Our investigation establishes two key steps in the development of a therapeutic strategy for delivering tissue-specific gene therapy. First, we demonstrate that circulating BOECs target tumour tissue and augment vessel growth through incorporation into vessel endothelium. Results of real-time PCR and chromium-51 labelling of BOECs showed that these cells migrate to sites of active vascular growth, such as tumour, liver, and spleen ( and ). This observation is consistent with data reported by Jevremovic et al, who has also found that endothelial cell precursors migrate to the tumour vasculature. In addition, CD31-stained tumours from mice injected with BOECs showed an increase in tumour vasculature ( and ). Taken together, our results suggest that BOECs integrate into and stimulate tumour vasculature growth ().
Second, we demonstrate that inhibition of tumour growth in EBOEC-injected mice was attributable to the inhibitory effect of endostatin on tumour angiogenesis. We observed a 28% reduction in tumour size in mice receiving EBOEC injections (). In addition, the biological activity of the endostatin present in the supernatant of EBOECs was verified by its ability to inhibit specifically the proliferation and tube formation by HUVECs in vitro (). Both in vivo and in vitro results demonstrate that BOECs transfected with retrovirus containing the endostatin gene are capable of long-term secretion of endostatin.
Recent studies report conflicting results with regard to the extent of endothelial cell integration into the tumour vasculature. While our findings demonstrate that BOECs integrate into tumour vessels, the integration occurs at a very low level. This finding is consistent with recent reports, which show that bone marrow-derived endothelial precursor cells migrate to tumour vasculature at similarly low levels (De Palma et al, 2003
; Machein et al, 2003
; Droetto et al, 2004
; Larrivee et al, 2005
). However, these results contrast with a previous report demonstrating that 90% of blood vessels in B6RV2 tumours are composed of bone marrow-derived endothelial cells (Lyden et al, 2001
). These incongruous results might be explained by the existence of multiple factors directing endothelial cell migration, including initial tumour size, extent of vascularity, differences in detection time after cell injection, total number of injected cells, and differences in the microenvironment of marrow-derived cells vs
BOECs in vivo
. In addition, the composition of angiogenic factors varies widely between the microenvironment of a s.c. tumour model and orthotopic model (Yancopoulos et al, 2000
). A low level of endothelial cell integration into the tumour vasculature may pose a potential problem for the delivery of sufficient quantities of gene product to sites of tumour angiogenesis. We are currently exploring, however, whether our observation was not solely due to the decreased incorporation of human endothelial cells into a murine host.
Our findings provide a rationale for developing antiangiogenic BOECs as an approach to gene therapy-mediated cancer treatment. Other gene therapies utilise viral or non-viral delivery systems, and the main drawback of these strategies is the absence of long-term expression of therapeutic proteins due to the probable immune response by the host to foreign material (Harrington et al, 2002
). Early attempts to overcome these drawbacks involved delivery of genes to tumour sites via cell-based carriers. The use of T cells and macrophages has been extensively studied due to the homing properties of these immune cells. Recently, problems with sustained production of antiangiogenic proteins were overcome by adeno-associated virus-mediated intratumoural delivery (Ma et al, 2002b
) or systemic delivery through intramuscular injection of angiostatin for treatment of intracranial tumours (Ma et al, 2002a
) or endostatin for treatment of ovarian carcinoma (Subramanian et al, 2006
). Long-term survival of mice with intracranial human glioblastoma was also seen when the Sleeping Beauty transposon system was used for the transfer of a gene encoding soluble vascular endothelial growth factor receptor or a fusion gene for angiostatin–endostatin (Ohlfest et al, 2005
Our study demonstrates that BOECs can be engineered to produce antiangiogenic proteins in vivo on a continuous basis without the need for daily administration of recombinant protein. Two potential advantages of using BOECs as a delivery system in gene therapy include the ability to grow autologous BOECs from peripheral blood and the ease of manipulating them to express any gene of interest. A possible drawback of using BOEC for antiangiogenic gene therapy would be the potential for increased tumour vessel outgrowth and increased tumour growth, if silencing of anticancer genes in therapeutic BOECs were to occur in vivo.
Based on our findings with endostatin-transfected BOECs, we propose BOECs as an appropriate delivery vehicle for novel antiangiogenic proteins. This therapeutic strategy also has implications for the specific targeting of other classes of anticancer agents to the tumour microenvironment using autologous BOECs.