Introduction

Cancers located in the peritoneal cavity, including mesothelioma and ovarian cancer, are attractive indications for the application of adenoviral anti-cancer treatments because they permit a regional administration of the vector into the peritoneal cavity with exposure of lesions to a locally high concentration of vector¹. For example, a number of preclinical studies in ovarian tumor models have been published using conditionally replicating adenovirus vectors^2,3 or replication-deficient vectors carrying anti-erbB2 single-chain antibody⁴, soluble flt-1⁵, p53^{6,7,8,9,10,11}, or HSV-TK^12,13,14,15. These vectors have shown varying degrees of success in various animal models of ovarian cancer. Some of these approaches have also advanced to clinical trials and have demonstrated the safety of adenovirus vectors for treating ovarian cancer^{16,17,18,19,20,21,22,23,24,25,26,27}.

There are two potential hurdles for using adenovirus to treat cancer²⁸. First, the tumors may not express adequate levels of adenovirus receptors to permit efficient adenovirus entry into these cells. Second, expression of native adenovirus receptors on healthy tissue may lead to vector- or transgene-related toxicity through the entry of vector into nontarget cells. These potential disadvantages highlight the importance of both understanding the roles of native adenovirus receptors in gene transfer in vivo and developing targeted adenovirus vectors with improved efficiency and specificity of gene transfer.

Entry of Ad into cells is mediated by two of its coat proteins, fiber and penton base. The fiber mediates primary attachment to the cell via the coxsackie–adenovirus receptor (CAR)^29,30, a cell–cell adhesion protein involved in tight junction formation. Following attachment, an RGD tripeptide motif in the penton base protein binds to _v integrins, which mediate cellular internalization^31,32. In addition to CAR and integrins, recent studies suggest that other receptors, including heparan sulfate proteoglycans, may play a role in adenovirus entry into cells^33,34.

Increasing evidence supports the idea that native receptor expression may limit adenovirus entry into cancer cells. For example, a number of ovarian cancer cell lines have been found to be refractory to transduction by adenovirus vectors^{12,24,35,36,37,38,39,40,41}. In addition, a recent study examining CAR and integrin expression in 37 ovarian tumors found that although most tumors were positive for CAR, a majority of the tumors exhibited considerable heterogeneity of CAR expression within the tumor¹⁶. Such heterogeneity of native adenovirus receptor expression might seriously compromise the efficacy of adenovirus-based gene therapy in ovarian cancer. In fact, the functional importance of CAR in gene transfer in vivo has been recently demonstrated, whereby introduction of CAR expression into refractory ovarian cancer cells increased gene transfer⁴².

Adenovirus tropism has been modified via changes to the fiber protein to increase the efficiency of gene transfer to cancer. For example, insertion of an RGD motif in the Ad5 fiber or replacement of the Ad5 fiber with the Ad3 fiber has been shown to increase transgene expression in ovarian cancer cell lines and primary ovarian tumor cells^{35,37,43,44,45}. Interestingly, these same modifications have also been shown to sustain the transduction of primary ovarian cancer cells in the presence of neutralizing antibodies from the ascites fluid of ovarian cancer patients^37,43,45. However, these tropism modifications have also been shown to increase gene transfer to nontarget mesothelial cells and it is not yet clear how these modifications function in an in vivo ovarian cancer model. Toward the goal of increasing both efficiency and specificity, a bispecific antibody conjugate that simultaneously blocks CAR binding and redirects the Ad5 vector to the ovarian cancer epitope, TAG-72, has been shown to achieve increased gene transfer in vitro to ovarian cancer cells while reducing gene transfer to mesothelial cells³⁶.

The contribution of CAR and _v-integrins in the biodistribution and tissue transduction by adenovirus vectors following intraperitoneal administration has not been elucidated. However, previous studies have begun to clarify the roles of CAR and integrins in liver transduction following intravenous administration. Using previously identified mutations^{46,47,48,49,50}, vectors have been developed that are ablated for binding to CAR, _v-integrins, or both CAR and _v-integrins. One study found that ablating both CAR and integrin binding was necessary to reduce significantly the transduction of the liver and other organs⁵¹. This study also showed that ablation of CAR binding alone did reduce liver transduction to some extent; however, other studies have shown that liver transduction is not affected by this single modification^52,53,54 or by the ablation of both CAR and integrin binding⁵⁵. The reasons for the differences in these findings remain unclear. In addition to the roles of CAR and integrins in liver transduction following intravenous administration, recent data have shown that the fiber shaft appears to play an important role in liver transduction following intravenous administration^{33,34,56,57,58}. Nakamura et al. have shown that an adenoviral vector with an unmodified penton base and a shorter fiber protein from adenovirus type 40 (12 shaft repeats for the Ad40 fiber versus 22 for Ad5 fiber) that does not bind CAR reduced liver transduction by 64-fold compared to a vector with an Ad5 fiber (F5)⁵⁶. However, replacement of the Ad40 short fiber shaft with the long shaft from the Ad5 fiber restored liver transduction to the same levels observed with the full-length Ad5 fiber. Two other recent studies have found similar results with respect to the ability of vectors with short, non-CAR-binding fibers to reduce liver transduction by 10- to 100-fold compared to F5 or CAR-binding-ablated F5^57,58. These results strongly suggest that either the length or the composition of the Ad5 fiber shaft plays a significant role in liver transduction. A putative heparin binding motif, KKTK, which is present in the Ad5 fiber but not in non-subgroup C adenovirus serotypes^33,34, may play a role in mediating liver transduction.

In this study, vectors ablated for CAR binding or both CAR and integrin binding were used as tools to investigate the roles of CAR and integrins in the biodistribution of vector particles and gene expression following intraperitoneal delivery. Using these vectors, we demonstrate that (i) ablation of CAR and integrin binding in adenovirus vectors dramatically reduces transgene expression in all tissues examined compared to an unmodified vector, (ii) mesothelial cells lining peritoneal organs are transduced in a CAR-dependent manner directly from particles present in the peritoneal cavity, (iii) liver and spleen parenchyma are transduced at high vector doses in an integrin-dependent manner from particles that enter the bloodstream, and (iv) ablation of CAR binding together with increased vector dose was correlated directly with a large increase in the levels of vector circulating in the blood following intraperitoneal administration. These findings suggest that adenovirus vectors ablated for native receptor binding and redirected to tumor-specific receptors will increase the specificity of gene transfer to tumors present in the peritoneal cavity. In addition, the increase in circulating vector levels following intraperitoneal administration of CAR-binding-ablated vectors suggests improved methodologies for treating disseminated cancer using targeted adenovirus vectors.

Results

Double-Ablated Vector Avoids Nontarget Tissue Transduction

We were interested in understanding the role of CAR and integrin interactions in the transduction, biodistribution, and pharmacokinetics of adenovirus vectors following intraperitoneal administration. Toward this aim, we utilized a panel of vectors that contain intact CAR and integrin interactions (Ad), mutations that ablate binding to CAR (Ad.F^*), or mutations that ablate binding to both CAR and integrins (Ad.PB^*F^*) (Fig. 1). We designed an initial experiment to determine the effect of ablating both the CAR and the integrin interactions on the kinetics and overall level of transduction following intraperitoneal administration. We compared transduction by the conventional vector, AdL, to that of a CAR- and integrin-binding-ablated vector, AdL.PB^*F^*. We used in vivo image analysis to detect luciferase expression under noninvasive conditions to determine the gross biodistribution of transduction over time. One day after administration of AdL, we detected a large amount of luminescence over the length and breadth of the abdominal side (Fig. 2A). The expression level was decreased by about 50-fold at day 4 postadministration and converged at the level of the double-ablated vector at day 7 (Fig. 2C). In the case of AdL.PB^*F^* the expression level was about 1000-fold lower than that of AdL at day 1, and the intensities were conserved throughout the period of measurement (Figs. 2B and 2C). Some of the mice showed slight expression around the spleen and injection site (Fig. 2B).

Figure 1.

Schematic diagram of Ad vectors with altered cell surface interactions. CAR and integrins mediate cellular transduction by adenovirus vectors. The single-ablated vector (Ad.F^*) has a modified fiber in which binding to CAR is abolished. In addition to the modification of the fiber, the double-ablated vector (Ad.PB^*F^*) has a modified penton base in which the integrin-binding RGD sequence is deleted. The HA tag, which is recognized by the artificial receptor (HA), was inserted into the HI loop of the fiber knob of both ablated vectors.

Full figure and legend (74K)

Figure 2.

Noninvasive imaging of luciferase expression. After the intraperitoneal administration of (A) conventional AdL or (B) double-ablated AdL.PB^*F^* at a dose of 1 10¹¹ pu, photons emitted from the abdominal side were collected with an ultrasensitive photon detector (exposure time: 30 s). Luciferin was administered into the peritoneal cavity just before the image acquisition. (C) Digitalization of photons around the lower abdominal area was performed on the images acquired at indicated time points (, AdL; , AdL.PB^*F^*). Luciferin was administered intraperitoneally just before the image acquisition at each time point. Data show averages standard deviations (n = 5). FOI, field of interest.

Full figure and legend (169K)

Mesothelial Cells are the Primary Targets of the Conventional Vector

We used green fluorescent protein (GFP) expression vectors with (Adf) and without (Adf.PB^*F^*) native receptor binding to identify the organs that were transduced following intraperitoneal administration. At day 3 postadministration of Adf, we observed most GFP expression in the squamous epithelial cells of the mesothelium that lines all of the organs and tissues in the peritoneal cavity (Fig. 3). Although the majority of the expression was observed on the surface of intestine and diaphragm, sliced specimens of the liver and spleen showed some limited transduction of the parenchyma (data not shown). These observations indicated that the mesothelium is the primary target of the conventional vector following intraperitoneal administration. For the Adf.PB^*F^* vector, little if any GFP expression was observed in either the mesothelium or the parenchyma, indicating that transduction of these tissues is strongly dependent on CAR and/or integrins.

Figure 3.

Green fluorescent protein (GFP) expression followed by intraperitoneal administration. Adf or Adf.PB^*F^* GFP-expression vectors were injected into the peritoneal cavity of CD-1 nude mice (2 10¹¹ pu/0.4-ml dose, n = 3). Three days after the administration, tissues were resected for image acquisition. Pictures are the GFP expression on the surface of tissues (peritoneum, liver, spleen, and kidney) acquired using a fluorescence stereomicroscope. Representative images are shown.

Full figure and legend (210K)

Effects of CAR or CAR plus Integrin Binding Ablation on the Biodistribution of Transduction and Vector Genome

Based upon the above results, we carried out experiments to distinguish between the roles of CAR and integrins in the transduction and biodistribution of adenovirus vectors following intraperitoneal vector administration. To accomplish this, we determined luciferase expression and adenoviral genome content in each tissue following the administration of the three vectors, i.e., conventional (AdL), CAR-binding-ablated (AdL.F^*), and CAR- and integrin-binding-ablated (AdL.PB^*F^*) vectors. In addition, each of the three vectors was administered at two dosages, 1 10¹⁰ or 1 10¹¹ pu (Fig. 4). At the low dose, the transduction of all the measured abdominal tissues was significantly lower for AdL.F^* compared to AdL, indicating that transduction of these tissues is strongly CAR-dependent. We observed additional reductions with AdL.PB^*F^* for all tissues and the transduction levels by each of the vectors were correlated with the vector DNA levels. However, by comparison of AdL and AdL.F^* at the high dose, only the peritoneum continued to show CAR-dependent transduction. In addition, the increases in peritoneal transduction and vector DNA for each of the vectors were correlated with the 10-fold increase in dose. Surprisingly, in contrast to the 100-fold reduction in liver transduction by the CAR-ablated vector at the lower dose, the ablation of CAR binding dramatically enhanced the transduction in the liver by approximately 30-fold compared to the conventional vector at the high dose. Transduction of the lung by the CAR-ablated vector was also enhanced compared to the conventional vector at the high dose, whereas transduction of the spleen by the CAR-ablated vector was equal to the conventional vector. An additional unexpected finding was that the Ad genome content in lung, which is located outside of the peritoneal cavity, increased by 500-fold going from the low to the high vector dose. We also observed this enhancement of Ad genome content in the liver and spleen with both ablated vectors at the level of 100-fold differences. These results suggested that the dissemination of adenoviral particles into the bloodstream was dramatically enhanced going from the 10¹⁰-pu dose to the 10¹¹-pu dose, especially when the ablated vectors were administered. Regardless of dose, the double-ablated vector showed the lowest transduction among the three vectors. These results indicate that the ablation of both CAR and integrin binding is necessary for reducing transduction of the above tissues at doses up to 10¹¹ pu, whereas the ablation of CAR binding alone is insufficient to reduce transduction of these tissues at the high dose of vector.

Figure 4.

Luciferase expression and Ad DNA content in tissues. Conventional (AdL), single-ablated (AdL.F^*), or double-ablated (AdL.PB^*F^*) vectors were administered intraperitoneally at 1 10¹⁰ and 1 10¹¹ doses into C57Bl/6 mice (0.5-ml dose, n = 5). One day after administration, harvested tissues were snap-frozen and ground into powder. Aliquoted powder was lysed with CCLR (Promega), and the supernatant was assayed for luciferase activities with the Luciferase Assay System (Promega). Protein content in supernatant was used for normalization. Black column shows luciferase activities (RLU/mg protein, left axis). DNA was isolated from aliquoted powder using the DNeasy Tissue Kit (Qiagen). Ad genome content was analyzed by quantitative PCR (ABI Prism 7700), using primers to the pIX region of adenovirus as described under Materials and Methods. Normalization was performed using the endogenous ribosomal RNA genome content in each preparation originating from host cells (Applied Biosystems). Gray column shows Ad genome contents (copies/g DNA, right axis). Left side of graph, 1 10¹⁰-pu dose; right side of graph, 1 10¹¹-pu dose. Each column shows the geometric mean standard deviation (n = 5).

Full figure and legend (226K)

Uptake of Adenoviral Vectors into Bloodstream Following Intraperitoneal Administration

To test the hypothesis that the 10¹¹-pu dose of vector resulted in large amounts of vector entering the bloodstream following intraperitoneal injection, we collected blood at 30, 90, and 180 min following vector administration. Adenoviral particles in the plasma were determined by titering particles in the plasma using the pseudo-receptor-expressing A232 (A232-HA) cells (Fig. 5A). After administration at the lower 10¹⁰ dose, we detected less than 1% of the injected dose in the bloodstream at any time for any of the vectors. After administration of the single-ablated vector (1 10¹⁰ dose), we detected 0.5, 0.9, and 0.2% of injected particles at 30, 90, and 180 min, respectively. The kinetics of the double-ablated vector was similar, although differences between the single- and the double-ablated vector grew larger over time. We found only 0.03% of the conventional vector in the bloodstream after 30 min. After 90 min, the blood level of the conventional vector fell to less that 0.01%. For all the vectors, we observed a much higher vector percentage in the bloodstream following administration at the 10¹¹ dose. After the administration of the single-ablated vector (1 10¹¹ dose), we detected 1.8% of injected particles at 30 min. The amount of particles was increased up to 18% at 90 min and sustained over the 10% range until 180 min. Although its blood levels were slightly less, the double-ablated vector behaved similar to the single-ablated vector. In contrast, only a small proportion (0.1%) of the conventional vector was detected in the bloodstream at 30, 90, and 180 min following intraperitoneal injection.

Figure 5.

Analysis of vector kinetics in the bloodstream. Luciferase-expressing vectors were injected into the peritoneal cavity of C57Bl/6 mice (1 10¹⁰ or 1 10¹¹ pu/0.5 ml). After 30, 90, and 180 min of administration, blood was collected by retro-orbital bleed using heparinized capillary tubes. (A) Adenoviral content in the plasma was determined by reporter gene expression on A232-HA cells as described under Materials and Methods. (AdL, ; AdL.F^*, ; AdL.PB^*F^*, ; solid line, 1 10¹¹ dose; dashed line, 1 10¹⁰ dose.) Results are shown using the geometric means + standard deviations (n = 4). Three of four plasma specimens collected at 180 min post-conventional vector (1 10¹⁰) administration showed less expression than the detection limit. (B) Area under the curve between 30 and 180 min was calculated applying the trapezoidal rule, multiplying the particle concentration by its retention time. (Gray column, 1 10¹⁰ pu/0.5 ml dose; black column, 1 10¹¹ pu/0.5 ml dose.) Each column shows the geometric mean standard deviation (n = 4).

Full figure and legend (76K)

We converted the above results for all of the vectors at both doses to area under the curve (AUC_30–180) values by multiplying the concentration and retention time (Fig. 5B). For all of the vectors, this analysis showed that a 10-fold increase in vector dose from 10¹⁰ to 10¹¹ resulted in a much larger than 10-fold increase in the AUC_30–180. These increases in AUC_30–180 between the two doses for the conventional, single-ablated, and double-ablated vectors were 50-, 200-, and 100-fold, respectively. The analysis also showed that regardless of dose, the ablation of CAR binding (either CAR binding alone or both CAR and integrin binding) dramatically increased the amount of vector detected in the bloodstream. At the 1 10¹¹ dose, the AUC_30–180 for the single- and double-ablated vectors was approximately 100- and 40-fold higher, respectively, than the AUC_30–180 for the conventional vector (P < 0.005).

Transduction of the Liver Parenchyma is Correlated with Integrin Binding

The conventional vector had been found to transduce the mesothelium present on the exterior of organs efficiently via the intraperitoneal fluid (Fig. 3). In contrast, the high concentrations of ablated vectors found in the blood suggested that these vectors primarily transduce the interior of organs via the blood. To test the above hypothesis, we performed X-gal staining of sectioned tissues after the administration of -galactosidase-expressing vectors at a dose of 1 10¹¹ (Fig. 6). While the conventional vector (AdZ) transduced only the mesothelial cells of liver and spleen, as indicated before, neither the single-ablated (AdZ.F^*) nor the double-ablated vector (AdZ.PB^*F^*) showed significant transduction of the mesothelium. However, X-gal staining of sectioned tissue showed significant transduction by the single-ablated vector in the interior of the liver and spleen. Because the AUC_30–180 values were similar between the single-ablated and the double-ablated vector (Fig. 5), the efficient transduction of the interior liver parenchyma by the single-ablated vector, but not the double-ablated vector, suggested that transduction of the liver parenchyma by the single-ablated vector was integrin-mediated.

Figure 6.

Transduction of mesothelium and/or parenchyma. -Galactosidase-expressing vectors were administered into the peritoneal cavity of C57Bl/6 mice (1 10¹¹ pu/0.5 ml). At 3 days postadministration, the liver and spleen were resected. To show the transduction of the mesothelium (exterior area), whole tissues were stained with X-gal. For the staining of parenchyma (interior area), tissues were sliced with a razor blade and immersed into X-gal solution. After the staining, images were acquired using a SPOT-RT digital camera connected to a fluorescence stereomicroscope under bright field. Representative images are shown. (Bar, 200 m.) Note: In the surface staining procedure for visualizing transduced mesothelial cells, the X-gal stain would penetrate into and stain parenchymal tissues within these organs. This staining is observed in the surface-stained samples for AdZ.F^*. However, comparison of the surface-stained AdZ.F^* samples with the surface-stained AdZ samples clearly shows that mesothelial cells are not transduced by AdZ.F^*.

Full figure and legend (249K)

Feasibility of Double-Ablated Vector as Targeted Vector in Vivo

The above results demonstrated that ablation of CAR and integrin binding significantly reduced transduction of the mesothelial cells lining the organs in the peritoneal cavity. We designed a study to test whether the double-ablated vector could be redirected to these cells when modified to express a nonnative target receptor. Since the conventional vector primarily transduced mesothelial cells following intraperitoneal administration, the study design involved using a conventional vector to express a target receptor for the double-ablated vector on mesothelial cells. We performed a primary intraperitoneal administration of the conventional vector, Ad(HA). This vector contains the gene encoding the artificial receptor for the hemagglutinin (HA) tag present in the double-ablated vector. Based on the above results, we expected Ad(HA) administration to result in expression of the artificial receptor for the HA tag on the mesothelial cells. Two days later, we administered the double-ablated vector, Adf.PB^*F^*. The double-ablated vector has an HA tag in the HI loop of the fiber knob. We performed primary administrations using three dosages (1 10¹⁰, 1 10¹¹, and 3 10¹¹ pu) of Ad(HA) or a control null vector (Ad.null, 3 10¹¹ pu). We performed secondary administration of the double-ablated vector, Adf.PB^*F^*, at a dose of 1 10¹¹, 2 days following the primary administration (Fig. 7A). We sacrificed the animals 2 days after the secondary administration of the double-ablated vector and examined them for GFP expression in the peritoneal cavity. When null vector was preadministered, the transduction by the double-ablated vector was indistinguishable from the single administration of double-ablated vector as shown in Fig. 3. However, following preadministration of Ad(HA), GFP was expressed in the mesothelium in a manner proportionate to the dose of Ad(HA). These results demonstrate that the double-ablated vector can be targeted to cells in vivo using specific ligand–receptor systems.

Figure 7.

Targeting to an artificial receptor in vivo. The feasibility of retargeting the Adf.PB^*F^* vector to cells in the peritoneal cavity was tested in vivo. (A) The pseudo-receptor expression vector, Ad(HA), or control vector, Ad.null, was injected at the indicated doses (0.5 ml) into the peritoneal cavity of C57Bl/6 mice. The administration of Ad(HA) was expected to cause the artificial receptor expression on mesothelial cells. At 2 days postadministration, 1 10¹¹ pu of double-ablated GFP-expression vector, Adf.PB^*F^*, was administered intraperitoneally. Three days later, GFP expression on the surface of abdominal tissues (liver, peritoneum, and spleen) was captured using a fluorescence stereomicroscope. (Bar, 200 m.) (B) AE25-HA tumor cells were injected into the peritoneal cavity of nude mice. Adf.PB^*F^* or Adf was injected intraperitoneally 30 min later at a dose of 1 10¹¹ pu. Two days later, GFP expression in tumor foci and on the surface of abdominal tissues was captured using a fluorescence stereomicroscope. Representative images are shown. (Bar, 500 m.)

Full figure and legend (352K)

We designed a final study to test whether the double-ablated vector could be redirected to tumor cells modified to express a nonnative target receptor. We stably transduced AE25 cells to express the artificial receptor to the HA tag. This tumor cell line is derived from the lung adenocarcinoma cell line, A549, and expresses high levels of CAR and _v-integrins. We injected the resulting AE25-HA tumor cells into the peritoneal cavity of nude mice and 30 min later the mice received an intraperitoneal injection of either Adf.PB^*F^* or Adf. We sacrificed the mice 2 days later and monitored GFP expression in the peritoneal cavity. The results of this study showed that both the Adf.PB^*F^* and the Adf vectors efficiently transduced the AE25-HA cells as evidenced by strong GFP expression in tumor foci (Fig. 7B). While both vectors efficiently transduced the target AE25-HA cells, the Adf vector efficiently transduced nontarget mesothelial cells while the Adf.PB^*F^* vector did not. These results demonstrate the feasibility of retargeting a vector ablated for CAR and integrin binding to a specific receptor located on tumor cells within the peritoneal cavity.

Discussion

These studies have shown that both CAR and integrins play important and somewhat surprising roles in controlling the biodistribution of gene expression following intraperitoneal delivery. The results of these studies have also demonstrated the feasibility of in vivo targeting of adenovirus vectors to cells in the peritoneal cavity. Overall, these findings strongly suggest that these vectors will be useful for increasing the selectivity of gene transfer to tumors present in the peritoneal cavity. An unanticipated outcome from these studies, namely that intraperitoneal administration of CAR- and integrin-binding-ablated vectors increases their persistence in the bloodstream, has important implications for targeting disseminated tumors located throughout the body.

The data support a model with at least two pathways of adenoviral vector transduction following intraperitoneal administration, i.e., CAR-dependent transduction of the mesothelium via direct contact in the intraperitoneal space and integrin-dependent transduction of parenchymal cells via the bloodstream (Fig. 8). The degree of transduction by either pathway was found to strongly depend not only on the CAR- and integrin-binding status of the vector, but also on the vector dose. At the low dose the mesothelium is the primary target for transduction as has been previously shown following intraperitoneal administration of adenovirus vectors⁵⁹. Transgene expression and vector DNA content in abdominal tissues were strongly reduced through the ablation of CAR binding, with some additional reductions achieved through the ablation of both CAR and integrin binding. These data essentially reflect previously reported in vitro data using CAR- and integrin-expressing cells⁵¹. Ablation of CAR binding does significantly increase the level of vector detected in the blood. However, at the low dose only a small fraction of the original dose is detected in the bloodstream.

Figure 8.

Model for two pathways of transduction following intraperitoneal administration. Schema of adenoviral biodistribution after ip administration (peritoneum, gray frame; blood circulation, red line; transduction, black arrow; viral uptake into circulation, gray arrow; green, transduced cells). (A) Conventional vector transduces mesothelium of abdominal tissues in CAR-dependent manner. Uptake of viral particles into bloodstream is lower than that of ablated vectors. (B) CAR-binding ablated vector does not transduce the mesothelium and exudes into bloodstream. Parenchyma is transduced by integrin interaction. (C) Double-ablated vector exudes to bloodstream as well as single-ablated vector; however, it does not transduce on parenchyma. (D) Transduction mediated by specific receptor–ligand system. Pseudo-receptor expression on mesothelia redirects doubly ablated vector to mesothelium.

Full figure and legend (168K)

At the high dose, transduction of the mesothelium remains the primary site of transduction for the CAR-binding vector, although much higher levels of vector enter the bloodstream compared to the lower dose. In contrast, ablation of CAR binding results in a dramatic change in the location and level of transduction in the liver and spleen. Parenchymal cells are the major sites of transduction. With the 1 log increase in the dose of single-ablated vector, the overall level of transduction increases between 100- and 1000-fold, while the vector DNA level increases by approximately 100-fold. These dramatic increases are correlated directly with an approximately 100-fold increase in the amount of vector present in the bloodstream. This increase is apparently responsible for the observed transduction of parenchymal cells and for the large increases in liver transduction. The efficient transduction of liver hepatocytes by the single-ablated vector is consistent with previous reports showing that liver transduction is not strongly affected by ablation of CAR binding when the vector is administered intravenously. As for the single-ablated vector, when both the CAR and the integrin binding are ablated, viral particles are likewise observed in the bloodstream at high concentrations (Fig. 5A). However, in contrast to the single-ablated vector, transduction of parenchyma is greatly reduced compared to either the conventional or the single-ablated vector. These data suggest an important role for integrin-mediated adenoviral transduction in the liver and are consistent with previous findings showing that the double-ablated vector is severely attenuated for transduction of the liver following intravenous injection. Based on our study, it is likely that the double-ablated vector remains in organs without expression. However, further analysis will be required to unravel their ultimate fate in vivo. In contrast to the liver and spleen, the overall transduction of the peritoneum remained CAR-dependent at the high dose. This finding may result from the vascular endothelial cells adjacent to peritoneal cells having small intercellular spaces unlike those of liver and spleen, which have open fenestrated capillaries⁶⁰. Consequently, this physical barrier may prevent the transduction of parenchymal tissue underlying the peritoneal membrane.

In the context of the present intraperitoneal studies it is important to note that the precise mechanisms governing liver transduction by adenovirus vectors remain to be determined. For example, in previous studies that have examined the effect of simultaneous ablation of CAR and integrin binding on liver transduction following intravenous delivery, one study found a significant reduction in liver transduction⁵¹, whereas another study found no reduction in liver transduction⁵⁵. These studies used different knob modifications and somewhat different routes of administration. The different results suggest two important conclusions: First, the route of administration and/or subtle differences between vectors may alter the extent of liver gene transfer. Second, CAR and integrin interactions may not be the only interactions that are involved in uptake by the liver. Additional interactions of adenovirus vectors with heparin sulfate proteoglycans^33,34 and with opsonizing proteins in the blood, as well as potential interactions with blood cells⁶¹, may play a role in uptake by the liver. In addition, a number of studies support a role for the length or composition of the fiber shaft in controlling liver gene transfer^{56,57,58,62,63}. It is also probable that the dominance of any particular interaction in controlling uptake by the liver may change depending on the route of administration and/or genetic differences between vectors that include the mutations used to ablate CAR, fiber shaft composition or length, or the size or composition of ligands inserted into the vector. The results obtained in the present studies clearly highlight that the administration route can alter liver gene transfer, biodistribution, and bloodstream persistence by tropism-modified adenovirus vectors.

The present studies have suggested important roles for CAR and integrin interactions in the location of transduction (mesothelial vs parenchymal), the overall levels of transduction in individual organs and the whole animal, and the levels of vector in the bloodstream following intraperitoneal administration. However, a number of questions remain that are related to the mechanisms by which particles enter the bloodstream. First, why are fewer CAR-binding particles present in the bloodstream compared to the CAR-ablated particles? One likely reason for this finding is that CAR-dependent uptake of the standard vector by the mesothelium or lymphatics prevents its efficient entry into the bloodstream. A second possible reason is that CAR-binding particles do rapidly enter the bloodstream from the peritoneal cavity but are then rapidly cleared by the liver in a CAR-dependent manner. However, if this were the case, it would be expected that transduction of liver parenchyma would be observed, which is not the case.

A second unanswered question is why there are roughly 2 logs higher blood levels of all the vectors with only a 1-log increase in vector dose. This finding suggests that innate vector clearance mechanisms that are present in the liver and/or in the peritoneal cavity become saturated at higher vector doses. In fact, Kupffer cells in the liver have been implicated in the rapid clearance of adenovirus vectors from the circulation following intravenous administration^64,65. Recent evidence suggests that the dramatic increase in gene expression observed in the mouse liver going from a dose of 10¹⁰ to 10¹¹ particles (the so-called threshold effect) is due to saturation of uptake by the Kupffer cells⁶⁶. The impaired clearance at the higher dose then permits vectors to transduce hepatocytes efficiently. It is possible that either liver Kupffer cells or peritoneal macrophages become unable to clear significant amounts of the vectors at the high doses, resulting in increased vector concentrations in the blood. The mechanism by which adenovirus clearance is inhibited could include direct saturation of receptor binding sites used for clearance or saturation of some soluble factor that binds to the adenovirus and tags it for clearance by cells in the peritoneum or the liver.

Perhaps the most intriguing question remaining is why the single- and double-ablated vectors persist for long periods in the bloodstream following intraperitoneal administration. This result is particularly surprising given our finding that both the conventional and the double-ablated vectors similarly display rapid clearance from the circulation immediately following intravenous administration of 1 10¹¹ pu with a half-life of approximately 2 min (data not shown), in agreement with earlier data on clearance of conventional adenovirus vectors⁵². Given the rapid clearance of these vectors by the liver following intravenous administration, it is difficult to explain why CAR-ablated vectors administered via the intraperitoneal route persist in the bloodstream at such high concentrations. The augmentation of bloodstream persistence following intraperitoneal administration could result from an extended release from the cavity or from an altered recognition of the vector by cells in the liver. It has been reported that the exudation of macromolecules from peritoneal cavity to circulation is mediated by stomata, which are openings into the diaphragmatic lymphatics^67,68. Nagy et al. reported that 5–10% of injected dextran particles, which are commonly used as a plasma volume expander, were drained into the blood compartment by 60 min post-intraperitoneal administration⁶⁹. The approximately 5–20% levels of ablated vectors detected in blood for 30–90 min at a dose of 1 10¹¹ suggest that they are behaving as biologically "inactive" molecules. Although the mechanism needs to be elucidated further, these findings suggest that intraperitoneal administration of double-ablated vector may have an advantage for obtaining prolonged blood circulation. Prolonged bloodstream circulation would likely dramatically improve the ability to target adenovirus vectors to disseminated cancer present outside the peritoneal cavity.

These studies suggest that adenovirus vectors devoid of their native CAR and integrin binding will significantly improve the safety of gene therapy for the treatment of cancers present in the peritoneal cavity, including disseminated colon or stomach cancer, mesothelioma, and ovarian cancer. Ablation of native tropism was found to cause dramatic reductions in gene transfer to a number of nontarget tissues, including the peritoneum, liver, lung, and spleen. Interestingly, ablation of CAR binding alone does not appear to be adequate, since nontarget tissue transduction can actually increase, rather than decrease, at the higher doses of CAR-ablated vectors. For the double-ablated vector, overall transduction in the whole animal was reduced by 2–3 orders of magnitude. In addition to demonstrating a reduction of nontarget tissue transduction by this vector, we have demonstrated that this vector retains its transducing function in vivo and can be redirected to cells that express a target receptor. This is an important milestone toward applying targeted vectors to the treatment of peritoneal cancers. Other studies have shown that many primary ovarian cancers and ovarian cell lines express low levels of CAR that limit gene transfer using conventional vectors. It has also been shown that directing tropism-expanded vectors to receptors that are significantly expressed on ovarian cells can significantly enhance gene transfer. Together, these studies and the present study suggest that a CAR- and integrin-ablated vector that is redirected to a tumor-specific marker will dramatically improve both the safety and the efficacy of gene therapy for cancers located in the peritoneal cavity.

Materials and methods

Cells

A232 cells (GenVec, Gaithersburg, MD) are derived from 293 cells obtained from ATCC (Manassas, VA). These cells express open reading frame 6 under the control of a metallothionein promoter as previously described⁷⁰. AE25 cells are an A549-based E1-complementing cell line that will be described in detail elsewhere (D. E. Brough, A. Lizonova, and I. Kovesdi, unpublished results, 1999). Pseudo-receptor-expressing cells, A232-HA and AE25-HA, were obtained as described before⁷¹. Briefly, the cells were transfected with a retrovirus encoding a membrane-anchored single-chain antibody that recognizes the influenza virus hemagglutinin epitope, as previously described^51,71. Stably expressing clones were selected using Geneticin (Life Technologies, Gaithersburg, MD).

Vectors

The standard (AdL), CAR-binding-ablated (AdL.F^*), and CAR- and integrin-binding-ablated (AdL.PB^*F^*) vectors were constructed as previously described⁵¹. Briefly, AdL is an E1- and E3-deleted recombinant Ad5-based vector containing the firefly luciferase transgene in the E1 region under the control of the CMV promoter. AdL.F^* has a modified AB loop in the fiber (R412S, A415G, E416G, and K417G) that abolishes CAR binding. AdL.F^* also has an insertion of the HA tag, which is recognized by the anti-HA pseudo-receptor. AdL.PB^*F^* has a modified penton base in which the RGD sequence is deleted and replaced with an HA tag, plus the CAR-binding-ablating mutation and HA tag as in AdL.F^*. Adf and Adf.PB^*F^* contain the marker gene encoding GFP, and AdZ, AdZ.F^*, and AdZ.PB^*F^* contain the marker gene encoding -galactosidase under the control of the CMV promoter. Their capsid modifications are the same as those of the luciferase-expressing vectors described above. Ad(HA) is an E1- and E3-deleted recombinant Ad5-based vector containing the anti-HA pseudo-receptor transgene in the E1 region under the control of the CMV promoter. The anti-HA gene is the same as used above to construct the A232-HA cells. Ad.null is an E1- and E3-deleted recombinant Ad5-based vector containing the CMV promoter in the E1 region without any transgene.

Animals

For the in vivo analysis, 6- to 8-week-old female CD-1 nude and C57Bl/6 mice were purchased from Charles River Laboratories (Wilmington, MA). Vectors were administered intraperitoneally in a volume of 0.4 or 0.5 ml. Blood collection and noninvasive in vivo imaging were performed under isoflurane anesthesia. Tissues were harvested after euthanasia with CO₂ inhalation.

Noninvasive imaging of luciferase expression in vivo

For the noninvasive imaging, 150 mg/kg body weight of D-firefly luciferin (Xenogene, Alameda, CA) was administered intraperitoneally just prior to image acquisition. Five to 30 min after the administration of substrate, mice were transferred to the light-shielding imaging chamber and hooked up to isoflurane inhalators. Photons emitted from the abdominal side of the mice were accumulated for 30 s.

Determination of luciferase expression in tissues

The liver, spleen, lung, and peritoneal lining were collected at day 1 post-vector administration and immediately frozen in liquid nitrogen. After being ground with pestle and mortar, aliquoted powder was lysed with Cell Culture Lysis Reagent (CCLR; Promega, Madison, WI), and the supernatant was separated by centrifugation. Luciferase activity in the supernatant was determined using the Luciferase Assay System (Promega) and a TR717 microplate luminometer (Perkin–Elmer). Protein concentration, measured by Bradford Assay (Bio-Rad Protein Assay), was used to normalize the luciferase expression.

Quantitative PCR

DNA was isolated from ground tissues using DNeasy Tissue kits (Qiagen, Valencia, CA). Tissue powder was incubated 3 h in proteinase K and DNA was purified by minicolumns and quantified by spectroscopy. Adenoviral DNA content was determined by ABI Prism 7700 (Applied Biosystems, Foster City, CA) real-time PCR using the following primers and probe directed to the pIX region: CGCGGGATTGTGACTGACT (sense), GCCAAAAGAGCCGTCAACTT (antisense), and FAM-AGCAGTGCAGCTTCCCGTTCATCC-TAMRA. Endogenous genome, rRNA gene content was used for normalization (Applied Biosystems).

Quantification of adenoviral particles in plasma

After the intraperitoneal administration of adenoviral vector at 1 10¹⁰- or 1 10¹¹-pu doses, blood was collected by retro-orbital bleeding at 30, 90, and 180 min using heparinized capillary tubes. Separated plasma (2 l) or aliquoted vector (serial dilution between 1 10⁷ and 170 pu/well) was incubated on A232-HA cells in 96-well plates. After 24 h incubation at 37°C under 5% CO₂, the cells were lysed with 50 l of CCLR and assayed for luciferase activity (Promega) as described above. Luciferase levels corresponding to each sample were converted to particle numbers in plasma using the "RLU–particle number standard curve" obtained for each vector. The total volume of plasma was estimated as 1.3 ml.

Histological studies

Three days post-vector administration, sliced tissues were stained with the HistoMark X-Gal Substrate Set (KPL, Gaithersburg, MD). Image acquisition for X-gal staining and GFP expression was performed under an MZ FLIII stereomicroscope (Leica Microsystems, Heerbrugg, Switzerland) together with a SPOT-RT digital camera (Diagnostic Instruments, Inc., Sterling Heights, MI).

in vivo tumor targeting

The artificial receptor-expressing tumor cells, AE25-HA, were prepared by detaching cells by EDTA treatment followed by resuspension in PBS. Twenty million cells were inoculated into the peritoneal cavity of CD-1 nude mice 30 min before vector administration. GFP-expressing vector was injected at a dose of 1 10¹¹ pu intraperitoneally. Two days later, mice were sacrificed and images were acquired under a fluorescence stereomicroscope as described above.

References

1. Wolf, J. K. and Jenkins, A. D. (2002). Gene therapy for ovarian cancer. Int. J. Oncol. 21: 461–468. | PubMed | ChemPort |
2. Bauerschmitz, G. J., et al. (2002). Treatment of ovarian cancer with a tropism modified oncolytic adenovirus. Cancer Res. 62: 1266–1270. | PubMed | ChemPort |
3. Heise, C., Ganly, I., Kim, Y. T., Sampson-Johannes, A., Brown, R. and Kirn, D. (2000). Efficacy of a replication-selective adenovirus against ovarian carcinomatosis is dependent on tumor burden, viral replication and p53 status. Gene Ther. 7: 1925–1929. | Article | PubMed | ChemPort |
4. Deshane, J., et al. (1995). Targeted tumor killing via an intracellular antibody against erbB-2. J. Clin. Invest. 96: 2980–2989. | PubMed | ChemPort |
5. Mahasreshti, P. J., et al. (2001). Adenovirus-mediated soluble flt-1 gene therapy for ovarian carcinoma. Clin. Cancer Res. 7: 2057–2066. | PubMed | ChemPort |
6. Kim, J., Hwang, E. S., Kim, J. S., You, E. H., Lee, S. H. and Lee, J. H. (1999). Intraperitoneal gene therapy with adenoviral-mediated p53 tumor suppressor gene for ovarian cancer model in nude mouse. Cancer Gene Ther. 6: 172–178. | Article | PubMed | ChemPort |
7. Hwang, E. S., et al. (1998). The effects of the adenovirus-mediated wild-type p53 delivery in human epithelial ovarian cancer cell line in vitro and in vivo. Int. J. Gynecol. Cancer. 8: 27–36. | Article | PubMed |
8. Modesitt, S. C., Ramirez, P., Zu, Z., Bodurka-Bevers, D., Gershenson, D. and Wolf, J. K. (2001). in vitro and in vivo adenovirus-mediated p53 and p16 tumor suppressor therapy in ovarian cancer. Clin. Cancer Res. 7: 1765–1772. | PubMed | ISI | ChemPort |
9. Song, K., Li, Z., Seth, P., Cowan, K. H. and Sinha, B. K. (1997). Sensitization of cis-platinum by a recombinant adenovirus vector expressing wild-type p53 gene in human ovarian carcinomas. Oncol. Res. 9: 603–609. | PubMed | ChemPort |
10. Song, K., Cowan, K. H. and Sinha, B. K. (1999). In vivo studies of adenovirus-mediated p53 gene therapy for cis-platinum-resistant human ovarian tumor xenografts. Oncol. Res. 11: 153–159. | PubMed | ChemPort |
11. Kanamori, Y., et al. (1998). A newly developed adenovirus-mediated transfer of a wild-type p53 gene increases sensitivity to cis-diamminedichloroplatinum (ii) in p53-deleted ovarian cancer cells. Eur. J. Cancer. 34: 1802–1806. | Article | PubMed | ISI | ChemPort |
12. Rancourt, C., et al. (1998). Basic fibroblast growth factor enhancement of adenovirus-mediated delivery of the herpes simplex virus thymidine kinase gene results in augmented therapeutic benefit in a murine model of ovarian cancer. Clin. Cancer Res. 4: 2455–2461. | PubMed | ISI | ChemPort |
13. Rosenfeld, M. E., et al. (1996). Adenoviral-mediated delivery of herpes simplex virus thymidine kinase results in tumor reduction and prolonged survival in a scid mouse model of human ovarian carcinoma. J. Mol. Med. 74: 455–462. | Article | PubMed | ISI | ChemPort |
14. Tong, X., et al. (1999). Comparison of long-term survival of cytomegalovirus promoter versus Rous sarcoma virus promoter-driven thymidine kinase gene therapy in nude mice bearing human ovarian cancer. Hybridoma. 18: 93–97. | PubMed | ChemPort |
15. Tong, X. W., et al. (1997). Human epithelial ovarian cancer xenotransplants into nude mice can be cured by adenovirus-mediated thymidine kinase gene therapy. Anticancer Res. 17: 811–813. | PubMed | ISI | ChemPort |
16. Zeimet, A. G., et al. (2002). Determination of molecules regulating gene delivery using adenoviral vectors in ovarian carcinomas. Gene Ther. 9: 1093–1100. | Article | PubMed | ChemPort |
17. Alvarez, R. D., et al. (2000). A cancer gene therapy approach utilizing an anti-erbB-2 single-chain antibody-encoding adenovirus (ad21): a phase I trial. Clin. Cancer Res. 6: 3081–3087. | PubMed | ISI | ChemPort |
18. Alvarez, R. D. and Curiel, D. T. (1997). A phase I study of recombinant adenovirus vector-mediated intraperitoneal delivery of herpes simplex virus thymidine kinase (HSV-tk) gene and intravenous ganciclovir for previously treated ovarian and extraovarian cancer patients. Hum. Gene Ther. 8: 597–613. | PubMed | ISI | ChemPort |
19. Alvarez, R. D. and Curiel, D. T. (1997). A phase I study of recombinant adenovirus vector-mediated delivery of an anti-erbB-2 single-chain (SFV) antibody gene for previously treated ovarian and extraovarian cancer patients. Hum. Gene Ther. 8: 229–242. | PubMed | ISI | ChemPort |
20. Barnes, M. N., Coolidge, C. J., Hemminki, A., Alvarez, R. D. and Curiel, D. T. (2002). Conditionally replicative adenoviruses for ovarian cancer therapy. Mol. Cancer Ther. 1: 435–439. | PubMed | ISI | ChemPort |
21. Buller, R. E., et al. (2002). A phase I/II trial of rad/p53 (Sch 58500) gene replacement in recurrent ovarian cancer. Cancer Gene Ther. 9: 553–566. | Article | PubMed | ISI | ChemPort |
22. Buller, R. E., et al. (2002). Long term follow-up of patients with recurrent ovarian cancer after Ad p53 gene replacement with Sch 58500. Cancer Gene Ther. 9: 567–572. | Article | PubMed | ISI | ChemPort |
23. Deshane, J., et al. (1997). Transductional efficacy and safety of an intraperitoneally delivered adenovirus encoding an anti-erbB-2 intracellular single-chain antibody for ovarian cancer gene therapy. Gynecol. Oncol. 64: 378–385. | Article | PubMed | ChemPort |
24. Hasenburg, A., et al. (2002). Adenovirus-mediated thymidine kinase gene therapy for recurrent ovarian cancer: expression of coxsackie–adenovirus receptor and integrins alphavbeta3 and alphavbeta5. J. Soc. Gynecol. Invest. 9: 174–180. | Article | ChemPort |
25. Hasenburg, A., et al. (2001). Adenovirus-mediated thymidine kinase gene therapy in combination with topotecan for patients with recurrent ovarian cancer: 2.5-year follow-up. Gynecol. Oncol. 83: 549–554. | Article | PubMed | ISI | ChemPort |
26. Hasenburg, A., et al. (2000). Thymidine kinase gene therapy with concomitant topotecan chemotherapy for recurrent ovarian cancer. Cancer Gene Ther. 7: 839–844. | Article | PubMed | ChemPort |
27. Hasenburg, A., et al. (1999). Thymidine kinase (tk) gene therapy of solid tumors: valacyclovir facilitates outpatient treatment. Anticancer Res. 19: 2163–2165. | PubMed | ChemPort |
28. Wickham, T. J. (2003). Ligand-directed targeting of genes to the site of disease. Nat. Med. 9: 135–139. | Article | PubMed | ChemPort |
29. Bergelson, J. M., et al. (1997). Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science. 275: 1320–1323. | Article | PubMed | ISI | ChemPort |
30. Tomko, R. P., Xu, R. and Philipson, L. (1997). HCAR and MCAR: The human and mouse cellular receptors for subgroup C adenoviruses and group B coxsackieviruses. Proc. Natl. Acad. Sci. USA. 94: 3352–3356. | Article | PubMed | ChemPort |
31. Wickham, T. J., Mathias, P., Cheresh, D. A. and Nemerow, G. R. (1993). Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell. 73: 309–319. | Article | PubMed | ISI | ChemPort |
32. Bai, M., Harfe, B. and Freimuth, P. (1993). Mutations that alter an Arg-Gly-Asp (RGD) sequence in the adenovirus type 2 penton base protein abolish its cell-rounding activity and delay virus reproduction in flat cells. J. Virol. 67: 5198–5205. | PubMed | ISI | ChemPort |
33. Dechecchi, M. C., Melotti, P., Bonizzato, A., Santacatterina, M., Chilosi, M. and Cabrini, G. (2001). Heparan sulfate glycosaminoglycans are receptors sufficient to mediate the initial binding of adenovirus types 2 and 5. J. Virol. 75: 8772–8780. | Article | PubMed | ISI | ChemPort |
34. Dechecchi, M. C., Tamanini, A., Bonizzato, A. and Cabrini, G. (2000). Heparan sulfate glycosaminoglycans are involved in adenovirus type 5 and 2-host cell interactions. Virology. 268: 382–390. | Article | PubMed | ISI | ChemPort |
35. Dmitriev, I., et al. (1998). An adenovirus vector with genetically modified fibers demonstrates expanded tropism via utilization of a coxsackievirus and adenovirus receptor-independent cell entry mechanism. J. Virol. 72: 9706–9713. | PubMed | ISI | ChemPort |
36. Kelly, F. J., et al. (2000). Selectivity of tag-72-targeted adenovirus gene transfer to primary ovarian carcinoma cells versus autologous mesothelial cells in vitro. Clin. Cancer Res. 6: 4323–4333. | PubMed | ISI | ChemPort |
37. Hemminki, A., et al. (2001). An adenovirus with enhanced infectivity mediates molecular chemotherapy of ovarian cancer cells and allows imaging of gene expression. Mol. Ther. 4: 223–231. | Article | PubMed | ISI | ChemPort |
38. Kim, J., et al. (2002). Targeting adenoviral vectors by using the extracellular domain of the coxsackie–adenovirus receptor: improved potency via trimerization. J. Virol. 76: 1892–1903. | Article | PubMed | ISI | ChemPort |
39. Kim, J. S., Lee, S. H., Cho, Y. S., Choi, J. J., Kim, Y. H. and Lee, J. H. (2002). Enhancement of the adenoviral sensitivity of human ovarian cancer cells by transient expression of coxsackievirus and adenovirus receptor (CAR). Gynecol. Oncol. 85: 260–265. | Article | PubMed | ChemPort |
40. Kanerva, A., et al. (2002). Targeting adenovirus to the serotype 3 receptor increases gene transfer efficiency to ovarian cancer cells. Clin. Cancer Res. 8: 275–280. | PubMed | ISI | ChemPort |
41. You, Z., Fischer, D. C., Tong, X., Hasenburg, A., Aguilar-Cordova, E. and Kieback, D. G. (2001). Coxsackievirus–adenovirus receptor expression in ovarian cancer cell lines is associated with increased adenovirus transduction efficiency and transgene expression. Cancer Gene Ther. 8: 168–175. | Article | PubMed | ISI | ChemPort |
42. Kim, M., et al. (2002). The therapeutic efficacy of adenoviral vectors for cancer gene therapy is limited by a low level of primary adenovirus receptors on tumour cells. Eur. J. Cancer. 38: 1917–1926. | Article | PubMed | ISI | ChemPort |
43. Kanerva, A., et al. (2002). Gene transfer to ovarian cancer versus normal tissues with fiber-modified adenoviruses. Mol. Ther. 5: 695–704. | Article | PubMed | ISI | ChemPort |
44. Vanderkwaak, T. J., et al. (1999). An advanced generation of adenoviral vectors selectively enhances gene transfer for ovarian cancer gene therapy approaches. Gynecol. Oncol. 74: 227–234. | Article | PubMed | ISI | ChemPort |
45. Blackwell, J. L., et al. (2000). Using a tropism-modified adenoviral vector to circumvent inhibitory factors in ascites fluid. Hum. Gene Ther. 11: 1657–1669. | Article | PubMed | ISI | ChemPort |
46. Bewley, M. C., Springer, K., Zhang, Y. B., Freimuth, P. and Flanagan, J. M. (1999). Structural analysis of the mechanism of adenovirus binding to its human cellular receptor, CAR. Science. 286: 1579–1583. | Article | PubMed | ISI | ChemPort |
47. Kirby, I., et al. (1999). Mutations in the DG loop of adenovirus type 5 fiber knob protein abolish high-affinity binding to its cellular receptor CAR. J. Virol. 73: 9508–9514. | PubMed | ChemPort |
48. Kirby, I., et al. (2000). Identification of contact residues and definition of the CAR-binding site of adenovirus type 5 fiber protein. J. Virol. 74: 2804–2813. | Article | PubMed | ChemPort |
49. Roelvink, P. W., Mi Lee, G., Einfeld, D. A., Kovesdi, I. and Wickham, T. J. (1999). Identification of a conserved receptor-binding site on the fiber proteins of CAR-recognizing adenoviridae. Science. 286: 1568–1571. | Article | PubMed | ISI | ChemPort |
50. Santis, G., et al. (1999). Molecular determinants of adenovirus serotype 5 fibre binding to its cellular receptor CAR. J. Gen. Virol. 80: (Pt. 6): 1519–1527. | PubMed | ChemPort |
51. Einfeld, D. A., et al. (2001). Reducing the native tropism of adenovirus vectors requires removal of both CAR and integrin interactions. J. Virol. 75: 11284–11291. | Article | PubMed | ISI | ChemPort |
52. Alemany, R. and Curiel, D. T. (2001). CAR-binding ablation does not change biodistribution and toxicity of adenoviral vectors. Gene Ther. 8: 1347–1353. | Article | PubMed | ISI | ChemPort |
53. Leissner, P., et al. (2001). Influence of adenoviral fiber mutations on viral encapsidation, infectivity and in vivo tropism. Gene Ther. 8: 49–57. | Article | PubMed | ISI | ChemPort |
54. Smith, T., et al. (2002). In vivo hepatic adenoviral gene delivery occurs independent of the coxsackievirus–adenovirus receptor. Mol. Ther. 5: 770–779. | Article | PubMed | ISI | ChemPort |
55. Martin, K., Brie, A., Saulnier, P., Perricaudet, M., Yeh, P. and Vigne, E. (2003). Simultaneous CAR- and alpha v integrin-binding ablation fails to reduce Ad5 liver tropism. Mol. Ther. 8: 485–494. | Article | PubMed | ISI | ChemPort |
56. Nakamura, T., Sato, K. and Hamada, H. (2003). Reduction of natural adenovirus tropism to the liver by both ablation of fiber-coxsackievirus and adenovirus receptor interaction and use of replaceable short fiber. J. Virol. 77: 2512–2521. | Article | PubMed | ISI | ChemPort |
57. Schoggins, J. W., Gall, J. G. and Falck-Pedersen, E. (2003). Subgroup B and F fiber chimeras eliminate normal adenovirus type 5 vector transduction in vitro and in vivo. J. Virol. 77: 1039–1048. | PubMed | ChemPort |
58. Vigne, E., et al. (2003). Genetic manipulations of adenovirus type 5 fiber resulting in liver tropism attenuation. Gene Ther. 10: 153–162. | Article | PubMed | ChemPort |
59. Setoguchi, Y., Jaffe, H. A., Chu, C. S. and Crystal, R. G. (1994). Intraperitoneal in vivo gene therapy to deliver alpha 1-antitrypsin to the systemic circulation. Am. J. Respir. Cell Mol. Biol. 10: 369–377. | PubMed | ISI | ChemPort |
60. Braet, F. and Wisse, E. (2002). Structural and functional aspects of liver sinusoidal endothelial cell fenestrae: a review. Comp. Hepatol. 1: 1. | Article | PubMed |
61. Cichon, G., et al. (2003). Titer determination of Ad5 in blood: a cautionary note. Gene Ther. 10: 1012–1017. | Article | PubMed | ChemPort |
62. Shayakhmetov, D. M. and Lieber, A. (2000). Dependence of adenovirus infectivity on length of the fiber shaft domain. J. Virol. 74: 10274–10286. | Article | PubMed | ISI | ChemPort |
63. Smith, T. A., et al. (2003). Adenovirus serotype 5 fiber shaft influences in vivo gene transfer in mice. Hum. Gene Ther. 14: 777–787. | Article | PubMed | ISI | ChemPort |
64. Lieber, A., et al. (1997). The role of Kupffer cell activation and viral gene expression in early liver toxicity after infusion of recombinant adenovirus vectors. J. Virol. 71: 8798–8807. | PubMed | ISI | ChemPort |
65. Wolff, G., Worgall, S., van Rooijen, N., Song, W. R., Harvey, B. G. and Crystal, R. G. (1997). Enhancement of in vivo adenovirus-mediated gene transfer and expression by prior depletion of tissue macrophages in the target organ. J. Virol. 71: 624–629. | PubMed | ChemPort |
66. Tao, N., et al. (2001). Sequestration of adenoviral vector by Kupffer cells leads to a nonlinear dose response of transduction in liver. Mol. Ther. 3: 28–35. | Article | PubMed | ISI | ChemPort |
67. Azzali, G. (1999). The lymphatic vessels and the so-called "lymphatic stomata" of the diaphragm: a morphologic ultrastructural and three-dimensional study. Microvasc. Res. 57: 30–43. | Article | PubMed | ChemPort |
68. Ohtani, Y., Ohtani, O. and Nakatani, T. (1995). Microanatomy of the rat diaphragm with special reference to the lymphatics and mesothelial stomata. Ital. J. Anat. Embryol. 100: (Suppl. 1): 143–153. | PubMed |
69. Nagy, J. A., Herzberg, K. T., Masse, E. M., Zientara, G. P. and Dvorak, H. F. (1989). Exchange of macromolecules between plasma and peritoneal cavity in ascites tumor-bearing, normal, and serotonin-injected mice. Cancer Res. 49: 5448–5458. | PubMed | ChemPort |
70. Brough, D. E., Lizonova, A., Hsu, C., Kulesa, V. A. and Kovesdi, I. (1996). A gene transfer vector-cell line system for complete functional complementation of adenovirus early regions E1 and E4. J. Virol. 70: 6497–6501. | PubMed | ISI | ChemPort |
71. Einfeld, D. A., Brough, D. E., Roelvink, P. W., Kovesdi, I. and Wickham, T. J. (1999). Construction of a pseudoreceptor that mediates transduction by adenoviruses expressing a ligand in fiber or penton base. J. Virol. 73: 9130–9136. | PubMed | ISI | ChemPort |

Acknowledgements

We thank Leslie West and Randy Osborne for assistance in the work involving animals; Angela Appiah, Barb Aughtman, Katarina Ujhazy, and Alena Lizonova for their help in providing cells and vectors; and our colleagues at FUSO Pharmaceutical, Ltd., for their support of this work.