Introduction

Ocular gene therapy in the anterior segment of the eye has received considerable interest for applications such as enhancement of corneal wound healing associated with excimer laser photorefractive keratectomy,¹ enhancement of conjunctival wound healing associated with pemphigoid, trachoma and glaucoma filtration,¹ correction of allograft rejection after corneal transplantation² and suppression of lacrimal gland autoimmunity associated with Sjögren's syndrome.³ Adenovirus (Ad) vectors, leading candidates for gene therapy to the anterior portion of the eye, do not require cycling of the target cell for gene transfer, and are renowned for mediating gene transfer to nondividing cells, particularly mucosal epithelial cells, with high efficiency.^{4, 5, 6} On the other hand, Ad is intrinsically a lytic virus and is well known to express proteins that effect alterations in a wide variety of cellular processes, including host cell cycle, protein synthesis and survival. Even the conventional E1-, E3-deleted Ad vectors in current use have been reported to express adenoviral coding sequences at a level sufficient to elicit a robust Class I immune response against transduced cells in vivo, exhibit cytotoxicity at high doses, and have been responsible for the only reported instance of a fatality associated with a clinical gene therapy trial to date.^{7, 8, 9, 10, 11, 12, 13} A particularly attractive feature of the eye within the context of Ad-mediated gene therapy is its relatively 'immune-privileged' status, although some examples of immunogenicity associated with Ad-mediated ocular gene therapy have been reported.^{12, 13}

Although the potential for therapeutic gene transfer by replication-deficient Ad has resulted in extensive investigations of its toxicity and immunogenicity, less information is available on the cellular effects associated with Ad binding, internalization and trafficking to the nucleus, particularly in epithelial cells that represent normal targets for Ad infection. In addition to the use of Ad-based viral vectors, such information is relevant to feasibility and safety studies for next-generation gene therapy vectors, which attempt to overcome issues associated with viral immunogenicity and toxicity by incorporating Ad capsid proteins into nonviral gene delivery systems.¹⁴ Analysis of the cellular effects of Ad internalization in host cells also represents an alternative approach to elucidating virus–host interactions and to designing new therapeutic approaches to prevent Ad-induced pathology. Within the context of ocular gene therapy with Ad and Ad-derived vectors, it is of paramount importance to understand their effects in cell types responsible for ocular homeostasis.

The acinar epithelial cells of the lacrimal gland are secretory epithelial cells, responsible for the production and release (basal and secretagogue-stimulated) of a variety of tear proteins (eg, antibodies, lysosomal hydrolases, growth factors and lactoferrin) into ocular fluid.^{15, 16, 17} The production and release of tear fluid of appropriate composition is essential for maintaining corneal and conjunctival integrity, protecting against infection and preserving visual acuity. Deficiencies or changes in tear fluid abundance or composition may lead to corneal dessication and ulceration, corneal or conjunctival infection and, in severe cases, blindness. Diseases characterized by major deficits in lacrimal gland function range in severity from keratoconjunctivitis sicca (dry eye) to the autoimmune disease, Sjögren's syndrome.^{18, 19} As well, the dryness associated with decreased corneal sensation and concomitant reduction in tear flow following refractive surgery can complicate corneal recovery.²⁰

Since preservation of tear production and release is essential for ocular health, we questioned whether exposure to Ad would alter the functioning of this essential gland. As an initial step toward this goal, we investigated the effect of replication-defective adenovirus serotype 5 (Ad5) containing either green fluorescent protein (GFP) or -galactosidase (LacZ) on the organization and function of the lacrimal acinar secretory pathway in reconstituted acinus-like structures formed in vitro from acinar cells from rabbit lacrimal gland.^{21, 22, 23} At conditions associated with high-efficiency transduction (>80%), Ad elicited specific changes in the organization and function of the stimulated secretory pathway, changes that could be largely recapitulated by ultraviolet light (UV)-inactivated Ad. These findings suggest that ocular gene therapy utilizing Ad or Ad-derived products may alter secretory membrane-trafficking functions in epithelial cells of the ocular surface.

Results

Ad-mediated gene transfer into lacrimal acinar epithelial cells

Previous studies have defined conditions for culture of isolated rabbit lacrimal acini such that the acinar cells reassociate into acinus-like structures exhibiting central lumena. Individual cells within these structures display distinct apical and basal–lateral domains, a polarized cytoskeleton, secretory vesicles and also release protein stores in response to secretagogues.^{21, 23} To determine the transduction efficiency of the recombinant Ad vectors, acinar cells were analyzed by fluorescence-activated cell sorting (FACS). The addition of Ad 5-containing green fluorescent protein (Ad-GFP) to reconstituted acini over a range of multiplicity of infections (MOIs) demonstrated that exposure at an MOI of 5 for 16–18 h resulted in high transduction efficiency. FACS analysis showed that 80.16.1% (n=4) of Ad-GFP-transduced cells were GFP positive, whereas non-transduced cells yielded little to no fluorescence in the emission channel associated with GFP fluorescence (0.50.2%) (Figure 1a). These results were confirmed by fluorescence and phase microscopy of Ad-GFP- and Ad5-containing -galactosidase (Ad-LacZ)-transduced acini, respectively (Figure 1b). Although Ad exposure elicited some loss of cells, the viability of the remaining cells on the plate assessed from Trypan blue exclusion was only modestly reduced relative to non-transduced acini (79% for Ad-LacZ and 86% for Ad-GFP, n=4).

Figure 1.

Ad-mediated high-efficiency gene transfer into primary rabbit lacrimal acini. (a) FACS analysis of nontransduced (control) and Ad-GFP-transduced acinar cells. Flow cytometric analysis was performed on a FAC-Scan™ using the FL1 channel to quantitate GFP-expressing cells. The indicated gate is calibrated for GFP-positive events, which deviate more toward the green fluorescence (FL1) channel (x-axis) and away from the red (FL2) channel (y-axis). (b) After transduction as described in Materials and methods, primary rabbit lacrimal acini were fixed and processed for the detection of reporter genes. For analysis of GFP content, nontransduced (control) and Ad-GFP-transduced acini were imaged by fluorescence microscopy (left column). For analysis of -galactosidase expression, nontransduced (control) and Ad-LacZ-transduced acini were imaged by phase microscopy after staining with X-gal. The dark color in the phase images indicates -galactosidase expression. Bar, 20 m.

Full figure and legend (125K)

Ad transduction does not alter cytoskeleton or biosynthetic/endosomal membrane organization

Actin filaments in lacrimal acini are highly enriched in apical/lumenal regions (indicated by ^*). As shown by the confocal fluorescence microscopy images in Figure 2a, the enrichment of actin filaments beneath the apical membrane relative to the basolateral membrane was comparable in acini with and without transduction with Ad-GFP. Likewise, the confocal fluorescence microscopy images in Figure 2b showed that the anchorage of individual microtubules (MTs) beneath the apical plasma membrane and extension toward the basolateral plasma membrane was comparable in acini with and without transduction with Ad-GFP. The organization of the underlying cytoskeleton, a fundamental determinant of cell polarity, therefore appears unaffected by Ad transduction.

Figure 2.

Primary rabbit lacrimal acini transduced with Ad retain normal cytoskeletal and biosynthetic membrane organization. Nontransduced (control) and transduced (Ad-GFP) rabbit lacrimal acini were fixed and processed for the detection of (a) actin filaments (red), (b) MTs (red) or (c) Golgi apparatus (red, labeled with -adaptin antibody). GFP fluorescence is shown in green in all panels. A region of interest (box) in the lower magnification images in (a and b) was enlarged and presented to resolve details of the organization of cytoskeletal filaments. Bars in all panels, 10 m; ^*, apical/lumenal region; arrows, individual MTs extending from a subapical organizing center; arrowheads, characteristic perinuclear Golgi apparatus organization.

Full figure and legend (342K)

The accurate segregation of secretory proteins within the Golgi apparatus plays a fundamental role in the establishment of cell polarity and in secretory vesicle formation and maturation in acinar epithelial cells. We probed the organization of the Golgi apparatus in lacrimal acini, using -adaptin immunofluorescence to delineate Golgi membranes.²⁴ As shown in Figure 2c, -adaptin immunofluorescence (red) in untransduced lacrimal acini is detected in punctate spots often arrayed in a semicircular conformation, consistent with the perinuclear organization of Golgi stacks. No changes in Golgi organization were detected in acini transduced with Ad-GFP. A similar lack of effect was seen when we evaluated the effects of Ad transduction on the distributions of rab6, rab11, rab4 and polymeric immunoglobulin A receptor (pIgAR) – trans-Golgi network, endosomal, endosomal and transcytotic markers, respectively (data not shown).

Ad transduction selectively alters the organization of the stimulated secretory pathway

Rab3D is the principal isoform of the rab3 family of small GTPases associated with the large pool of mature secretory vesicles beneath the apical plasma membrane in lacrimal gland as well as pancreas and parotid gland.^{25, 26, 27} Stimulation of lacrimal acini causes the loss of subapical rab3D concomitant with the fusion of mature secretory vesicles with the apical plasma membrane.²⁷ As shown in the triply-labeled nontransduced and Ad-GFP transduced lacrimal acini in Figure 3, rab3D (red) in resting, nontransduced acini is enriched in the subapical cytoplasm beneath the actin-enriched (blue) apical plasma membrane surrounding the lumen (^*) in association with large (1–3 m) vesicular structures (arrows). Transduction with Ad-GFP (green) caused a striking change in rab3D immunofluorescence, with labeling dispersed throughout the cytoplasm (arrowheads) in GFP-labeled acini rather than beneath the apical plasma membrane.

Figure 3.

Apical enrichment of rab3D is dispersed by Ad-GFP transduction of rabbit lacrimal acini. Representative images of nontransduced (control) and transduced (Ad-GFP) rabbit lacrimal acini fixed and processed using appropriate primary and secondary antibodies and affinity label for the detection of actin (blue), rab3D (red) and GFP (green). Arrow, accumulation of secretory vesicles enriched in rab3D beneath the apical plasma membrane; arrowheads, dispersed rab3D fluorescence in the cytoplasm of transduced cell; bar, 10 m; ^*, apical/lumenal region.

Full figure and legend (374K)

We quantified this apparent change in cellular localization of rab3D from a predominantly apical to a dispersed distribution across multiple randomly selected lumena from nontransduced (control) and transduced lacrimal acini. The transduction efficiencies of Ad-GFP and Ad-LacZ were comparably high (>80%) such that most acini organized around lumena were transduced (Figure 1b). Since we were not selecting single images but scoring of hundreds of images, we simplified these experiments to use dual rather than triple labeling. Figure 4a shows representative images of rab3D labeling in control acini and acini transduced with Ad-LacZ, showing the characteristic dispersal of the normally apical rab3D (green) immunofluorescence labeling pattern previously seen in the Ad-GFP-transduced acini in Figure 3. We identified three distinct patterns for rab3D accumulation in lacrimal acini: an apical distribution with the majority of label localized in the apical-most quadrant of the cell in all cells organized around the lumen; a half-apical distribution with the majority of label localized in the apical-most quadrant of at least half of the cells organized around the lumen; a dispersed distribution with the majority of label distributed randomly throughout the cytoplasm in cells organized around the lumen. These patterns are indicated schematically in Figure 4b, which also shows quantitative results from visual scoring of the rab3D labeling pattern in single sections of reconstituted acini optimized for maximal resolution of lumenal actin. Percentage values in each category were obtained by visual quantitation of the morphology of labeling in Control (404 lumena) and Ad-LacZ-treated (392 lumena) acini from three to four preparations. Analysis of lumena in control acini showed that 81% of lacrimal acini retained rab3D in either an entirely apical (44%) or half-apical (37%) distribution. In contrast, in lumena in Ad-LacZ-transduced acini, only 55% of lacrimal acini retained rab3D in an apical distribution; remarkably, only 17% of these acini exhibited an entirely apical rab3D distribution.

Figure 4.

Apical enrichment of rab3D is reduced by Ad transduction of rabbit lacrimal acini. (a) Representative images of nontransduced (control) and transduced (Ad-LacZ) rabbit lacrimal acini fixed and processed for immunofluorescent detection of rab3D (green) in parallel with rhodamine–phalloidin to label actin filaments (red). 2^o only shows control acini labeled with rhodamine–phalloidin and FITC-conjugated secondary antibody alone and acquired at comparable contrast/gain settings, revealing an absence of background staining. Arrows, accumulation of subapical secretory vesicles enriched in rab3D; bar, 10 m; ^*, apical/lumenal region. (b). Visual scoring of rab3D labeling in nontransduced and Ad-LacZ-transduced acini at defined lumenal regions in single confocal sections optimized for the resolution of lumenal actin into three categories: apical, majority of label localized in apical-most quadrant of the cell in all cells organized around the lumen; half-apical, majority of label localized in the apical-most quadrant of at least half of the cells organized around the lumen; dispersed, majority of label distributed randomly throughout the cytoplasm in cells organized around the lumen. rab3D content was scored in 392 and 404 lumenal regions in Ad-LacZ-treated and control acini, respectively, from three to four separate preparations. (c) Western blot of rabbit lacrimal acinar cell lysate (50 g protein/lane). The left lane, labeled 'Anti-rab3D' was blotted with a rabbit polyclonal antibody to rab3D and a goat anti-rabbit secondary antibody conjugated to IRDye™800, while the right lane, labeled '2° only' was blotted with secondary antibody alone. (d) Representative Western blot of rab3D content of lysates from nontransduced (control) or transduced (Ad-LacZ or Ad-GFP) rabbit lacrimal acini resolved by 7.5% SDS-PAGE before transfer to nitrocellulose and development with appropriate primary and secondary antibodies. (e) Quantitation of rab3D signal intensity per equivalent protein content, n=4.

Full figure and legend (185K)

We also quantified apical versus dispersed rab3D labeling in z-plane sections from multiple nontransduced (control) and Ad-LacZ-transduced acini using digital image analysis. Simply, the pixel intensity associated with total, apical and dispersed rab3D fluorescence was calculated in z-plane sections, and values from sections acquired at 1–1.5 m intervals within an acinus were pooled, before final calculation of apical/dispersed (A/D) ratios. Figure 5a shows how the apical and total pixel intensity within individual acinar sections were defined, while Figure 5b shows composite results from multiple acini from three experiments. Digital quantitation confirmed that Ad transduction caused a significant (P0.05) decrease in the A/D ratio, consistent with Figures 3 and 4.

Figure 5.

Quantitation of rab3D labeling pattern by digital image analysis. (a) Representative confocal sections labeled with rhodamine-phalloidin (red) and rab3D (green) in nontransduced (control) and Ad-LacZ-transduced lacrimal acini. ^*, lumenal regions; area outlined in bold white line, lumenal rab3D labeling; area outlined in dashed white line, entire acinus. Bar, 10 m; ^*, apical/lumenal region. (b) Metamorph^® image analysis of rab3D labeling pattern in z-plane sections. Fluorescence intensity associated with rab3D in the apical/lumenal area of the acinus (A) versus the entire acinus (D) was analyzed as described in Materials and methods. Seven images/acinus were acquired at 1–1.5 m intervals; five acini and 105 images were acquired per treatment; n=3 separate preparations. A/D rab3D labeling ratios (A/D) for control and transduced acini were compared using a paired t-test and ^*, P0.05. Error bars represent s.e.m.

Full figure and legend (146K)

Ad does not alter rab3D expression or partitioning with membranes

To determine whether Ad altered the expression of rab3D, we measured rab3D content in cell lysates. As shown in Figure 4d and e, rab3D content was unaffected by Ad transduction as evidenced by Western blotting with our monospecific polyclonal antibody (Figure 4c). We also investigated whether the partitioning of rab3D between membrane compartments and cytosol was affected by Ad (Figure 6). Marker proteins, pIgAR and -tubulin, were used to confirm the integrity of the isolated soluble and membrane fractions from Control (nontransduced) and Ad-LacZ-transduced acini. As expected, most transmembrane pIgAR detected in control cell lysates was associated with membrane fractions, while most tubulin detected in control cell lysates was associated with soluble fractions. Most rab3D in control cell lysates was also associated with membrane fractions, consistent with the punctate vesicular labeling pattern detected by confocal fluorescence microscopy (Figures 3, 4 and 5). As expected, Ad-LacZ transduction had no change on the partitioning of -tubulin and pIgAR; however, Ad-LacZ also had no effect on the partitioning of rab3D between soluble and membrane phases.

Figure 6.

Rab3D dispersal by Ad is not associated with dissociation from cellular membranes. Rabbit lacrimal acini were cultured for 2 days and then transduced with Ad-LacZ at an MOI of 5 for 16–18 h. Cells were fractionated into Pi and Si phases as described in Materials and methods. (a) Western blots from representative preparations of control and Ad-LacZ-transduced lacrimal acini showing pIgAR, -tubulin and rab3D contents in each fraction. Sample (50 g) was loaded per lane. Blots used appropriate primary and Alexa Fluor^® 680 or IRDye™800-conjugated secondary antibodies. (b–d) Summary of changes in pIgAR, -tubulin and rab3D recovery in each fraction from control and Ad-LacZ-transduced acini expressed as a percentage of the total cellular content of each protein. Results were obtained from three separate preparations. Error bars represent s.e.m.

Full figure and legend (61K)

Dispersal of apical rab3D is not due to changes in cellular signaling

Vesicle-associated membrane protein 2 (VAMP2) is a SNARE protein associated with secretory vesicles in lacrimal gland as well as pancreas, parotid gland and submandibular gland.^{27, 28} In lacrimal gland, rapid transport of VAMP2-enriched membranes into the subapical cytoplasm following carbachol (CCH) stimulation occurs along MTs and is driven by the MT-based motor protein, cytoplasmic dynein;²⁷ VAMP2 and elements of the dynein complex such as dynein intermediate chain and p150^Glued, a dynein effector, are normally colocalized in the subapical cytoplasm of CCH-stimulated acini. We examined the distributions of VAMP2 and cytoplasmic dynein in nontransduced and transduced acini. The redistribution and colocalization of these proteins in response to CCH was unaffected by Ad transduction (data not shown), indicating that the signaling pathways responsible for the recruitment of secretory membranes to the apical surface remain intact.

Ad transduction does not decrease protein synthesis

Changes in rab3D-enriched secretory vesicle content associated with Ad transduction might be due to reduced protein synthesis, leading to decreased availability of secretory proteins for packaging into vesicles. To determine whether protein synthesis was affected, transduced and nontransduced rabbit lacrimal acini were labeled with [³⁵S]Translabel. As shown in Figure 7, the incorporation of [³⁵S] into total cellular protein in transduced cells was slightly but not significantly elevated to 11427 and 15617% of nontransduced (control) following transduction with Ad-LacZ or Ad-GFP, respectively. Analysis of [³⁵S] incorporation into proteins resolved by SDS-PAGE revealed a comparable pattern, with the additional expression of GFP and -galactosidase (arrows) in transduced acini.

Figure 7.

New protein synthesis is not reduced following Ad transduction of primary rabbit lacrimal acini. Primary rabbit lacrimal acini without (control) or with transduction (Ad-GFP or Ad-LacZ) were pulse labeled with [³⁵S]Translabel to detect newly synthesized proteins. (a) Shows the spectrum of proteins labeled with [³⁵S] in nontransduced (control) and transduced (Ad-GFP or Ad-LacZ) cells after resolution by 1 % SDS-PAGE. Analysis of 10 g of each lysate revealed that the banding patterns and intensities were comparable with the addition of 28 kDa protein present in Ad-GFP lysates identified as GFP by Western blotting (data not shown) and the addition of a 150 kDa protein present in Ad-LacZ lysates identified as -galactosidase (-gal) by its migration on SDS-PAGE. (b) Shows the incorporation of [³⁵S] into TCA-precipitated protein in extracts from nontransduced (100% of control) or transduced (Ad-LacZ or Ad-GFP) lacrimal acini. The small changes seen with Ad-LacZ and Ad-GFP transduction were not statistically significant at P0.05 and n=4 experiments.

Full figure and legend (45K)

UV-inactivated Ad diminishes rab3D-enriched secretory vesicles

We considered two explanations for the effects of Ad on rab3D-enriched secretory vesicles. Exposure to the Ad capsid proteins, which mediate virus internalization through interaction with elements of the host trafficking machinery, might alter other trafficking/sorting effects required for appropriate secretory vesicle formation, maturation or retention in the infected acini. Alternatively, although the Ad is replication defective, leaky transcription of some viral genes might persist, with these gene products able to modulate secretory vesicle formation, maturation or retention. To test the latter hypothesis, we exposed replication-defective Ad to UV light in the presence of 8-methoxypsoralen. This process has previously been shown to render the virus genome transcriptionally inactive.^{29, 30} We processed Ad-LacZ and Ad-GFP in the absence of 8-methoxypsoralen/UV light and used these constructs in parallel (mock Ad-LacZ or mock Ad-GFP) to control for any effects of dilution and purification on viral titer. As shown in Figure 8a, the high transduction efficiency seen for Ad-LacZ in Figure 1 was retained for mock-treated Ad-LacZ. However, <10% of cells expressed -galactosidase when exposed to equivalent amounts of UV Ad-LacZ, consistent with successful UV inactivation of virus. Comparable findings were obtained for mock Ad-GFP and UV Ad-GFP (Figure 9).

Figure 8.

UV-inactivated Ad depletes rab3D-enriched secretory vesicles in lacrimal acini. (a) Lacrimal acini cultured for 2 days were exposed without (control) or with mock Ad-LacZ or UV Ad-LacZ at an MOI of 5 for 16–18 h. Top row shows reporter gene expression for -galactosidase (phase image); middle and bottom rows depict rab3D (green, middle) or Ad5 (green, bottom) immunoreactivity in parallel with actin filaments (red) under these conditions (fluorescence images). (b) Visual scoring of rab3D labeling in nontransduced acini or acini transduced with mock Ad-LacZ or UV Ad-LacZ into one of the three categories as previously defined in Figure 4: apical, half-apical and dispersed. Rab3D content was scored in 231 (nontransduced), 191 (mock Ad-LacZ) and 263 (UV Ad-LacZ) acini, respectively, from three separate preparations.

Full figure and legend (389K)

Figure 9.

Ad modulation of CCH-stimulated release of protein and -hexosaminidase from primary rabbit lacrimal acini. (a) Shows phase/fluorescence overlaid images in nontransduced acini and acini exposed to either mock Ad-GFP or UV Ad-GFP at an MOI of 5 for 16–18 h. (b–c) Primary rabbit lacrimal acini cultured on Matrigel-coated 24-well plates without (control) or with transduction (mock Ad-GFP or UV Ad-GFP) were treated without or with CCH (100 M, 30 min) and then processed for the measurement of basal, total and stimulated protein (b) and -hexosaminidase (c) release as described in Materials and methods. The plots of protein and -hexosaminidase depict the amounts recovered in 25 or 10 l of culture medium, respectively, before normalization to mg cell protein in the well. Results were obtained from six experiments. (^*), different from untreated at P0.05; (#), different from mock Ad-GFP at P0.05; AU, arbitrary fluorescence units.

Full figure and legend (161K)

When the effects of UV Ad-LacZ were assessed on rab3D-enriched mature secretory vesicles, we detected a marked redistribution of their normally apical immunofluorescence relative to controls, although this dispersal was not as complete as that elicited by mock-treated Ad-LacZ (Figure 8a). Percentage values for apical, half-apical and dispersed lumenal rab3D labeling (defined in Figure 4) were obtained by visual quantitation of the morphology of labeling in control (231 lumena), mock Ad-LacZ-treated (191 lumena) and UV Ad-LacZ-treated (263 lumena) acini from three separate preparations. In control acini, 63% of the total lumena showed an entirely apical rab3D labeling pattern, while this percentage was reduced to 9 and 25% of the total lumena in acini treated with mock Ad-LacZ and UV Ad-LacZ, respectively (Figure 8b). As shown in Figure 8a, both mock Ad-LacZ and UV Ad-LacZ exposure resulted in the detection of punctate Ad5 immunoreactivity distributed throughout the acini. Exposure to fluorescently conjugated secondary antibody in the absence of the anti-Ad5 primary antibody showed no background fluorescence (data not shown). These findings suggest that (1) cellular Ad5 immunoreactivity is largely associated with the Ad capsid proteins, since the extent of labeling seen for UV Ad-LacZ is approximately equal to that elicited by mock-treated Ad-Lac Z and (2) Ad capsid proteins are highly persistent within the cytosol since immunoreactivity is detected 16–18 h after the addition of Ad.

Ad exposure results in the inhibition of CCH-stimulated secretion

The consequences of the redistribution of rab3D-enriched secretory vesicles caused by replication-defective or UV-inactivated Ad to lacrimal acinar secretory functions was investigated. Figure 9 shows the effects of mock-treated Ad-GFP and UV Ad-GFP on the basal and CCH-stimulated release of the secretory protein, -hexosaminidase and on total protein. Mock Ad-GFP and UV Ad-GFP significantly inhibited the CCH-stimulated release of -hexosaminidase and protein, while not significantly affecting the basal release of these products. The magnitude of the inhibition of CCH-stimulated -hexosaminidase release was comparable for UV Ad-GFP and mock Ad-GFP, but the magnitude of inhibition of CCH-stimulated protein release was significantly less for UV Ad-GFP relative to mock Ad-GFP.

Discussion

Recombinant Ad vectors can efficiently transfer and express genes in a wide variety of cells including those comprising ocular tissues.^{1, 2, 3, 31, 32, 33} In this study, we have assessed the consequences of replication-defective and UV-inactivated Ad on secretory compartment organization and function in the acinar epithelial cells of the lacrimal gland. Exposure of lacrimal acini to replication-defective Ad constructs at an MOI of 5 for 16–18 h elicited a marked dispersal of rab3D from its normally apical enrichment (Figures 3, 4 and 5), a change independent of altered rab3D expression (Figure 4) or membrane association (Figure 6). This dispersal of apical rab3D occurred independently of effects on the cytoskeleton or other membrane compartments (Figure 2) or decreased protein synthesis (Figure 7). Transduction with replication-defective Ad caused a significant decrease in the CCH-stimulated release of protein and the secretory product, -hexosaminidase (Figure 9). Exposure of lacrimal acini to UV-inactivated Ad also depleted rab3D-enriched secretory vesicles in parallel with the inhibition of CCH-stimulated release of protein and -hexosaminidase (Figures 8 and 9), although the extent of the dispersal and inhibition, respectively, by UV-inactivated virus was slightly less than that caused by mock UV-inactivated Ad.

We propose that the Ad-induced dispersal of rab3D-enriched secretory vesicles is responsible for the reduction in CCH-stimulated secretion in our in vitro studies. Essentially, if fewer fusion-competent vesicles are available for release in response to CCH, then the amount of secretory protein released is decreased. We further suggest that the primary source of these inhibitory effects on secretory functions is the Ad capsid, since UV-inactivated Ad elicited comparable effects. As the changes elicited by UV-inactivated Ad were consistently less in magnitude (60–70%) relative to those elicited by mock UV-inactivated Ad, it is possible that residual leaky gene expression, perhaps even of the capsid proteins, contributes to some of the observed effects on lacrimal acinar secretory responses. Alternatively, virus inactivation by UV irradiation in the presence of an alkylating agent may structurally modify Ad capsid proteins, so their effects on host responses are blunted.

The effects of Ad in this study were elicited by a fairly prolonged exposure to replication-defective or UV-inactivated Ad for 16–18 h. Exposure of lacrimal acini to replication-defective Ad-LacZ for only 4 h at an MOI of 5, a treatment which also resulted in 80% transduction efficiency, also caused a 50% reduction in apical rab3D labeling,²⁷ although this change was less than the 75–85% reduction elicited by the longer exposure to Ad constructs in this study (Figures 4 and 8). The effects on acinar secretory functions may therefore increase as the duration of exposure to the Ad capsid lengthens.

How might Ad transduction alter lacrimal acinar rab3D-enriched secretory vesicle maturation or exocytosis? Rab3D is a member of the rab family of small GTP binding proteins that participate in membrane trafficking in eukaryotic cells.³⁴ The GTP-bound forms of these proteins are membrane associated, with GTP hydrolysis promoting release of the GDP-bound rab from the membrane cargo. The cycle of GTP binding, GTP hydrolysis and release/exchange of GDP for GTP is tightly regulated by a number of different proteins, specific for the individual rabs. GTP-bound rabs are thought to recognize their designated membrane cargo by association with specific extrinsic or transmembrane cargo proteins, with GTP hydrolysis concomitant with docking of donor and acceptor membranes. The observation that >90% of total rab3D in Ad-treated acini remains associated with the membrane pool of unstimulated acini (Figure 6) suggests that those features that regulate rab3D membrane binding and maintenance of its GTP-bound state are unaffected by Ad. Trafficking factors participating in the maturation of rab3D-associated membranes from the trans-Golgi network into mature secretory vesicles represent likelier targets of Ad modulation. Two specific proteins known to participate in secretory vesicle maturation represent viable targets for modulation by Ad: cytoplasmic dynein and clathrin-associated adaptors.

Ad internalization in most cells is thought to involve binding to the ubiquitous coxsackievirus and Ad receptor on the target cell surface via the fiber capsid protein;^{4, 35} the penton capsid protein then triggers viral endocytosis via ligation and clustering of _v integrins,^{4, 36} an event involving the activation of a host of signaling pathways including those mediated by Rac and Cdc42.^{37, 38, 39} Once in endosomes, Ad undergoes endosomolysis, allowing capsid proteins to interact with cytosolic factors. Ad capsid proteins are known to interact with MTs and cytoplasmic dynein during transit to the nucleus.^{40, 41} We have previously shown that inhibition of cytoplasmic dynein activity in lacrimal acini severely impairs the formation of rab3D-enriched secretory vesicles.²⁷ However, direct dynein inhibition also prevents the recruitment of VAMP2-enriched secretory transport vesicles to the apical membrane in response to CCH, while this response is unaffected by Ad (data not shown), suggesting that some dynein-mediated traffic to the apical membrane can proceed normally in Ad-treated acini. Changes in the recruitment of other trafficking proteins such as adapter proteins and clathrin by Ad may also impair secretory vesicle formation. It is of interest to note that the penton proteins in the Ad capsid contain a consensus endosomal sorting motif, the dileucine (LL) motif. Proteins that interact with the LL motif include clathrin-associated adapter proteins,^{42, 43} effectors required for normal secretory vesicle maturation in some specialized secretory cells.⁴⁴ Findings that the HIV Nef protein utilizes this same motif to recruit clathrin-associated adapters suggest that this motif in other viral proteins is functional.^{45, 46} Future experiments will specifically probe the consequences of Ad and Ad capsid exposure to these trafficking pathways in acinar epithelial cells, as well as test other scenarios for mechanisms of Ad modulation of secretion including chronic changes in the signaling environment or inability to retain mature secretory vesicles.

By using either replication-incompetent or UV-inactivated virus, we show for the first time here that the incoming Ad capsid alters epithelial secretory pathways. The success of Ad-mediated gene delivery in various systems has prompted the design of a number of next generation nonviral gene delivery vehicles largely based on the use of Ad capsid proteins,^{14, 47, 48} many of which could elicit comparable changes in epithelial secretory functions. Clearly, further studies are in order to determine whether these effects of the Ad capsid also extend to other types of epithelial cells in addition to the secretory epithelia evaluated here. Adeno-associated virus (AAV) has also been extensively considered as an alternative to Ad-mediated gene therapy, largely due to its considerably decreased immunogenicity.¹³ Recent findings suggest that AAV utilizes mechanisms comparable to Ad5 to enter host cells, including receptor-mediated endocytosis through _v integrin binding, and MT-based trafficking within the endosomal pathway.⁴⁹ These findings suggest some shared interactions between Ad and AAV capsid proteins and host cells, leading to the possibility that AAV capsid proteins may likewise influence epithelial secretory functions.

It should be further mentioned that, although Ad 8, 19 and 37 are the most common Ad serotypes implicated in the highly contagious and largely untreatable epidemic keratoconjunctivitis, a total of 19 different Ad serotypes have been implicated cases of sporadic and epidemic keratoconjunctivitis.^{50, 51, 52} Little is known about the conservation of cellular entry mechanisms between Ad serotypes. If capsid proteins associated with pathogenic Ad serotypes elicit comparable changes in lacrimal secretory functions, persistent ocular Ad infections may trigger changes in intracellular traffic of secretory products, promoting further ocular surface pathology in susceptible individuals.

Materials and methods

Reagents

CCH, rhodamine–phalloidin, mouse anti- tubulin antibody and goat anti-mouse secondary antibody conjugated to FITC were purchased from Sigma Chemical Co (St Louis, MO, USA). Mouse anti--adaptin and anti-p150^Glued antibodies were from Transduction Laboratories (Lexington, KY, USA). Polyclonal antibody to rab3D was either the gift from Dr James Jamieson (Yale University, New Haven, Connecticut) or was generated in rabbits against recombinant rab3D produced in Escherichia coli (Antibodies, Inc., Davis CA, USA) and purified by chromatography over protein A/G agarose. The Ad5 antibody is a rabbit polyclonal generated against whole Ad5 purchased from Access Bio Medical (San Diego, CA, USA). The sheep polyclonal antibody to pIgAR was the gift from Dr Curtis Okamoto (University of Southern California, Los Angeles, California). Goat anti-rabbit and anti-mouse IRDye™800-conjugated secondary antibodies were purchased from Rockland (Gilbertsville, PA, USA). Goat anti-mouse and anti-rabbit horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence reagents were from Amersham (Arlington Heights, IL, USA). Matrigel and dispase were from Collaborative Biochemicals (Bedford, MA, USA). ProLong antifade mounting media, Alexa Fluor^® 350-phalloidin and goat anti-rabbit secondary antibody conjugated to Alexa Fluor^® 568 were from Molecular Probes (Eugene, OR, USA). Paraformaldehyde was purchased from Polysciences (Warrington, PA, USA). Cell culture reagents were obtained from Life Technologies. All other chemicals were reagent grade and obtained from standard suppliers.

Cell isolation and culture

Female New Zealand white rabbits weighing between 1.8 and 2.2 kg were obtained from Irish Farms (Norco, CA, USA). Lacrimal gland acinar cells were isolated and cultured for 2–3 days as described previously.^{21, 22, 23}

Generation of Ad vectors

The recombinant Ad-expressing GFP (Ad-GFP) or -galactosidase (Ad-LacZ) were derived from Ad5 with deletions in E1 and E3 regions. Ad-GFP was constructed using the AdEasy Vector System (generously provided by Dr B Vogelstein, Johns Hopkins University, Baltimore, Maryland). Briefly, the shuttle plasmid, pAdTrack, which contains a CMV-driven GFP marker gene and two arms of homology to the left and right ends of the Ad5 genome flanking a plasmid backbone containing the kanamycin resistance gene, was linearized and coelectroporated into the recombinogenic E. coli BJ5183 strain, along with a large 30 kb supercoiled plasmid, pAdEasy, that contains an adenoviral genome in an ampicillin-resistant plasmid backbone, recreating the replication-deficient Ad genome. The transformants were selected on kanamycin plates, mini-prep DNA from the resistant colonies was screened by restriction digest, and clones showing the correct restriction pattern were retransformed into the more stable DH10 strain to prevent further recombination events. A different shuttle plasmid, pShuttle, was used to generate Ad-expressing -galactosidase (LacZ) by homologous recombination as described. Ad-LacZ contains an expression cassette including the E. coli -galactosidase gene with a eukaryotic nuclear translocation signal under the transcriptional control of the CMV promoter.

Virus stocks were produced in 293 cells expressing the E1A and E1B proteins, which support the replication of the E1-defective adenoviral mutants, followed by serial passaging and purification of the harvested virus by cesium chloride ultracentrifugation and dialysis. The titer of the purified viral stocks was determined using plaque assays (PFU/ml). Transduction efficiency in reconstituted lacrimal acini was determined by flow cytometry and fluorescence microscopy (Ad-GFP) or colorimetric production with X-Gal as substrate (Ad-LacZ). Optimal gene expression was obtained by transducing 2-day acinar cultures at an MOI of 5 for 16–18 h.

UV-inactivated Ad-GFP and Ad-LacZ (UV Ad-GFP and UV Ad-LacZ) were prepared according to established methods.^{29, 30} Briefly, purified virus was resuspended in 0.33 mg/ml 8-methoxypsoralen (Sigma Chemical Company, St Louis, MO, USA) and then exposed to a 365 nm UV light source on ice 4 cm from the lamp for 30 min. Virus was removed from excess alkylating agent over Protein Desalting Spin Columns (Pierce, Rockford, IL, USA). Mock-inactivated Ad (no exposure to 8-methoxypsoralen or UV) was processed in parallel and used as a control.

Flow cytometry analysis

Control and Ad-GFP-transduced acinar cells cultured on Matrigel-coated coverslips in 12-well plates were washed with Dulbecco's phosphate-buffered saline (DPBS), treated with Dispase at 37°C for 1.5 h, EDTA was added to the mixture and cells collected by centrifugation at 800 rpm for 5 min. The cell pellet was then exposed to a cocktail of collagenase, hyaluronidase and DNAse^{21, 22} at 37°C for 10 min, cells were again rinsed with DPBS and collected by centrifugation as described above. Cells were resuspended in DPBS, and GFP-positive cells were quantified by a Becton Dickinson FAC-Scan™ analyzer (Becton Dickinson, Franklin Lakes, NJ, USA) using a 15 mW air-cooled argon laser set at 488 nm and recorded with a 530 nm emission filter in the FL1 emission channel. Green fluorescence was plotted against total fluorescence in the FL2 (red) emission channel. The advantage of this two-dimensional gating method is that even a few GFP-positive cells can be visualized inside the FL1⁺FL2^- gate.

Membrane fractionation

Lacrimal acini were cultured on 150 mm dishes for 2 days, transduced with Ad-LacZ at an MOI of 5 for 16–18 h, and harvested by gentle scraping with a rubber policeman. After washing in cold DPBS, acini were resuspended in 1 ml PMEE buffer (35 mM PIPES, 5 mM MgSO₄, 1 mM EGTA, 0.5 mM EDTA, pH 7.4) containing protease inhibitors as described.²² Cells were lysed by 20 passages through a 20-gauge syringe needle followed by 40 passages through a Balch cell press (H&Y Enterprises, Redwood City, CA, USA). The lysate was centrifuged for 10 min at 10 000 rpm in an Eppendorf 5415 C centrifuge at 4°C. The supernatant, designated So, was centrifuged at 92 000 g for 30 min at 4°C in a Sorvall S120-AT2 rotor. The supernatant fraction, containing the soluble phase, was designated Si. The pellet, containing the membrane phase, was designated Pi. The distribution of proteins of interest was determined by SDS-PAGE and Western blotting utilizing secondary antibodies conjugated to IRDye™800 and quantified using an Odyssey Scanning Infrared Fluorescence Imaging System (Li-Cor, Lincoln Nebraska, Lincoln, Nebraska).

Confocal fluorescence microscopy

Control and Ad-transduced lacrimal acinar cells were cultured on Matrigel-coated glass coverslips seeded in 12-well plates. For analysis of the cellular distribution of rab3D in parallel with actin filaments, acini were rinsed with DPBS, then fixed and permeabilized with ethanol at -20°C for 10 min before rehydration in DPBS as described.²⁷ For analysis of the cellular distribution of Ad proteins, -tubulin or -adaptin in parallel with actin filaments or for triple labeling experiments, acini were fixed with 4% paraformaldehyde for 15 min and permeabilized in 0.1% Triton X-100 for 10 min as described.^{23, 53} Samples were blocked with 1% bovine serum albumin, incubated with appropriate primary and FITC- or Alexa Fluor^® 568-conjugated secondary antibodies and Alexa Fluor^® 350- or rhodamine–phalloidin. Confocal images from dual-labeled specimens were acquired on a Nikon PCM Confocal System equipped with Argon ion (488 nm) and HeNe (543 nm) lasers attached to a Nikon TE300 Quantum inverted microscope; confocal images from triple-labeled specimens were acquired on a Leica confocal microscope TCS SP2 with AOBS equipped with a violet diode UV (405 nm), argon ion (488 nm) and HeNe (543, 594 and 633 nm) lasers. Images were compiled in Adobe Photoshop 7.0. (Adobe Systems Inc, Mountain View, CA, USA).

Digital image analysis and evaluation of A/D rab3D labeling ratios in multiple planes from z-stacks acquired from acini was as described previously.⁵⁴ Serial sections from confocal fluorescence microscopy images of nontransduced and Ad-LacZ-transduced acini of comparable size and thickness labeled to detect actin filaments and rab3D were acquired at equal gain and contrast intensities, pinhole size, image section thickness (0.6 m) and zoom ( 2 zoom magnification, 60 objective). Initial gain and contrast levels were established using both samples to ensure that fluorescence levels were not saturated. Seven z-plane sections, per acinus, were acquired, with intervals representing 1–1.5 m. The first and last sections were discarded to avoid imaging very small surface areas of the acinus (top) and/or stress fibers (bottom). Green fluorescence (rab3D) was quantified utilizing MetaMorph^® Image Analysis Software (Universal Imaging Company, West Chester, PA, USA). Total rab3D fluorescence intensity (T) was calculated by measuring the pixel intensity of the entire acinus normalized to the acinus perimeter. Each z-plane image within an acinus contained 2–6 lumena. Individual lumena were identified based on their enrichment in actin filaments. After the lumena were outlined, a region extending 1–2 m into the cell from the lumena was outlined and the fluorescence intensity associated with rab3D accumulation in this apical region (A) was calculated by normalizing the pixel intensity within that region to its perimeter. The fluorescence intensity of the dispersed rab3D present in the remainder of the cytoplasm (D) was calculated by subtracting (A) from (T) and final A/D rab3D labeling ratios are represented as A/D.

Secretion assays

Secretion assays were based on those described previously.^{23, 53} Control, mock Ad-GFP-transduced or UV Ad-GFP-transduced lacrimal acini in Matrigel-coated 24-well plates were incubated in fresh medium for 2 h before the removal of a small aliquot of medium for measurement of initial protein content or -hexosaminidase activity. After treatment with or without CCH (100 M, 30 min), a second aliquot of medium was removed for the measurement of final protein content or -hexosaminidase activity. The amounts of protein or -hexosaminidase present under each condition were calculated from three replicate wells/treatment in each assay, and values normalized to total cellular protein before comparison across treatments using a paired t-test with P0.05. Protein contents were determined on 96-well plates with the Micro BCA Protein Assay (Pierce) with bovine serum albumin as standard, and -hexosaminidase activity was assessed on 96-well plates using methyumbelliferyl--D-glucosaminide as substrate. A Tecan GENios Plus UV/Visible/fluorescence plate reader (Phenix Research Products, Hayward, CA, USA) was used to measure reaction products.

Protein synthesis

Acini were pulse labeled with [³⁵S]Translabel (6.6 Ci/million cells) in sulfur amino-acid-free medium for 1 h at 37°C as described.⁵³ After rinsing, cellular proteins were precipitated in 10% TCA, rinsed again with 5% TCA and the precipitated cellular proteins were dissolved in 0.5 N NaOH. Scintillation counting measured [³⁵S], while protein content was determined using the Pierce Micro-BCA protein Assay Reagent Kit with bovine serum albumin as standard. For analysis of [³⁵S] incorporation into cellular proteins, [³⁵S]-labeled acini were lysed with RIPA buffer. Lysates containing equivalent amounts of protein from control and transduced cells were separated by SDS-PAGE and the gel was treated with DMSO-PPO for visualizing [³⁵S]-labeled proteins following by Coomassie Brilliant Blue G250 staining prior to autoradiography for 1–2 days.⁵⁵

References

1. Cordeiro MF et al. Molecular therapy in ocular wound healing. Br J Ophthalmol 1998; 83: 1219–1224.
2. Borras T. Recent developments in ocular gene therapy. Exp Eye Res 2003; 76: 643–652. | Article | PubMed | ChemPort |
3. Zhu Z et al. TNF-inhibitor gene expression suppresses lacrimal gland immunopathology and ocular surface disease in a rabbit model of autoimmune dacryoadenitis. Cornea 2003; 22: 25–31. | Article | PubMed |
4. Wickham TJ, Mathias P, Cheresh DA, Nemerow GR. Integrins v3 and v5 promote adenovirus internalization but not virus attachment. Cell 1993; 73: 309–319. | Article | PubMed | ISI | ChemPort |
5. Greber UF, Willetts M, Webster P, Helenius A. Stepwise dismantling of adenovirus 2 during entry into cells. Cell 1993; 75: 477–486. | Article | PubMed | ISI | ChemPort |
6. Mathias P, Wickham T, Moore M, Nemerow G. Multiple adenovirus serotypes use v integrins for infection. J Virol 1994; 68: 6811–6814. | PubMed | ISI | ChemPort |
7. Kass-Eisler A et al. The impact of developmental stage, route of administration and the immune system on adenovirus-mediated gene transfer. Gene Therapy 1994; 1: 395–402. | PubMed | ChemPort |
8. Dai Y et al. Cellular and humoral immune responses to adenoviral vectors containing factor IX gene: tolerization of factor IX and vector antigens allows for long-term expression. Proc Natl Acad Sci USA 1995; 92: 1401–1405. | Article | PubMed | ChemPort |
9. Barr D et al. Strain-related variations in adenovirally mediated transgene expression from mouse hepatocytes in vivo: comparisons between immunocompetent and immunodeficient strains. Gene Therapy 1995; 2: 151–155. | PubMed | ISI | ChemPort |
10. Yang Y, Xiang Z, Ertl HCJ, Wilson JM. Upregulation of class 1 MHC antigens by interferon- is necessary for the T cell-mediated elimination of recombinant adenovirus infected hepatocytes in vivo. Proc Natl Acad Sci USA 1995; 92: 7257–7261. | Article | PubMed | ChemPort |
11. Yang Y, Haecker SE, Su Q, Wilson JM. Immunology of gene therapy with adenoviral vectors in mouse skeletal muscle. Hum Mol Genet 1996; 5: 1713–1726. | Article | PubMed | ISI | ChemPort |
12. Reichel MB et al. Immune responses limit adenovirally mediated gene expression in the adult mouse eye. Gene Therapy 1998; 5: 1038–1046. | Article | PubMed | ISI | ChemPort |
13. Bennett J. Immune response following intraocular delivery of recombinant viral vectors. Gene Therapy 2003; 10: 977–982. | Article | PubMed |
14. Medina-Kauwe LK. Endocytosis of adenovirus and adenovirus capsid proteins. Adv Drug Deliv Rev 2003; 55: 1485–1496. | Article | PubMed | ISI | ChemPort |
15. Dartt DA. Signal transduction and activation of the lacrimal gland. In: Albert DM, Jacobiec FA (eds) Principles and Practice of Ophthalmology, 2nd edn. Saunders: Philadelphia, 1994, pp 458–465.
16. Fullard R. Tear proteins arising from lacrimal tissue. In: Albert DM and Jacobiec FA (eds) Principles and Practice of Ophthalmology, 2nd edn. Saunders: Philadelphia, 1994, pp 473–479.
17. Mircheff AK. Water and electrolyte secretion and fluid modification. In: Albert DM and Jacobiec FA (eds) Principles and Practice of Ophthalmology, 2nd edn. Saunders: Philadelphia, 1994, pp 466–472.
18. Fox RI, Stern M, Michelson P. Update in Sjögren's syndrome. Curr Opin Rheumatol 2000; 12: 391–398. | Article | PubMed |
19. Pflugfelder SC, Solomon A, Stern ME. The diagnosis and management of dry eye: a twenty-five-year review. Cornea 2000; 19: 644–649. | Article | PubMed | ISI | ChemPort |
20. Ang RT, Dartt DA, Tsubota, K. Dry eye after refractive surgery. Curr Opin Ophthalmol 2001; 12: 318–322. | Article | PubMed |
21. Gierow JP, Lambert RW, Mircheff AK. Fluid phase endocytosis by isolated rabbit lacrimal gland acinar cells. Exp Eye Res 1995; 60: 511–525. | PubMed |
22. Hamm-Alvarez SF et al. Cholinergic stimulation of lacrimal acinar cells promotes redistribution of membrane-associated kinesin and the secretory protein, -hexosaminidase, and increases kinesin motor activity. Exp Eye Res 1997; 64: 141–156. | Article | PubMed |
23. da Costa SR et al. Microtubules facilitate the stimulated secretion of -hexosaminidase in lacrimal acinar cells. J Cell Sci 1998; 111: 1267–1276. | PubMed |
24. Robinson MS. Coats and vesicle budding. Trends Cell Biol 1997; 7: 99–102. | Article | ISI |
25. Ohnishi H et al. Rab3D localizes to zymogen granules in rat pancreatic acini and other exocrine glands. Am J Physiol 1996; 271: G531–G538. | PubMed |
26. Valentijn JA et al. Rab3D localizes to secretory granules in rat pancreatic acinar cells. Eur J Cell Biol 1996; 70: 33–41. | PubMed |
27. Wang Y et al. Cytoplasmic dynein participates in apically-targeted stimulated secretory traffic in primary rabbit lacrimal acinar epithelial cells. J Cell Sci 2003; 116: 2051–2065. | Article | PubMed |
28. Fujita-Yoshigaka J et al. Vesicle-associated membrane protein 2 is essential for cAMP-regulated exocytosis in rat parotid acinar cells. J Biol Chem 1996; 271: 13130–13134. | Article | PubMed |
29. Cotton M et al. High-efficiency receptor-mediated delivery of small and large (48 kilobase) gene constructs using the endosome-disruption activity of defective or chemically inactivated adenovirus particles. Proc Natl Acad Sci USA 1992; 89: 6094–6098. | Article | PubMed | ChemPort |
30. Yang Y, Greenough K, Wilson JM. Transient immune blockage prevents formation of neutralizing antibody to recombinant adenovirus and allows repeated gene transfer to mouse. Gene Therapy 1996; 3: 412–420. | PubMed | ChemPort |
31. Larkin DF et al. Adenovirus-mediated gene delivery to the corneal endothelium. Transplantation 1996; 61: 363–370. | Article | PubMed |
32. Hoffman LM, Maguire AM, Bennett J. Cell-mediated immune response and stability of intraocular transgene expression after adenovirus-mediated delivery. Invest Ophthalmol Vis Sci 1997; 38: 2224–2233. | PubMed | ISI | ChemPort |
33. Kamata Y et al. Adenovirus-mediated gene therapy for corneal clouding in mice with mucopolysaccharidosis type VII. Mol Ther 2001; 4: 307–312. | Article | PubMed | ChemPort |
34. Pfeffer SR. Rab GTPases: specifying and deciphering organelle identity and function. Trends Cell Biol 2001; 11: 487–491. | Article | PubMed | ISI | ChemPort |
35. Bergelson JM et al. Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science 1997; 275: 1320–1323. | Article | PubMed | ISI | ChemPort |
36. Chiu CY, Mathias P, Nemerow, GR, Stewart PL. Structure of adenovirus complexed with its internalization receptor, v5 integrin. J Virol 1999; 73: 6759–6768. | PubMed | ISI | ChemPort |
37. Li E, Stupack D, Bokoch GM, Nemerow GR. Adenovirus endocytosis requires actin cytoskeleton reorganization mediated by Rho family GTPases. J Virol 1998; 72: 8806–8812. | PubMed | ChemPort |
38. Li E et al. Adenovirus endocytosis via v integrins requires phophoinositide-3-OH Kinase. J Virol 1998; 72: 2055–2061. | PubMed | ISI | ChemPort |
39. Li E et al. Association of p130CAS with phosphatidylinositol-3-OH kinase mediates adenovirus cell entry. J Biol Chem 2000; 275: 14729–14735. | Article | PubMed | ChemPort |
40. Suomalainen M et al. Microtubule-dependent plus- and minus-end directed motilities are competing processes for nuclear targeting of adenovirus. J Cell Biol 1999; 144: 657–672. | Article | PubMed | ISI | ChemPort |
41. Leopold PL et al. Dynein and microtubule-mediated translocation of adenovirus serotype 5 occurs after endosomal lysis. Human Gene Ther 2000; 11: 151–165. | Article |
42. Heilker R, Spiess M, Crottet P. Recognition of sorting signals by clathrin adaptors. BioEssays 1999; 21: 558–567. | Article | PubMed | ISI | ChemPort |
43. Kirchhausen T. Adaptors for clathrin-mediated traffic. Annu Rev Cell Dev Biol 1999; 15: 705–732. | Article | PubMed | ISI | ChemPort |
44. Tooze SA, Martens GJM, Hutter WB. Secretory granule biogenesis: rafting to the SNARE. Trends Cell Biol 2001; 11: 116–122. | Article | PubMed | ISI | ChemPort |
45. Greenberg ME et al. Co-localization of HIV-1 Nef with the AP-2 adaptor protein complex correlates with Nef-induced CD4 down-regulation. EMBO J 1997; 16: 6964–6976. | Article | PubMed | ISI | ChemPort |
46. Craig HM, Pandori MW, Guatelli JC. Interaction of HIV-1 Nef with the cellular dileucine-based sorting pathway is required for CD4 down-regulation and optimal viral infectivity. Proc Natl Acad Sci USA 1998; 95: 11229–11234. | Article | PubMed | ChemPort |
47. Fender P et al Adenovirus dodecahedron, a new vector for human gene transfer. Nat Biotechnol 1997; 15: 52–56. | Article | PubMed | ISI | ChemPort |
48. Medina-Kauwe LK, Kasahara N, Kedes LH. 3PO, a novel nonviral gene delivery system using engineered Ad5 penton proteins. Gene Therapy 2001; 8: 797–803.
49. Lai CM, Lai YKY, Rakoczy PB. Adenovirus and adeno-associated virus vectors. DNA Cell Biol 2002; 21: 895–913. | Article | PubMed | ChemPort |
50. Gordon JS, Aoki K, Kinchington PR. Adenovirus keratoconjunctivitis. In: Pepose JS, Holland GN, Wilhelmus KR (eds). Ocular Infection and Immunity. Mosby: St Louis, 1996, pp 877–894.
51. Ford E, Nelson KE, Warren D. Epidemiology of epidemic keratoconjunctivitis. Epidemiol Rev 1987; 9: 244–261. | PubMed |
52. Trousdale MD et al. Studies of adenovirus-induced eye disease in the rabbit model. Invest Ophthamol Vis Sci 1995; 36: 2740–2748.
53. Zhang L et al. Protein phosphatase inhibitors alter cellular microtubules and reduce carbachol-dependent protein secretion in lacrimal acini. Curr Eye Res 2000; 20: 373–383. | Article | PubMed |
54. da Costa SR et al. Impairing actin filament or syndapin functions promotes accumulation of clathrin-coated vesicles at the apical plasma membrane of polarired cells. Mol Cell Biol 2003; 14: 4397–4413. | Article |
55. Laskey RA, Mills AD. Quantitative film detection of 3 H and 14C in polyacrylamide gels by fluorography. Eur J Biochem 1975; 56: 335–341. | Article | PubMed | ISI | ChemPort |

Acknowledgements

This work was supported by NIH Grants EY-13949 to SHA, EY12689 to MDT, P30 DK-48522 (Confocal Microscopy and Viral Vector Subcores, USC Center for Liver Diseases) and Doheny Eye Institute Core EY03040. Additional salary support to SHA was from NIH Grants EY-11386, EY-05081, NS-38246 and GM-59297. We thank Drs Austin Mircheff and Silvia da Costa for helpful discussions.