Antiangiogenic cancer therapy: monitoring with molecular US and a clinically translatable contrast agent (BR55)

Abstract

Purpose: To develop and test human kinase insert domain receptor (KDR)-targeted microbubbles (MBs) (MB(KDR)) for imaging KDR at the molecular level and for monitoring antiangiogenic therapy in a human colon cancer xenograft tumor model in mice.

Materials and methods: Animal studies were approved by the Institutional Administrative Panel on Laboratory Animal Care. A heterodimeric peptide that binds to human KDR with low nanomolar affinity (K(D) = 0.5 nmol/L) was coupled onto the surface of perfluorobutane-containing lipid-shelled MBs (MB(KDR)). Binding specificity of MB(KDR) to human KDR and cross-reactivity with murine vascular endothelial growth factor (VEGF) receptor 2 (VEGFR2) were tested in cell culture under flow shear stress conditions (at 100 sec(-1)). In vivo binding specificity of MB(KDR) to VEGFR2 was tested in human LS174T colon cancer xenografts in mice with a 40-MHz ultrasonographic (US) transducer. Targeted contrast material-enhanced US imaging signal by using MB(KDR) was longitudinally measured during 6 days in tumors with (n = 6) and without (n = 6) antiangiogenic treatment (anti-VEGF antibody). Ex vivo VEGFR2 staining and microvessel density analysis were performed. Significant differences were evaluated (t, Mann-Whitney, or Wilcoxon test).

Results: Cell culture experiments showed four times greater binding specificity of MB(KDR) to human KDR and cross-reactivity to murine VEGFR2 (P < or = .01). In vivo imaging signal was more than three times higher (P = .01) with MB(KDR) compared with control MBs and decreased significantly (approximately fourfold lower, P = .03) following in vivo receptor blocking with anti-VEGFR2 antibody. One day after initiation of antiangiogenic therapy, imaging signal was significantly decreased (approximately 46% lower, P = .02) in treated versus untreated tumors; it remained significantly lower (range, 46%-84% decreased; P = .038) during the following 5 days. Microvessel density was significantly reduced (P = .04) in treated (mean, 7.3 microvessels per square millimeter +/- 4.7 [standard deviation]) versus untreated tumors (mean, 22.0 microvessels per square millimeter +/- 9.4); VEGFR2 expression was significantly decreased (>50% lower, P = .03) in treated tumors.

Conclusion: Human MB(KDR) allow in vivo imaging and longitudinal monitoring of VEGFR2 expression in human colon cancer xenografts.

Figure 1a:

Cell culture binding assay of MBs to human KDR and mouse VEGFR2 by using a parallel plate flow chamber. (a) Schematic diagram of flow chamber apparatus setup used to flow different types of MBs over cells grown on a microscope slide at a wall shear rate of 100 sec⁻¹. (b) Phase-contrast bright-field micrographs (scale bar = 20 µm) show binding of human MB_KDR (KDR-MB) and negative control MB_NT (NT-MB) seen as white spheres (arrows) to KDR- and/or VEGFR2-positive and KDR- and/or VEGFR2-negative cells. Binding of MB_KDR was substantially higher to both KDR- and VEGFR2-expressing cells compared with negative control cells and could be substantially blocked by incubation of cells beforehand with an anti-KDR and anti-VEGFR2 antibody. (c) Box and whisker plot of attached MBs per cell number counted from micrographs (as in b). Whiskers are standard deviations. SVR and 4T1 are cell lines. Means with standard deviations are listed in Table 1.

Figure 1b:

Cell culture binding assay of MBs to human KDR and mouse VEGFR2 by using a parallel plate flow chamber. (a) Schematic diagram of flow chamber apparatus setup used to flow different types of MBs over cells grown on a microscope slide at a wall shear rate of 100 sec⁻¹. (b) Phase-contrast bright-field micrographs (scale bar = 20 µm) show binding of human MB_KDR (KDR-MB) and negative control MB_NT (NT-MB) seen as white spheres (arrows) to KDR- and/or VEGFR2-positive and KDR- and/or VEGFR2-negative cells. Binding of MB_KDR was substantially higher to both KDR- and VEGFR2-expressing cells compared with negative control cells and could be substantially blocked by incubation of cells beforehand with an anti-KDR and anti-VEGFR2 antibody. (c) Box and whisker plot of attached MBs per cell number counted from micrographs (as in b). Whiskers are standard deviations. SVR and 4T1 are cell lines. Means with standard deviations are listed in Table 1.

Figure 1c:

Cell culture binding assay of MBs to human KDR and mouse VEGFR2 by using a parallel plate flow chamber. (a) Schematic diagram of flow chamber apparatus setup used to flow different types of MBs over cells grown on a microscope slide at a wall shear rate of 100 sec⁻¹. (b) Phase-contrast bright-field micrographs (scale bar = 20 µm) show binding of human MB_KDR (KDR-MB) and negative control MB_NT (NT-MB) seen as white spheres (arrows) to KDR- and/or VEGFR2-positive and KDR- and/or VEGFR2-negative cells. Binding of MB_KDR was substantially higher to both KDR- and VEGFR2-expressing cells compared with negative control cells and could be substantially blocked by incubation of cells beforehand with an anti-KDR and anti-VEGFR2 antibody. (c) Box and whisker plot of attached MBs per cell number counted from micrographs (as in b). Whiskers are standard deviations. SVR and 4T1 are cell lines. Means with standard deviations are listed in Table 1.

Figure 2:

Timeline of both longitudinal US scanning and antiangiogenic treatment schedule. Human colon cancer LS174T cells were injected subcutaneously in the right flank region of nude mice, and tumors were imaged 7 days later to obtain baseline molecular US imaging. Antiangiogenic treatment (anti-VEGF antibody [B20 treatment]) was initiated in a group of six tumor-bearing mice, whereas another group of six tumor-bearing mice remained untreated. Anti-VEGF therapy continued every day, and a final US scan was obtained at 144 hours, after which tumors were excised for ex vivo analysis.

Figure 3a:

Transverse US images of a subcutaneous human colon cancer xenograft (yellow arrows) imaged following intravenous administration of (a) control MB_NT and (b) MB_KDR in the same mouse. Targeted US imaging signal, shown as green signal (white arrows) overlaid on gray-scale brightness-mode image, was substantially higher with MB_KDR than with control MBs.

Figure 3b:

Transverse US images of a subcutaneous human colon cancer xenograft (yellow arrows) imaged following intravenous administration of (a) control MB_NT and (b) MB_KDR in the same mouse. Targeted US imaging signal, shown as green signal (white arrows) overlaid on gray-scale brightness-mode image, was substantially higher with MB_KDR than with control MBs.

Figure 4a:

Longitudinal study monitoring antiangiogenic treatment with molecular US and MB_KDR. (a) Graph of tumor volumes measured over time in untreated and B20-treated mice. (b) Box and whisker plot of MB_KDR targeted US signal (video intensity normalized to that at baseline at 0 hours) measured in tumors of untreated and B20-treated mice. Whiskers are standard deviations. Average signals are listed in Table 2. (c) Representative micrographs (scale bar = 100 µm) of tumor tissue slices from untreated tumor (top row) and treated tumor (bottom row). Slices were immunofluorescently stained for CD31 (green, left images) and VEGFR2 (red, middle images) expression. Merged images (right images) demonstrate colocalization of VEGFR2 expression (yellow) on tumor endothelial cells.

Figure 4b:

Longitudinal study monitoring antiangiogenic treatment with molecular US and MB_KDR. (a) Graph of tumor volumes measured over time in untreated and B20-treated mice. (b) Box and whisker plot of MB_KDR targeted US signal (video intensity normalized to that at baseline at 0 hours) measured in tumors of untreated and B20-treated mice. Whiskers are standard deviations. Average signals are listed in Table 2. (c) Representative micrographs (scale bar = 100 µm) of tumor tissue slices from untreated tumor (top row) and treated tumor (bottom row). Slices were immunofluorescently stained for CD31 (green, left images) and VEGFR2 (red, middle images) expression. Merged images (right images) demonstrate colocalization of VEGFR2 expression (yellow) on tumor endothelial cells.

Figure 4c:

Longitudinal study monitoring antiangiogenic treatment with molecular US and MB_KDR. (a) Graph of tumor volumes measured over time in untreated and B20-treated mice. (b) Box and whisker plot of MB_KDR targeted US signal (video intensity normalized to that at baseline at 0 hours) measured in tumors of untreated and B20-treated mice. Whiskers are standard deviations. Average signals are listed in Table 2. (c) Representative micrographs (scale bar = 100 µm) of tumor tissue slices from untreated tumor (top row) and treated tumor (bottom row). Slices were immunofluorescently stained for CD31 (green, left images) and VEGFR2 (red, middle images) expression. Merged images (right images) demonstrate colocalization of VEGFR2 expression (yellow) on tumor endothelial cells.