Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Lung cancer is the leading cause of cancer mortality in the United States in both men and women. It results in over 150,000 deaths annually, and with current multimodality therapy, the overall 5-yr survival is less than 15%. A decrease in mortality will likely arise from strategies that enable disease prevention (e.g., smoking cessation), earlier detection, or better treatment. Of the latter, treatment advances may potentially originate from a variety of rational therapeutic strategies that are currently under development, including some gene therapy approaches (1). For most gene therapy paradigms, efficient gene transfer into target cells in vivo is a prerequisite for the successful translational gene therapy of lung cancer. In turn, efficient in vivo gene transfer and expression are functions of vector entry into the target cell. The ability of any vector to approach the target cell is contingent on the vector mode of delivery and the anatomic (e.g., endothelium for intravascular delivery) (1) and functional (e.g., neutralizing antibodies) (5) barriers that must be overcome. Subsequent transduction of target cells is mediated by vector target cell interactions, which may use vector-specific attachment, internalization, and intracellular transport processes to effect gene delivery to the nucleus (9).

Clinical trials for intrathoracic malignancies conducted recently in the United States and Europe using the currently most efficient in vivo gene therapy vehicle, the adenoviral (Ad) vector, suggest that although Ad delivery is safe, in situ transduction is inefficient (8, 14). Although the phase 1 clinical trials for malignant mesothelioma suggest that in situ gene transfer can be detected by immunohistochemistry (14) using high doses of Ad vectors, the most reliable measurements have been made using molecular criteria (8, 14). Because of the tremendous sensitivity of polymerase chain reaction-based assays, therapeutic efficacy directly attributable to gene transfer is arguable (15) and may represent epiphenomena resulting from mode of vector delivery (generally intratumoral injections) or immune-mediated adjuvant effects, or effects specific to the vector rather than the transgene. Thus, in vivo Ad gene transfer into lung cancer or malignant mesothelioma, although feasible, is inefficient and therefore may be nonefficacious.

We have been investigating intrapleural malignancy as a general "proof of concept" model for the gene therapy of non-small cell lung cancer (NSCLC). As a result, our preclinical studies have targeted malignant pleural effusions as the clinical correlate and have used cell cultures and effusions derived from patients. Malignant pleural effusions from NSCLC arise in 10% of all cases of lung cancer, and mostly in patients with disseminated disease (16, 17). The pathogenesis of these effusions stems from the ineffectiveness of lymphatic clearance, and characteristically, the effusions are exudative (16). Importantly, because of diminished clearance from the pleural cavity, the effusion is a particularly rich reservoir of soluble factors (derived from plasma filtrate, matrix and metabolic byproducts secreted by tumor, stromal, and immune cells, and the residual debris of cells that have died in this milieu). Previously, we identified specific soluble factors, namely chondroitin sulfate proteoglycans and glycosaminoglycans, that inhibited gene transfer by a variety of retroviral vectors (RV) in malignant pleural effusions (18). In an analogous manner, we determined that soluble factors in effusions inhibit gene transfer by Ad vectors. Here we describe characteristics of these factors mediating this inhibition.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Lines

NCI-H226 and NCI-H1703 cells (gifts from Dr. Herbert Oie, National Cancer Institute, Bethesda, MD) are derived from intrapleural metastases of NSCLC and are maintained in RPMI 1640 (GIBCO-BRL, Grand Island, NY) with 10% fetal bovine serum (Irvine Scientific, Irvine, CA) and penicillin (100 U/ml)/streptomycin (100 µg/ml) (growth medium). These cells bind adenoviruses in a specific manner and exhibit an efficient transduction profile by Ad vectors in vitro (19). Consequently, these target cells are uniformly transduced in growth medium at Ad multiplicity of infectious particles (MOI) of 100 (1-h exposure, 37°C).

Viral Vectors

Ad vectors were constructed in the Vector Core at the Gene Therapy Center of the University of North Carolina School of Medicine (Chapel Hill, NC). Ad5LacZ is E1a/E1b and partially E3-deleted, and expresses the reporter LacZ gene under the control of the cytomegalovirus promoter region. [³⁵S]-radiolabeled Ad5 vectors were also produced in the Vector Core at concentrations of ~ 10¹² particles/ml (specific activity, 1 to 3 × 10³ cpm/ particle) and were used to quantitate cellular binding and internalization as previously described (19). Ad vectors were typically purified and concentrated with double CsCl ultracentrifugation and stored at 20°C in a nonfreezing solution containing 25% glycerol, 0.05% bovine serum albumin, 4 M CsCl, 50 mM NaCl, 0.5 mM MgCl₂, and 5 mM Tris buffer. Immediately before use, vectors were gel filtered for desalting (G-50 Sephadex; Boehringer Mannheim, Indianapolis, IN) and eluted into growth medium for transduction studies as previously described (19).

Effusions

Pleural effusions were collected using aseptic technique from symptomatic patients with known intrapleural malignancy at the time of hospital admission by an institutionally approved protocol. All effusions studied were exudative and pathologically positive for tumor as depicted in Table 1. The cells were pelleted and the supernatants were sterile-filtered (to remove microscopic cellular debris) through 0.45-µm filters (Schleicher and Schuell, Keene, NH). The fluid components of the effusions were then either used immediately for transduction bioassays or were stored at 70°C for further analyses.

Transduction Protocols and Reporter Gene Expression Analysis

A total of 2.5 × 10⁵ cells in six-well (9.6 cm²) tissue culture plates (Costar, Cambridge, MA) was exposed to Ad vectors admixed in test (effusion component containing 50% vol/vol, total volume 1 ml/well) or control (without effusions) growth medium for 2 h in humidified chambers at 37°C. In some experiments, cells were exposed to unmodified fluid component of the effusions for 2 h, after which the material was aspirated and the cells washed twice with phosphate-buffered saline (PBS) before vector exposure. Ad-binding studies with radiolabeled vector were performed at 4°C as described previously (19). LacZ expression was determined histochemically for intracellular 5-bromo-4-chloro-3-indoyl--D-galactopyranoside (X-gal; 5 prime-3 prime, Boulder, CO) hydrolysis in fixed (0.5% glutaraldehyde for 10 min, 4°C) target cells. The percent positive (blue) cells were determined by counting a total of at least 400 cells from each group under a hematocytometer (19).

Chromatography and Immunoadsorption

Aliquots (10 ml) of acellular and sterile-filtered effusions were passed with centrifugation (3000 × g for 4 h, 4°C) through microporous membranes that enabled concentration of molecules greater than 30 or 100 kD (Amicon Centriplus-30 or 100; Millipore Corporation, Bedford, MA). The eluted and retained fractions were isolated and used in transduction bioassays after replacing their total volume to 10 ml with PBS and repeat sterile-filtering. To extract immunoglobulin (Ig), 1.6 ml suspension (200 µg/ml PBS; sufficient to extract approximately 50 mg of Ig) of sepharose A beads (Amersham Pharmacia Biotech, Uppsala, Sweden) was added to 2.0 ml of sterile-filtered effusion and incubated overnight at 4°C. Antibody-depleted specimens were then used in transduction bioassays where the control effusion specimens were appropriately diluted (vol/vol) with PBS (1.6 ml of PBS added to 2 ml filtered effusions). The efficiency of Ig depletion was assessed by Western blot analysis. Briefly, 5 µg of reduced and denatured (100 mM dithiothreitol, heated at 95°C for 5 min) protein in unmodified and Ig-depleted effusion specimens was separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (Bio-Rad, Hercules, CA). Proteins were transferred to nitrocellulose (Hybond; Amersham Pharmacia Biotech) in transfer buffer (39 mM glycine, 48 mM Tris base, 0.037% SDS, 20% methanol), and the membrane was blocked in 5% (wt/vol) nonfat milk in 20 mM Tris base, 137 mM NaCl, and 0.1% Tween (TTBS). Human Ig was detected with rabbit antihuman Ig-Fc (Sigma, St. Louis, MO) (1:15,000 dilution in TTBS, 30 min, room temperature), the membrane washed with TTBS and exposed to horseradish peroxidase (HRP)-linked antirabbit-IgG (1:7000 dilution in TTBS, 1 h, room temperature; Sigma). After rinsing in TTBS, the labeled proteins were detected using enhanced chemiluminescence Western blotting kit after exposure to Amersham hyperfilm (Amersham Pharmacia Biotech).

Soluble Extracellular Matrix Components and Enzymes

Soluble fibronectin, hyaluronic acid, chondroitin sulfate (CS), glycosaminoglycans (GAG), and bovine testicular hyaluronidase (BTH) were purchased from Sigma. The fibronectin and GAGs were resuspended in PBS and added to vector-containing medium to test for their inhibition to transduction. BTH was resuspended in 50 mM Na acetate/PBS (pH 6.0) as described (18) and 250 U of enzyme was added into each milliliter of sterile-filtered effusion. This enzymatic treatment was previously determined to reverse the CS-mediated inhibition to retroviruses in malignant effusions (18). The BTH-treated effusions were then admixed with vector-containing medium and compared with control specimens in which the virus alone was exposed to similar concentrations of BTH.

Statistical Methods

All experiments were performed using at least three distinct specimens a minimum of two separate times. To determine statistical significance, results were indexed to Ad MOI and comparisons made using Kruskal-Wallis analysis of variance on ranks, followed by Bonferroni group comparisons. A statistically significant difference was defined as P < 0.05 between the groups compared.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The fluid component of all effusion specimens tested had the capacity to inhibit Ad gene transfer, and the majority of the nine pleural effusions studied contained factor(s) that block Ad gene transfer into target cells (NCI-226 cells [as shown] and NCI-H1703 cells [data not shown]) despite high Ad MOIs (100 to 1,000) (Figure 1). Transduction efficiency was not restored by heating the effusions to 56°C (Figure 1), indicating that complement proteins are not the inhibitors. This finding was expected because Ad vectors are not encapsulated. Boiling (100°C, 10 min) converted the effusions to a consistency of boiled-egg white, and when this material was ultracentrifuged (100,000 × g, 10 min), the resultant aqueous supernatant lost the inhibitory activity, suggesting that boiling denatures or precipitates irreversibly the inhibitory factors (data not shown). Freeze-thawing the fluid component of the effusions did not affect the inhibitory factors (data not shown); thus, effusions may be stored at 70°C for future analyses. Lastly, because the inhibition to Ad transduction could be overcome by increasing the Ad MOI (Figure 1) or by decreasing the amount of effusion in the transduction media (Figure 2), the Ad particle likely remained intact and infectious within the effusions and apparently was not degraded by biochemical mechanisms.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Inhibition of Ad gene transfer in malignant pleural effusions. NCI-H226 cells are exposed to AdLacZ vector in the absence (closed bars) or presence (open bars) of sterile-filtered malignant pleural effusion supernatants, and transduction efficiency is measured by counting blue cells after X-gal histochemistry. Ad MOIs (10, 100, or 1000) are depicted on the abscissa, and as labeled, one set of effusion supernatants are heated to 56°C for 30 min to inactivate complement (56*C). Results are depicted as mean ± standard error of the mean (SEM) of transduction efficiency in nine separate effusions, and asterisk denotes P < 0.005.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Inhibition of Ad gene transfer varies with the amount of effusion in the transduction media. NCI-H226 cells are exposed to AdLacZ vector in the presence of varying volume percent (0, 10, or 50) of effusion supernatants in the transduction media, and transduction efficiency is measured by counting blue cells after X-gal histochemistry. Ad MOI is fixed at 100 PFU/cell. Results are depicted as mean ± SEM of transduction efficiency in four separate effusions, and asterisks denote P < 0.05.

To determine the relative molecular size of the species responsible for this inhibition, three malignant pleural effusions were selected for fractionation by size using porous membranes. These three were selected based on their high inhibitory activity as defined by the absence of transduction after exposure to cells at an Ad MOI of 1,000. In all three effusions tested, the native inhibitory factors were large (> 100 kD) components of the effusions that fractionate with the volume retained by the filters (Figure 3). This observation suggested that the inhibitory factor(s) is either a single species of large macromolecules or is a complex of smaller, divergent molecules that coassociate in native malignant pleural effusions.

View larger version (19K): [in this window] [in a new window]   Figure 3.   The inhibitory factors in malignant pleural effusions are > 100 kD in size. Aliquots, 10 ml, of malignant pleural effusions are filtered through microporous membranes that retain molecules > 30 kD or > 100 kD in size. Control (0, vector containing medium without effusion components), effusion supernatants (Effusion), filtrates (30K Filtrates or 100K Filtrates), and retentates (30K Retentates or 100K Retentates) are then tested for inhibition of Ad transduction of NCI-H226 cells. Results are presented as mean ± SEM of transduction efficiency in three separate effusions, and asterisks denote P < 0.001.

Given the large size of the native inhibitory factors and our prior observation that CS-proteoglycans (PG)/GAGs mediate inhibition to retroviral gene transfer (18), we studied whether these molecules could similarly account for the inhibition to Ad transduction. However, hyaluronic acid (HA) or, in contrast to the previous studies with retroviral gene transfer, CS did not inhibit Ad gene transfer. The inhibition to Ad transduction was not reversed by pretreatment of the effusions with mammalian hyaluronidase (BTH) at doses that reversed inhibition of retroviral transduction (Figure 4A) (18). Furthermore, exogenous addition of soluble HA or CS-GAGs (at concentrations of up to 500 µg/ml) into the transduction media did not abrogate Ad transduction efficiency (Figure 4B). By comparison, the CS-PG/GAG content in typical malignant pleural effusions is approximated to be 30 to 60 µg/ml (18). Thus, large HA or CS-GAGs did not appear to account for the inhibition of Ad transduction in malignant pleural effusions.

View larger version (12K): [in this window] [in a new window]   View larger version (30K): [in this window] [in a new window]   Figure 4.   (A) Inhibition to Ad transduction in malignant effusions is not reversed by treatment of the effusions with mammalian hyaluronidase. Ad transduction of NCI-H226 cells is studied in the presence of sterile-filtered effusion supernatants that have been pretreated (Ad + Effusion + BTH) or untreated (Ad + Effusion) with mammalian hyaluronidase (BTH, 250 U/ml) at concentrations that can reverse the inhibition exhibited by effusions to RV transduction. AdLacZ-control (Ad + BTH) is similarly treated with BTH for comparison. Results are presented as mean ± SEM of transduction efficiency in six separate effusions and asterisks denote P < 0.001. (B) Ad transduction is not inhibited by the addition of soluble GAGs in the transduction medium. HA or CS-GAGs (CSA, CSB, CSC) (concentrations of 0, 100, or 500 µg/ml) are admixed into the transduction medium containing AdLacZ at an MOI of 10 (solid bars) or 100 (open bars) that are exposed to NCI-H226 cells. Data are presented as means of transduction efficiency in two separate experiments.

To determine the mechanisms of action of the inhibitory factors, we investigated whether the components of Ad entry could be inhibited. The first interaction of the Ad particle with the cell surface is postulated to be binding mediated by the Ad fiber knob with a high-affinity cellular receptor, including Coxsackie B-virus and adenovirus receptor (CAR) (9). Next, Ad vectors are internalized using pathways reserved for receptor-mediated endocytosis, a process believed to be triggered after Ad fiber knob attachment to the cell surface and that involves interactions of the Ad penton base with cell-surface _v3,5 integrins via Arg-Gly-Asp (RGD) peptides (10, 11, 20). We investigated the inhibition to adenovirus both in terms of binding and internalization. Binding of radiolabeled adenovirus to its target cell was inhibited in the presence of malignant pleural effusions (Figure 5A). This inhibition occurred within the fluid component of the effusions rather than on the cell surface, and washing the effusions off the target cells before Ad exposure restored basal Ad binding to the cell membrane (Figure 5B). With respect to internalization, the _v3,5 integrins also mediate high-affinity cellular interactions with components of the extracellular matrix containing RGD moieties, including fibronectin. Because high concentrations of fibronectin (approximately 200 µg/ ml) are found within malignant effusions (21), we questioned whether soluble fibronectin is an inhibitory soluble factor that is competing for Ad interaction with cell-surface integrins. However, fibronectin, at concentrations ranging from 100 to 400 µg/ml, did not alter Ad transduction (Figure 5C) and excluded this factor as a competitive inhibitor of Ad gene transfer. In conclusion, these series of experiments suggested that the inhibition to Ad transduction was mediated at least in part by impaired Ad binding to the cell surface. In addition, though subsequent internalization may also be impaired, our studies suggest that fibronectin, a known soluble component of malignant pleural effusions, did not inhibit Ad transduction.

View larger version (10K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   View larger version (26K): [in this window] [in a new window]   Figure 5.   (A) Ad binding to the cell is inhibited in the presence of effusions. NCI-H226 cells are exposed to radiolabeled adenovirus at 4°C for 2 h in the presence of effusions (50% vol/vol in transduction media). The cell-associated [35S] counts are collected after alkaline lysis. Results are presented as mean bound/total cpm ± SEM in six separate test effusions, and asterisk denotes P < 0.001. (B) Ad binding to the cell is restored if effusions are washed off the target cells. NCI-H226 cells are exposed to the same effusions tested in A for 2 h at 4°C. Following, the effusions are washed off and the cells are exposed to radiolabeled adenovirus at 4°C for 2 h and the cell-associated counts are collected after alkaline lysis. Results are presented as mean bound/total cpm ± SEM after exposure to six separate test effusions. (C ) Ad transduction is not inhibited by soluble fibronectin. Fibronectin (at concentrations of 0 [closed bars], 100, 200, or 400 µg/ml [open bars]) is admixed into transduction medium containing AdLacZ at an MOI of 10 or 100 before exposure to NCI-H226 cells. Transduction efficiency is measured by counting blue cells after X-gal histochemistry, and the data are presented as the mean of two separate experiments.

The recognized soluble barriers to Ad vectors proximal to its interaction with the target cell are chiefly comprised of anti-Ad antibodies (Ab). The major neutralizing Ab in immune sera to the adenovirus are to the hexon-capsid protein, the penton base, and the fiber. Of these, only the latter (anti-fiber Ab) inhibit binding to the target cell surface, whereas anti-hexon and anti-penton Abs attenuate Ad infectivity by blocking endosomal escape postentry (22). Given that preformed Ad-neutralizing Ab may be responsible for inhibiting Ad transduction in malignant effusions, we tested whether adsorption of Ig out of effusions (using protein A-sepharose) results in increased gene transfer. The effectiveness of the antibody-depletion strategy was validated using Western blotting that detected the Fc portion of Ig in the effusions before and after immunoadsorption (Figure 6A). Ig adsorption out of the effusions restored gene transfer by Ad vectors when the target cells were exposed to high Ad MOI (Figure 6B; transduction efficiency of NCI-H226 cells at Ad MOI of 1,000 was not significantly different between unmodified and antibody-depleted effusions). However, at Ad MOIs of less than 1,000, significant inhibition remained to Ad transduction (Figure 6B). Thus, even after substantial depletion of Ig from effusions (Figure 6A), inhibitory activity persisted, suggesting that preformed neutralizing Ab constitute a significant part but not the entirety of the observed inhibition to Ad gene transfer in malignant effusions.

View larger version (41K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 6.   (A) Western blot analysis verifies the efficacy of Ig depletion. To extract Ig, sepharose A beads in suspension were incubated overnight at 4°C with effusion specimens and the efficiency of Ig depletion was assessed by Western blot analysis. A total of 5 µg of reduced and denatured protein in unmodified (1 and 2) and Ig-depleted [Ab()] effusion specimens was separated by 10% SDS-PAGE, transferred to nitrocellulose, and the human Ig was detected with rabbit antihuman IgFc. The labeled proteins were detected using ECL after exposure to HRP-linked antirabbit IgG. Depicted are the apparent molecular weights (kD) on the ordinate, the test-effusion specimens (lanes 2-5), and the marker (M) lane. (B) Ig depletion from malignant pleural effusions partially overcomes the block to Ad transduction. NCI-H226 cells are exposed to AdLacZ for 2 h in the absence (squares) or presence of either native effusions (triangles) or effusions that have undergone total Ig depletion by adsorption to sepharose A (circles). Transduction efficiency is measured by counting blue cells after X-gal histochemistry 48 h later. Results are depicted as mean ± SEM of transduction efficiency in three separate effusions. *P < 0.05, and its placement reflects the control value from which the test value is significantly different.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Because in vivo gene transfer for inherited and acquired pulmonary diseases using Ad vectors is generally poor, a number of studies have endeavored to characterize the molecular and cellular basis of that inhibition in an effort to define specific strategies to overcome inefficient gene transfer. As a result, the relative paucity of receptors mediating Ad entry in target cells has been implicated as being an initial, immediate hurdle to efficient Ad gene transfer (25). Strategies to overcome this initial resistance have typically sought to molecularly retarget the Ad vector to the airway epithelium or target cell surface (29). In addition, after Ad delivery into the host, an acute, nonspecific (innate) immune response is generated and is implicated in diminishing Ad gene delivery over a period of hours to a few days (33). Pretreatment with corticosteroids and inhibiting macrophage-phagocytic function has resulted in overcoming the resistance to Ad gene transfer in this context (35). Three to seven days after Ad exposure, a specific immune response follows and leads to a humoral block of the viral vector. In the same time frame, a cell-mediated immune response has been implicated in eliminating the transduced cells in most tissues (38, 39). Strategies to overcome the humoral block have included the use of adenoviruses of alternate serotypes (40), and a variety of strategies including the coutilization of immunomodulatory or cytotoxic therapy and the construction of "gutted" Ad vectors have been proposed to extend the longevity of transduced cells in immunocompetent hosts (41).

Another component of the block to Ad vectors, one that is immediate and comprised in part by pre-existing humoral immunity, has not been studied in detail with respect to its impact on adenovirus-mediated gene transfer. In this regard, for targeting cancer in the pleural space, the Ad vector has to transgress barriers that are proximal to its encounter with the target cells. This report demonstrates that included among these barriers is a soluble block to Ad transduction in the fluid component of malignant pleural effusions (Figures 1 and 2). We have further determined that the inhibitory activity is filterable, sensitive to boiling, and can be titrated. Thus, decreasing the volume percent of effusions in the transduction media or increasing the concentration of the virus results in increased transduction efficiency. The latter observation suggests that the Ad particle is not biochemically inactivated but likely remains intact and infectious in the effusions. In addition, the data suggest that the preponderance (if not all) of the native inhibitory activity lies in the high molecular weight (> 100 kD) fractions of effusions (Figure 3). Thus, in native effusions, the inhibitory factors are either large molecules or aggregates of molecules that fractionate with molecular species that are greater than 100 kD. Unlike the soluble inhibitors to retroviral transduction, complement proteins and CS-PGs and GAGs are not implicated (Figures 1 and 4). One mechanism whereby these soluble inhibitors operate is by limiting Ad entry into target cells. For example, specific binding of adenovirus by target cells is inhibited in the presence of effusions, and this inhibition occurs in solution rather than on the target cell surface (Figures 5A and 5B). By comparison, Ad internalization by _v3,5 integrins is probably not inhibited by soluble fibronectin in malignant effusions (Figure 5C). Lastly, although preformed anti-Ad antibodies appear to account for a portion of the observed inhibition, other unidentified factors are also important, especially at low adenoviral:target cell ratios (Figure 6).

It is self-evident that identifying specific molecular interactions that preclude targeted Ad gene delivery in vivo may foster strategies for either inactivating these inhibitors or for re-engineering the Ad capsid to avoid these interactions. Because different effusions from different patients exhibit common features, further study may enable us to isolate the inhibitory activity and characterize another general hurdle to Ad transduction that may be relevant in other clinical contexts. Based on studies to date, we speculate that the inhibition that is observed to Ad gene transfer has both specific and nonspecific components. For example, factors that are specific to Ad vectors may include neutralizing immunoglobulins that are not excluded by our antibody-depletion protocol, soluble CAR, or other cellular adherence receptors that are shed or have entered the soluble component of effusions from dead cells. To identify other high-affinity molecular interactions that function to inhibit Ad transduction, we propose to quantify the residual inhibitory activity in antibody-depleted effusions (as exemplified in Figure 6). Subsequently, we propose to fractionate the remaining inhibitory components by gel electrophoresis and attempt to identify specific Ad-binding partners using viral overlay assays (45). Elution and sequence analysis of specific bands may prove useful for their identification. Alternatively, it is possible that we may not be able to fractionate or specifically categorize the remaining inhibition. In either case, if the inhibitory activity is entirely soluble and is without an interstitial or tissue-based reservoir, washing the pleural cavity with saline before vector instillation may enhance gene transfer to cells within this space.