A. Human Lung Cancer
Lung cancer is the leading cause of cancer-associated mortality in both men and women. Although susceptibility to environmental carcinogens may be predetermined and follow a pattern of autosomal dominant Mendelian inheritance [2, 3], lung cancer results from an accumulation of acquired genetic mutations [4-6]. In fact, it is suggested that 10-20 genetic mutations may be necessary for the development of lung cancer , although the discrete steps for the progression of a hyperplastic bronchial lesion to metaplasia and anaplasia have not been uncovered. Tobacco use is the strongest epidemiologic risk for the development of lung cancer and it is anticipated that approximately 10% of all smokers will develop lung cancer over their lifetime . Current paradigms predict that lung cancer results from the widespread exposure of the carcinogen, leading to a process of "field cancerization," whereby the entire aerodigestive track is exposed to the offending agents and leads to the occurrence of synchronous and metachronous tumors . The tobacco carcinogens apparently invoke the multiple clonal chromosomal abnormalities found throughout the airways and alveoli of smokers [10, 11]. Following, the series of genetic mutations likely results in patterned aberrancies in signal transduction and cell-cycle pathways, eventuating in malignant and metastatic phenotypes . The general pattern of genetic changes are characteristic but not specific for pathologic subtypes of lung cancer (see below). Overall, K-ras mutations are observed in 20-50% , p53 mutations are present in 50% , 60% exhibit reduced expression of pl6-ink4a [15, 16], and 30% show deletion of Rb. Small-cell lung cancers (SCLC) display a greater proclivity to c-myc-amplification and a greater degree of p53 (80%) and rb mutations (90%). Chromosome 3p deletions, occurring at a chromosomal fragile site that includes the FHIT locus, are found in 50% of non-small-cell lung cancers (NSCLC) and in 90% of SCLC primary tumors . Overexpression of the tyrosine growth factor receptor erbB2-neu is seen in 10-30% and overexpression of bcl-2  in 10-25% of NSCLC tumors .
Clinically, lung cancer is discriminated into SCLC and NSCLC categories by histopathology or cytopatholgy and by their characteristic clinical presentations and divergent responses to conventional cytoreductive therapies. NSCLC may be further subclassifed pathologically into squamous cell (SCCa), adenocarcinoma, broncho-alveolar cell carcinoma (BAC), adenosquamous (mixed pathology), or large-cell carcinoma. As noted above, the progression of lung cancer from a premalignant state to the clinical/pathological entity that is diagnosed in the vast majority of patients is unknown. This is because although the disease is prevalent, it is typically diagnosed when it has already spread outside the lungs and is pathologically advanced. Not surprisingly, because of the late stage of diagnosis, progressive genetic instability confers marked genetic and phenotypic heterogeneity within lung cancers, even in individual patients. The late stage of diagnosis also results in an absolute lack of premalignant material, making it difficult to assign specific roles for the genetic mutations in the systematic progression of lung cancer. Recently, however, some of the characteristic genetic mutations of lung cancer (e.g., loss of heterozygosity at chromosome 3p, p53 mutations) are being identified in microdissected dysplas-tic epithelium . Similar observations are implicating the characteristic K-ras abnormalities in lung cancer as a correlate of mucinous differentiation , A precursor to lung adenocarcinoma, a lesion pathologically termed alveolar atypical hyperplasia or AAH, is being advanced. AAH is described by increased cellular proliferation when compared to adjacent normal parenchyma and by immunohistochemical evidence of p53 stabilization, K-ras mutations, and c-erb-B2 overexpression [22-24]. The presence of these mutations in AAH may explain why such mutations may be detectable in sputum cytology specimens that predate the onset of clinical lung cancer . Identification of these early events are a particular focus of study because they may serve as genetic markers for malignant progression, or as targets of specific genetic or chemopreventative approaches. More relevant to this discussion, perhaps, these early events may be better modeled in murine models than late stage lung cancer (see below). Thus, there exists an inherent complexity in human lung cancer, and to precisely recapitulate the disease process in animals is not practical.
Due to time of model development, ease of experimentation, and cost restraints, murine models of disease are the accepted standards. However, there are generic shortcomings in this approach. For example, cigarette smoke, which is a strong epidemiological risk for the development of human lung cancer and is proximally responsible for approximately 85-90% of lung cancer cases in humans , is only weakly carcinogenic in mice [27, 28]. In addition, although both mouse and human lung adenocarcinomas may share common molecular defects , the histopathological repertoire of spontaneous or induced tumors in mice is very limited [29, 30], and morphologically, nearly all mouse lung tumors bear structural similarities only to BAC or well-differentiated adenocarcinomas. Consequently, whereas humans typically die from lung cancer of "late stage" metastatic disease, mice succumb to respiratory failure following the diffuse involvement of their lungs by "early stage" carcinoma in situ ,
Spontaneous lung cancer develops in 3% of wild mice [31, 32] with strain-dependent sensitivity. Clones have been isolated from spontaneously arising tumors, and established as cultures in vitro. These cultures now serve as a readily available source for the generation of transplantable allografts. Many investigators, including our group [33-38] have extensively utilized line 1 alveolar carcinoma (L1C2), a murine lung cancer cell line that is syngeneic to BALB/c, and 3LL (Lewis lung cancer), which is syngeneic to C57B1/6. Usually, these cell lines are utilized to generate transplantable heterotopic (referring to a location outside of the organ of origin, typically subcutaneous) tumors in syngeneic mice. Our group has utilized these models to investigate, in general, the interplay between the immune system and the host. Both L1C2 and 3LL tumors are relatively "nonimmunogenic," as is human lung cancer, and immunogenetic strategies that modulate the immune system to generate an anti-tumor immune response can be systematically investigated in these models. However, other lung tumor-allografts, especially when cells are selected to express "marker antigens" to enable their easy detection in culture systems, may indeed become immunogenic. Notably, the transplantable allograft system is artificial, and all recipient hosts have a "stress" response to the implanted tumor that cannot be recapitulated in control animals. In addition, extrapolating anti-tumor responses in mice to humans is not a straightforward proposition, and many therapies that reliably "cure" tumors in inbred strains of mice are not as effective in humans. In part, these differences may be attributable to differences in immune responses in the two hosts. For example, cluster determinants (CD-antigens) in murine strains may not have homologous or functional cellular analogs in the human host.
Laboratory animals used for medical experimentation are genetically inbred strains with reliable phenotypic characteristics. Although this feature imposes a generic limitation on the extrapolation of results in lab animal studies to outbred populations, and thus, human disease, there are significant advantages that need to be considered. The inbred nature of laboratory animals enable investigators to reliably establish disease in an animal host, and subsequently to study that disease process in controlled subsets. With respect to tumorigenesis, murine-A/J and SWR strains are the most sensitive, BALB/c is of intermediate sensitivity, and DBA and C57BL/6 are the most resistant. Crosses between susceptible and resistant inbred mouse strains may allow for the mapping of modifier loci for the development of lung cancer . For example, it is reported that the propensity of strains to develop lung tumors correlates with a polymorphism in the second intron of K-ras . Practical experience suggests that there are common genetic alterations affecting known tumor suppressor genes and proto-oncogenes occur during mouse lung carcinogenesis. Molecular abnormalities may also be shared with human lung cancer, and K-ras activation is a conspicuous example . Human adenocarcinomas commonly carry K-ras mutations; most of these mutations are in codon 12 and are transversions of GGT to either TGT or GTT. It is postulated that these mutations occur early in lung cancer pathogenesis since they can be detected in sputum samples of smokers prior to the clinical diagnosis of lung cancer. Analogously, 80 to 90% of both spontaneous and chemically induced murine lung tumors contain K-ras mutations. Moreover, K-ras mutations also occur early in murine lung tumorigenesis, and remarkably, codon 12 is the site of genetic change induced by many chemical carcinogens . Furthermore, a consistent loss of mouse chromosome region 4, an area that contains the mouse homolog of the human pl6-ink4a [42, 43], has been described to result in an allelic loss of the pl6-ink4a seen in 50% of mouse adenocarcinomas. Similarly, p53 mutations are found, albeit infrequently , although mouse chromosomal regions containing p53 and Rb more commonly exhibit LOH . Reduced expression of Rb and pl6 and increased c-myc expression  have also been reported. These commonalties have suggested some to conclude that mouse and human lung carcinomas are sufficiently similar for the murine model to be informative , and have formed the rationale for the testing of chemo-preventative strategies  in mice. Analogously, these commonalties may be advanced to form the basis for the testing of genetic therapies in murine tumors as well.
Mice strains also vary with respect to inducible-tumorigenesis. Generally, mice that are sensitive to the development of spontaneous lung tumors are also at the highest risk for chemically induced tumors  and form the basis for the quantitative carcinogenecity bioassays. Although a variety of agents, including urethane, metals, and concentrated components of tobacco smoke such as pol-yaromatic hydrocarbons and nitrosoamines [45, 46], can induce lung cancer in mice, tobacco smoke per se is only weakly carcinogenic . Murine lung tumors histologically resemble early lesions that originate peripherally (from type 2 alveolar cells or Clara cells) and simulate papillary or bronchioloalveolar cell cancer (BAC). In contrast, the bulk of human tumors are bronchogenic (arise in the airways) and, as described above, display a broad histopathologic variation. In fact, individual human lung cancers may be histologically heterogeneous; i.e., they often display mixed morphologies within the same tumor specimen. So how does one reconcile these differences between murine lung cancer and human lung cancer, and moreover, can one generalize observations and results from one species to another, or even from one human being to another? When considered in the context of adenoviral gene delivery, there is a limiting paucity of in vivo data to generate any broad conclusion. On the contrary, our observations in vitro suggest that gene transfer into subtypes of human lung cancer is highly variable, and strategies directed toward achieving intratumoral gene transfer may require patient or disease-specific vector formulations .
The biological heterogeneity of human lung cancer drives our investigations along specified pathways, utilizing many different models and strategies to come up with viable treatment approaches. For instance, we believe that a systematic assessment of the efficiency and optimal route of adenoviral gene delivery in vivo into murine lung tumors and transplanted human xenografts needs to be performed. Researchers are beginning to identify the Ad-cellular attachment receptor (termed the Coxackievirus-adenovirus receptor (CAR) ) as a major determinant underlying efficient transduction . Along these lines, the scope and "polarity" of CAR expression in tumors in vivo needs to be defined. Thus, one focus of our program is to systematically evaluate gene transfer into these model systems using conventional and retargeted adenoviral vectors with the aim of optimizing a vector system and a mode of delivery. This focus evolves from the premise supported by our in vitro data that the histological heterogeneity of lung tumors may be a harbinger of variable responsiveness to both adenoviral entry and/or the efficacy of adenoviral gene therapy . Because uniform targeting of tumor in vivo may be unattainable, we have also generated protocols in which the Ad-vector is used in precisely controlled ex vivo "dosing" approaches to genetically modify antigen-presenting cells (APCs) or tumor cells to vaccinate the host against their tumor .
2. Murine Models That Spontaneously Develop Lung Cancer
Murine models of lung cancer include strains susceptible to chemically induced tumors and transgenic strains that express viral and cellular oncogenes. The simian virus-40 large T-Ag (SV40-TAg) has been commonly used to produce tumors in transgenic mice [50, 51]. SV40-TAg binds and incapacitates the cell cycle checkpoint and DNA-binding capabilities of the p5 3
and Rb gene products, resulting in uncontrolled cellular proliferation . To develop a murine model of lung cancer, Wikenheiser and colleagues chose to express the SV40-TAg under the transcriptional control of the lung-specific human surfactant protein C (SP-C) promoter in transgenic mice [53, 54]. They demonstrated that these mice consistently developed multifocal lung adenocarcinomas that had pathological features similar to some human lung adenocarcinomas, and that the mice succumbed to respiratory distress by age 4-5 months. As expected, the transgenic animals developed no tumors in any other organ systems, although some nonmalignant tissue also expressed the transgene . Within the lungs, tumors consistently involved the bronchi-olar and alveolar regions of the lung while sparing the large airways. The tumors of these mice also varied with respect to the expression of the large TAg, suggesting perhaps that SV-40 TAg may contribute to transformation, but continued expression may not be necessary for tumor progression. Likewise, organ-specific expression of SV40-TAg using the regulatory regions of uteroglobin  and the Clara cell-specific Mr 10,000 protein (CC-10) also results in the induction of lung tumors . Uteroglobin is a marker protein for the nonciliated epithelial Clara cells, the source of xenobiotic metabolism in the lung, lining the respiratory and terminal bronchioli of the lungs. In animals expressing SV40-TAg under the uteroglobin promoter control, the pulmonary epithelium was morphologically normal at 2 months, dysplastic by 4 months, and transgenic animals were described as developing multifocal pulmonary adenocarcinoma present in various stages of differentiation by 5 months of age. In situ hybridization studies suggested that tumors did not contain the transcripts of the uteroglobin gene, and again, late stage tumors lost expression of the large T-Ag. Tumors also formed in the urogenital tract where uteroglobin is also expressed.
Transgenic mice were also generated using the CC10 kDa promoter driving SV-40 large T-Ag , and it is in this model that we have chosen to test the immunomodulatory capacity of secondary lymphoid chemokine or sic . In the 7736 mouse line, CC-lOTAg-transgenic mice develop multifocal pulmonary adenocarcinomas and succumb to respiratory failure at 16-20 weeks of age. Pathology is localized to the lungs, and the tumors express the large T-Ag in normal Clara cells and in transformed tumor cells. Pathological progression is similar to that described above, with the lungs appearing morphologically normal at 2 months of age, a number of tumor foci are grossly discernable by 3 months, and the majority of the lung is replaced by coalesced nodules by 4 months of age. As tumor progresses, the expression of endogenous CC10 expression diminishes, and there is increased nuclear p53 expression, suggesting binding and stabilization of the protein by the large T-Ag . From our standpoint, we have found that the reliable progression of lethal tumors in these transgenic mice enable us the test a number of hypotheses, dosing schemes, and dosing routes. Importantly, the effects of immunomodulation by the gene transfer of specific cytokines and chemokines into tumor cells in vivo can be determined. Moreover, one can compare this direct-delivery strategy with alternative approaches, including ex vivo modification of autologous APCs using recombinant Ad-vectors. The subsequent reintroduction of gene-modified APCs back into the tumor environment overcomes the inability of dendritic cells to maturate in the presence of tumor in vivo  by providing functional APCs that are capable of processing and presenting tumor antigens to cytolytic T cells in vivo .
Xenotransplantation of human tumors into immunocompromised mice began in the late 1960s  following the discovery of the nude mouse in 1962 and its characterization as an athymic mutant in 1968 . The morphologic and karyotypic stability of tumors serially passaged in nude mice was described , and it was established that xenotransplanted tumors in nude mice often retained distinctive phenotypic and functional characteristics found in the human host . However, the "tumor-take" rate for nude mouse xenotransplants is tumor-specific, and generally, carcinomas are more difficult to establish than melanomas or sarcomas . Thus, progressive tumor growth from inoculated primary tumors (i.e., cultured directly from the patient) is observed in only 33% for lung cancers [61, 63] and is virtually nil for primary breast or prostate cancers. In addition to properties inherent to the tumor, nude-mouse-related factors also impact on tumor take. For example, mice infected with the mouse hepatitis virus do not accept xenotransplants, presumably because of enhanced NK-cell activity , In this regard, it is important to recall that although nude mice lack functionally mature T cells, they are capable of mounting normal humoral responses to T-cell-independent antigens  and they exhibit high NK-cell activity , and these properties probably impact negatively on the tumor-take rate of xenotransplants. The high NK-cell activity also abrogates the metastatic potential of implanted tumors, and the incidence of metastasis is higher in mice with lower NK cell activity, e.g., young (3-week-old) syngeneic mice or the beige (bg'/ bg>) mutants derived from the C57BL/6 mice [67, 68].
The discovery of a severe combined immunodeficiency in mice  offered yet another option for hosting human tumor xenografts. The scid/scid mice are characterized by the virtual absence of functional T and B lumphocytes due to aberrancies in the rearrangement of antigen receptor genes . The first successful engraftment of human solid organ tumors into scid mice began with the subcutaneous inoculation using the A549 lung adenocarcinoma cell line . Since that time, a variety of human solid-organ cancers, both from cell lines and primary tumor specimens, have been successfully engrafted . The higher rates of successful engraftment, presumably because of the lack of residual B-cell function in scid mice, have led many investigators to prefer scid/scid mice over nu/nu mice as the host recipients of human xenograft tumors. Xenografts are still impacted upon by the scid host's innate immunity, and NK and monocyte/macrophage activities can be upregulated in these hosts. For specific needs, selective breeding of other available mutants (beige mutants with reduced NK-cell activity and osteopetrosis with altered macrophage differentiation) enables the generation of strains that harbor overlapping defects in immune function . Furthermore, genetic engineering and genetargeting technology has helped create murine-mutants with exquisitely specific immune defects, including mice in which CD4 or CD8 T cells are deleted  and mice which lack p-2 microglobulin and thus do not express transplantation antigens .
Xenotransplants have many advantages, the primary being that they provide a replenishable source of human tumor. This enables the genetic characterization and gene discovery of tumor-specific phenotypes and, in rare occasions, the progression toward an advanced or metastatic phenotype of the tumor (e.g., from an androgen-dependent prostate tumor to one that is androgen independent, see below). Xenografts incorporating human tumor cells in immune-deficient mice are plentiful. For example, we have developed a novel animal model mimicking intrapleural malignancy that allows for a controlled, focal dosing of reagents and evaluation of therapeutic benefit . The model is composed of 2.5-cm segments of rat intestine that is denuded, and then everted so that the serosal surface is converted into the lumenal surface of a tube. Lung cancer cells are instilled into the lumen via a polyethylene cannula on day -1, allowed to adhere to the serosal surface overnight, and this tubular xenograft is implanted into the interscapular subcutaneous tissue of a nude mouse on day 0 . The graft simulates metastatic tumor growth on the pleural surface basal lamina both grossly and histopathologically and enables robust quantitation of tumor kinetics . The appearance of tumor on this surface is nodular, and these nodules coalesce over time with intervening fibrous stroma. Neovascularization is evident on histological exam of the graft, and tumor growth is continuous with a variety of NSCLC cell lines. We have found this model to offer certain tangible advantages. For example, with respect to the transduction characteristics of tumor, the value of this model is evidenced by the following: (l)the cells are representative of human lung cancer; (2) the location of the tumor is precisely known and tumor is directly accessible; (3) the vast majority of cells that repopulate the graft are derived from those instilled (host leukocytes and fibrocytes comprise the remaining minority); (4) the mode of delivery of reagents (fluid inoculation rather than intratumoral injection) is designed so as to be clinically applicable for installation into pleural space; (5) the size of the xenograft enables quantitative assessments of transgene expression and morphometry simultaneously, containing human tumor into nude mice .
Abnormalities at the cell surface (e.g., erbB2), signal transduction (e.g., ras -oncogene), gene regulation and cell cycle control (e.g., p53, Rb, c-myc-oncogene), or apoptosis (e.g., p53, BCL-2) are all implicated in the process of transformation and can serve as targets for rational therapeutic intervention. For example, to overcome the deficits due to mutated p53, one strategy for lung cancer gene therapy has opted to replace the mutated p53 gene with a normal copy , Restoring p53 function in these cells has led to decreased tumorigenicity of human cancer cells in vitro and in animal models [78, 79]. Based on these preliminary studies, the first clinical gene therapy trial for human NSCLC also utilized a /?53-gene transfer strategy . In this study, nine patients with advanced NSCLC were treated with either bronchoscopic or percutaneous CT-guided injections with a retroviral p53 expression vector (a genetically reengineered retrovirus that is designed to integrate into the cell genome and express the normal p53-protein). Of the seven patients evaluated, three showed evidence of tumor regression at the treatment site and six showed increased apoptosis of tumor cells on posttreatment biopsies. Importantly, there was no significant toxicity associated with the therapy, and in situ gene transfer was achieved. However, limited therapeutic efficacy was observed and the mechanisms responsible for the anti-tumor effects are still under study. For example, although it was originally believed that mutated p53 function would have to be compensated in each and every cell for restoration of the normal apoptosis-program, the results suggested otherwise. Because there was substantive tumor regression despite poor in situ gene transfer, mechanisms for the observed "bystander effect" were hypothesized . The term "bystander effect" refers to the ability of gene-modified tumor cells to mediate killing of neighboring nontransfected cells. One plausible explanation is that wild-type p53 induces release of angiostatic factors, thus undermining the blood supply to the tumor . In addition, the expression of p53 may also contribute to an immune-mediated response [82, 83]. These issues have led to more mechanism-based bench and animal studies, as well as other phase 1 clinical trials using Ad vectors encoding the p53 gene for a variety of cancers, including lung tumors .
Because of the high frequency of p53 mutations, another strategy that uses replication-competent viruses has been hypothesized to be ideally suited for lung cancer. This approach employs adenoviruses (mutant dll520 or ONYX-O15) that are suggested to selectively replicate in /?53-mutated (therefore, selectively in cancer) cells [85, 86]. Consequently, these mutant viruses are promoted as "magic bullets" that kill tumor cells and leave normal tissues intact. This particular approach has generated considerable controversy both in terms of its reputed efficacy as well as its proposed mechanism of action [87-89]. In brief, its effectiveness in both in vitro and in vivo models of lung cancer needs to be confirmed. Nevertheless, the approach represents a prime example of a novel hypothesis-driven strategy that attempts to exploit the biology of a mutant virus to clinical advantage.
Effective immunotherapy has the potential for systemic eradication of disease, a payoff that is especially enticing for the treatment of lung cancer. Previous, largely unsuccessful immunomodulatory campaigns utilized nonspecific immune strategies (e.g., BCG adjuvants). Increasingly, the interest now is in developing specific immune interventions for lung cancer. The major obstacle for effective immunotherapy of lung cancer has been a meager understanding of the immunobiology of this disease. However, a better understanding of the reciprocal interaction between the tumor and the immune system is starting to emerge, lending itself to plausible hypotheses for intervention. We realize that an effective anti-tumor response may either provoke the immune system to recognize and attack the tumor, or conversely, it may serve to reduce the immunosuppression encumbered upon the host by the tumor.
Specific and effective anti-tumor immunity requires both adequate tumorantigen presentation and the subsequent generation of effector lymphocytes. A variety of cytokines have been investigated to implement such a program in situ [90-97], and many of the studies have utilized the Ad-vector for gene delivery. For several reasons, our efforts have focused on IL-7, IL-12, and more recently on the chemokine sic, for the treatment of lung cancer. The rationale underlying the use of these particular cytokines and chemokines is that they all optimize conditions for tumor antigen processing and presentation by the host's APCs, and they help appropriately localize and sustain the effector lymphocytes response [36, 37, 97, 98].
Although the cellular infiltrates differ depending on the cytokine and model used, many studies indicate that tumor cells that have been transfected with cytokine genes can generate specific and systemic antitumor immunity in vivo. Based on these promising animal studies, what prevents these strategies from being translated into successful and curative human clinical trials? One major problem in human cancer patients may be that although lung cancers express tumor antigens , they are ineffective as APCs . Tumor cells cannot function as APCs because (i) they lack costimulatory molecules, (ii) they are unable to adequately process Ag, and (iii) they secrete a variety of inhibitory peptides which promote a state of specific T-cell anergy. Thus, even for highly immunogenic tumors, professional APCs are required for antigen presentation . As described above, local augmentation of IL-7 and IL-12 may help to overcome some of these defects . In addition, one may bring into the tumor environment professional APCs to orchestrate a satisfactory immune response against the tumor. In this regard, dendritic cells (DCs) are potent APCs that are ideal for interacting with and activating naive T cells to generate antigen-specific immunity [102, 103]. Recent advances in the isolation and in vitro propagation of DC has stimulated great interest in the use of these cells for clinical cancer therapy [104, 105]. In such approaches, DC may be envisioned to serve as vehicles for genes expressing antigens  or expressing cytokines in lung cancer gene therapy . In addition, DC-based immunogenetic therapies may be used in combination with other strategies that have been optimized for Ag presentation [34, 37]. Importantly, of the various approaches tested to gene modify the DCs, our colleagues at UCLA have determined that the Ad-vector is best suited for DC-transduction .
Overcoming metastatic disease is paramount for effective lung cancer therapy, and the biology underlying metastasis is gaining clarity. Metastasis is a process involving several complimentary yet distinct elements, including the capacity for tumor cells to invade and traverse the basement membrane, and to reestablish viable tumor foci in distant organs. Each step in this process may serve as a point for therapeutic intervention in lung cancer. As the molecular biology becomes better understood, the opportunity to incorporate specific genes into vector systems invariably materializes. The initial step, tumor invasion, requires proteolysis, which has been suggested to be mediated by an overexpression and secretion of matrix metalloproteinases (MMPs) by lung cancer cells [108-111]. Therapeutically, gene transfer strategies have incorporated tissue inhibitors of metalloproteinases (TIMP) to inhibit invasion and metastasis , or have utilized antisense abrogation of MMPs to inhibit tumorigenicity .
Similarly, angiogenesis (induced growth of blood vessels) is suspected to be critical for tumor survival and progression at each stage of metastasis . Angiogenic progression in lung cancer is felt to be due to an imbalance of angiogenic and angiostatic factors, and the risk of metastasis in NSCLC directly correlates with the extent of tumor-derived angiogenesis . Thus, strategies that inhibit of angiogenic mediators or restore angiostatic factors have potential utility for all stages of lung cancer [115-119]. The important mediators implicated in promoting or inhibiting angiogenesis lend themselves favorably for inclusion into gene therapy strategies. For example, recent studies indicate that vascular endothelial growth factor (VEGF) is an important angiogenic factor produced by a variety of tumors, including lung cancer. Lymph nodes with NSCLC metastases express significantly higher levels of VEGF than do normal, uninvolved nodes , consistent with the speculation that VEGF plays an important role in the metastasis of lung cancer. In addition to VEGF, recent studies have also implicated CXC chemokines in the abnormal angiogenic/angiostatic balance in NSCLC . Members of this family containing the ELR motif (e.g., IL-8) are angiogenic, whereas those that lack this motif (e.g., interferon-inducible protein 10; IP-10) are angiostatic. Accordingly, neutralizing antibodies to IL-8 reduce angiogenesis and consequently the growth of human lung tumors in scid mice .
Other molecular strategies to specifically target angiogenic vessels are also being developed. For example, the adhesion protein av^3 is relatively specific for angiogenic vessels where it mediates endothelial cells interaction with extracellular matrix components  and enables cell motility . Importantly, its blockade can promote tumor regression in vivo in lung cancer models by inducing apoptosis of tumor-associated blood vessels . More recently, phage-display peptide libraries, which are used to screen the specific binding of a massive array of peptides, have isolated small peptides which selectively bind to receptors (including avPj) on angiogenic vessels. Conjugating these peptides to chemotherapeutic agents have enabled investigators to specifically target tumor vasculature and abrogate tumor growth .
4. Adjuvants to Conventional Therapeutic Approaches for Lung Cancer
Conventional multimodality therapy for lung cancer incorporates surgery, radiation, and chemotherapy using a variety of clinical protocols dictated by the subtype and extent of disease. Theoretically, gene therapies may play important synergistic roles in augmenting the effectiveness of conventional approaches. For many such strategies, there already exists a scientific rationale to test them in combination with conventional multimodality therapy. For example, one may enhance the radiation-sensitivity or chemosensitivity of tumor cells (e.g., p53 or kBa gene therapy) [127, 128] or modify normal tissue susceptibility to cytoablative therapy (e.g., mucosal/tissue protection: by virtue of MDR-1 or bFGF gene transfer). Examples of synergism with the suicide gene therapy approaches have also been studied. The HSV thymidine kinase gene/ganciclovir system induces radiation sensitivity into transduced tumor cells , suggesting that these two forms of therapy can be combined to potentiate antitumor responses . Similarly, tumor cells transduced with the cytosine deaminase transgene exhibit enhanced radiation sensitivity following pretreatment with 5-fluorocytosine . Because the loss of p53 function can result in tumor resistance to ionizing radiation , restoring p53 function may restore apop-totic pathways and promote effective radiation or chemotherapy. In fact, gene transfer of wild-type p53 has been shown to enhance radiation sensitivity  and can act synergistically with c/s-platinum-based chemotherapy to augment cytotoxicity ,
Many of the approaches outlined above as being strategies for gene therapy of 'lung cancer" are generic; these approaches can be generalized to a variety of malignancies since transformed cells have in common the same aberrant growth regulatory and signal transduction pathways. The molecular and cellular pathogenesis of tumor invasion and immune evasion are also similar between tumors originating in diverse organ systems. Unfortunately, this commonality may not confer a broad-based advantage when gene therapy strategies are advanced clinically. In this respect, vectors need to provide both efficient gene delivery as well as tumor specificity, and as a result, the gene transfer strategies have to become "disease specific." Targeted vectors (as discussed elsewhere in this compilation) have to incorporate features rendering them capable of selective cell surface adherence or entry or, alternatively, express their therapeutic transgenes under tumor-specific regulation. Unfortunately, a lung cancer-specific cell surface target (for transductional targeting) has not been identified, and one is left trying to use targets that are generally over-expressed in tumor cells or tumor-induced endothelium [135, 136], Similarly, lung cancer also does not express a specific tumor marker. Thus, transcriptional targeting approaches largely utilize elements that are "tissue-specific" rather than "cancer-specific." Accordingly, constructs where transgene expression is regulated by tissue-specific promoters (e.g., SLPI, SP-A, CC-10) are being actively developed and tested.
Was this article helpful?
Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.