INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

E1-deleted replication-defective adenovirus (Ad) vectors are attractive candidates for gene transfer because of their ability to transduce a wide variety of dividing and nondividing tissues in vivo (4, 14, 16, 17, 19, 30). We and others have used such Ad vectors for gene transfer to the respiratory epithelia of experimental animals and patients with cystic fibrosis (CF) (3, 9, 14, 24, 28-30). Early studies from several investigators have suggested that administration of high doses of E1-deleted Ad vector results in only transient gene expression in vivo (4, 5, 23, 26, 27, 33). Results of experiments carried out with a variety of immunodeficient and immunocompetent strains of mice have suggested that the transience of gene expression is due, at least in part, to the destruction of vector-transduced cells by host cellular immune responses (predominantly CD8⁺ cytotoxic T cells) directed against viral proteins (4, 5, 23, 26, 27, 33). Reduction of this cellular immune response with second-generation Ad vectors with modification or deletion of the E2 and E4 regions (5, 21, 24) has been reported. However, interpretation of these studies is complicated because of the immunogenic nature of the transgenes such as Escherichia coli -galactosidase and luciferase, which were used in these experiments.

More recent studies have demonstrated persistent expression in several strains of mice following intramuscular injection of an Ad vector encoding mouse erythropoietin (19). Other studies have shown that Ad vectors expressing human alpha 1-antitrypsin or human factor IX as the transgene can give rise to long-term expression when the vectors are delivered intravenously to the livers of C57BL mice but not with other strains (2, 11-13, 20). The prolonged expression in all these studies appears to correlate with the absence of antibodies to the secreted transgene product (11, 12). To date, there have been no reports of an Ad vector capable of persistent transgene expression in the airways of adult immunocompetent animals. Here we describe the construction and in vivo characterization of Ad vectors which encode a therapeutic gene, the human CF transmembrane conductance regulator (hCFTR), and give persistent transgene expression in the lungs of normal immunocompetent mice and functional CFTR expression in the nasal epithelia of CF mutant mice.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Ad vectors. Ad2/CFTR-2 is an Ad2-based vector with most of the E1 region (nucleotides 357 to 3328) deleted and replaced with the CFTR expression cassette (9). Ad2/CFTR-2 contains a PGK promoter driving hCFTR as the transgene, followed by a bovine growth hormone poly(A) signal and retains wild-type (wt) E2 and E3 regions. The E4 transcription unit has been replaced with open reading frame 6 (ORF6) of E4.

Cytotoxic T-cell assay. The detailed protocol for cytotoxic T lymphocyte (CTL) assays was essentially as described previously (8, 15, 20). Briefly, spleen cells from animals treated with Ad2/CFTR-16 were pooled and stimulated in vitro with syngeneic fibroblasts infected with Ad2/CFTR-16 at a multiplicity of infection of 100. Cytolytic activity was assayed after 6 days of culture. Target fibroblasts were infected with different Ad2/CFTR vectors at a multiplicity of infection of 100, labeled with ⁵¹Chromium (⁵¹Cr; NEN) overnight (30 µCi/10⁵ cells), and added to the wells of a round-bottom 96-well plate in a 100-µl volume (5 × 10³ fibroblasts/well). Effector cells were added in a 100-µl volume at various effector/target cell ratios in triplicate. After 5 h of incubation at 37°C with 5% CO₂, 20 µl of cell-free supernatant was collected from each well and counted in a Microbeta Trilux liquid scintillation counter (Wallac Inc., Gaithersburg, Md.). The percentage of lysis was calculated as follows: % lysis = {[(sample counts per minute)  (spontaneous counts per minute)]/[(total counts per minute)  (spontaneous counts per minute)]} × 100.

RNA extraction and analysis. Total RNA was extracted from lung tissue with acid guanidium thiocyanate-phenol-chloroform and treated with DNase (RQ1 Dnase; Promega, Madison, Wis.). Levels of hCFTR mRNA were determined by quantitative reverse transcriptase (RT)-PCR essentially as described previously (10). RNA samples from animals in the same group were pooled. The RT reaction was performed with a cDNA kit (Invitrogen). Following reverse transcription, cDNA was amplified with 2.5 U of Taq polymerase (Perkin-Elmer) in a mixture containing a final concentration of 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, 0.01% gelatin, 200 µM (each) dTTP, dATP, dGTP, 150 µM dCTP, 500 nM (each) primer, and 0.5 µl [³³P]dCTP (3,000 Ci/mmol). Amplification was carried out after denaturation for 1 min at 94°C by 30 amplification cycles consisting of 60 s at 92°C, 60 s at 62.5°C, and 60 s at 72°C, with a final extension step of 5 min at 72°C. Amplified products were separated by gel electrophoresis in a 1.5% agarose gel and stained with ethidium bromide. The gel was dried, analyzed on a PhosphorImager (Molecular Dynamics), and quantitated with image analysis software (ImageQuant).

-Galactosidase assay. Lungs from individual animals were homogenized, and -galactosidase activity was measured by the Galactolight assay (Tropix). The protein concentration in lung homogenate was measured with the Bio-Rad DC reagent, and -galactosidase activity is expressed as relative light units per microgram of lung protein.

In vivo measurement of nasal PD. FABP-hCFTR double transgenic CF (/) mice (32) were obtained from the Jackson Laboratory. Ad2/CFTR-16 was administered to the nasal mucosa of these mice as described previously (7). Briefly, mice were anesthetized with an intraperitoneal injection of 2,2,2-tribromoethanol (0.4 mg/g of body weight) and t-amyl alcohol (0.4 µl/g) in 0.9% NaCl. The solution containing Ad2/CFTR-16 (10⁹ IU in 100 µl of phosphate-buffered saline) was perfused into the nasal cavity with a catheter (pulled from PE20 tubing) connected to a 1-ml syringe mounted in a microperfusion pump at a rate of 1.6 µl/min. By continuous perfusion, the mouse nasal epithelium was constantly exposed to the adenovirus vectors over 1 h. The potential difference (PD) across the nasal epithelia of the CF mice was measured as described previously (6, 7, 31).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Construction of new Ad2/CFTR vectors. A primary goal in designing new Ad2 vectors for CF gene therapy was to obtain vectors that are capable of directing persistent hCFTR gene expression in vivo. The CMV enhancer-promoter was chosen to direct expression of hCFTR cDNA because we have shown previously that this promoter can direct prolonged transgene expression in the lung, the target organ for CF gene therapy (1). Sustained CMV promoter-driven expression is dependent on E4 gene activity; thus, the wt E4 region was included in the new vectors (1). The Ad E3 region is known to encode several proteins involved in evasion from host immune surveillance (22). In Ad2/CFTR-16, the gp19K coding sequences from E3A were retained on the basis of reports that expression of the gp19K protein may improve the longevity of gene expression from Ad vectors, presumably by inhibiting major histocompatibility complex (MHC) class I antigen presentation (13). The E311.6K (Ad death protein [18]) coding region was deleted from Ad2/CFTR-16. All of the E3B coding sequences were deleted, since these sequences have not been shown clearly to improve Ad vector performance in vivo, and because some viral sequences had to be deleted to allow for efficient vector packaging. No modifications were made in the E2 region.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Persistence of hCFTR expression in nude mice. Nude C57BL/6 mice were instilled intranasally with 2 × 109 IU of Ad2/CFTR-16 (). Levels of hCFTR mRNA in the lung were measured at different time points by quantitative RT-PCR as described in Materials and Methods. Lung tissues from four mice at each time point were pooled, and the results shown are the average of duplicate RT-PCR assays.

Persistence of Ad2/CFTR vector-directed gene expression in the lungs of immunocompetent mice. The ability of Ad vectors to direct prolonged transgene expression in vivo has been correlated primarily with the absence of an immune response to the transgene product (11, 12, 19). In several studies of Ad vector-directed human CFTR gene transfer in mice in our laboratory, we have tested for but have not observed immune responses to hCFTR (data not shown). It appears that the levels of hCFTR expressed from the CMV enhancer-promoter, combined with the routes of administration used, are insufficient to provoke a detectable immune response. Thus hCFTR, approximately 89% identical to the mouse CFTR coding region at the amino acid sequence level, can be considered a weakly immunogenic transgene in conventional inbred strains of mice.

View larger version (17K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   FIG. 2.   Persistence of hCFTR expression in immunocompetent mice. Different strains of mice were instilled intranasally with 2 × 109 IU of Ad2/CFTR-2, Ad2/CFTR-5, or Ad2/CFTR-16. Levels of hCFTR mRNA expressed in the lung at different time points were measured by quantitative RT-PCR as described in Materials and Methods. Lung tissues from animals in the same treatment group (four mice/group/strain) were pooled, and results shown are the average of duplicate RT-PCR assays. (a) , Ad2/CFTR-5. (b) , Ad2/CFTR-2. , Ad2/CFTR-16.

View larger version (21K): [in this window] [in a new window]   FIG. 3.   Qualitative RT-PCR analysis. Ad2/CFTR-5 or Ad2/CFTR-16 (2 × 109 IU) was instilled intranasally into C57BL/6 (a) or BALB/c (b) mice, and lung RNA was isolated on different days postinstillation. RT-PCR was performed as described in Results. The PCR products were digested with MspI, which cleaves only hCFTR, and analyzed by gel electrophoresis.

Functional CFTR expression in nasal epithelia of CFTR mutant mice. It is clear from the above-mentioned experimental results that Ad2/CFTR-16 can direct long-term expression of hCFTR mRNA in the lungs of immunocompetent mice. Under the experimental conditions used, the majority of gene transfer is to the small airways within the lung, the desired location for CF gene therapy. However, the small airways of the mouse are inaccessible to current techniques for measurement of chloride secretion. Previously, we have measured the ability of Ad2/CFTR-5 to correct the CF chloride secretion defect in the nasal passage of a strain of CFTR knockout mouse (7). We found that Ad2/CFTR-5 can correct the nasal defect on day 2 postadministration but that this correction had largely vanished by 7 days postadministration (7). To measure the ability of Ad2/CFTR-16 to correct the CF chloride secretion defect, similar gene transfer experiments were carried out in the nasal cavities of mice defective for CFTR activity. The mice employed in these experiments bear a knock-out deletion within the endogenous CFTR gene but are transgenic for hCFTR cDNA controlled by an intestine-specific promoter (32). This genetic composition renders the animals functionally defective for CFTR within the respiratory tract, including the nasal epithelium.

View larger version (26K): [in this window] [in a new window]   FIG. 4.   Functional CFTR expression in nasal epithelia of CF mutant mice. Ad2/CFTR-16 (109 IU) was perfused over the nasal epithelia of CF mutant mice over a 60-min period. Basal nasal PD (a) and change in PD in response to low Cl substitution (b) were measured on days 2, 7, and 15 posttreatment. Data are expressed as mean ± standard errors of the mean (n = 4).

Anti-Ad vector CTLs are present in immunocompetent mice treated with Ad2/CFTR vectors. A CTL response against E1-deleted Ad vectors has been demonstrated by several investigators following vector delivery to the lungs or livers of immunocompetent mice, and it has been shown that viral proteins expressed from E1-deleted Ad vectors are targets for Ad-specific CTL in vitro (10, 11, 20, 23, 26, 27). We have shown previously that these vector-specific CTL do not necessarily limit vector expression from transfected cells in the liver (20). To test whether the longevity of expression from Ad2/CFTR vectors could be explained by an unforeseen failure to provoke a CTL response, leading to an escape from immune surveillance, we investigated the CTL response following intranasal delivery of Ad2/CFTR-16. Consistent with previous studies from our laboratory and others, we were able to demonstrate a vector-specific CTL response in C57BL/6 (data not shown) and BALB/c mice (Fig. 5). We do not find any significant differences in cytolysis with target cells infected with Ad2/CFTR-16, Ad2/CFTR-5, or Ad2/CFTR/E3. In several studies conducted in our laboratory, we have not been able to demonstrate a CTL response specific for hCFTR in Ad/CFTR vector-treated mice.

View larger version (15K): [in this window] [in a new window]   FIG. 5.   CTL response to Ad2/CFTR vectors. BALB/c mice were instilled with 2 × 109 IU of Ad2/CFTR-16 on day 0. On day 21, spleen cells were collected, restimulated in vitro with Ad2/CFTR-16-infected syngeneic fibroblasts, and tested for cytolytic activity against target cells infected with different Ad/CFTR vectors. Results shown are the mean percentages of lysis from triplicate wells at various effector/target ratios. Targets: uninfected (); Ad2/CFTR-5 (); Ad2/CFTR-16 (); Ad2/CFTR/E3 ().

Repeat administration of Ad2/CFTR-16. Our results demonstrate that the CTL response against an Ad vector expressing a weakly immunogenic transgene is ineffective after a single administration of the vector to the lung, as has been demonstrated previously for administration to the liver (20). In terms of persistence of transgene expression, an important issue for effective use of gene therapy vectors for the treatment of CF is the CTL response following repeated administration. Effective repeated administration of Ad vectors can be achieved through the use of immunomodulatory drugs such as antibody to CD40 ligand (15, 25) or deoxyspergualin (8). The use of such agents depresses the neutralizing antibody response and allows for repeated gene transfer but also obscures investigation of the nature of the cellular immune response to repeated vector administration. To allow investigation of the effect of a CTL response to repeated vector treatment on the longevity of Ad2/CFTR-16 expression, we employed a strategy designed to prime the immune system to Ad and CFTR antigens and to deliver a second vector dose before neutralizing antibodies could rise to a level that would prevent administration. BALB/c mice were treated with a priming dose (5 × 10⁸ IU) of Ad2/CFTR-16 or an empty Ad vector lacking the hCFTR cDNA (Ad2/EV) on day 0 and were then challenged with a dose of Ad2/CFTR-16 on day 14. During this time interval, a CTL response to vector antigens is stimulated, but neutralizing antibodies do not reach inhibitory levels (data not shown). On day 0, a parallel group of animals were treated with a dose of 10⁹ IU of Ad2/CFTR-16, and the vector-derived hCFTR mRNA levels were measured until day 59 (Fig. 6). There was no indication that previous exposure of animals to either Ad2/CFTR-16 or Ad2/EV had any impact on the duration of expression from the challenge dose of Ad2/CFTR-16 compared to that in animals receiving only a single vector dose.

View larger version (17K): [in this window] [in a new window]   FIG. 6.   Repeat dosing with Ad2/CFTR-16. BALB/c mice were instilled with 5 × 108 IU of Ad2/EV or Ad2/CFTR-16 on day 0 and challenged with 109 IU of Ad2/CFTR-16 on day 14. RNA isolation and RT-PCR analysis were performed as described in Materials and Methods. , CF16 on day 0; , CF16 on days 0 and 14; , EV on day 0 and CF-16 on day 14.

Role of E3 in persistence. Poller et al. (13) have reported that the E3A region plays an important role in persistence of a human factor IX expressing Ad vector in the livers of immunocompetent mice. To determine the role of E3 in the persistence of transgene expression with Ad2/CFTR-16 in the lungs of immunocompetent mice, normal BALB/c mice were instilled intranasally with 5 × 10⁸ IU of Ad2/CFTR-16 or Ad2/CFTR/E3, a vector that is completely deleted for the E3 region. hCFTR mRNA expression was measured over time in total lung RNA by quantitative RT-PCR. Expression of hCFTR mRNA was essentially undiminished with both vectors until 70 days, the last time point analyzed (Fig. 7a), indicating that the endogenous E3 region does not play a significant role in the persistence of these vectors expressing a weakly immunogenic transgene. Parallel groups of immunocompetent C57BL/6 mice were instilled intranasally with 10⁹ IU of Ad2/CMVgal-1 (wt for E3 and E4) or Ad2/CMVgal/E3 (E3 completely deleted and wt for E4). -Galactosidase activity was measured over time by the Galactolight assay. Expression of -galactosidase declined rapidly to background levels by day 21 with both vectors (Fig. 7b), indicating that the inclusion of the endogenous E3 region in Ad vectors does not improve persistence of an Ad vector expressing a highly immunogenic transgene.

View larger version (12K): [in this window] [in a new window]   FIG. 7.   Role of E3 region in persistence. (a) Mice were instilled with 5 × 108 IU of Ad2/CFTR-16 () or Ad2/CFTR/E3 (). RT-PCR analysis was performed as described in Materials and Methods. (b) C57BL/6 mice were instilled with 109 IU of Ad2/CMVgal-1 () or Ad2/CMVgal/E3 (). -Galactosidase (gal) expression was measured as described in Materials and Methods. RLU, relative light units.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Prominent among the potential limitations of Ad vectors for gene therapy is the belief that in vivo expression of Ad vector-encoded transgenes is inherently transient in nature. Early reports highlighted the transient nature of in vivo Ad vector-mediated gene transfer such that gene expression was seen to plummet within 2 to 3 weeks to levels representing only a small fraction of that measured just after administration (4, 24, 26, 33). In principle, any of a number of events can result in truncated gene expression: vector-induced cytotoxicity, vector- or transgene-specific CTL attack, complete loss of vector DNA, or promoter shutoff. Persistent gene expression can occur only in the absence of each of these potential limitations. The vast majority of studies on gene expression have relied on the use of foreign reporter genes, the protein products of which are highly immunogenic and capable of provoking strong cellular and/or humoral immune responses; this immune response appears to be responsible for extinguishing expression from the Ad vector under study. The realization that Ad vector-specific CTL may not necessarily be effective in eliminating transfected cells came from the demonstration that long-term gene expression could be achieved in immunocompetent animals under conditions in which there were no immune responses to the vector-encoded reporter gene product (2, 11-13, 19, 20). The conclusions of those studies were limited to liver and muscle and were mouse strain dependent. The current study employs human CFTR as the transgene, which, although potentially immunogenic in mouse, has failed to provoke a detectable immune response as expressed from our Ad vectors. Persistent gene expression in the lung has been achieved through the use of a specific vector genotype, the CMV enhancer-promoter for transgene expression in conjunction with a wt E4 region (1).

Having demonstrated that human CFTR was weakly immunogenic in normal mice and that the CMV promoter combined with a wt E4 region could direct prolonged -galactosidase expression in nude mice, new vectors incorporating these features were constructed. Although it was predicted that these new Ad/CFTR vectors would escape immune surveillance and that its gene expression cassette would remain active, the possibility remained that prolonged hCFTR expression in a normal mouse might elicit immune and/or inflammatory responses sufficient to extinguish gene expression. However, the results of the current study demonstrate clearly and consistently that Ad2/CFTR vectors can direct expression of hCFTR mRNA at undiminished levels for at least 70 days. Moreover, functional correction of the CF chloride secretion defect in the nasal epithelium was undiminished for up to 15 days, the longest interval tested.

Persistent gene expression was observed despite the presence of an immediate nonspecific response to vector administration characterized by inflammatory cells and cytokines and a vector-specific cellular immune response. These responses have been implicated broadly as representing intrinsic limitations to the effectiveness of Ad vectors for gene therapy. The results of this study indicate clearly that in the three immunocompetent inbred mouse strains tested, conventional Ad vectors of the appropriate genotype can direct long-term gene expression in the lung. However, a similar study performed by Yang et al. in the mouse lung (24) with an E1-deleted Ad5/CFTR vector gave rise to CFTR expression that declined to undetectable levels by 21 days. An important difference between the vector used in that study (24) and the Ad2/CFTR vectors used in this study is in the promoter used to drive hCFTR expression, a crucial determinant for maintenance of persistent transgene expression (1). The Ad vector used by Yang et al. utilized a CMV enhancer and chicken -actin promoter (24).

Poller et al. (13) recently reported that incorporation of E3A improved the persistence of a human factor IX encoding Ad vector in the livers of C57BL/6 mice. In the context of the mouse lung, we find that endogenous E3 does not play a significant role in the persistent transgene expression from Ad vectors expressing weakly immunogenic transgenes such as hCFTR. However, if a highly immunogenic transgene like -galactosidase is used, the inclusion of the E3 region does not improve persistence. This does not rule out the possibility that overexpression of E3 proteins may have benefits in the mouse or other experimental systems (6a). Ad vectors do provoke a CTL response in C57BL/6 and BALB/c mice (10), and the E3 gp19K protein does bind to the MHC I heavy chains encoded by these strains. However, the C3H mouse exhibits a weak CTL response to Ad vector (10), and the MHC I heavy chain from this strain binds only weakly to the E3 gp19K protein.

While the results of this study are encouraging in the sense that vector-directed gene expression was achieved for a period of several weeks, a key question that remains is the half-life of the respiratory epithelial cell. Since the results of numerous studies indicated that respiratory epithelial cells in a number of species are replaced over a period of 60 to 100 days, we arbitrarily terminated our gene expression studies at 70 days. The issue of longevity of respiratory epithelial cells in the CF patient's lung clearly cannot be resolved in the rodent model and must be addressed in the context of a clinical study.

The treatment of certain genetic diseases such as CF will require repeated administrations throughout the lifetime of the patient. It is known that Ad vector-based gene delivery gives rise to a dose-dependent humoral immune response leading to the development of neutralizing antibodies to adenoviruses (2, 4, 9, 23, 28) which reduces the efficiency of repeated gene transfer to a fraction of that achieved initially. Switching of vector serotypes may circumvent the neutralizing antibody issue, although practical considerations such as development, approval, and manufacturing of multiple vectors for a single disease indication may limit this approach. Recent studies have shown that transient immunosuppression can effectively block the humoral response and allow repeated Ad vector administration to the mouse lung (8, 15, 25). This strategy could have undesirable side effects and remains untested in humans.

Our demonstration of Ad vector-mediated long-term transgene expression in the murine lung taken together with the transient immunosuppression strategies to block the anti-Ad neutralizing antibody response provides hope that gene therapy for CF may be clinically feasible in the future. Further studies of this nature are clearly warranted in nonhuman primates.