Recurrence of bronchioloalveolar carcinoma in transplanted lungs.

BACKGROUND: Bronchioloalveolar carcinoma is a distinctive subtype of typical adenocarcinoma of the lung that tends to metastasize widely throughout the lungs but less commonly elsewhere. Because conventional therapies for intrapulmonary metastatic bronchioloalveolar carcinoma are generally ineffective, we treated seven patients who had intrapulmonary metastatic bronchioloalveolar carcinoma with lung transplantation. METHODS: Seven patients with biopsy-proved bronchioloalveolar carcinoma and no evidence of extrapulmonary disease received transplants of either one or two cadaveric lungs. At transplantation, all native lung tissue was removed and replaced with a donor lung or lungs. The patients received the usual post-transplantation care given at the institution. RESULTS: Four of the seven patients had recurrent bronchioloalveolar carcinoma within the donor lungs; the recurrences appeared from 10 to 48 months after transplantation. All recurrences were limited to the donor lungs. Histologic and molecular analyses showed that the recurrent tumors in three patients originated from the recipients of the transplants. CONCLUSIONS: Lung transplantation for bronchioloalveolar carcinoma is technically feasible, but recurrence of the original tumor within the donor lungs up to four years after transplantation was common.

Coacervate microspheres as carriers of recombinant adenoviruses.

The therapeutic utility of recombinant adenoviruses (rAds) is limited in part by difficulties in directing the viruses to specific sites and by the requirement for bolus administration, both of which limit the efficiency of target tissue infection. As a first step toward overcoming these limitations, rAds were encapsulated in coacervate microspheres comprised of gelatin and alginate followed by stabilization with calcium ions. Ultrastructural evaluation showed that the microspheres formed in this manner were 0.8-10 microM in diameter, with viruses evenly distributed. The microspheres achieved a sustained release of adenovirus with a nominal loss of bioactivity. The pattern of release and the total amount of virus released was modified by changes in microsphere formulation. Administration of the adenovirus-containing microspheres to human tumor nodules engrafted in mice showed that the viral transgene was transferred to the tumor cells. It is concluded that coacervate microspheres can be used to encapsulate bioactive rAd and release it in a time-dependent manner.

Enhanced efficacy of a novel controlled release paclitaxel formulation (PACLIMER delivery system) for local-regional therapy of lung cancer tumor nodules in mice.

The efficacy of systemic chemotherapy for non-small cell lung cancer (NSCLC) has improved with newer agents. However, the response rates and prolonged survival times achieved by chemotherapy remain modest, and these small gains are obtained at the cost of significant toxicity. In this study, the efficacy of a controlled release formulation of paclitaxel was compared with conventional paclitaxel in animals with human lung cancer xenografts. Paclitaxel (10%) was encapsulated in a proprietary polymer in the form of microspheres (PACLIMER Delivery System). Tumor nodules comprised of two different cell lines (A549 and H1299) were treated by a single i.p. or intratumoral administration of conventionally formulated paclitaxel or a single intratumoral injection of the PACLIMER Delivery System. In vitro testing demonstrated that paclitaxel was released slowly from the microspheres with >80% released after 90 days. Direct comparison of the highest dose for all formulations (24 mg/kg) showed that for nodules comprised of either NSCLC cell line, growth of the PACLIMER Delivery System-treated nodules were inhibited significantly more than the groups treated with conventional paclitaxel or the vehicle controls. Tumor volume doubling times for A549 and H1299 nodules treated with PACLIMER Delivery System were 60 and 35 days, respectively, compared with 10 and 11 days, respectively, in the nodules treated with the conventional paclitaxel by intratumoral administration. We conclude that intratumoral administration of the PACLIMER Delivery System may substantially increase the efficacy of paclitaxel for the therapy of local-regional NSCLC.

Use of a tissue-specific promoter for targeted expression of the herpes simplex virus thymidine kinase gene in cervical carcinoma cells.

Molecular chemotherapy strategies have been developed for a number of epithelial malignancies based on selective delivery and expression of a toxin-encoding gene into the cancer cells. To date, these strategies have not been explored in the context of carcinoma of the cervix, despite the fact that a variety of factors suggest this as an appropriate disease for this gene therapy approach. One limitation in this respect is that appropriate tissue-specific promoters for selective toxin gene expression have not been defined for cervical carcinoma. In this regard, the secretory leukoprotease inhibitor (SLPI) gene has been shown to be constitutively expressed in many epithelial carcinoma cells including the uterine cervix. Thus, we investigated the utility of the SLPI gene as a tissue-specific promoter for regulatory control of the herpes simplex virus thymidine kinase gene for in vitro treatment of cervical carcinoma cells. For this analysis, a gene construct was derived with the herpes simplex virus thymidine kinase gene under regulatory control of the 5' upstream regions of the SLPI gene. Transient transduction of three human cervical carcinoma cell lines with the SLPI-thymidine kinase (TK) construct was followed by treatment with the prodrug ganciclovir. Crystal violet staining was subsequently used to assess cell viability. In this analysis, it was shown that the SLPI-TK construct directed TK-mediated killing in two of three tested cervical cell lines, with the two cell lines being positive for SLPI. In addition, mixing experiments established that cervical carcinoma cells could exhibit a bystander effect which potentially augments the efficacy of molecular chemotherapy approaches. These findings may allow for the development of efficacious, target-specific, toxin gene therapy strategies for cervical carcinoma in human patients.

trans E1 component requirements for maximal replication of E1-defective recombinant adenovirus.

Strategies that enable E1-defective recombinant adenoviruses to selectively undergo replication in neoplastic tissue may be useful for future investigations or therapies of malignancies. A growing body of evidence suggests that some molecular alterations commonly associated with malignancies, such as p53 mutations, can modify the specific E1 requirements for replication of human serotype adenoviruses. In the studies reported here, a panel of human non-small cell lung cancer cell lines with previously defined p53 status were characterized for basal interleukin-6 (IL-6) and bcl-2 content because previous studies have indicated both proteins can functionally substitute for the replication requirements provided by native E1 viral proteins. Cell lines were infected with E1-defective adenovirus 5 and simultaneously transfected with different combinations of E1 plasmids, or a bcl-2 expression plasmid, and adenovirus present in the cells was quantified 6 days later. These assays demonstrated that E1A with both 19- and 55-kDa E1B-encoding plasmids were required for maximal adenoviral replication, independent of the varying p53/IL-6/basal bcl-2 phenotypes of the host cell lines. E1A was required for maximal replication enablement, independent of the basal IL-6 content of these cell lines, and exogenous IL-6 also did not obviate the E1A requirement. Interestingly, the bcl-2 expression plasmid did not consistently substitute for the 19-kDa expression plasmid in the context of this replication complementation assay. These results suggest that (1) basal levels of IL-6 greater than that present in these cell lines are necessary for functional replacement of the E1A replication function and (2) bcl-2 does not predictably substitute for the 19-kDa E1B replication function in the context of trans complementation.

Overexpression of adenovirus-encoded transgenes from the cytomegalovirus immediate early promoter in irradiated tumor cells.

Efficient expression of therapeutic genes in irradiated tumor cells would facilitate the conversion of a malignant tumor nodule into a cancer vaccine in situ. We reported previously that transgene expression from an adenoviral vector could be markedly enhanced by treating transduced tumor cells with butyrate. In this study, we demonstrated that a similar butyrate effect could be achieved in irradiated tumor cells. In addition, irradiating cells at doses of 2-40 Gy prior to transduction could also amplify recombinant adenoviral transgene products in a cell-type-specific manner. This suggests that adenovirus-mediated gene therapy, radiation therapy, and butyrate-mediated cancer therapy may potentially be formulated into one synergistic protocol for cancer treatment.

Amplification of recombinant adenoviral transgene products occurs by inhibition of histone deacetylase.

n-Butyrate (butyrate) has been shown to amplify transgene expression in cells infected with E1-defective adenoviruses. The present studies were undertaken in order to better define the actions of butyrate in the context of adenovirus gene expression, and to attempt to elucidate the mechanism by which butyrate mediates the transgene amplification. It was found that butyrate amplified viral transgene expression over a concentration range of 0.5-5 mM, and that the amplification required an exposure of 12-24 hr for maximal effect. Western blot analysis of representative viral proteins showed that butyrate treatment amplified DNA-binding protein, but not fiber protein. A transient adenoviral replication system suggested that butyrate had a modest inhibitory effect on replication of the E1-defective adenovirus. Use of a specific inhibitor of histone deacetylase, trichostatin A (TSA), reproduced the amplification of the viral transgene product achieved with the butyrate. In contrast, adenoviral transgene expression could not be amplified by TSA treatment in a cell line known to have a TSA-resistant histone deacetylase. Butyrate amplified steady-state gene expression of the viral transgene, but had no detectable effects on either DNA-binding protein or fiber steady-state gene expression. Nuclear run-off experiments showed that both butyrate and TSA caused an increase in the viral transgene transcription. It was concluded that inhibitors of histone deacetylase amplify adenoviral transgene expression at the transcriptional level.

Modulation of adenovirus-mediated gene transfer by nitric oxide.

We assessed the role of .NO in recombinant adenovirus-mediated gene transfer both in vitro and in vivo. NIH3T3 fibroblasts, stably transfected with the human inducible nitric oxide synthase, but lacking tetrahydrobiopterin (NIH3T3/iNOS [inducibile nitric oxide synthase]), were infected with replication-deficient adenovirus (E1-deleted), containing either the luciferase or the Lac Z reporter genes (AdCMV-Luc and AdCMV-Lac Z; 1-10 plaque forming units [pfu]/cell). Incubation of infected cells with sepiapterin (50 microM), a precursor of tetrahydrobiopterin, progressively increased nitrate/nitrite levels in the medium and decreased both luciferase and beta-galactosidase protein expression to approximately 60% of their corresponding control values, 24 h later. NIH3T3/iNOS cells had normal ATP (adenosine 5'-triphosphate) levels and did not release LDH(lactic dehydrogenase) into the medium. Pretreatment of these cells with N(G)-monomethyl-L-arginine (L-NMMA; 1 mM), an inhibitor of iNOS, prevented the sepiapterin-mediated induction of .NO and restored gene transfer to baseline values. Incubation of NIH3T3/iNOS with 8-bromo-cGMP (400 microM) in the absence of sepiapterin, or exposure of AdCMV-Luc to large concentrations of .NO, did not alter the efficacy of gene transfer. .NO produced by NIH3T3/iNOS cells also suppressed beta-galactosidase expression in NIH3T3 cocultured cells stably transfected with beta-galactosidase gene, suggesting .NO inhibited gene expression at either the transriptional or posttranscriptional levels. To investigate the effects of inhaled .NO on gene transfer in vivo, CD1 mice received an intratracheal instillation of AdCMV-Luc (4 x 10(9) pfu in 80 microl of saline) and exposed to .NO (25 ppm in room air) for 72 h. At that time, no significant degree of lung inflammation was detected by histological examination. However, lung luciferase activity decreased by 53% as compared with air breathing controls (P < 0.05; n > or = 8). We concluded that overproduction of .NO decreases the efficiency of adenovirus-mediated gene transfer in lung cells in the absence of cytotoxicity or inflammation.

E1A RNA transcripts amplify adenovirus-mediated tumor reduction.

Previous work by this group has established that E1-defective, recombinant adenoviruses can be replication-enabled by the codelivery of a plasmid encoding the deleted E1 functions, a strategy now designated conditional replication-enablement system for adenovirus (CRESA). In the studies reported here, the original replication-enabling plasmid was replaced by two separate plasmids that encoded the necessary E1A and E1B functions, respectively. An RNA transcript encoding the requisite E1A functions was shown to substitute functionally for the E1A plasmid without significant loss of new adenovirus production in in vitro experiments. No replication competent adenovirus was detectable in the cells treated with the plasmids, or the RNA and plasmid combinations. Subcutaneous human tumor nodules containing a fraction of cells cotransduced with the replication-enabling RNA + DNA and an adenovirus containing a herpes simplex virus thymidine kinase (HSVtk) expression cassette were reduced to a greater extent than control nodules containing the same fraction of cells cotransduced with the virus and an irrelevant plasmid. These experiments show that an E1-defective adenovirus can be conditionally replication-enabled by an RNA transcript encoding the required E1 functions, and that the replication-enablement is sufficient to produce an augmentation of an adenovirus-mediated therapeutic effect in vivo.

Quantitative and in vivo activity of adenoviral-producing cells made by cotransduction of a replication-defective adenovirus and a replication-enabling plasmid.

Achieving limited recombinant viral replication may provide a means of amplifying viral-mediated gene transfer in vivo. We have previously shown that cotransduction of an E1-defective adenovirus with a plasmid containing the deleted E1 functions into prostate carcinoma cells resulted in E1-defective virus production by those cells. The studies described here have extended these findings to more firmly establish the capacity of the trans complementation approach to achieve amplification of recombinant viral transgene expression. The recombinant virus used for all the studies was AdCMV-luc which contained a luciferase expression cassette; the replication-enabling plasmid, pE1, encoded the E1 functions deleted from AdCMV-luc. Quantitative in vitro studies with the HeLa cell line showed that for each plaque forming unit of AdCMV-luc originally exposed to the cells, 0.54 x 10(3) new replication-defective viruses were detected in supernatants and lysates over the following 4 days. Multiple cell lines were shown to support new virus production following cotransduction of AdCMV-luc and pE1. Small numbers of replication-competent viruses were detected in the lysates and supernatants from the cotransduced cells such that for every 10(5) replication-defective viruses approximately two replication-competent viruses were produced. Tumor nodules produced by engrafting a mixture of AdCMV-luc/pE1-cotransduced HeLa cells with uninfected HeLa cells yielded much higher levels of luciferase expression than control tumors containing mixtures of cells infected with AdCMV-luc alone. In total, these results demonstrate new virus production by cells receiving a replication-defective adenovirus and a replication-enabling plasmid are capable of amplifying recombinant viral transgene expression in vivo.

Supernatant rescue assay vs. polymerase chain reaction for detection of wild type adenovirus-contaminating recombinant adenovirus stocks.

The certification of recombinant adenoviruses prepared for clinical use requires the exclusion of contaminating, replication-competent adenovirus (wild type virus). Polymerase chain reaction (PCR)-based detection assays have been developed that detect the presence of viral sequences present only in wild type adenoviruses. As an alternative, this report describes a novel bioassay, designated the 'supernatant rescue assay', that detected minimal amounts of wild type virus mixed with high numbers of recombinant adenoviruses. This assay is based on the observation that minimal numbers of wild type adenovirus can be rescued from the supernatants of cells exposed to high multiplicities of infection (mois) of recombinant virus mixed with a known, minimal number of wild type virions. This assay was highly reproducible and routinely detected the presence of a single wild type virus mixed within 10(9) recombinant viruses. Under the experimental conditions employed, the supernatant rescue assay was significantly more sensitive than all three different PCR-detection assays.

Bystander killing of melanoma cells using the human tyrosinase promoter to express the Escherichia coli purine nucleoside phosphorylase gene.

We used a gene transfer-based system to generate highly toxic purine bases in tumor cells transfected with the Escherichia coli purine nucleoside phosphorylase (PNP) gene. Because these toxic purines are membrane permeant, they mediate effective killing of neighboring cells that do not express E. coli PNP ("bystander" toxicity). In mixed cultures containing increasing percentages of cells with gene expression, 100% cancer cell growth arrest and total population killing was demonstrated when as few as 1-2% of cells expressed E. coli PNP. We used E. coli PNP to test bystander killing of human melanoma cells. A 529-bp region upstream of the human tyrosinase gene start site was shown to direct melanoma-specific expression in human cell lines. When this human tyrosinase regulatory region was used to control E. coli PNP expression, profound toxicity was observed in melanoma cells after treatment with the relatively nontoxic substrate 6-methylpurine-deoxyriboside, which is converted by E. coli PNP into the highly toxic purine base 6-methylpurine. Bystander toxicity was estimated as at least 100 cells killed for each cell expressing E. coli PNP, a level substantially higher than that of other tumor sensitization genes currently being used in clinical trails. These results suggest that the high bystander activity of the system could lead to significant antimelanoma responses in vivo.

Trans complementation of an E1A-deleted adenovirus with codelivered E1A sequences to make recombinant adenoviral producer cells.

Replication-incompetent adenovirus is conventionally produced by cells that supply replication-enabling proteins from viral sequences present in trans. As an alternative means of recombinant adenovirus production, replication-enabling E1A sequences were cotransduced into human prostate carcinoma cells infected with an E1A-deleted adenovirus containing a luciferase expression cassette. The replication-enabling plasmid was cotransduced by ionic linkage to the recombinant adenovirus exterior. Cells cotransduced with the replication-enabling plasmid made new adenovirus with titers up to 8 x 10(6) in the supernatants 72-120 hr after transduction. Like the parent virus, the virus present in the cotransduced supernatants and lysates was capable of transferring luciferase activity to new cells. The virus present in the cotransduced cell supernatants was amplified and shown to be identical to the parent virus by genomic analysis. It was concluded that simultaneous addition of a replication-defective adenovirus and a replication-enabling gene sequence in a trans configuration converts some of the cotransduced cells into recombinant adenovirus-producing cells.

Midkine and pleiotrophin expression in normal and malignant breast tissue.

BACKGROUND: Some growth factors may promote tumor growth by affecting tumor angiogenesis. The angiogenic growth factor, pleiotrophin, was demonstrated previously in human breast carcinoma tissues; however, the pattern of pleiotrophin expression in normal breast tissues has not been established. METHODS: The expression of pleiotrophin and the related growth factor, midkine, was examined by polymerase chain reaction amplification of reverse transcriptase copies of RNA transcripts (RT-PCR) from freshly resected normal and malignant human breast tissues. Northern blot analysis of midkine expression was performed on a limited number of the specimens and on human and canine breast carcinoma cell lines. Clinicopathologic variables from the breast cancer patients were examined in relation to the growth factor expression patterns. RESULTS: The majority of both malignant and normal breast tissues expressed pleiotrophin. In contrast, midkine was expressed frequently in the malignant breast tissues but in only one of the normal specimens. Northern blot analysis of the breast carcinoma cells lines showed that they commonly expressed midkine transcripts. The only correlation of the growth factor expression patterns with the other clinical variables was the finding that the three midkine-negative breast carcinoma specimens also had low estrogen receptor levels. CONCLUSIONS: By this analysis, the expression of pleiotrophin was equivalent in both malignant and normal human breast tissues. Midkine, on the other hand, exhibited increased expression in the breast carcinomas but showed much lower expression in the normal breast tissue. Although the cellular source of the midkine expression was not determined by the RT-PCR assay, the Northern blot analysis showed that isolated populations of breast cancer cells commonly express this growth factor. This is the first example of a tissue simultaneously expressing high amounts of both pleiotrophin and midkine, a finding of unclear pathophysiologic significance.

Receptor-mediated gene delivery employing lectin-binding specificity.

Selective targeting of malignant cells will be necessary to implement many of the gene therapy strategies being designed to combat cancer. Targeting can be achieved by transductional or transcriptional approaches. Transductional targeting can be accomplished by exploiting differences in the molecules or receptors expressed on the cell surface of malignant versus normal cells. Given that malignant cells can be distinguished from normal by differences in the expression of cell surface carbohydrates, we hypothesized that transductional targeting would be feasible by molecular conjugate vectors which achieve cell binding by virtue of lectins directed against the cell surface glycocalyx. We have shown that gene transfer can be accomplished by these novel lectin-targeted molecular conjugate vectors and that lectin binding specificities may serve as a means for potential targeting of cancer cells for the purposes of gene therapy.

Strategy for achieving selective killing of carcinomas.

Carcinomas are malignancies derived from epithelial cells that frequently respond poorly to conventional chemotherapy. Selective expression or transduction of toxin genes to carcinomas, i.e. molecular chemotherapy, may offer important advantages over conventional chemotherapy. As one approach to developing a means of selectively expressing toxin genes, the transcriptional regulatory sequences of a gene expressed in multiple carcinomas were used to direct expression of the herpes simplex virus thymidine kinase (HSVtk) coding sequences. The secretory leukoprotease inhibitor (SLPI) gene was found to be expressed in lung, breast, oropharyngeal, bladder, endometrial, ovarian and colorectal carcinomas. The tissue-specific transcriptional regulatory sequences were isolated and used to construct a chimeric gene in which the SLPI sequences directed HSVtk expression. SLPI-expressing carcinomas were reduced in number by transduction of the SLPI-directed toxin plasmid plus ganciclovir, but the same treatment had no effect on a cell line that did not express SLPI. These results suggest that SLPI-directed therapeutic genes could be used for directing toxicity to carcinoma tissues, especially if combined with other targeting strategies.

Surfactant protein A-directed toxin gene kills lung cancer cells in vitro.

Human surfactant protein A (SPA) expression is considered a marker of respiratory epithelial differentiation. Non-small cell lung cancers (NSCLC) are respiratory epithelial derivatives, and it was previously shown that a minority of these cancers expressed SPA, presumably a consequence of their respiratory epithelial origin. In the studies reported here, SPA-I gene transcriptional regulatory sequences were localized to a 2.75-kb genomic 5'-flanking region fragment obtained by screening a human genomic library. The 2.75-kb fragment was used to direct a luciferase coding sequence transcriptionally within a plasmid construct. In plasmid transduction experiments, the SPA-directed luciferase plasmid produced significant luciferase activity in the SPA-expressing NSCLC cell line, H441, but only background levels in the non-SPA-expressing A549 cells. Because Northern blot analysis of resected NSCLC showed that the majority expressed SPA, an SPA-transcriptional targeting strategy was investigated using chimeric toxin genes comprising the coding sequence for herpes simplex virus thymidine kinase (HSV-TK) under transcriptional control of SPA or SV40 regulatory sequences. As expected, transduction of the constitutive, SV40-directed plasmid followed by ganciclovir treatment reduced numbers of both the A549 and H441 cells. In contrast, the SPA-directed plasmid reduced only the SPA-expressing H441 cells and had no significant effect on the A549 cells. The results of these in vivo experiments suggest the concept of transcriptionally directing toxin genes with SPA can produce targeted toxicity in NSCLC.

Reciprocal expression of pleiotrophin and midkine in normal versus malignant lung tissues.

Abundant evidence suggests that growth factors are important mediators of non-small cell lung cancer (NSCLC) growth. Although multiple growth factors have been found to be produced by NSCLC tissues, little is known about possible differences in growth factor expression between malignant and adjacent normal lung tissues. Variation in growth factor expression between normal and malignant lung tissues could be potentially useful diagnostically and therapeutically. In studies reported here, the expression of the angiogenic growth factor pleiotrophin (PTN) and homolog midkine (MK) was assessed in resected normal and malignant lung tissues. Primers specific for the two growth factors were used to amplify reverse transcriptase-produced DNA copies of RNA transcripts harvested from the tissues. This analysis revealed that all normal lung tissues examined (n = 17) expressed PTN but only two expressed MK. Conversely, all of the resected lung cancers (n = 20) expressed MK but only one expressed PTN. These results demonstrated a striking reciprocal expression pattern of MK and PTN in normal versus malignant lung tissue.

Transforming growth factor beta within fibrotic scleroderma lungs.

PURPOSE, PATIENTS, AND METHODS: Since transforming growth factor beta (TGF beta) has been implicated as an important mediator of pulmonary fibrosis, we measured TGF beta protein and gene expression in alveolar epithelial lining fluid (ELF) of fibrotic scleroderma lungs sampled by bronchoalveolar lavage (BAL). TGF beta protein was qualitatively examined by Western blot analysis, and quantitatively by radioreceptor assays. Gene expression was evaluated in BAL mononuclear cells by Northern blot analysis with quantification of relative gene expression by densitometric analysis of the autoradiograms. RESULTS: Normal and scleroderma subjects had a 24-kd protein that comigrated with defined human TGF beta 1 and immunoreacted with anti-TGF beta antibody. The normal population had a significantly higher average TGF beta concentration (705 pM) compared with the scleroderma subjects (177 pM). The TGF beta 1 gene was expressed in amounts that did not significantly differ between the scleroderma and normal groups. On an individual subject basis, the TGF beta concentration variability did not correlate with variations in BAL cellularity or TGF beta 1 gene expression within the recovered mononuclear cells. CONCLUSIONS: It is concluded that both normal and fibrotic lungs have TGF beta 1 present at the alveolar epithelial surface. However, in the fibrotic scleroderma lungs, TGF beta protein content and gene expression were not increased at the alveolar epithelial surface. The simultaneous analysis of TGF beta protein content, gene expression, and cellular constituents within individual ELF specimens showed that the cellular components of the ELF do not appear to be major determinants of TGF beta protein concentration at the alveolar epithelial surface.