and nonantigen-specific responses to the therapy lead to toxicity, loss of transgene expression, and difficulties with vector readministration. We have created new cell lines that allowed the isolation of more disabled adenovirus vectors that have both E1 and E4 deletions. Studies with murine models of liver-directed gene therapy indicated that the E1- and E4-deleted vector expresses fewer virus proteins and induces less apoptosis, leading to blunted host responses and an improved safety profile. The impact of the E4 deletion on the stability of vector expression was confounded by immune responses to the transgene product, which in this study was beta-galactosidase. When transgene responses were eliminated, the doubly deleted vector was substantially more stable in mouse liver than was the E1-deleted construct. These studies indicate that adenovirus vectors with both E1 and E4 deletions may have advantages in terms of safety and efficacy over first-generation constructs for liver-directed gene therapy.     

Effect of the E4 region on the persistence of transgene expression from adenovirus vectors.

The utility of adenovirus vectors for gene therapy is limited by the transience of expression that has been observed in various in vivo models. Immunological responses to viral targets can eliminate transduced cells and cause the loss of transgene expression. We previously described the characterization of an E4 modified adenovirus, Ad2E4ORF6, which is replication defective in cotton rats. We reasoned that gene transfer vectors based on Ad2E4ORF6 would have a reduced potential for viral gene expression in vivo which might be beneficial for achieving persistence of transgene expression. E1 replacement vectors expressing the cystic fibrosis transmembrane regulator or beta-galactosidase were constructed as series of vectors that differed with respect to the E4 region. Vectors containing a wild-type E4 region, E4 open reading frame 6, or a complete E4 deletion were compared in the lungs of BALB/c mice for persistence of expression. Results obtained with nude mice indicate that nonimmunological factors have a major influence on the longevity of transgene expression. Expression was transient from the E1a promoter with all vectors but persisted from the cytomegalovirus promoter only with a vector containing a wild-type E4 region. Transience of expression did not correlate with the disappearance of vector DNA, suggesting that promoter down-regulation may be involved. Coinfection studies indicate an E4 product(s) could be supplied in trans to allow persistent expression from the cytomegalovirus promoter. In summary, the choice of promoter is important for achieving persistence of expression; in addition, some promoters are highly influenced by the context of the vector backbone.     

Activation of transgene expression by early region 4 is responsible for a high level of persistent transgene expression from adenovirus vectors in vivo.

The persistence of transgene expression has become a hallmark for adenovirus vector evaluation in vivo. Although not all therapeutic benefit in gene therapy is reliant on long-term transgene expression, it is assumed that the treatment of chronic diseases will require significant persistence of expression. To understand the mechanisms involved in transgene persistence, a number of adenovirus vectors were evaluated in vivo in different strains of mice. Interestingly, the rate of vector genome clearance was not altered by the complete deletion of early region 4 (E4) in our vectors. The GV11 (E1- E4-) vector genome cleared with a similar kinetic profile as the GV10 (E1-) vector genome in immunocompetent and immunocompromised mice. These results suggest that the majority of adenovirus vector genomes are eliminated from transduced tissue via a mechanism(s) independent of T-cell, B-cell, and NK cell immune mechanisms. While the levels of persistence of transgene expression in liver or lung transduced with GV10 and GV11 vectors expressing beta-galactosidase, cystic fibrosis transmembrane conductance regulator, or secretory alkaline phosphatase were similar in immunocompetent mice, a marked difference was observed in immunocompromised animals. Levels of transgene expression initially from both GV10 and GV11 vectors were the same. However, GV11 transgene expression correlated with loss of vector genome, while GV10 transgene expression persisted at a high level. Coadministration and readministration of GV10 vectors showed that E4 provided in trans could activate transgene expression from the GV11 vector genome. While transgene expression activity per genome from the GV10 vector is clearly activated, expression from a cytomegalovirus promoter expression cassette in a GV11 vector appeared to be further inactivated as a function of time. Understanding the molecular mechanisms underlying these expression effects will be important for developing persistent adenovirus vectors for chronic applications.     

Molecular characterization of replication-competent variants of adenovirus vectors and genome modifications to prevent their occurrence.

Adenovirus (Ad) vectors for gene therapy are made replication defective by deletion of E1 region genes. For isolation, propagation, and large-scale production of such vectors, E1 functions are supplied in trans from a stable cell line. Virtually all Ad vectors used for clinical studies are produced in the 293 cell, a human embryonic kidney cell line expressing E1 functions from an integrated segment of the left end of the Ad type 5 (Ad5) genome. Replication-competent vector variants that have regained E1 sequences have been observed within populations of Ad vectors grown on 293 cells. These replication-competent variants presumably result from recombination between vector and 293 cell Ad5 sequences. We have developed Ad2-based vectors and have characterized at the molecular level examples of replication-competent variants. All such variants analyzed are Ad2-Ad5 chimeras in which the 293 cell Ad5 E1 sequences have become incorporated into the viral genome by legitimate recombination events. A map of Ad5 sequences within the 293 cell genome developed in parallel is consistent with the proposed recombination events. To provide a convenient vector production system that circumvents the generation of replication-competent variants, we have modified the Ad2 vector backbone by deleting or rearranging the protein IX coding region normally present downstream from the E1 region such that the frequency of recombination between vector and 293 cell Ad5 sequences is greatly reduced. Twelve serial passages of an Ad2 vector lacking the protein IX gene were carried out without generating replication-competent variants. In the course of producing and testing more than 30 large-scale preparations of vectors lacking the protein IX gene or having a rearranged protein IX gene, only three examples of replication-competent variants were observed. Use of these genome modifications allows use of conventional 293 cells for production of large-scale preparations of Ad-based vectors lacking replication-competent variants.     

Diminishing adenovirus gene expression and viral replication by promoter replacement.

The adenovirus E4 promoter was replaced by a synthetic promoter composed of a minimal TATA box and five consensus 17-mer yeast GAL4-binding-site elements. The viral vectors, which also contained human factor IX (hFIX) cDNA driven by Rous sarcoma virus long terminal repeat in the E1 region, were then constructed and expanded in 293 cells permanently expressing GAL4/VP16 fusion protein. Viral replication and expression of adenovirus E4 genes and late genes (hexon and fiber) were evaluated in vitro in the human lung carcinoma cell line H1299. Viral replication and viral gene expression were dramatically reduced in the cells transduced by vectors with a replaced E4 promoter compared to the levels in the cells transduced by vectors with the wild-type E4 promoter. The levels of transgene (hFIX) expression remained similar between vectors with or without E4 promoter replacement. These results indicate that diminution of viral gene expression and viral replication is achievable by promoter replacement.     

Production and Characterization of Improved Adenovirus Vectors with the E1, E2b, and E3 Genes Deleted

Adenovirus (Ad)-based vectors have great potential for use in the gene therapy of multiple diseases, both genetic and nongenetic. While capable of transducing both dividing and quiescent cells efficiently, Ad vectors have been limited by a number of problems. Most Ad vectors are engineered such that a transgene replaces the Ad E1a, E1b, and E3 genes; subsequently the replication-defective vector can be propagated only in human 293 cells that supply the deleted E1 gene functions in trans. Unfortunately, the use of high titers of E1-deleted vectors has been repeatedly demonstrated to result in low-level expression of viral genes still resident in the vector. In addition, the generation of replication-competent Ad (RCA) by recombination events with the E1 sequences residing in 293 cells further limits the usefulness of E1-deleted Ad vectors. We addressed these problems by isolating new Ad vectors deleted for the E1, E3, and the E2b gene functions. The new vectors can be readily grown to high titers and have several improvements, including an increased carrying capacity and a theoretically decreased risk for generating RCA. We have also demonstrated that the further block to Ad vector replication afforded by the deletion of both the E1 and E2b genes significantly diminished Ad late gene expression in comparison to a conventional E1-deleted vector, without destabilization of the modified vector genome. The results suggested that these modified vectors may be very useful both for in vitro and in vivo gene therapy applications.     

A helper virus-free packaging system for recombinant adeno-associated virus vectors

Adeno-associated virus (AAV) is a human parvovirus that is currently receiving widespread attention for its potential use as a gene therapy vector. Construction of the recombinant AAV vector (rAAV) involves replacing most of the viral genome with a transgene of interest and then packaging this recombinant genome into an infectious virion. Most current protocols for generating rAAV entail the co-transfection of a vector plasmid and a packaging plasmid that expresses the viral replication and structural genes onto adenovirus (Ad) infected cells growing in culture. Limitations of this procedure include (1) contamination of rAAV with the Ad helper virus, (2) low yields of rAAV and (3) production of replication-competent AAV. In this report we describe new helper plasmids (pSH3 and pSH5) that eliminate the Ad co-infection requirement. The helper plasmids express the AAV rep and cap genes and the Ad E2A, VAI and E4 genes. When the helper plasmids are co-transfected onto human 293 cells with a vector plasmid in the absence of Ad infection, the rAAV vector yield is up to 80-fold greater than those obtained with the pAAV/Ad packaging plasmid. Moreover, replication competent AAV in the rAAV preparations is less than 0.00125%. The major advantages of this system are (1) the absence of infectious adenovirus and (2) the use of only two plasmids, which enhances transfection efficiencies and hence vector production. We believe that this two-plasmid transfection system will allow for more widespread use of the AAV vector system because of its simplicity and high yields. This system will be especially useful for preclinical analyses of multiple rAAV vectors.     

Development of a complementing cell line and a system for construction of adenovirus vectors with E1 and E2a deleted.

Although adenovirus vectors offer many advantages, it would be desirable to develop vectors with improved expression and decreased toxicity. Toward this objective, an adenovirus vector system with deletion of both the El and E2a regions was developed. A 5.9-kb fragment of the adenovirus type 5 (Ad5) genome containing the E2a gene and its early and late promoters was transfected into 293 cells. A complementing cell line, designated 293-C2, expressed the E2a mRNA and protein and was found to complement the defect in Ad5 viruses with temperature-sensitive or deletion mutations in E2a. A deletion of 1.3 kb removing codons 40 to 471 of the 529 amino acids of E2a was introduced into plasmids for preparation of viruses and vectors. An Ad5 virus with disruption of the El gene and deletion of E2a grew on 293-C2 cells but not on 293 cells. Vectors with E1 and E2a deleted expressing Escherichia coli beta-galactosidase or human alpha1-antitrypsin were prepared and expressed the reporter genes after intravenous injection into mice. This vector system retains sequences in common between the complementing cell line and the vectors, including 3.4 kb upstream and 1.1 kb downstream of the deletion. These vectors have potential advantages of increased capacity for insertion of transgene sequences, elimination of expression of E2a, and possibly reduction in expression of other viral proteins. Although the titers of the vectors with deleted are about 10- to 30-fold below those of vectors with E2a wild-type regions, the former vectors are suitable for detailed studies with animals to evaluate the effects on host immune responses, on duration of expression, and on safety.     

Generation of an Adenovirus Vector Lacking E1, E2a, E3, and All of E4 except Open Reading Frame 3

Toxicity and immunity associated with adenovirus backbone gene expression is an important hurdle to overcome for successful gene therapy. Recent efforts to improve adenovirus vectors for in vivo use have focused on the sequential deletion of essential early genes. Adenovirus vectors have been constructed with the E1 gene deleted and with this deletion in combination with an E2a, E2b, or E4 deletion. We report here a novel vector (Av4orf3nBg) lacking E1, E2a, and all of E4 except open reading frame 3 (ORF3) and expressing a beta-galactosidase reporter gene. This vector was generated by transfection of a plasmid carrying the full-length vector sequence into A30.S8 cells that express E1 and E2a but not E4. Production was subsequently performed in an E1-, E2a-, and E4-complementing cell line. We demonstrated with C57BL/6 mice that the Av4orf3nBg vector effected gene transfer with an efficiency comparable to that of the Av3nBg (wild-type E4) vector but that the former exhibited a higher level of beta-galactosidase expression. This observation suggests that E4 ORF3 alone is able to enhance RNA levels from the beta-galactosidase gene when the Rous sarcoma virus promoter is used to drive transgene expression in the mouse liver. In addition, we observed less liver toxicity in mice injected with the Av4orf3nBg vector than those injected with the Av3nBg vector at a comparable DNA copy number per cell. This study suggests that the additional deletion of E4 in an E1 and E2a deletion background may be beneficial in decreasing immunogenicity and improving safety and toxicity profiles, as well as increasing transgene capacity and expression for liver-directed gene therapy.     

Construction of an adenovirus type 7a E1A- vector.

A strategy for constructing replication-defective adenovirus vectors from non-subgroup C viruses has been successfully demonstrated with adenovirus type 7 strain a (Ad7a) as the prototype. An E1A-deleted Ad7a reporter virus expressing the chloramphenicol acetyltransferase (CAT) gene from the cytomegalovirus promoter enhancer was constructed with DNA fragments isolated from Ad7a, an Ad7a recombination reporter plasmid, and the 293 cell line. The Ad7a-CAT virus particle transduces A549 cells as efficiently as Ad5-based vectors. Intravenous infections in a murine model indicate that the Ad7a-CAT virus infects a variety of tissues, with maximal levels of CAT gene expression found in the liver. The duration of Ad7a-CAT transgene expression in the liver was maximally maintained 2 weeks postinfection, with a decline to baseline activity by the week 4 postinfection. Ad7a-CAT represents the first example of a non-subgroup C E1A- adenovirus gene transfer vector.     

Long-term gene delivery into the livers of immunocompetent mice with E1/E4-defective adenoviruses.

We have compared the in vitro and in vivo behaviors of a set of isogenic E1- and E1/E4-defective adenoviruses expressing the lacZ gene of Escherichia coli from the Rous sarcoma virus long terminal repeat. Infection of tumor-derived established cell lines of human origin with the doubly defective adenoviruses resulted in (i) a lower replication of the viral backbone that correlated with reduced levels of E2A-specific RNA and protein, (ii) a significant shutoff of late gene and protein expression, and (iii) no apparent virus-induced cytotoxicity. Independently of the extent of the deletion, the additional inactivation of E4 from the viral backbone therefore drastically disabled the virus in vitro, with no apparent effect on transgene expression. A lacZ-transgenic model was used to compare the different recombinant adenoviruses in the livers of C57BL/6 mice. The immune response to the virally encoded beta-galactosidase was minimal in this model, as infusion of the E1-defective adenovirus resulted in a time course of transgene expression that mimicked that in immunodeficient (nu/nu) mice, with very little inflammation and necrosis in the liver. Administration of a doubly defective adenovirus to the transgenic animals led to long-term extrachromosomal persistence of viral DNA in the liver, with no detectable methylation of CpG dinucleotides. However, transient transgene expression was observed independently of the extent of the E4 deletion, suggesting that the choice of the promoter may be critical to maintain transgene expression from these attenuated adenovirus vectors.     

Several E4 region functions influence mammary tumorigenesis by human adenovirus type 9

Among oncogenic adenoviruses, human adenovirus type 9 (Ad9) is unique in eliciting exclusively estrogen-dependent mammary tumors in rats and in not requiring viral E1 region transforming genes for tumorigenicity. Instead, studies with hybrid viruses generated between Ad9 and the closely related nontumorigenic virus Ad26 have roughly localized an Ad9 oncogenic determinant(s) to a segment of the viral E4 region containing open reading frame 1 (E4-ORF1), E4-ORF2, and part of E4-ORF3. Although subsequent findings have shown that E4-ORF1 codes for an oncoprotein essential for tumorigenesis by Ad9, it is not known whether other E4 region functions may similarly play a role in this process. We report here that new results with Ad9/Ad26 hybrid viruses demonstrated that the minimal essential Ad9 E4-region DNA sequences include portions of both E4-ORF1 and E4-ORF2. Investigations with Ad9 mutant viruses additionally showed that the E4-ORF1 protein and certain E4-ORF2 DNA sequences are necessary for Ad9-induced tumorigenesis, whereas the E4-ORF2 and E4-ORF3 proteins are not. In fact, the E4-ORF3 protein was found to antagonize this process. Also pertinent was that certain crucial nucleotide differences between Ad9 and Ad26 within E4-ORF1 and E4-ORF2 were found to be silent with respect to the amino acid sequences of the corresponding proteins. Furthermore, supporting a prominent role for the E4-ORF1 oncoprotein in Ad9-induced tumorigenesis, an E1 region-deficient Ad5 vector that expresses the Ad9 but not the Ad26 E4-ORF1 protein was tumorigenic in rats and, like Ad9, promoted solely mammary tumors. These findings argue that the E4-ORF1 oncoprotein is the major oncogenic determinant of Ad9 and that an undefined regulatory element(s) within the E4 region represents a previously unidentified second function likewise necessary for tumorigenesis by this virus.     

Persistence of recombinant adenovirus in vivo is not dependent on vector DNA replication.

Recombinant adenovirus vectors represent an efficient means of transferring genes into many different organs. The first-generation E1-deleted vector genome remains episomal and, in the absence of host immunity, persists long-term in quiescent tissues such as the liver. The mechanism(s) which allows for persistence has not been established; however, vector DNA replication may be important because replication has been shown to occur in tissue culture systems. We have utilized a site-specific methylation strategy to monitor the replicative fate of E1-deleted adenovirus vectors in vitro and in vivo. Methylation-marked adenovirus vectors were produced by the addition of a methyl group onto the N6 position of the adenine base of XhoI sites, CTCGAG, by propagation of vectors in 293 cells expressing the XhoI isoschizomer PaeR7 methyltransferase. The methylation did not affect vector production or transgene expression but did prevent cleavage by XhoI. Loss of methylation through viral replication restores XhoI cleavage and was observed by Southern analysis in a wide variety of, but not all, cell culture systems studied, including hepatoma and mouse and macaque primary hepatocyte cultures. In contrast, following liver-directed gene transfer of methylated vector in C57BL/6 mice, adenovirus vector DNA was not cleaved by XhoI and therefore did not replicate, even after a period of 3 weeks. Although replication may occur in some tissues, these results show that stabilization of the vector within the target tissue prior to clearance by host immunity is not dependent upon replication of the vector, demonstrating that the input transduced DNA genomes were the persistent molecules. This information will be useful for the design of optimal adenovirus vectors and perhaps nonviral episomal vectors for clinical gene therapy.     

Construction of adenoviral vectors

Recombinant adenovirus vectors have proven to be useful tools in facilitating gene transfer. Construction of such vectors requires a knowledge of the adenovirus genome structure and its life cycle. A commonly used recombinant adenovirus involves deletion of the E1 region; such a recombinant is traditionally produced by overlap recombination after contransfection of 293 cells with a plasmid shuttle vector and a large right-end restriction fragment of viral DNA. The shuttle vector contains a cassette for a transgene placed in region E1 and flanking sequences from adenovirus for recombination. Normally, a high background of parental virus results because of the difficulty in separating right-end restriction fragment length DNA from uncut DNA. This paper describes a negative selection based on the traditional cotransfection method using viral DNA from an E1-deleted adenoviral recombinant that expresses green fluorescent protein (GFP). In situ fluorescent microscopy is used to distinguish the recombinant plaques (white or nonfluorescent) from the parental virus plaques (green or fluorescent). In addition, this system allows for the detection of contaminating parental virus at later stages when production lots of the recombinant vector are being made.     

A new type of adenovirus vector that utilizes homologous recombination to achieve tumor-specific replication

We have developed a new class of adenovirus vectors that selectively replicate in tumor cells. The vector design is based on our recent observation that a variety of human tumor cell lines support DNA replication of adenovirus vectors with deletions of the E1A and E1B genes, whereas primary human cells or mouse liver cells in vivo do not. On the basis of this tumor-selective replication, we developed an adenovirus system that utilizes homologous recombination between inverted repeats to mediate precise rearrangements within the viral genome resulting in replication-dependent activation of transgene expression in tumors (Ad.IR vectors). Here, we used this system to achieve tumor-specific expression of adenoviral wild-type E1A in order to enhance viral DNA replication and spread within tumor metastases. In vitro DNA replication and cytotoxicity studies demonstrated that the mechanism of E1A-enhanced replication of Ad.IR-E1A vectors is efficiently and specifically activated in tumor cells, but not in nontransformed human cells. Systemic application of the Ad.IR-E1A vector into animals with liver metastases achieved transgene expression exclusively in tumors. The number of transgene-expressing tumor cells within metastases increased over time, indicating viral spread. Furthermore, the Ad.IR-E1A vector demonstrated antitumor efficacy in subcutaneous and metastatic models. These new Ad.IR-E1A vectors combine elements that allow for tumor-specific transgene expression, efficient viral replication, and spread in liver metastases after systemic vector application.     

Mechanism of enhancement of DNA expression consequent to cointernalization of a replication-deficient adenovirus and unmodified plasmid DNA.

Given the knowledge that replication-deficient adenoviruses can mediate the delivery of unlinked plasmid DNA into eukaryotic cells (K. Yoshimura, M. A. Rosenfeld, P. Seth, and R. G. Crystal, J. Biol. Chem. 268:2300-2303, 1993), this study focuses on the role of receptor-mediated endocytosis in this process. AdCFTR (an E1- E3- adenovirus type 5-based replication-deficient adenovirus containing the 4.5-kb human cystic fibrosis transmembrane conductance regulator cDNA) was added to Cos-7 cells together with plasmid pRSVL (containing the Rous sarcoma virus long terminal repeat promoter followed by the luciferase cDNA), and luciferase activity was quantified as a measure of the expression of the plasmid DNA. When AdCFTR was bound to Cos-7 cells at 4 degrees C and the cells were subsequently incubated at 37 degrees C in the presence of pRSVL, the expression of luciferase activity was increased in proportion to the amount of AdCFTR added, reaching > 10(4)-fold at 3,000 PFU per cell. AdCFTR-mediated increase in pRSVL was inhibited by addition of purified adenovirus fiber but not hexon, suggesting cell surface adenovirus receptors were involved in the cointernalization process. Cell lines with a high number of adenovirus receptors (Cos-7 and HeLa) showed significant AdCFTR-dependent pRSVL expression, while cell lines with low numbers of adenovirus receptors (NIH 3T3 and U-937) showed little. AdCFTR-mediated increase in the expression of pRSVL was prevented when AdCFTR was heat treated and exposed to antibody against adenovirus or when the cointernalization process was evaluated in the presence of chloroquine, conditions all known to prevent adenovirus-mediated disruption of endocytic vesicles. In contrast, the uptake of AdCFTR into Cos-7 cells was not affected by any of these conditions. When AdCFTR was exposed to UV light, its ability to grow in 293 cells was obviated, but AdCFTR-dependent increase in pRSVL expression was minimally reduced. Finally, empty capsids of AdCFTR were able to enhance the delivery and expression of plasmid pRSVL into Cos-7 cells, suggesting that the adenovirus genome is not required for AdCFTR-mediated plasmid cointernalization. Together, these observations suggest that the ability of a replication-deficient recombinantly adenovirus to mediate the cointernalization and expression of plasmids is mediated by the receptor-mediated endocytosis pathway.     

E1/E3缺失型腺病毒载体引起细胞周期G_2/M阻滞

腺病毒载体广泛应用于基因治疗和转基因研究 ,目前常用的E1 E3缺失型复制缺陷腺病毒载体虽然失去了病毒复制必需的E1基因 ,但载体上的其它病毒基因仍能在宿主细胞内表达 .为研究这些基因对细胞的毒性作用 ,选择了 3种携带没有明显细胞毒性外源基因的腺病毒载体 ,观察感染 2种肿瘤细胞后细胞核形态改变 ,并用流式细胞仪检测细胞周期及凋亡情况 .发现大剂量重组缺陷型腺病毒感染细胞后引起细胞变圆 ,核增大 ,细胞周期阻滞于G2 M期 ,继而染色质凝聚 ,细胞发生坏死或凋亡 ;各种腺病毒载体造成G2 M阻滞所需感染量不同 ,但都随时间延长和感染量增加而加重 .这些结果提示腺病毒基因对细胞的影响是多方面的 ,在以此类病毒载体进行基因转移和基因治疗的研究中 ,精确滴定病毒滴度和转导效率非常重要 ,腺病毒基因表达造成的毒副作用给此类研究增加了变数     

Tissue-specific, tumor-selective, replication-competent adenovirus vector for cancer gene therapy

We have previously described two replication-competent adenovirus vectors, named KD1 and KD3, for potential use in cancer gene therapy. KD1 and KD3 have two small deletions in the E1A gene that restrict efficient replication of these vectors to human cancer cell lines. These vectors also have increased capacity to lyse cells and spread from cell to cell because they overexpress the adenovirus death protein, an adenovirus protein required for efficient cell lysis and release of adenovirus from the cell. We now describe a new vector, named KD1-SPB, which is the KD1 vector with the E4 promoter replaced by the promoter for surfactant protein B (SPB). SPB promoter activity is restricted in the adult to type II alveolar epithelial cells and bronchial epithelial cells. Because KD1-SPB has the E1A mutations, it should replicate within and destroy only alveolar and bronchial cancer cells. We show that KD1-SPB replicates, lyses cells, and spreads from cell to cell as well as does KD1 in H441 cells, a human cancer cell line where the SPB promoter is active. KD1-SPB replicates, lyses cells, and spreads only poorly in Hep3B liver cancer cells. Replication was determined by expression of the E4ORF3 protein, viral DNA accumulation, fiber synthesis, and virus yield. Cell lysis and vector spread were measured by lactate dehydrogenase release and a "vector spread" assay. In addition to Hep3B cells, KD1-SPB also did not express E4ORF3 in HT29.14S (colon), HeLa (cervix), KB (nasopharynx), or LNCaP (prostate) cancer cell lines, in which the SPB promoter is not expected to be active. Following injection into H441 or Hep3B tumors growing in nude mice, KD1-SPB caused a three- to fourfold suppression of growth of H441 tumors, similar to that seen with KD1. KD1-SPB had only a minimal effect on the growth of Hep3B tumors, whereas KD1 again caused a three- to fourfold suppression. These results establish that the adenovirus E4 promoter can be replaced by a tissue-specific promoter in a replication-competent vector. The vector has three engineered safety features: the tissue-specific promoter, the mutations in E1A that preclude efficient replication in nondividing cells, and a deletion of the E3 genes which shield the virus from attack by the immune system. KD1-SPB may have use in treating human lung cancers in which the SPB promoter is active.