Edgardo Marecos

Imaging of tumour neovasculature by targeting the TGF-β binding receptor endoglin

In vivo imaging of endothelial markers in intact tumour neovasculature would have applications in assessing the efficacy of anti-angiogenic agents in clinical trials. Although a variety of different endothelial markers have been described, few have been evaluated as imaging markers. The transforming growth factor-beta (TGF-beta) binding receptor endoglin is a proliferation-associated endothelial marker. We hypothesised that endoglin would be an ideal target for imaging since it is strongly upregulated in proliferating endothelial cells of the tumour neovasculature. We used a radiolabelled monoclonal anti-endoglin antibody and compared its neovascular binding, accumulation and in vivo behaviour to an isotype-matched control IgG(2a). Our data show that the probe binds specifically and rapidly within minutes in vivo and that correlative autoradiography and immunohistology support the in vivo imaging findings. Imaging of abundantly expressed endothelial targets circumvents delivery barriers normally associated with other tumour targeting strategies, and can potentially be used to quantitate molecular angiogenic markers.

Non-invasive in vivo mapping of tumour vascular and interstitial volume fractions

Non-invasive measurement of haemodynamic parameters and imaging of neovasculature architecture is of importance in determining tumour prognosis, in directing tissue sampling and in assessing treatment efficacy. In the current research we investigated a dual tracer nuclear magnetic resonance (NMR) technique to map the tumour vascular (VVF) and interstitial volume fraction (IVF) non-invasively in vivo. We hypothesised that a NMR signal emanating after intravenous administrations of a vascular paramagnetic probe (MPEG-PL-GdDTPA) can be maximised so that additional signal after administration of a second interstitial probe (GdDTPA) would only reflect the IVF but not the VVF. The method and its assumptions were verified and experimental conditions optimised both in phantoms and in C6 glioma bearing rats. Data derived from in vivo studies show tumoral VVF and IVF values that are consistent with histology data and literature values; the relative ranking order of values was tumour > muscle > brain. Image maps showed intratumoral and intertumoral heterogeneity of both parameters at submillimetre pixel resolution. The method is applicable to a wide variety of tumour models and can theoretically be performed repeatedly to study tumour growth or involution during therapy.

In vivo assessment of vascular endothelial growth factor-induced angiogenesis.

To determine whether vascular endothelial growth factor (VEGF)‐induced tumor microvascularity is detectable by in vivo NMR imaging, an experimental study was conducted in nude mice. Human breast cancer cells (MCF‐7) and MCF‐7 cells stably transfected with the cDNA for the VEGF165 isoform (MV165) were grown in nude mice and models were characterized by RT‐PCR, Western blotting, ELISA, immunohistochemistry and NMR imaging using a novel synthetic protected graft copolymer (PGC) as a vascular probe. MV165 tumors showed a 1.6‐fold higher microvascular density by histology. Both tumors showed identical MR signal intensities on non‐contrast and Gd‐DTPA enhanced images. PGC enhanced MR imaging of tumoral vascular volume fraction (VVF), however, revealed significant differences between the 2 tumor types (MV165: 8.9 ± 2.1; MCF‐7: 1.7 ± 0.5; p < 0.003), as expected from histology. VVF changes were more heterogeneous in the MV165 model both among tumors as well as within tumors as determined 3‐dimensionally at submillimeter resolutions. Our results have potential applications for non‐invasive assessment of angiogenesis by in vivo imaging and for clinical monitoring during angiogenic therapies. Int. J. Cancer 83:798–802, 1999.

Mechanism of gadophrin-2 accumulation in tumor necrosis.

The molecular mechanism by which gadophrin‐2 targets necrotic tumor tissue was investigated. Biodistribution studies and magnetic resonance imaging (MRI) and histologic/autoradiographic correlation were performed in xenograft mouse models bearing human tumors (HT 29, WiDr, LX 1). Binding of gadophrin‐2 to DNA, lipids, or proteins was determined by fluorescence spectrophotometry. Protein binding was determined by dialysis and gel electrophoresis. Accumulation of gadophrin‐2 was low (<0.7% injected dose/g tissue at 24 hours after injection) in viable tumor but higher in necrotic tumor regions and was readily detectable by MRI. Within a given tumor, the agent preferentially localized in the periphery of necrotic areas. Within these regions gadophrin‐2 was bound to interstitial albumin and not other proteins, lipids, or DNA. Tumoral accumulation of gadophrin‐2 most likely occurs through its binding to plasma albumin and subsequent slow extravasation into the tumor interstitium.J. Magn. Reson. Imaging 1999;9:336–341.

In Vivo localization of diglycylcysteine-bearing synthetic peptides by nuclear imaging of oxotechnetate transchelation

A phenomenon of in vivo transchelation of oxotechnetate from a complex with glucoheptonic acid to synthetic peptides bearing oxotechnetate-binding motifs and a technique for in vivo visualization of these peptides are described. Using two model peptides bearing two tandem diglycylcysteine (GGC) motifs (P1) or three GGC motifs (P2), we demonstrated that: (i) these peptides efficiently transchelated oxo-[99mTc]technetate from a complex with glucoheptonic acid in vitro (a complex with peptides was stable at least 24 h; radiochemical purity exceeded 95% by high performance liquid chromatography); (ii) injection of peptides into the rectus femoris muscle (at 0.5-1 micromol of SH groups) followed by an intravenous injection of 99mTc-glucoheptonate (0.25-0.5 mCi per animal) yielded visualization of the injected muscle by nuclear imaging within 1 h after injection; (iii) the experimental/control (contralateral) thigh muscle ratio was 1.80 +/- 0.05 for peptide P1 and 3.0 +/- 0.1 for P2; (iv) the injection of a control peptide P2 with SH groups covalently modified with N-ethylmaleimide resulted in a ratio of 1.4 +/- 0.2. These findings argue for specific association of oxo-[99mTc]technetate with free thiols within the binding motif of injected peptides in vivo. In vivo transchelation of oxo-[99mTc]technetate may be useful for the purpose of noninvasive imaging of gene expression, i.e., when the expression product bears GGC motifs.