In Vivo Imaging Of Human Dormant Tumors

Traditionally, various in vivo imaging techniques have been used for the detection and quantification of tumors implanted orthotopically or ectopically (i.e., outside their orthotopic site). However, some of these techniques can be employed for the in vivo detection of microscopic dormant tumors. By definition, nonangiogenic, dormant tumors are microscopic in size. Therefore, they are usually undetectable by palpation (limited to tumor sizes 50 mm3 and smaller) when located in the subcutaneous space or mammary fat pad. It is an even greater challenge to detect microscopic tumors located in internal organs. In the originally published dormancy model of osteosarcoma (MG-63), the presence of dormant tumors in a fraction of the inoculated mice was revealed through careful examination of the hair growth overlying the original tumor inoculation site.48 The inner side of the skin in the area associated with hair growth contained a microscopic white lesion, from which a histology section showed a viable tumor. Although this detection method clearly reveals this interesting phenomenon, it is terminal (i.e., the animal has to be euthanized) and does not provide longitudinal quantitative information about the tumor size.

Stable infection of tumor cells with fluorescent proteins (such as green fluorescent protein [GFP] and red fluorescent protein [RFP]) or luciferase allows for in vivo longitudinal detection of tumors even at a microscopic size. GFP-expressing tumor cells can be visualized noninva-sively from the skin surface by directed blue light (488 nm) epi-illumination. Submillimeter tumors can be localized using this method. The utility of fluorescence visualization of dormant tumors has been reported by Udagawa et al.,48 using osteosarcoma (MG-63 and SAOS-2) and gastric cancer (ST-2) dormancy models. Tumor-associated blood vessels appear dark against the background of a fluorescent tumor tissue, allowing for morphological (e.g., vessel diameters, tortuousity, branching) and even functional (e.g., red blood cell velocity) quantification of angiogenesis.49 More recently, this labeling technique was used to determine the minimum number of human tumor cells necessary to form a nonangiogenic, dormant microscopic tumor in mice (Naumov et al., unpublished). However, detecting fluorescently labeled tumors has its limitations. Microscopic tumors in internal organs can only be visualized ex vivo. Certain procedures, such as in vivo videomicroscopy, can be used for visualization of liver and lung metastases.50'51 However, in the brain, excitation or emission of fluorescently labeled tumor cells is not only limited by tissue depth, but also by light penetration through the skull.

Infection of tumor cells with the luciferase reporter gene allows for the reliable detection in mice of a signal from tumors that are < 1 mm in diameter (as verified by histology) in all internal organs, including the brain. Almog et al.43 has used the luciferase method for the detection of dormant human liposarcoma tumors in the renal fat pad of mice. The method can also be used to monitor the growth of microscopic human glioblastomas stereotactically after tumor cells are inoculated in the brains of mice (Naumov et al., unpublished work). Following intravenous injection of the luciferine substrate, the enzymatic activity of luciferase is rapid and transient. Only viable and metabolically active tumor cells can be detected by luminescence. The transient effect of the enzymatic reaction allows for real-time detection of tumor cells and for monitoring their viability during the dormancy period, as well as at times throughout the angiogenic switch. The persistent luciferase signal during the dormancy period of microscopic human tumors confirms the previous conclusion (based on histology) that dormancy does not result from tumor cell cycle arrest or eradication. Although the intensity of the luciferase signal directly correlates with the size of tumor, this imaging modality does not provide a clear tumor boundary or an anatomical outline of the tumor. However, small animal magnetic resonance imaging (MRI) provides a clear anatomical definition of a microscopic tumor, and it can be effectively used in combination with luciferase imaging (Naumov et al., unpublished work). Recent reports have demonstrated that single cancer cells can be detected in a mouse brain using MRI.52 Individual tumor cells trapped within the brain microcirculation were detected using MRI and validated using high-resolution confocal microscopy. Graham et al.53 demonstrated that three-dimensional, high-frequency ultrasound can quantitatively monitor the growth of liver micrometastases as small as 0.5 mm in diameter.

Collectively, these recent advances in animal imaging modalities enable, in most cases, noninvasive, real-time, longitudinal observations of single cancer cell trafficking and detection of nonangiogenic microscopic tumors in vivo during the dormancy period and as they switch to the angiogenic phenotype. Quantitative imaging of tumors throughout their progression to the angiogenic phenotype can be used for evaluating the efficacy of antiangio-genic therapy in primary and metastatic tumors.

0 0

Post a comment