Skip to main content
NCBI home page
Molecular Therapy logoLink to Molecular Therapy
. 2010 Feb 23;18(5):921–928. doi: 10.1038/mt.2010.5

Use of Adenoviral Vectors to Target Chemotherapy to Tumor Vascular Endothelial Cells Suppresses Growth of Breast Cancer and Melanoma

Pingchuan Li 1, Yanzheng Liu 1, Jonathan Maynard 1, Yucheng Tang 1, Albert Deisseroth 1
PMCID: PMC2890101  PMID: 20179680

Abstract

To target chemotherapy to tumor vascular endothelial cells (TVECs), we created the AdTie2RprCDFibknob−RGD+ vector by inserting into an AdEasy adenoviral vector (Ad) backbone: (i) the cytosine deaminase (CD) gene driven by the Tie2 receptor promoter (Tie2Rpr) into the E1 region of Ad; (ii) mutations that reduce binding of the fiber knob to the Coxsackie adenovirus receptor (CAR); and (iii) the RGD peptide into the H1 loop of fiber for binding to the αVβ3 integrin receptors on TVECs. To reduce uptake of the AdTie2RprCDFibknob−RGD+ by reticuloendothelial (RE) and liver cells, we intravenously (i.v.) injected Hetastarch and low-dose Ad (one million vector particles (VPs)) prior to i.v. injection of a therapeutic dose (one billion VPs) of the AdTie2RprCDFibknob−RGD+ vector. This treatment induced regressions of N202 breast cancer and B16 melanoma without toxicity to normal tissues. We showed that the tumor regression was induced by infection of the TVECs and not by the infection of tumor cells by the AdTie2RprCDFibknob−RGD+ vector.

Introduction

One factor limiting the success of chemotherapy is the damage it causes to normal tissues. A second problem is the difficulty in delivering chemotherapy across the vascular wall into the extravascular cancer tissues. The only way to increase transfer of drugs to the extravascular tumor tissue is by increasing the dose injected intravenously that increases toxicity to the normal tissues. To solve this problem, we changed the target of therapy from the tumor cells to the luminal membrane of the tumor vascular endothelial cells (TVECs).

We created the AdTie2RprCDFibknob−RGD+ vector1 so that it could be injected intravenously, bind selectively to the TVECs, and program them to produce the cytosine deaminase (CD) protein. CD protein2,3,4,5,6,7 converts the nontoxic prodrug 5-fluorocytosine (5FC) into the chemotherapy agent 5-fluorouracil (5FU). This would destroy the tumor vessels, eventually destroying the tumor cells themselves.

In order to avoid damage to normal tissues, we used the Tie2 receptor promoter (Tie2Rpr), which is active primarily in proliferating endothelial cells,8,9,10,11,12,13,14,15,16,17,18,19,20 as a transcriptional promoter for the CD gene. The Tie2prCD transcription unit was placed in the E1 region of the AdTie2RprCDFibknob−RGD+ vector. Because the biggest difference between the endothelial cells of vessels in normal tissues versus cancer tissue is that 0.05% of endothelial cells of vessels in normal tissue are proliferating, whereas 30% of the endothelial cells in tumor blood vessels are proliferating,21,22 this vector would damage vascular endothelial cells in tumor tissue but not in blood vessels of normal tissue.

The Tie2Rpr is active in angiogenic TVECs.8,9,10,11,12,13,14,15,16,17,18,19,20,21,22 De Palma et al.23 demonstrated that there were three distinct cell populations that expressed the Tie2 receptor in tumor nodules: Tie2 receptor expressing TVECs (TETVECs), Tie2 receptor expressing monocytes (TEMs), which represent 3% of all monocytes, and the Tie 2 receptor expressing pericyte precursors of mesenchymal origin (TEPMOs), which represent <5% of all pericyte cells. Remarkably, in De Palma's work,23 destruction of all Tie2R expressing cells including the TEMs completely prevented human glioma neovascularization and induced substantial degrees of tumor regression in the Tie2tkKO.Tg mouse. This group has recently used “pro-angiogenic monocytes” that express the Tie2R gene to deliver interferon-α therapeutic transcription units to tumor cells.24

The αVβ3 integrin receptor is expressed on the luminal membrane of TVECs. The RGD peptide has been used to deliver nanoparticles, liposomes, transdominant-negative inhibitors of oncogenes and chemotherapy to TVECs through its ability to bind the αVβ3 integrin receptors.25 We therefore inserted the RGD peptide into the H1 loop of the Ad fiber. We also introduced two loss-of-function mutations (S408E and P409A) into the knob domain of the end of the fiber protein. These mutations reduce binding of the fiber knob to Coxsackie adenovirus receptor (CAR) that is on most mammalian cells.26

We then tested the effect of i.v. administration of the AdTie2RprCDFibknob−RGD+ vector on the subcutaneous (SC) growth of N202 murine breast cancer cells and B16 mouse melanoma cells. In order to increase the levels of the AdTie2RprCDFibknob−RGD+ vector in the intravascular space following i.v. injection, we injected two things prior to the administration of therapeutic doses of the AdTie2RprCDFibknob−RGD+ vector: (i) Hetastarch [which saturates the endocytosis pathway in reticuloendothelial (RE) cells] and (ii) A low dose of the AdTie2RprCDFibknob−RGD+ vector, which reduces uptake of the therapeutic dose of the AdTie2RprCDFibknob−RGD+ vector by the liver cells from the blood stream. The results of these experiments showed that the AdTie2RprCDFibknob−RGD+ vector induced a regression of SC deposits of either the N202 breast cancer cell line or the B16 melanoma cell line, and that this effect was due to infection of the TVECs but not due to infection of the tumor cells.

In addition, the combination of the Tie2Rpr for driving the expression of the CD therapeutic transcription unit coupled with the introduction of loss-of-function mutations into the fiber knob and the insertion of the RGD into the fiber protein spared the normal tissues.

Results

Receptors for Ad infection are found on VEGF-treated HUVECs but not on cancer cells or normal cells

The data in Table 1 show that the level of CAR, the αVβ3 receptors, and the αVβ5 receptors were all high on vascular endothelial growth factor (VEGF)–stimulated HUVECs but low or undetectable on N202 mouse breast cancer cells, on B16 mouse melanoma cells, on phorbol myristate acetate–treated U937 monocyte cells, and on pericyte-like C3H/10T1/2 cells.

Table 1.

Percentage of cells positive for the CAR, αVβ3, and αVβ5 receptors as measured by FACS in cancer cell lines (N202 and B16), in a monocytic leukemia cell line (U937) and in a mesenchymal cell line that resembles pericytes (H10T1/2)

graphic file with name mt20105t1.jpg

Open in a new tab

The AdTie2RprGFPFibknob−RGD+ vector infects VEGF-treated HUVECs but not cancer cells nor normal monocytes or pericyte-like cells

We next tested the specificity of the AdTie2RprGFPFibknob−RGD+ vector against proliferating endothelial cells. As shown in Table 2, 18.9% of the VEGF-treated HUVECs were positive for green fluorescent protein (GFP) following exposure to the AdTie2RprGFPFibknob−RGD+ vector and 22% of the VEGF-treated HUVECs were GFP+ when exposed to the AdCMVGFPFibknob−RGD+ vector. Roughly half as many VEGF-treated HUVECs (7.6 and 11.5%) were GFP+ following exposure to the AdTie2RprGFPFibknob+RGD− and AdCMVGFPFibknob+RGD− vectors carrying the native fiber protein (see Table 2). Thus, the RGD modified and mutated fiber protein (Fibknob−RGD+) is functional and the Tie2R promoter is active in the proliferating HUVECs. This suggests that the expression of the vector transgenes in the HUVECs is limited more by receptor availability than by the activity of the transcriptional promoter driving the viral transgenes.

Table 2.

FACS analysis of GFP expression in the HUVECs, N202, B16, C3H/10T1/2, and U937 lines infected with the indicated Ad vectors at MOIs of 30 and 100, engineered to target TVECs

graphic file with name mt20105t2.jpg

Open in a new tab

We then exposed the following cell lines to the AdTie2RprGFPFibknob−RGD+ and the AdTie2RprGFPFibknob+RGD− vectors: the mouse breast cancer cell line N202, the B16 mouse melanoma cell line, the human U937 monocyte cell line, and the human C3H10T1/2 pericyte cell line. Less than 1% of the cells exposed to either of these two vectors contained GFP+ cells. Less than 0.2% GFP+ cells were detected in U937 or C3H10T1/2 cells exposed to the AdTie2RprGFPFibknob+RGD− or AdTie2RprGFPFibknob−RGD+ vectors (see Table 2). These studies indicate that these two cell lines do not contain cells positive for the Tie2 receptor. Only small numbers of GFP+ cells were found in C3H10T1/2 human pericyte-like cells and human monocyte-like U937 cells exposed at 30 MOI to the AdCMVGFPFibknob−RGD+ and AdCMVGFPFibknob+RGD− vectors carrying the CMV promoter-driven GFP transcription units (Table 2). This is possibly due to the low level of the CAR and integrin receptors on these cells (see Table 1).

When we exposed at 30 MOI the N202 breast cancer cell line (αVβ3 positive) to vectors carrying the Fibknob−RGD+ in which the knob had been inactivated and the RGD had been introduced into the H1 loop of the fiber protein (e.g., the AdCMVGFPFibknob−RGD+ vector), only low levels of GFP+ breast cancer cells were found (see Table 2). These data suggest that the VEGF-treated HUVECs were infectable by the AdCMVGFPFibknob−RGD+ vector, but that the N202, B16, U937, and 10T1/2 cells were only minimally infectable if at all by the AdCMVGFPFibknob−RGD+ vector.

In vivo 5FU sensitivity (IC50) of VEGF-treated HUVECs compared to N202 cells and B16 cells

CD protein catalyzes the conversion of 5FC into 5FU. The serum levels of 5FU achieved when this drug is administered intravenously (5 µmol/l) kill only dividing cells. In contrast, when AdCD vectors infect cells that are then exposed to 5FC, the intracellular 5FU levels attained in AdCD-infected cells (200 µmol/l) kill nondividing cells (as well as dividing cells) due to disruption of translation of protein from RNA.27 We therefore tested the sensitivity of the N202 and the B16 cells to toxicity induced by 5FU. B16, which is a very rapidly proliferating mouse melanoma cell line, is very sensitive to in vitro exposure to 5 µmol/l 5FU, whereas VEGF-treated HUVECs that are very slowly dividing are less sensitive to the levels of 5FU that kill dividing cells through binding to thymidylate synthase (see Table 3). The sensitivity of N202, which is intermediate between B16 and HUVECs in proliferation, is less than B16 but more than HUVECs (see Table 3).

Table 3.

IC50 of the 5FU in cell lines, HUVEC, N202, and B16

graphic file with name mt20105t3.jpg

Open in a new tab

Injection of Hetastarch and low doses of the AdTie2RprGFPFibknob−RGD+ vector before injection of therapeutic doses of the AdTie2RprGFPFibknob−RGD+ vector increase the serum levels of this vector in tumor-bearing mice

We then tested whether the levels of the AdTie2RprGFPFibknob−RGD+ vector in the serum following i.v. injection of 1 × 10 9 viral particles (VPs) were higher in mice preinjected i.v. with 5% Hetastarch and 1 × 106 VPs of the AdTie2RprGFPFibknob−RGD+ vector. The purpose of the injection of the Hetastarch is to saturate the phagocytic activity of the RE system so as to reduce the extent of uptake of the adenoviral vectors by the RE system following i.v. administration.7 The preinjection of the “low-dose” (1 × 106 VPs) of the adenoviral vector is designed to saturate the transport mechanisms that result in uptake of the Ad vector into the hepatic cells. We first injected 5 × 105 N202 cells subcutaneously (SC) into nude mice. Once the tumor nodules reached a size that was palpable (150 mm3), we injected i.v. 200 µl of 5% Hetastarch followed by one i.v. injection of a “low-dose” AdTie2RprGFPFibknob−RGD+ vector (1 × 106 VPs). After 2 hours, we injected i.v. a therapeutic dose (1 × 109 VPs) of the AdTie2RprGFPFibknob−RGD+ vector. As shown in Figure 1, the preinjection of the Hetastarch and low-dose vector increased the blood levels generated by injection of the therapeutic dose of the AdTie2RprGFPFibknob−RGD+ vector at 30 minutes and at 180 minutes by four- and eightfold, respectively.

Figure 1.

Figure 1

Open in a new tab

The effect of Hetastarch and low-dose vector preinjection on the titer of the AdTie2RprGFPFibknob−RGD+ vector in peripheral blood. Preinjection mice (white histogram): tail vein injection of 200 µl Hetastarch followed by two injections of the AdTie2RprGFPFibknob−RGD+ vector delivery (the first at one million viral particles and the second 2 hours later at one billion particles, respectively). No preinjection mice (solid black histogram): tail vein injection of the AdTie2RprGFPFibknob−RGD+ vector (one billion particles) without preinjection of the Hetastarch nor the low-dose vector. The blood was collected at 5, 30, and 180 minutes following the completion of the above injections, and then the level of the vector in the serum was tittered by Hexon assay with a permissive cell indicator cell line (293). The preinjection of the Hetastarch and low-dose AdTie2RprGFPFibknob−RGD+ vector increased the blood levels of the AdTie2RprGFPFibknob−RGD+ vector at 30 and at 180 minutes by four- and eightfold, respectively (*P = 0.008). There was no difference at 5 minutes between the two groups. ifu, infectious units.

TEVECs but not tumor cells nor normal cells are GFP+ following in vivo exposure to the AdTie2RprGFPFibknob−RGD+ vector in tumor-bearing mice

De Palma et al.23 had reported the presence of TETVECs, TEMs, and Tie2R expressing mesenchymal progenitor cells (TEMPOs) in SC tumor nodules in mice that had been made transgenic for a transcription unit containing the GFP gene driven by the Tie2R promoter. In order to test whether this was true in our mouse model, we injected tumor-bearing mice i.v. with the AdTie2RprGFPFibknob−RGD+ vector. One of the important objectives of this experiment was to test whether the i.v. administered AdTie2RprGFPFibknob−RGD+ vector was infecting TVECs and not the tumor cells themselves. As shown in Figure 2, the level of TETEVCs that were yellow (double positive for the GFP protein (green) and for the CD31 marker for endothelial cells (red)), or infected by the AdTie2RprGFPFibknob−RGD+ vector was 6% in the periphery of tumor nodules (see panel A of Figure 2). In contrast, the percentage of yellow or infected endothelial cells in the vessels in the center of a tumor nodule was <1%. These numbers of double-positive cells were obtained by dividing the number of yellow cells in panel A of Figure 2 by the total number of red cells. One thousand cells were counted. The observation that the markers for angiogenesis are seen on endothelial cells only at the periphery of a tumor nodule has been made previously by our laboratory.28

Figure 2.

Figure 2

Open in a new tab

Multiparameter immunofluorescence confocal microscopic analysis of the Tie2R expressing cells in tumor tissue. Frozen sections of N202 tumors grown in nude mice into which the AdTie2RprGFPFibknob−RGD+ vector had been injected intravenously were studied by fluorescence confocal microscopy. Green in all panels = green fluorescent protein; blue in all panels = DAPI staining of nuclei. Panel A: Tie2R expressing tumor vascular endothelial cells (TETVECs). Red = CD31 antibody (pan vascular antibody that detects endothelial cells). Dashed line shows the tumor area. There are about 6% cells in the periphery of the tumor nodule that are yellow which means they are double positive for both GFP green color and CD31 red color, compared with only <1% in the central region of the tumor nodule portion (1,000 cells were counted). Panel B: Tumor Tie2R expressing monocytes (TEMs). Red = CD11b antibody, which detects monocytes. Panel C: Tie2R expressing mesenchymal progenitor cells (TEMPOs). Red = CD13 antibody, which detects mesenchymal progenitor cells or pericytes. Panel D: Staining control. Arrows show the double positive (yellow) that are positive for GFP (green) and the red color of the following antibodies: CD31 = pan vascular; CD11b = monocytes; CD13 = mesenchymal progenitors or pericytes. Original magnification ×400.

In contrast, the percentages of the TEMs and TEMPOs in the tumor nodules that were infected by intravenously administered AdTie2RprGFPFibknob−RGD+ vector (yellow divided by red for each cell type in panels B and C of Figure 2) were <1% range. No GFP+ cells were found in the following normal organs: heart, liver, spleen, brain, and kidney. In the lung, which is the organ of the body which has the second greatest content/tissue weight of macrophages, there were a few GFP+ cells. These cells were found to be endothelial cells (CD31+), monocytes (CD11b+), and pericytes (CD13+) by multiparameter immunofluorescence confocal microscopy as described above (data not shown). These data indicated that the AdTie2RprGFPFibknob−RGD+ vector can target the GFP gene to TVECs.

Efficacy and safety of the AdTie2RprCDFibknob−RGD+ vector in the tumor-bearing mice

We next compared the growth of SC deposits of tumor derived from SC injection of either the N202 or the B16 cell lines into mice following the i.v. injection of the AdTie2RprCDFibknob−RGD+ vector or the AdCMVCDFibknob−RGD+ vector followed by i.p. administration of 5FC. As shown in Figure 3a,b, the i.v. injection of the AdTie2RprGFPFibknob−RGD+ vector or the AdCMVCDFibknob−RGD+ vector on days 3, 8, and 17 following SC injection of the tumor cells induced a statistically significant reduction in the growth of N202-derived tumor nodules (Figure 3a) and in the growth of the B16-derived tumor nodules (Figure 3b), compared to the control mice (P < 0.05). The degree of suppression was greater in the case of the N202 cell line than in that of the B16 melanoma cell line (see Figure 3). A statistically significant suppressive effect of the AdTie2RprGFPFibknob−RGD+ vector or the AdCMVCDFibknob−RGD+ vector on the growth of the B16 was apparent at a later stage in which the tumor nodules were very large. In contrast, in the case of N202, there was significant tumor volume reduction at earlier as well as later time points (see Figure 3).

Figure 3.

Figure 3

Open in a new tab

Effect of i.v. administration of the AdTie2RprCDFibknob−RGD+ (diamonds) and AdCMVCDFibknob−RGD+ (squares) vectors on the growth of (a) N202 and (b) B16 tumor cells in nude mice. Vectors were delivered i.v. at 3, 8, and 17 days after tumor inoculation (arrow). 5FC was then injected intraperitoneally every other day from 4 to 22 days. Triangles represent control mice (six mice) that received phosphate-buffered saline instead of vectors. Each data point represents the mean ± SD of seven or eight mice per treatment group. The data are representative of three experiments. (a) From 10 days after tumor inoculation, the N202 tumor size in both groups of vector-treated mice was statistically significantly smaller than in the control group at each time point (P < 0.05). (b) In the case of the B16 cell line, statistically significant differences in the SC tumor volume in both groups of vector-treated and the control mice were seen after 25 days. *Statistically significant differences. The tumor growth in untreated control mice is defined by open triangles; the tumor growth in AdTie2RprCDFibknob−RGD+ treated mice is defined by the open diamonds; the tumor growth in AdCMVCDFibknob−RGD+ treated mice is defined by the open squares. (c) Four weeks after AdCMVCDFibknob−RGD+ vector plus 5FC treatment, the N202-bearing mice were thin, exhibited lethargy and hunched posture. (d) The mice treated with the AdTie2RprCDFibknob−RGD+ appeared healthy and active. i.v., intravenous.

As shown in Figure 3c, the mice treated with i.v. injection of the AdCMVCDFibknob−RGD+ vector followed by i.p. 5FC looked emaciated, whereas the mice treated with i.v. injection of the AdTie2RprCDFibknob−RGD+ vector followed by 5FC injections was very healthy and did not appear emaciated (see Figure 3d). This may be due to the fact that the AdTie2RprCDFibknob−RGD+ vector has a very restricted spectrum of cells in which the CD gene is expressed. In contrast, the i.v. injection of the AdCMVCDFibknob−RGD+ vector results in CD expression in all proliferating cells throughout the body (liver cells, RE cells, gastrointestinal mucosal epithelial cells, and hematopoietic cells) that are displaying the αVβ3 and αVβ5 integrin receptors. This observation underscores the value of using the Tie2R promoter to regulate the expression of the CD gene in the therapeutic vector.

CD protein is present in TEVECs but not in tumor cells or normal tissues following i.v. administration of the AdTie2RprCDFibknob−RGD+ vector

In Figure 4, the antibody used to detect CD protein was conjugated to a green reporter molecule in all panels. The blue color is DAPI staining of cell nuclei in all panels. In panel A, the red color identifies endothelial cells; in panel B, the red color identifies monocytes; and in panel C, the red color identifies pericytes. The percentage of CD+ cells of each cell type is calculated by the ratio of the yellow cells (cells which are both positive for CD (green) to the red marker for endothelial cells, monocytes, or pericytes) divided by the cells positive for red marker. As shown in panel A of Figure 4, the percentage of endothelial (red) cells that were yellow (and therefore positive for the CD protein) was between 0.05 and 1% in all panels.

Figure 4.

Figure 4

Open in a new tab

Multiparameter immunofluorescence confocal microscopic analysis of the cytosine deaminase (CD) expression in tumor tissue. The N202 tumors were grown SC in nude mice that were treated intravenously with AdTie2RprCDFibknob−RGD+ plus 5FC intraperitoneally (i.p.). Green in all panels = CD; blue in all panels = DAPI staining of nuclei. Panel A: tumor Tie2R expressing vascular endothelial cells (TETVECs). Red = CD31 antibody (pan vascular antibody that detects endothelial cells). Panel B: tumor Tie2R expressing monocytes (TEMs). Red = CD11b antibody, which detects monocytes. Panel C: Tie2R expressing mesenchymal progenitor cells (TEMPOs) or pericytes. Red = CD13 antibody, which detects mesenchymal progenitor cells. Panel D: staining control. Arrows show the double positive (yellow) for CD (green) and the red color of the following antibodies: CD31 = pan vascular; CD11b = monocytes; CD13= mesenchymal progenitors. Original magnification ×400.

One can now ask why the percentage of endothelial cells that were GFP+ in the TVECs of tumor-bearing mice injected i.v. with the AdTie2RprGFPFibknob−RGD+ vector was 6%, whereas the percentage of endothelial cells in tumor-bearing mice injected with the AdTie2RprCDFibknob−RGD+ vector was <1%. One possible explanation is that the endothelial cells infected with the AdTie2RprCDFibknob−RGD+ vector are undergoing cell death due to the therapeutic effect of the vector and the 5FC, whereas in the experiment with the GFP vectors, cell death of the infected cells is not occurring. Another possible explanation is that the GFP protein may be more stable than the CD protein and thus, it is more stable in the TVECs.

Discussion

The experimental results summarized in this article document that the i.v. administration of the AdTie2RprCDFibknob−RGD+ vector followed by intraperitoneal administration of 5FC induces regression of SC deposits of the N202 mouse breast cancer cell line and of the B16 melanoma cell line through infection of the TVECs rather than by infection of the tumor cells. There are several lines of evidence that suggest that the action of the AdTie2RprCDFibknob−RGD+ vector is mediated by infection of the mouse angiogenic endothelial cells in vasculature of tumor tissue rather than by infection of the tumor cells.

1. We found (see Figure 1) GFP+ cells in populations of HUVECs exposed in vitro to the AdTie2RprGFPFibknob−RGD+ vector but not in any of the other vector exposed cell lines tested: N202 (mouse breast cancer), B16 (mouse melanoma), U937 (human monocyte), and C3H/10T1/2 cells (mouse pericyte-like cells).

2. When we injected the AdTie2RprGFPFibknob−RGD+ vector i.v., we found GFP+ cells (see Figure 4) in 6% of the angiogenic endothelial cells in the tumor vessels, in 1% of the monocytes in tumor tissue, and in 1% of the pericytes around tumor vessels, but not in tumor tissue nor in the normal organs of the body with the exception of the lung in which a small number of cells were infected.

3. When we injected the AdTie2RprCDFibknob−RGD+ vector i.v. following preinjections of Hetastarch and low doses of Ad vector, cells positive for the CD were found (see Figure 4) in the angiogenic endothelial cells in tumor vessels, in monocytes in tumor tissue, and in the pericytes around tumor vessels, but not in the N202 or the B16 tumor cells themselves.

The engineering of the AdTie2RprCDFibknob−RGD+ vector makes it possible for this vector to infect TVECs thereby inducing regressions of extravascular tumor cells without having to infect the tumor cells themselves. Thus, the original goal of this work, to circumvent the obstacles blocking success in cancer treatment by shifting the focus from the tumor cells to the TVECs, has been achieved.

The tumor regressions induced by the AdTie2RprCDFibknob−RGD+ vector occurred without side effects suggesting that the normal tissues and their vessels were spared from the damaging effect of the AdTie2RprCDFibknob−RGD+ vector and 5FC. This is an unambiguous demonstration that it is possible to use i.v. administered adenoviral vectors to deliver chemotherapy to tumor vasculature so as to treat cancer.

This success is built on two features of the AdTie2RprCDFibknob−RGD+ vector: the regulation of the GFP or CD genes by the Tie2Rpr, which limits the expression of the transcription unit to angiogenic vascular endothelial cells, and the engineering of the fiber protein so that infection is dependent on the presence of the αVβ3 and αVβ5 integrin receptors on angiogenic dividing endothelial cells, which are present at 100-fold higher levels in the angiogenic vessels (tumor vessels) than is the case of the blood vessels in normal tissues.

The mutational inactivation of the knob region of the fiber protein as a binding ligand and its replacement is a major reason why the AdTie2RprCDFibknob−RGD+ vector may be attractive as a therapeutic in human beings. The replacement of the CAR by the RGD will be a major safety factor because the ability of the therapeutic AdTie2RprCDFibknob−RGD+ vector to infect normal parenchymal tissue cells and vascular cells of normal tissues will be vastly reduced compared to vectors carrying a wild-type fiber protein. The replacement of the CAR-binding domains of the fiber with the integrin-binding RGD accomplishes something even more important. Nagel et al.26 reported that Ad5 failed to agglutinate human erythrocytes following introduction of the same two mutations (S408E and P409A) in the AB loop of the fiber protein, which ablates its binding to human red cells which contain CAR receptors on their surface.29

The advantages of the AdTie2RprCDFibknob−RGD+ adenoviral vector are as follows: (i) reproducibility and ease of production of the therapeutic vector as compared to liposomes or nanoparticles, which are highly variable in terms of formulation; (ii) the multiplier effect of using an enzyme for a therapeutic: for every vector particle delivered to the target cell, thousands of CD protein enzyme molecules are produced within each cell, and each CD protein produces hundreds of thousands of 5FU molecules to generate an amplification effect, as compared to chemotherapy agents or recombinant biologicals, the effect of which is stoichiometric rather than enzymatic; (iii) the generation of the high levels of 5FU produces a “bystander effect” on the noninfected cells through release of unphosphorylated 5FU; and (iv) the high levels of 5FU achieved in the target cells produces cell death even in nondividing cells.

One may ask why the suppression of the growth of the N202 and B16 cells seen in Figure 3a,b is partial and not complete. One answer of particular relevance to the tumor vascular targeting mechanism is that the tumor vasculature in the center of a tumor does not express “angiogenic” markers. In particular, we have shown previously28 that the Tie2R promoter is not active in the TVECs in the center of a tumor, and the vessels at the center of a tumor do not express the integrin receptors. In contrast, the TVECs at the outer edge of a tumor nodule are proliferating, promote the expression of genes driven by the Tie2R promoter, and are positive on their plasma membrane for the integrin receptors. In order to apply tumor vascular targeting to the complete destruction of tumor nodules, repetitive applications of the tumor vascular targeting therapy must be carried out. After each application, as the other angiogenic layer of the tumor is stripped away by therapy, the vessels of the next layer toward the center of the tumor start to proliferate and thereby become sensitive to the tumor vascular targeting therapy.

The use of the AdTie2RprCDFibknob−RGD+ vector preceded by the i.v. injection of Hetastarch and a low dose of vector (one million particles) was shown to increase the blood levels of the vector due to decreased uptake of the vector by the RE system. Because human erythrocytes are CAR positive and would therefore bind the wild-type fiber of the Ad vector, the Fibknob−RGD+ mutant virus may be less subject to intravascular sequestration by binding to human red blood cells.24,25,26,27,28,29 The combination of the angiogenic endothelial cell–specific transcriptional promoter with the CD therapeutic transcription unit, the engineering of the fiber and the use of Hetastarch and low-dose vector intravenous preinjection prior to i.v. injection of a therapeutic dose (one billion particles) of the AdTie2RprCDFibknob−RGD+ vector, has led for the first time to a successful strategy for using adenoviral vectors for i.v. tumor vascular targeting therapy. Future directions for our studies of this vector will include the following: extensive safety testing in animal models, and combination of this AdTie2CDFibknob−RGD+ vector with vectors that are designed to target other steps in the angiogenesis pathway in tumor cells.

Materials and Methods

Adenoviral vector constructs. The plasmid pHHSDKXK was a gift from Thomas Sato's lab, which contains the murine Tie2R promoter [HindIII, 2.1 kilobases (kb)] and enhancer (XhoI–KpnI, 1.7 kb) sequences. The AdFibknob−RGD+ vector that contains the pAdeasy-1 sequence1 plus an additional RGD-4C sequence in the HI loop of the fiber protein and two mutations (S408E and P409A) in the AB loop of the fiber protein was constructed as described previously.7 Replication-deficient adenoviral vectors, AdTie2RprGFPFibknob+RGD−, AdTie2RprGFPFibknob−RGD+, AdCMVGFPFibknob+RGD−, and AdCMVGFPFibknob−RGD+, as well as the corresponding AdTie2RprCDFibknob−RGD+ and AdCMVCDFibknob−RGD+ vectors that carry E. coli CD gene, were constructed as outlined previously.7

The Tie2R promoter regulatory sequences that were inserted into the adenoviral vectors were made by AdEasy system.1 Briefly, both the Tie2R promoter (HindIII, 2.1 kb) and enhancer (XhoI–KpnI, 1.7 kb) fragments, which came from pHHSDKXK, were inserted into pShutter vector to generate pShutter-Tie2/e. Primer 1 (5′-ACGCGTCGACATGGTGAGCAAGGGC-3′, SalI) and primer 2 (5′-GGAAGATCTTAAGATACATTGATGAG-3′, BglII) were used to amplify GFP plus SV40poly(A) genes from plasmid pTrack.7 Primer 3 (5′-ACGCGTCGACGAGGCTAACAATGTC-3′, SalI) and primer 4 (5′-ACGCGTCGACTAAGATACATTGATGAG-3′, SalI) were used to amplify E. coli CD plus SV40poly(A) genes from plasmid pShutterCMVCD.7 Cloning of the GFP or CD expression cassette into the pShutter-Tie2Rpr/e resulted in pShutter-Tie2RprGFP or pShutter-Tie2CD, respectively, followed by homologous recombination with pAdeasy-1 or AdFibknob−RGD+ in the bacterial strain BJ5183. The recombined plasmids for the following vectors: AdTie2RprGFPFibknob+RGD−, AdTie2RprGFPFibknob−RGD+, AdTie2RprCD Fibknob−RGD+, and AdCMVCDFibknob−RGD+ vectors, were digested with PacI and transfected separately into 293 cells by Lipofectamine 2000 (Invitrogen, Carlsbad, CA), according to the manufacturer's protocol. The resulting viral vectors were named as follows: the AdTie2RprGFPFibknob+RGD− vector that is a vector carrying the Tie2R promoter driving the GFP transcription unit in a vector with a wild-type fiber protein; the AdTie2RprGFPFibknob−RGD+ vector, which carries a GFP transcription unit driven by the Tie2R promoter in a vector with a fiber carrying an inactive functionless knob domain and a H-1 loop with a RGD peptide inserted; the AdTie2RprCDFibknob−RGD+; and the AdCMVCDFibknob−RGD+, which are the corresponding vectors carrying the CD gene.

All viruses were double plaque-purified, expanded on HEK 293 cells, purified on a cesium chloride gradient, titered with Adeno-X Rapid Titer Kit (Clontech, Mountain View, CA) as infectious units (ifu)/ml and stored at −80 °C. The rapid RCA Assay Kit (Cell Biolabs, San Diego, CA) was used to rule out high levels of replication-competent adenovirus. The vectors were shown to contain <1 RCA in one billion vector particles.

Cells. HUVECs were kindly provided by Zhiwei Hu from Yale, New Haven, CT, cultured in M199 medium which is supplemented with fetal bovine serum (10%), endothelial cell growth supplement from Sigma, St Louis, MO (E2759), and VEGF from BD Pharmingen, San Jose, CA (cat. 551515, 5µg/ml). HEK 293, N202.1A, a rH2N breast cancer cell line isolated from a breast cancer of a FVB-rH2N,Tg mouse and the B16 murine mouse melanoma cell line were obtained from ATCC, Manassas, VA and cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Invitrogen). U937 (CRL-1593.2) and C3H/10T1/2 (CCL-226) also come from ATCC and follow the culturing protocol.

Fluorescence-activated cell sorting analysis of receptors on different cell lines. We used the following antibodies to determine the expression level of CAR as well as αVβ3 and αVβ5 integrin receptors on the N202, B16, U937, HUVECs, and the C3H/10T1/2 cell lines: mouse antihuman CAR mAb (sc-32795), PE-goat anti-mouse IgG (M30004-1; Invitrogen), PE-mouse antihuman αVβ3 mAb (sc-7312PE), and PE-mouse antihuman αVβ5 mAb (FAB2528P; R&D, Minneapolis, MN).

IC50 (half-maximal growth inhibition) determination of 5FU on different cell lines. The median-effect principle and MTT method were used in this assay. The MTT cell proliferation assay kit was purchased from ATCC and we followed its instructions. Briefly, we plated the optimal cell number in 96-well plates in 100 µl of culture medium. We included cell-only control wells. These were subjected to overnight incubation. These cultures were fed with fresh medium supplemented with various amounts of the 5FU. We used a serial twofold dilution of 5FU starting at 12 µmol/l. Following 2–3 days of incubation, the medium was exchanged for fresh medium without drug, including the medium-only control wells. The next day, 10 µl of MTT reagent were added to each well, including plate background control wells. Following this, we carried out a 4-hour incubation at 37 °C until a purple precipitate was visible. Following this, we added 100 µl detergent reagent and incubated at room temperature in the dark overnight. We then measured the absorbance at 570 nm. IC50 values were calculated as follows: inhibition rate = (1 − absorbance of treated cells/control cells) × 100. IC50 values were determined from the inhibition rates and are expressed as the median of three independent experiments (four replicates each).

Vector in vitro transduction of different cell lines. We cultured 5 × 104 of the following cell lines: HUVECs, N202, B16, U937, and C3H/10T1/2 in 6-well plates. After 1 day, we infected the cells with the following vectors at different multiplicities of infection (number of VPs per cell): AdTie2RprGFPFibknob+RGD−, AdTie2RprGFPFibknob−RGD+, AdCMV GFPFibknob+RGD−, and AdCMVGFPFibknob−RGD+ vectors. Mock-transduced cells served as the negative control. Infected cells were allowed to grow for 5 days before fluorescence-activated cell sorting analysis in order to reach steady state of GFP expression and to rule out pseudotransduction. Cells were washed and treated with Cellstripper (nonenzymatic cell dissociation solution from Cellgro, Manassas, VA). Single-cell suspensions of 10,000 were analyzed by fluorescence-activated cell sorting to count the GFP+ cells.

Animal experiment procedures. N202 (5 × 105) or B16 (2 × 105) cells [in 200 µl of phosphate-buffered saline (PBS)] were injected subcutaneously into female nude mice (6–8 weeks of age). For the GFP distribution assay, after the tumor size reached visible size, we injected Hetastarch i.v. (200 µl), followed by injection of 1 × 106 VPs of the AdTie2RprGFPFibknob−RGD+ vector, followed in 2 hours by an iv injection of 1 × 109 VPs of the AdTie2RprGFPFibknob−RGD+ vector, which is a therapeutic dose. We collected blood at 5, 30, and 180 minutes later and titered the level of the AdTie2RprGFPFibknob−RGD+ vector. Different organs and tumor tissue were collected and snap-frozen in OCT cryomatrix 3 days later. For the tumor treatment assay, vectors were delivered i.v. at 3, 8, and 17 days after the tumor cell inoculation. At each time, we first delivered Hetastarch (200 µl) i.v., followed by two times of injections of the AdTie2RprCDFibknob−RGD+ or the AdCMVCDFibknob−RGD+ vectors, (first at a dose of 1 × 106 and 2 hours later at a dose of 1 × 109 VPs) or PBS for the control group. Then the mice received 5FC by i.p. injection (a single dose of 500 µg/g weight every other day for 10 days). Tumor growth was measured by caliper every other day, and tumor size was calculated using the equation: volume = length × 0.5 width2.

Multiparameter immunofluorescence confocal microscopy. Six-micrometer cryostatic sections were quenched in the blue stuff for 20 minutes (blue stuff: 1 mmol/l CuSO4, 50 mmol/l NH4Ac) and blocked with 5% goat serum in PBS (0.1% Tween) for 20 minutes, washed in PBS; we then blocked the sections again in M.O.M. mouse Ig blocking reagent (vector MKB-2213) for 1 hour and washed with PBS.

For double immunofluorescence staining, we applied the rabbit anti-GFP (A11122; Molecular Probes, Carlsbad, CA) or the mouse antibacterial CD (16D8F2, BD Pharmingen) together with the following rat anti-mouse antibodies: CD31 (MEC13.3; BD Pharmingen), CD11b (M1/70; BD Pharmingen), and CD13 (R3-63; AbD Serotec, Raleigh, NC) for 1 hour or overnight at 4 °C. The normal rabbit and normal rat sera served as a staining control. Antibodies were diluted in blocking buffer (2% normal goat serum, 0.1% Tween, 1% bovine serum albumin in PBS 1:100). Slides were rinsed three times with PBS-Tween (0.1%).

For staining with secondary antibodies, we added Alexa Fluor 488 goat anti-rabbit IgG (A11008; Molecular Probes) or Cy5 goat anti-mouse IgG (Chemicon, Billerica, MA) and Alexa Fluor 546 goat anti-rat IgG (A11081; Molecular Probes) for 1 hour in the dark. Antibodies were diluted in blocking buffer (1:1,000). After washing, slides were mounted with DAPI medium (vector H-1500). Images were acquired with Nikon Eclipse E800 fluorescence microscope (Nikon, Melville, NY) and analyzed by MetaMorph software (version 6.3r5; Molecular Devices, Downingtown, PA).

Statistical analysis. Tumor growth was analyzed by Student's t-test.

Acknowledgments

We are grateful to the following funding agencies that have supported this work: the DOD (grants 17999457 and BC022063), the California Breast Cancer Research Program (CB CRP 12IB-0159), the NIH grant PPG (CA 104898), and the Breast Cancer Research Foundation. We thank Gosia Czarny for confocal microscopy assistance and Carina Rumold for FACS analysis. This work was supported also by grants from Richard and Kaye Woltman, the George and Barbara Bush Leukemia Research Fund, the Anthony Dewitt Frost Melanoma Fund, and the Sidney Kimmel Foundation. The authors declared no conflict of interest. The views expressed are the result of independent work and do not represent the views or findings of the US FDA or the US government.

REFERENCES

  1. He TC, Zhou S, da Costa LT, Yu J, Kinzler KW., and , Vogelstein B. A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci USA. 1998;95:2509–2514. doi: 10.1073/pnas.95.5.2509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ramnaraine M, Pan W, Goblirsch M, Lynch C, Lewis V, Orchard P, et al. Direct and bystander killing of sarcomas by novel cytosine deaminase fusion gene. Cancer Res. 2003;63:6847–6854. [PubMed] [Google Scholar]
  3. Kievit E, Bershad E, Ng E, Sethna P, Dev I, Lawrence TS, et al. Superiority of yeast over bacterial cytosine deaminase for enzyme/prodrug gene therapy in colon cancer xenografts. Cancer Res. 1999;59:1417–1421. [PubMed] [Google Scholar]
  4. Kievit E, Nyati MK, Ng E, Stegman LD, Parsels J, Ross BD, et al. Yeast cytosine deaminase improves radiosensitization and bystander effect by 5-fluorocytosine of human colorectal cancer xenografts. Cancer Res. 2000;60:6649–6655. [PubMed] [Google Scholar]
  5. Mahan SD, Ireton GC, Knoeber C, Stoddard BL., and , Black ME. Random mutagenesis and selection of Escherichia coli cytosine deaminase for cancer gene therapy. Protein Eng Des Sel. 2004;17:625–633. doi: 10.1093/protein/gzh074. [DOI] [PubMed] [Google Scholar]
  6. Dong Y, Wen P, Manome Y, Parr M, Hirshowitz A, Chen L, et al. In vivo replication-deficient adenovirus vector-mediated transduction of the cytosine deaminase gene sensitizes glioma cells to 5-fluorocytosine. Hum Gene Ther. 1996;7:713–720. doi: 10.1089/hum.1996.7.6-713. [DOI] [PubMed] [Google Scholar]
  7. Liu Y, Ye T, Maynard J, Akbulut H., and , Deisseroth A. Engineering conditionally replication-competent adenoviral vectors carrying the cytosine deaminase gene increases the infectivity and therapeutic effect for breast cancer gene therapy. Cancer Gene Ther. 2006;13:346–356. doi: 10.1038/sj.cgt.7700906. [DOI] [PubMed] [Google Scholar]
  8. Martin V, Liu D, Fueyo J., and , Gomez-Manzano C. Tie2: a journey from normal angiogenesis to cancer and beyond. Histol Histopathol. 2008;23:773–780. doi: 10.14670/HH-23.773. [DOI] [PubMed] [Google Scholar]
  9. Schlaeger TM, Bartunkova S, Lawitts JA, Teichmann G, Risau W, Deutsch U, et al. Uniform vascular-endothelial-cell-specific gene expression in both embryonic and adult transgenic mice. Proc Natl Acad Sci USA. 1997;94:3058–3063. doi: 10.1073/pnas.94.7.3058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Peters KG, Coogan A, Berry D, Marks J, Iglehart JD, Kontos CD, et al. Expression of Tie2/Tek in breast tumour vasculature provides a new marker for evaluation of tumour angiogenesis. Br J Cancer. 1998;77:51–56. doi: 10.1038/bjc.1998.8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Tanaka S, Sugimachi K, Yamashita Yi Y, Ohga T, Shirabe K, Shimada M, et al. Tie2 vascular endothelial receptor expression and function in hepatocellular carcinoma. Hepatology. 2002;35:861–867. doi: 10.1053/jhep.2002.32535. [DOI] [PubMed] [Google Scholar]
  12. Caine GJ, Blann AD, Stonelake PS, Ryan P., and , Lip GY. Plasma angiopoietin-1, angiopoietin-2 and Tie-2 in breast and prostate cancer: a comparison with VEGF and Flt-1. Eur J Clin Invest. 2003;33:883–890. doi: 10.1046/j.1365-2362.2003.01243.x. [DOI] [PubMed] [Google Scholar]
  13. Yu JL, Rak JW, Carmeliet P, Nagy A, Kerbel RS., and , Coomber BL. Heterogeneous vascular dependence of tumor cell populations. Am J Pathol. 2001;158:1325–1334. doi: 10.1016/S0002-9440(10)64083-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Brown LF, Dezube BJ, Tognazzi K, Dvorak HF., and , Yancopoulos GD. Expression of Tie1, Tie2, and angiopoietins 1, 2, and 4 in Kaposi's sarcoma and cutaneous angiosarcoma. Am J Pathol. 2000;156:2179–2183. doi: 10.1016/S0002-9440(10)65088-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Zadeh G, Qian B, Okhowat A, Sabha N, Kontos CD., and , Guha A. Targeting the Tie2/Tek receptor in astrocytomas. Am J Pathol. 2004;164:467–476. doi: 10.1016/S0002-9440(10)63137-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Puri MC., and , Bernstein A. Requirement for the TIE family of receptor tyrosine kinases in adult but not fetal hematopoiesis. Proc Natl Acad Sci USA. 2003;100:12753–12758. doi: 10.1073/pnas.2133552100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Arai F, Hirao A, Ohmura M, Sato H, Matsuoka S, Takubo K, et al. Tie2/angiopoietin-1 signaling regulates hematopoietic stem cell quiescence in the bone marrow niche. Cell. 2004;118:149–161. doi: 10.1016/j.cell.2004.07.004. [DOI] [PubMed] [Google Scholar]
  18. Müller A, Lange K, Gaiser T, Hofmann M, Bartels H, Feller AC, et al. Expression of angiopoietin-1 and its receptor TEK in hematopoietic cells from patients with myeloid leukemia. Leuk Res. 2002;26:163–168. doi: 10.1016/s0145-2126(01)00110-2. [DOI] [PubMed] [Google Scholar]
  19. Schliemann C, Bieker R, Padro T, Kessler T, Hintelmann H, Buchner T, et al. Expression of angiopoietins and their receptor Tie2 in the bone marrow of patients with acute myeloid leukemia. Haematologica. 2006;91:1203–1211. [PubMed] [Google Scholar]
  20. Wang J, Wu K, Zhang D, Tang H, Xie H, Hong L, et al. Expressions and clinical significances of angiopoietin-1, -2 and Tie2 in human gastric cancer. Biochem Biophys Res Commun. 2005;337:386–393. doi: 10.1016/j.bbrc.2005.09.051. [DOI] [PubMed] [Google Scholar]
  21. Liu Y., and , Deisseroth A. Tumor vascular targeting therapy with viral vectors. Blood. 2006;107:3027–3033. doi: 10.1182/blood-2005-10-4114. [DOI] [PubMed] [Google Scholar]
  22. Ferrara N., and , Kerbel RS. Angiogenesis as a therapeutic target. Nature. 2005;438:967–974. doi: 10.1038/nature04483. [DOI] [PubMed] [Google Scholar]
  23. De Palma M, Venneri MA, Galli R, Sergi Sergi L, Politi LS, Sampaolesi M, et al. Tie2 identifies a hematopoietic lineage of proangiogenic monocytes required for tumor vessel formation and a mesenchymal population of pericyte progenitors. Cancer Cell. 2005;8:211–226. doi: 10.1016/j.ccr.2005.08.002. [DOI] [PubMed] [Google Scholar]
  24. De Palma M, Mazzieri R, Politi LS, Pucci F, Zonari E, Sitia G, et al. Tumor-targeted interferon-alpha delivery by Tie2-expressing monocytes inhibits tumor growth and metastasis. Cancer Cell. 2008;14:299–311. doi: 10.1016/j.ccr.2008.09.004. [DOI] [PubMed] [Google Scholar]
  25. Hood JD, Bednarski M, Frausto R, Guccione S, Reisfeld RA, Xiang R, et al. Tumor regression by targeted gene delivery to the neovasculature. Science. 2002;296:2404–2407. doi: 10.1126/science.1070200. [DOI] [PubMed] [Google Scholar]
  26. Nagel H, Maag S, Tassis A, Nestlé FO, Greber UF., and , Hemmi S. The alphavbeta5 integrin of hematopoietic and nonhematopoietic cells is a transduction receptor of RGD-4C fiber-modified adenoviruses. Gene Ther. 2003;10:1643–1653. doi: 10.1038/sj.gt.3302058. [DOI] [PubMed] [Google Scholar]
  27. Armstrong RD, Lewis M, Stern SG., and , Cadman EC. Acute effect of 5-fluorouracil on cytoplasmic and nuclear dihydrofolate reductase messenger RNA metabolism. J Biol Chem. 1986;261:7366–7371. [PubMed] [Google Scholar]
  28. Tang Y, Borgstrom P, Maynard J, Koziol J, Hu Z, Garen A, et al. Mapping of angiogenic markers for targeting of vectors to tumor vascular endothelial cells. Cancer Gene Ther. 2007;14:346–353. doi: 10.1038/sj.cgt.7701030. [DOI] [PubMed] [Google Scholar]
  29. Lyons M, Onion D, Green NK, Aslan K, Rajaratnam R, Bazan-Peregrino M, et al. Adenovirus type 5 interactions with human blood cells may compromise systemic delivery. Mol Ther. 2006;14:118–128. doi: 10.1016/j.ymthe.2006.01.003. [DOI] [PubMed] [Google Scholar]

Articles from Molecular Therapy are provided here courtesy of The American Society of Gene & Cell Therapy

RESOURCES