Published ahead of print on December 30, 2004, doi:10.1165/rcmb.2004-0226OC
© 2005 American Thoracic Society DOI: 10.1165/rcmb.2004-0226OC Gene Transfer Mediated by Native versus Fibroblast Growth Factor–Retargeted Adenoviral Vectors into Lung Cancer CellsDepartment of Medicine and The Lung Cancer Research Program, Jonsson Comprehensive Cancer Center, David Geffen School of Medicine at University of California at Los Angeles; and Veterans Administration Greater Los Angeles Healthcare System, Los Angeles, California Correspondence and requests for reprints should be addressed to Raj K. Batra, M.D., David Geffen School of Medicine at UCLA and Division of Pulmonary and Critical Care Medicine, Veterans Administration Greater Los Angeles Health Care System, 111Q, 11301 Wilshire Blvd., Los Angeles, CA 90073. E-mail: rbatra{at}ucla.edu
We compared native Adenoviral (Ad) vectors to a basic Fibroblast Growth Factor–retargeted Adenovirus (FGF2-Ad) for gene delivery into a diverse panel of lung cancer cells in vitro and xenografts in vivo. Cells were first evaluated for vector-specific receptor expression. Marked variations of surface coxsackie-adenovirus receptor (CAR), but relatively similar levels of  v integrin and FGF receptor expression were evident. Transduction efficiency by Ad directly correlated (R = 0.77, 95% CI 0.28–0.94, P = 0.0085) with CAR, but not with v integrin expression. Transduction efficiency by FGF2-Ad did not correlate with the measured FGF receptor expression. Blocking studies indicated that gene transfer by FGF2-Ad occurred by a CAR-independent pathway, and could be inhibited by free FGF in a dose-dependent manner. Ad-antiserum inhibited FGF2-Ad gene transfer, suggesting that the Ad-component was needed for post-entry DNA-delivery. Soluble heparin sulfate proteoglycans (HSPG) or v integrin blockers marginally decreased FGF2-Ad transduction. Both Ad and FGF2-Ad equally transduced CAR-positive non–small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) cells. By contrast, FGF2-Ad had a distinct transduction advantage in CAR-deficient NSCLC cells. This improvement in transduction of CAR-deficient cells by FGF2-Ad persisted in vivo. These data justify the need for an improved FGF2-Ad vector for clinical use in CAR-deficient lung cancer.
Key Words: CAR • FGF2 • gene therapy • lung cancer • molecular conjugates
Our focus has been on the development of gene-based therapies for lung cancer. Lung cancers are diverse, categorized within four major types and over 40 subtypes that possess a wide range of biological phenotypes. These tumors likely exhibit heterogeneity of binding sites for potential gene transfer vectors. Consequently, a single gene transfer vector for generically targeting these tumors is unlikely to emerge. Expectedly, the efficiency of gene transfer into lung cancer cells by various viral vectors is highly variable (1). This inefficiency is attributable to a combination of factors, which are specific to the tumor cells or their microenvironment (2–6). For the Adenoviral (Ad) vector system, lung cancer cells exhibit a wide variability in terms of expressing the coxsackie-adenovirus receptor (CAR), the primary attachment receptor for Ad (1, 7). Whereas large doses (high multiplicity of infectious particles per cell, or MOI) or prolonged exposure of Ad vector to target cells can overcome deficiencies in CAR expression in vitro (1), these measures cannot overcome inefficient Ad transduction in vivo. Accordingly, in lung cancer xenograft models, Ad gene transfer in vivo is a function of CAR expression (7), and is exceedingly poor into CAR-deficient xenografts. The basic fibroblast growth factor (FGF2)-receptors are reported to be ubiquitously upregulated in numerous human tumor types, including lung cancers (8–11), and on tumor neovasculature in situ. FGF2 binds these receptors at picomolar affinities (12–15), and FGF2 retargeted vectors may improve tumor-specific attachment of cancer cells and/or cancer stroma. In fact, FGF2-retargeted adenoviral (FGF2-Ad) molecular conjugates have been formulated (16) and reported to achieve efficient gene transfer in ovarian and pancreatic adenocarcinomas (17–19), melanoma (20), and Kaposi's sarcoma (21). The reported success of these studies prompted our interest in assessing and optimizing this system for CAR-deficient lung cancer. To infect target cells, subtype C Ad vectors first bind to their cell-surface adherence receptor, termed the coxsackie-adenovirus receptor or CAR, using the knob domain of their fiber protein. There are 12 fiber proteins that emanate from each Ad capsid particle. The binding of the Ad5 fiber with CAR initiates processes that promote entry of the virus into the cell, probably using coated pit mechanisms that are triggered by the recruitment of cell surface integrins. Subsequently, other protein components of the Ad capsid facilitate transport of the virus/vector DNA to the nucleus (see below). By comparison, the mechanisms underlying FGF-receptor (FGFR)-mediated gene delivery are largely speculative. Sosnowski and coworkers (22) first demonstrated that cellular FGFR could be used to deliver genes. These initial plasmid DNA-based vectors evolved into FGF2-Ad conjugates (16). In this formulation, the Ad component serves to promote post-entry escape of the internalized vector out of cellular endosomes. Escape from the degradative endo-/lysosomal compartment allows a greater proportion of the vector to gain entry into the nucleus, and thus improves gene transfer efficiency. The synthesis of FGF2-Ad was accomplished by chemically linking a recombinant FGF2 molecule to an immunoglobulin-Fab' fragment derived from a blocking monoclonal antibody generated against the knob domain of the Ad5 fiber. Hypothetically, this covalent coupling jointly served to eliminate the native (CAR-dependent interaction with the Ad5 fiber knob domain) tropism of Ad, and replaced it with tropism for the cellular FGFRs. The conjugate was reported to uniformly increase the efficiency of gene transfer, in both Ad-sensitive and Ad-resistant cell lines (23). Follow-up studies suggested that retargeting the adenovirus to the FGFR pathway allowed efficient transduction of targeted tumor cells through unique vector entry mechanisms (23, 24). However, direct evidence to implicate FGFRs in FGF2-Ad entry is lacking. Instead, the conjugate is postulated to mediate target cell transduction in a manner analogous to FGF's known signaling properties. For cellular signaling, FGF2 is exported to the extracellular matrix and is sequestered there until it is presented to cell surface heparin sulfate proteoglycans (HSPG) (25). In response to injury or with malignancy, cellular FGFRs are upregulated, and the HSPG present the FGF to one of several high-affinity receptors that mediate FGF internalization. Thus, HSPGs serve to protect the FGF from proteolysis in the extracellular space, and heparin binding of FGF precedes its presentation to high-affinity cellular receptors. Whether this schema for signal transduction translates into a mechanism of FGF-retargeted Ad gene transfer requires additional study. We studied the ability of FGF2-Ad in mediating efficient gene transfer to (CAR-deficient) lung cancer cells, and examined whether receptor expression correlated with gene transfer efficiency by the native versus retargeted vectors.
Cell Cultures and Surface Phenotyping Non–small cell lung cancer (NSCLC) cell lines were a gift of Dr. Herbert Oie, NCI. Cells were maintained in RPMI 1640 (Irvine Scientific, Santa Ana, CA) with 10% fetal bovine serum (FBS; Gemini, Woodland, CA), 2 mM L-glutamine, and penicillin (100 U/ml)/streptomycin (100 µg/ml) (complete growth medium) at 37°C in 5% CO2 incubator (1). The small cell lung cancer (SCLC) cell lines (NCI-H82 and NCI-H146) were purchased from the American Type Culture Collection (HTB-175 and HTB-173; ATCC, Manassas, VA), and maintained in ATCC complete growth medium (RPMI 1640 with 2 mM L-glutamine adjusted to contain 1.5 g/liter sodium bicarbonate, 4.5 g/liter glucose, 10 mM HEPES, 1.0 mM sodium pyruvate, and 10% FBS) containing penicillin (100 U/ml)/streptomycin (100 µg/ml).
To detect CAR by flow cytometry, 1 x 106 cells were preincubated in 100 µl 0.1% BSA (in PBS, 20 min, room temperature) before primary antibody (RmcB; mouse IgG1, 1:100 dilution in PBS/0.1% BSA, 90 min, 4°C) was admixed on an orbital shaker. RmcB, a murine monoclonal antibody that recognizes the extracellular amino terminus region of CAR (26) was purified from hybridoma (CRL-2379; ATCC) SCID-mouse ascites using protein-G affinity chromatography (Amersham Pharmacia Biotech, Piscataway, NJ). Surface expression for the
Adenoviral Vectors and FGF2-Fab' Ad Conjugate The general methodology for conjugate preparation is based on published protocols (23, 24). The FGF2–Fab' conjugate components were obtained from Selective Genetics Inc. (San Diego, CA). Briefly, a bi-functional molecule was generated by conjugating a modified recombinant 155–amino acid form of FGF2 (28) to a neutralizing Fab directed against the knob domain of the Ad5 fiber protein (1D6.14 [29]) in a 1:1 molar ratio (16, 22). This FGF2–Fab' conjugate was then admixed with the Ad vector in 0.1 M NaCl, 20 mM HEPES pH 7.6 (binding buffer) for 30 min at room temperature to generate the FGF2-retargeted Ad. We empirically determined that a 100-fold Molar excess (indexed to the number of fibers on Ad particles) of the FGF2–Fab conjugate yielded the optimal formulation for gene transfer, and this formulation was subsequently used for further transduction studies in vitro and in vivo.
Receptor Blocking Experiments
Transduction Studies and Reporter Gene Expression Analysis
For studies in vivo, 4-wk-old female SCID/SCID were injected subcutaneously in the flanks with 8 x 106 cells/mouse (four mice per group). Animal studies were approved by the Institutional Animal Care and Use Committee and conformed to the guidelines set forth in the Guide for the Care and Use of Laboratory Animals (National Research Council; Washington, D.C.). When tumors reached a cross-sectional area between 0.5 and 1 x 0.5 and 1 cm2 (2–4 wk depending on the cell line), they were directly injected with Ad5LacZ or FGF-retargeted Ad at an MOI of 100 in a volume of 50 µl of PBS. The number of cells in the tumor (for determination of MOI) were estimated based on the following measurements and assumptions: (1) bisecting diameters of the tumors were measured and the tumor volume was approximated using the formula 0.4 (ab2), where a is the long measured axis of the tumor and b is the short measured axis (31); (2) the cells within the tumors were assumed to be spherical (volume 4/3
Statistical Analysis
A panel of ten different lung cancer cell lines, representing SCLC, NSCLC-adenocarcinoma (AdenoCa), squamous cell carcinoma (SCCa), large cell carcinoma, and bronchoalveolar carcinoma (BAC) subtypes, was evaluated for surface expression of CAR, FGF receptors, and v-integrin expression (Table 1). Lung cancer cells have a heterogeneous distribution of these receptors, especially CAR. The differences in surface CAR enabled us to clearly distinguish cells with high CAR expression (e.g., NCI-H1703, NCI-H157, and NCI-H226) versus relatively low CAR expression (e.g., NCI-H1437, NCI-H2122, and NCI-H460), but a similar categorization based on FGFR or v-integrin expression was not possible.
To determine whether transduction efficiency correlated with the surface expression of vector-specific attachment and entry receptors, gene transfer into the complete panel of these cells was performed, using native Ad versus FGF2-retargeted Ad vectors (Figure 1). Expectedly, native Ad transduction efficiency directly correlated with surface CAR expression (R = 0.77 with 95% CI between 0.28 and 0.94, two-tailed P value = 0.0085). Surprisingly, there was no significant correlation of Ad-transduction with v-integrin expression (R = 0.29, P = 0.425), or between FGF receptor expression and the FGF2-Ad conjugate–mediated transduction efficiency (R = –0.28, P = 0.44).
However, in line with expectations, the conjugate markedly increased infectivity in those cells that displayed a paucity of surface CAR expression (Figure 1A). Comparatively modest (but statistically significant) increases in transduction were evident for the BAC and SCLC subtypes as well (Figures 1C and 1D). In contrast, in cells that displayed high CAR expression, gene transfer was equivalent or better with the Ad vectors (Figure 1B), despite the fact that these cells had equivalent or greater expression of FGF receptors (Table 1). These results indicated that FGF retargeting does not confer a universal transduction efficiency advantage for lung cancer cells, but only in those that have a relative paucity of CAR.
To further study the importance of specific ligand–receptor interactions for gene transfer, blocking studies were performed. Cells were exposed to the FGF2-retargeted Ad in the presence or absence of specific inhibitors of the Ad component of the conjugate (Figure 2). Using gene transfer mediated by the conjugate in the absence of inhibitors as an index of optimal transduction efficiency, we determined that specific blockers (anti-Ad5 fiber knob, the monoclonal anti-CAR antibody RmcB, and excess Ad5 fiber knob) of the CAR–Ad fiber interaction did not significantly impede gene transfer into either CAR-positive or CAR-deficient lung cancer cells. The effect of blocking vector interaction with the
We next examined whether FGF2-Ad transduction was mediated through attachment with high- and/or low-affinity (HSPG) FGF receptors. Using the gene transfer mediated by the conjugate in the absence of inhibitors as an index of maximal transduction efficiency, soluble HSPG significantly antagonized conjugate mediated gene transfer (Figure 3A). However, the extent of this inhibition was marginal and relatively dose-independent in the range of soluble HSPG concentrations used. By comparison, the interaction of the FGF2-Ad conjugate with the high-affinity FGF receptors was competitively inhibited with soluble FGF in a dose-dependent manner (Figure 3B). Whereas free FGF did not impair gene transfer efficiency at 1:1 molar excess, significant inhibition was evident as the concentration of free FGF in the transduction medium was increased by 10- or 100-fold molar excess to that present in the conjugate. As a result, FGF2-Ad gene transfer in the presence of 100-fold molar FGF excess was similar to that exhibited by the Ad vector alone in the CAR-deficient target cells (Figure 3B). Collectively examining the data from Table 1, Figure 1, and Figure 3 gives the impression that although the measured cell surface FGF receptor expression could not predict the efficiency of conjugate gene transfer, FGF2-Ad does appear to use the high-affinity cellular FGFRs for cell transduction.
The crux of effective targeting lies in the vector's ability to achieve an advantage in gene transfer in vivo. To determine whether CAR-deficient lung cancer cells could be effectively transduced with FGF2-Ad in vivo, we evaluated gene transfer efficiency into NSCLC-tumor xenografts following a single injection controlled for equivalent volume and vector dose (Ad-MOI = 100) administered intratumorally. The results paralleled those determined by in vitro testing. FGF2-retargeted Ad accomplished gene delivery comparable to recombinant Ad in CAR (+) lung cancer cells, but was distinctly better at gene transfer into the CAR-deficient lung cancer xenografts (Figure 4). Parenthetically, in terms of percentage of cells transduced, this enhancement in gene delivery was generally associated with an increase from < 1% to 6 ± 3% transduction within the tumor xenografts generated using CAR-deficient cells.
The FGF2-retargeted Ad had been touted as being superior and generally more efficacious than the native Ad in several earlier reports. Our prior experience indicates that a single vector type is unlikely to be uniformly efficacious for transductional targeting (1, 3, 5–7, 33), and our inability to obtain efficient gene transfer to CAR-deficient lung cancer cells prompted us to test FGF2-Ad. We compared the efficacy of the FGF2-Ad vector versus native Ad for gene transfer into a panel of lung cancers representing various subtypes.
Given the heterogeneity that characterizes various lung cancer subtypes, it is not surprising that the panel of cells chosen for study showed variability in terms of surface CAR expression (Table 1). Conversely, pharmacologically relevant differences in
Basic FGF is reported to bind tightly to all four FGF receptors with picomolar affinity (34). In comparison, the interaction of CAR with the knob domain of the Ad5 fiber protein has a dissociation rate constant (Kd) of
What role do Considering many of the variables (e.g., surface expression of specific receptors and specific inhibitors of ligand–receptor interactions) that may impact Ad or FGF2-Ad entry into target cells, the in vivo studies were designed simply to determine whether FGF2-Ad conferred an advantage for gene delivery into CAR-deficient lung cancer cells in situ. This simplistic design was reasonable given the realization that the pharmacologic properties of the FGF2-Ad conjugate vectors are probably too imprecise for clinical use. Accordingly, a single vector dose was administered with a single intratumoral injection, to assess the relative importance of ligand–receptor interactions in the tumor microenvironment for mediating gene transfer in vivo. Recognizing that a number of variables (e.g., the intratumoral dispersion of vector, dwell time of vector within tumor bed, and/or clearance of vector by host circulation or reticuloendothelial cells) in vivo were not directly measured, gene transfer into CAR-deficient tumors was still determined to be greater with the FGF2-Ad than native Ad. Again, FGF2-Ad did not confer a significant advantage over native Ad in tumors generated from cells with relatively high surface CAR expression (Figure 4), as gene transfer efficiency was equivalent in this subset in vivo. These pilot studies indicate that the FGF2-retargeted Ad vectors are not a panacea for achieving high-efficiency transduction of all lung cancers. In addition, several improvements in their design (e.g., incorporating the FGF2-targeting moiety directly within the viral capsid) may be useful to overcome the cumbersome formulation of the conjugate that, dependent on Ad-particle to pfu ratios, may confer different pharmacologic properties on the vector from one viral preparation to another. Given that these vectors do provide a significant transduction advantage for CAR-deficient lung cancer cells, efforts to make such improvements in vector design are indicated. FGF2-retargeted Ad may have a distinct niche in the gene therapy of lung cancers.
The authors thank S. K. Tran for assistance with manuscript preparation.
This project has received support from the American Lung Association Career Investigator Award, NIH-R01-CA78654, the Veterans Administration-Medical Research Funds, NIH-1P50–90388, and the UCLA-Jonsson Comprehensive Cancer Center Flow Cytometry Facility. Conflict of Interest Statement: M.Q. received partial salary support from Titan Pharmaceuticals; B.E. received no industry support for his participation in this study; S.S. received no industry support for his participation in this study; and R.K.B. received reagent and supply support from Titan Pharmaceuticals for the initiation of this study, but Titan withdrew support after one year of the two-year contract, resulting in the continuation/completion of the study being undertaken by him and colleagues with ALA, NIH, and VA funds. Received in original form July 15, 2004 Received in final form December 2, 2004
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
r3) with a radius of 10 µm; and (3) the entirety of the tumor mass was assumed to be comprised of tumor cells. Thirty hours after vector injections, the tumors were extirpated, washed with ice-cold PBS, and an antiprotease cocktail mixture 10 µl/ml (Sigma), and homogenized on ice in reporter lysis buffer 1x (Promega, Madison, WI) containing antiproteases (10 µl/ml; Sigma). The insoluble debris was sedimented, and the supernatant was recovered and stored at –80°C. ß-Gal activity (as measured by absorbance at 420 nm after hydrolysis of o-nitrophenyl-ß-D-galactopyranoside over 30 min) was normalized to the total protein concentration (BCA assay; Pierce, Rockford, IL) within tumor homogenates, and measured in the batched supernatants. 
0.5–1 x 10–8 M (