Instrumentation and Probes for Molecular and Cellular Imaging

Aug 29, 03:49 AM

Current Headlines: By Lecchi, M Ottobrini, L; Martelli, C; Sole, A Del; Lucignani, G

Molecular and cellular imaging is a branch of biomedical sciences that combines the use of imaging instrumentation and biotechnology to characterize molecular and cellular processes in living organisms in normal and pathologic conditions. The two merging areas of research behind molecular and cellular imaging are detection technology, i.e. scanners and imaging devices, and development of tracers, contrast agents and reporter probes that make imaging with scanners and devices possible. Several in vivo imaging instruments currently used in human studies, such as computer tomography, ultrasound, magnetic resonance, positron emission tomography and single photon emission computed tomography, have been rescaled for small animal studies, while other methods initially used for in vitro evaluation, such as bioluminescence and fluorescence, have been refined for in vivo imaging. Conventional imaging relies on the use of non specific contrast agents and classical probes; however, newly developed targeted contrast agents and activable "smart" imaging probes for so-called "targeted imaging" have demonstrated high specificity and high signal to noise ratio in small animal studies. This review focuses on basic recent findings in the technical aspects of molecular and cellular imaging modalities (equipment, targeted probe and contrast agents and applied combinations of instrumentation and probe) with particular attention to the choice of the future: the multimodal imaging approach. KEY WORDS: Molecular imaging - Contrast agents - Emission tomography - Bioluminescence - Magnetic resonance imaging.

Molecular imaging can be broadly defined as the non-invasive repetitive imaging of targeted macromolecules and biological processes in living organisms. Cellular imaging, on the other hand, can be similarly defined as the non-invasive repetitive imaging of targeted cells and cellular processes in living subjects. Non- invasive imaging using different imaging modalities is being increasingly used role for the in vivo assessment of cellular and molecular processes.1

Clinical imaging technologies allow us to differentiate pathological from normal conditions. The enhanced specificity afforded by molecular and cellular imaging techniques has led to the development of novel procedures that permit an in-depth understanding of biological processes.2 This will results in an improvement of early detection and characterization of diseases, and treatment evaluation based on molecular and cellular processes assessment.

In the applied sciences, the advent of genetic engineering (production of specific animal models) has changed the ways new drugs are discovered and developed. Likewise, animal imaging procedures are being developed and exploited to provide new tools for preclinical studies.3

Molecular and cellular imaging techniques are primarily applied to directly assess specific processes in living organisms, including gene expression, proteinprotein interaction, dynamic cell tracking throughout the entire organism, and drug action analysis, thus contributing to our understanding of the physiology of living organisms and providing new means for drug target identification and preclinical testing to improve drug discovery. Molecular imaging protocols can achieve these goals non invasively, rapidly, quantitatively, and repetitively in the same animal, under different conditions and stimuli.

The technology behind molecular and cellular imaging is being pursued by merging 2 areas: detection technology, i.e. scanners and imaging devices, and the development of tracers, contrast agents and reporter probes that make imaging with scanners and other devices possible. Advances in engineering, computer sciences and physics drive the development of detection devices, while progress in biotechnology is vital for the evolution of tracers, contrast agents and probes.

Medical imaging began with two-dimensional (2D) X-ray images. In the early 1970s, the introduction of X-ray computed tomography (CT) provided threedimensional (3D) anatomical information for the first time. The advent of magnetic resonance (MR) imaging gave a major impetus to 3D visualization, leading to the development of a wide variety of applications in diagnostic medicine that are now being progressively used for research on small animals. Several in vivo imaging instruments currently used in human studies, including CT, ultrasound (US), MR, positron emission tomography (PET) and single photon emission computed tomography (SPECT) have been rescaled for studies in small animals, while other methods initially used for in vitro studies, such as nuclear magnetic resonance (NMR), bioluminescence (BL) and fluorescence (FL), have been refined for in vivo imaging.4- 5 For example, the distribution of fluorescent and bioluminescent sources can now be represented in 3D with dedicated tools.6

Molecular and cellular imaging techniques are often grossly divided into morphological and functional, depending on the type of information they deliver. This subdivision has somewhat uncertain boundaries, as some techniques provide different types of information in different proportions. Another reason for considering this division arbitrary is that upcoming generation instruments incorporate complementary technologies. PET/CT and PET/MR devices, for instance, can be used for morphological, functional and biochemical imaging studies. The combined use of different technologies is fundamental to detect and visualize molecular signals as a result of the overall imaging system's sensitivity, specificity and physical resolution.7

In addition to representing general morphology and function of an organ, molecular and cellular imaging techniques can identify within organs and systems the specific biochemical processes and cellular modifications that are characteristic of pathological conditions and indicative of disease progression and regression or drug action, making these changes primary imaging targets.

Both endogenous and exogenous molecules can be targeted in vivo to study a molecular process or the behavior of a specific cell population. Tailoring imaging strategies to an endogenous molecule is long and expensive. For this reason, exogenous molecules have been used to tag specific molecular processes or specific cell populations in order to generalize the imaging strategy to the marker of interest. For example, reporter genes are genetic markers that encode easily detectable proteins. Once located downstream from a specific promoter, these markers become extraordinary tools for reporting the activity of a specific promoter and the factors regulating its activity. While reporter genes have been widely utilized to study cell physiology, the recent generation of animals ubiquitously expressing a given reporter gene now allows direct in vivo analysis of gene expression and regulation.8

For many years imaging modalities relied on nonspecific contrast agents and probes. These are being increasingly replaced by targeted contrast agents and activable "smart" imaging probes endowed with high specificity and high signal to noise ratio for molecular and cellular "targeted imaging".2

The strategy of targeted imaging was initially based on the use of radionuclide imaging derived from research in radiopharmacology and radiopharmacy. Until recently, molecular imaging was mostly based on techniques such as PET, SPECT and optical imaging. For emission imaging the use of tracers is implicit in the methodology, thus the development of tracers for PET and SPECT for assessing innumerable variables is a continuous process. Optical imaging relies on reporter probes such as fluorescent proteins that require external excitation, or cellular enzymes that transform an exogenously added non-fluorescent substrate into its fluorescent derivative, or luciferases that emit light when provided with the appropriate substrate.9 To increase in vivo detection sensitivity, fluorescent proteins and luciferase mutant variants have been created to shift emission closer to red and infrared wavelengths.10

In other imaging areas such as MR imaging and US, targeted imaging has been pursued through the development of specific targeted contrast agents.11, 12 Selective ligands conjugated to MR contrast agents in molecular imaging studies induce differences in signal intensity between targeted and non-targeted tissues because of selective binding; in cellular imaging, specific cells can be labeled with MR contrast agents to make them stand out from the surrounding tissues. For such applications, superparamagnetic iron oxide (SPIO) particles currently appear to be the preferred agent.13 Similar concepts are applied to US imaging by using microbubbles as contrast agents.14

Molecular and cellular imaging techniques differ in complexity, degree of expertise required to interpret the images, cost of acquiring and building instrumentation and facilities, safety requirements and cost for performing studies.15

The main features of the molecular and cellular imaging techniques and applications are summarized in Table I and will be discussed below, focusing on recent advances in device development and biotechnology. Emission tomography imaging

Essentially functional imaging techniques, PET and SPECT do not provide any relevant morphologic information. PET and SPECT tomographs are devices that detect photons emitted from the subject under examination following the administration, generally by intravenous injection, of radioactively labeled tracers. After reaching the target organ, the tracers are retained there as a result of transport, binding to receptors and antigens, reaction with enzymes, biochemical or mechanical trapping, phagocytosis, or a combination of the above. PET and SPECT are both radionuclide imaging techniques, yet they differ substantially in the type of radionuclide employed to label the tracer and method of data acquisition.16

PET is based on the use of molecular probes labeled with a radionuclide that decays by positron emission, i.e., the ejection from the nucleus of a particle that has the same mass as an electron but the opposite charge. Proton-rich nuclei, such as ^sup 18^F, ^sup 11^C, ^sup 15^O, ^sup 64^Cu and ^sup 124^I, have very short half- lives ranging from about 4 days for ^sup 124^I to about 2 s for ^sup 15^O. The emitted positrons pass through the tissues and interact with the electrons and nuclei of nearby atoms. The positron range (about 0.6 mm for ^sup 18^F) poses a lower limit to spatial resolution of PET. Near the end of its trajectory, positron velocity is low and may combine with an electron. The pair annihilates and their mass is converted into high energy photons, each with an energy of 511 keV. The two photons are emitted simultaneously about 180[degrees] apart. In a PET acquisition event, both the 511 keV coincident photons emitted within the detection system are detected.17

Clinical PET scanners have a spatial resolution of about 5 mm, which is not very amenable for small animal preclinical imaging. However, the development and commercial availability of small animal PET tomographs with about 1.5 mm spatial resolution has made this technology attractive to biologists investigating a variety of animal models of disease.18 Despite its low spatial resolution, small animal PET imaging has a relatively high molecular sensitivity (10^sup -11^-10^sup -12^ mol/L) and it is independent of the location depth of the emission source.

The most widely used PET probe is the glucose analogue 2-[^sup 18^F]fluoro-2deoxy-D-glucose ([^sup 18^F]FDG), which is transported into the cells by glucose transporters and then phosphorylated to FDG-6-P by hexokinase.19 New probes and PET imaging tracers, such as [^sup 18^F]fluorothymidine ([^sup 18^FJFLT), [^sup 18^F]-annexin V, [^sup 18^F]fluoroazomycin arabinoside ([^sup 18^FJFAZA), [^sup 11^C]methionine ([^sup 11^C]MET), [^sup 18^F]fluoro-ethyl-spiperone ([^sup 18^F]FESP) and 9-([^sup 18^F-Fluoro-3-[hydroxymethyl] butyl)guanine ([^sup 18^F]FHBG), have been developed to investigate cellular proliferation and apoptosis, tissue oxygenation, receptor- ligand interactions and enzyme activity. More recently, they have been applied in animal model studies to assess protein-protein interactions, gene expression, adoptive cell therapy and somatic gene therapy.20,21

SPECT is based on the use of radiotracers labeled with a radionuclide, such as ^sup 99m^Tc, ^sup 123^I, ^sup 125^I, ^sup 111^In and ^sup 201^Tl, that decays with the emission of single photons with different energy (e.g., ^sup 111^In has two photons of 171 and 245 keV). To locate the source of an emitted photon, its direction of incidence into the detection system needs to be accurately fixed and its interaction location determined in the position sensitive detector. Lead or tungsten collimators are perforated plates that define the angle of photon incidence and are therefore positioned in front of the detector during SPECT acquisition.22 Collimator design is a compromise between spatial resolution and sensitivity: high resolution (200 [mu]m) SPECT images are now possible with micro-pinhole apertures.23

TABLE I.-Main non-invasive in vivo imaging modalities used in molecular and cellular imaging studies.

TABLE I.-Main non-invasive in vivo imaging modalities used in molecular and cellular imaging studies.

SPECT is the most widely used nuclear medicine technique in clinical practice. However, most clinically useful SPECT tracers are of limited interest for molecular and cellular imaging research applications. Instead, positron-emitting isotopes can usually be readily substituted for naturally occurring atoms, rendering PET a more robust technique than SPECT for imaging most molecular events. Leyton et al.24 showed the potential of [^sup 18^F]FLT to measure early cytostasis and cytotoxicity induced by cisplatin treatment of radiationinduced fibrosarcoma 1 tumor-bearing mice by demonstrating that [^sup 18^F]FLT was superior to [^sup 18^F]FDG-PET in the quantitative measurement of tumor cell proliferation. Solomon et al.25 demonstrated the effect of Gefitinib (Iressa) on modulating oxygenation in human epidermal growth factor receptor expressing A431 squamous cell carcinoma xenografts using in vivo small animal PET imaging with [^sup 18^F]FAZA. Yaqle et al.26 used [^sup 18^F]- annexin V to study cycloheximide-induced liver apoptosis in rats. Their PET data demonstrated that [^sup 18^F]-annexin V binds specifically to apoptotic tissues in this model of chemically induced apoptosis in the rat liver. The short physical half-life of [^sup 18^F]-annexin V and the rapid clearance of its metabolites via the urinary system suggest that [^sup 18^F]-annexin V may be useful in the early assessment of clinical response to cancer therapy in individual patients.

PET has been widely used to assess reporter gene expression in small animals.27-29 Dubey et al.,30 using Herpes simplex virus I- thymidine kinase mutant sr39 (HSV11-sr39tk) as a reporter gene in adoptively transferred lymphocytes, showed that antitumor responses due to adoptively transferred immune T lymphocytes could be quantified using PET in the entire animal. Su et al.31 reported for the first time the ability to estimate the number of cells from a region of interest by correlating PET signal intensity to cell number. Primary T lymphocytes were retrovirally transduced with the HSVI-sr39tk gene and detected by PET scanning following accumulation of [^sup 18^F]FHBG after intratumoral injection of different amounts of these cells. Xiong et al.32 used the HSVI-sr39TK reporter gene to study the capacity of a knob-modified adenoviral vec tor (modified- Adtk) to specifically transduce tumor cells in mice. PET imaging showed minimal non-specific uptake of [^sup 18^F]FHBG in the liver and tumors devoid of HSVI-sr39TK, whereas significantly higher tumor accumulation of [^sup 18^F]FHBG was visible in tumor expressing the reporter compared to the concurrent lower liver uptake. The important advantages of PET and SPECT over other molecular and cellular imaging techniques are that they allow to generate quantitative data and that the small probe mass and labeling strategies do not significantly perturb the biological processes under study. PET has a higher molecular sensitivity and a strong quantitative potential, whereas SPECT allows us to image multiple probes simultaneously,provided each is labeled with a different radionuclide.

Optical imaging

Optical imaging technologies analyse the propagation of non- ionizing radiation, light photons, through a medium such as tissue. Compared with other modalities, optical imaging techniques are relatively simple, do not entail the use of hazardous radiation and require very simple detecting devices. Furthermore, they can be used repeatedly to monitor rapidly cycling processes. For these reasons, optical imaging is the most extensively used functional technique in molecular and cellular imaging to detect protein-protein interactions, monitor the transcriptional and post-transcriptional regulation of specific genes, track and monitor in vivo cells and gene expression in transgenic animals.33

The emitted light is detected with a charge-coupled device camera. Low read-out noise images require prolonged acquisition times. However, very long acquisition increases thermal noise, which, in turn, can be reduced with a thermo-electric cooler.34 Images can be either planar (2D) or tomographic (3D), but this latter modality is still in its infancy.35, 36

Optical imaging is based on two different physical processes: BL and FL. BL requires administering a substrate that interacts with a reporter enzyme, while in FL the detection of exogenously administered fluorescent molecules does not necessarily require their interaction with a component of the organism under examination to generate the light signal, although it does need an excitated light to activate photon emission from the fluorophores. BL photons are weaker and produce images with lower spatial and temporal resolution than FL photons. However, bioluminescent probes are generally more sensitive to biological processes and less toxic. This latter feature makes them suitable for repeated and long-term whole-body studies.

BL exploits the emission of visible photons at specific wavelengths on the basis of energy-dependent reactions catalyzed by enzymes such as luciferases, a family of photoproteins that emit detectable photons in the presence of oxygen, ATP and Mg^sub 2^+ during the metabolism of such substrates (luciferin into oxyluciferin). Luciferase systems include, among others, the bacterial lux genes of terrestrial Photorhabdus luminescens and marine Vibrio harveyi bacteria, as well as eukaryotic luciferase luc and rluc genes from the firefly (Photinus pyralis) and the sea pansy (Renilla reniformis), respectively.37

FL exploits the emission of visible photons at specific wavelengths based on the return to the ground state of electrons excited by laser photostimulation. Substances having this property are defined as fluorophores.38 FL imaging generally requires the use of filters to select excitation and emission wavelengths and to reduce background noise due to autofluorescence and free non- specifically bound probes. Examples of proteic fluorophores include green fluorescent protein (GFP), red fluorescent protein and rhodamine. Besides these proteins, other fluorescent dyes can be used, including indiocyanine and indiocyanine green. The wavelength range of an emitted BL photon is between 485 nm and 613 nm, while that of fluorescent photons from dyes is between 442 nm and 800 nm. Only photons with a wavelength higher than 600 nm, i.e., close to or in the red spectrum, can penetrate tissues efficiently, whereas those emitted in the green spectrum or below are strongly attenuated in tissues and so are only poorly detected in vivo. The light intensity emerging from within the body depends on photon wavelength, source depth, shape and brightness, optical properties of the tissue. For example, in vivo regional differences in detection are largely due to the presence of hemoglobin that absorbs primarily in the blue-green region of the spectrum, and melanin in pigmented animals.39

The two optical imaging techniques have produced promising results for non-invasive repetitive imaging of targeted macromolecules and cellular processes in living organisms. Cao et al.40 have recently reported the use of luciferase BL imaging of hematopoietic stem cells following transplantation into irradiated mice. Donor stem cells were derived from either a luciferase or a luciferase/GFP transgenic mouse bred to express the reporter gene(s) in every cell line. Both luciferase and GFP reporter genes were used also by Schimmelpfennig et al.41 to track expanded dendritic cells in vivo trafficking and survival. At the University of Milan, Ciana et al. developed a mouse model (ERE-Luc) for the in vivo study of estrogen receptor activity by optical imaging using Luciferase gene expression controlled by an estrogen responsive promoter.42 Several studies43-45 demonstrated that in the ERE-Luc mouse luciferase expression reflects the state of ER transcriptional activity and that it can be visualized by optical imaging directly in vivo. The ERE-Luc mouse is a paradigmatic model because it allows assessment of specific receptor activity in all tissues in living animals, thus opening the way to the development of innovative protocols for physiological, pharmacological and toxicological studies. A description of the many applications of these pharmacological protocols is beyond the scope of this article.

Whole-body bioluminescent reporter imaging has also been used in oncological applications. Van der Pluijm et al.46 studied the detection, monitoring, and quantification in vivo of the progression of bone metastases induced by intracardiac or intraosseous injection of luciferase-transfected breast cancer cells (MDA-231-B/luc+) into nude mice.

FL imaging has also been used to study in vivo processes related to tumor growth, progression and therapy. Citrin et al.47 demonstrated the possibility to use near-infrared imaging to follow in vivo the distribution and tumor-specific localization of endostatin conjugated to Cy5.5 monofunctional dye.

Ray et al. described for the first time a method for the in vivo study of protein-protein interaction in living mice by BL imaging. This approach may have important implications for the study of protein-protein interactions and for the in vivo evaluation of new pharmaceuticals targeted to modulate protein-protein interactions.48

An emerging technique in optical imaging is nearinfrared fluorescence (NIRF). NIRF photons (wavelength ranges from 650 to 900 nm) have a high penetration due to minimal tissue absorbance and minimal autofluorescence, but higher scattering than fluorophores with lower wavelengths.

Besides organic fluorophores, fluorescent semiconductor nanocrystals (known as Quantum dots or Qdots) are also being considered for molecular and cellular imaging. Qdots are characterized by a narrow tunable emission, high resistance to degradation, as well as a high photobleaching threshold. Their emission wavelength can range from 400 to 1350 nm, and their size can vary from 1 to 10 nm. These crystals can be covalently bound to various types of biomolecules, including peptides, antibodies, nucleic acids and employed as fluorescent probes. An indepth review of the use of Qdots for in vivo imaging can be found elsewhere.49

As seen, optical imaging techniques are ideal for small animal studies, but their usefulness in human subjects is limited to studies of superficial lesions: optical photons are too weak to penetrate thick tissue layers (especially blood-rich tissues), making an optical analogue of X-ray CT, emission tomography or MR imaging difficult to achieve.

Magnetic resonance imaging

In clinical practice, magnetic resonance imaging (MRI) is chiefly used to reveal inner organ structure and assess soft tissue morphology (brain, musculoskeletal system and other parenchymal organs). However, several other variables can be assessed, including vascular volume and permeability, tissue perfusion, water diffusion, central nervous system function activation, metabolic spectroscopy, pH, pharmacokinetics and gene expression. By exploiting MRI's high spatial resolution and great soft-tissue contrast, various tissue types can be differentiated, as well as structural and volumetric changes determined in the course of disease processes. For these reasons, MR is both a morphological and a functional imaging modality.

In order to understand the basic principles of the MRI technique, we should start from the properties of the nuclei in a magnetic field. MRI works by the fact that about one in a million nuclei of the body tissues loses its random orientation and aligns its spin either parallel or antiparallel to the magnetic field when placed in a strong uniform magnetic field. The nuclei undergoing such an alignment produce a detectable change in the magnetic field. The nuclei precess around their axis at a frequency called the Larmor frequency. When a sample (e.g., a tissue specimen, an animal or a human subject) is temporarily exposed to radiofrequency (RF) pulses in a plane perpendicular to the magnetic field, the magnetically aligned nuclei with a Larmor frequency (i.e., their precession frequency) equal to the applied RF frequency undergo resonance: they jump temporarily in a high-energy state. When the RF pulse is turned off, the nuclei realign with the magnetic field, i.e. longitudinal relaxation, and the return to baseline is recorded as a change in electromagnetic flux. The emission of recordable signals during relaxation gives information about the environment of the resonating nuclei. Signal strength and frequency depend on the concentration and identity of the resonant atom and the strength of the local magnetic field. Two different time variables can then be defined and recorded. The first refers to the time the nuclei take to realign (i.e., longitudinal relaxation time or "Time 1", commonly known as T1). This is the determining variable of so-called T1-weighted imaging. The second is related to spin dephasing following application of the phase encoding gradient (i.e., transverse relaxation time or "Time 2" commonly known as T2). This is the determining variable of so-called T2-weighted imaging. It is important to appreciate that different tissues can have different T1 and T2 values and that the signal intensity in MR images can be made sensitive (weighted) to one or the other of these relaxation parameters. T1 and T2 become potential sources of contrast between different organs within the body and between healthy and diseased tissues. The recorded signals of the different sequences are then processed to produce a map of MR signal intensity versus position, i.e., a morphological image with a resolution varying between 1 mm^sup 3^ with current diagnostic MR scanners, and a few cubic micrometers ([mu]??3) with researchdedicated coils and probes. Thus, the advantage of MR imaging resides in its very high spatial resolution and ability to measure more than one physiological parameter using different pulse sequences.

MRI relies mostly on measuring ^sup 1^H atoms present in water, ubiquitously and in high concentrations in most tissues and organs, thus allowing tissues to be discriminated in relation to their water content. For example, fat tissues, which have little water and high concentrations of organic compounds, appear under appropriate imaging paradigms bright, whereas blood vessels or other fluid- filled areas appear dark. Detailed MR imaging of the brain is often based on the principle that gray matter has more fluid than white matter, making the distinction between the two possible. The presence of other atoms, including ^sup 3^He, ^sup 13^C, ^sup 23^Na, ^sup 31^P, can also be exploited for in vivo studies; with these, however, the MRI signal is small and most strongly coupled with the concentration of the nuclei in the sample. This explains why excellent body images can be obtained from the hydrogen present in water. Nuclei of other metabolites in vivo are typically found at much lower concentrations. So to obtain the same signal intensity with the same amount of ^sup 31^P as from ^sup 1^H, the smaller concentration of ^sup 31^P relative to ^sup 1^H needs to be compensated with a larger sample volume. In molecular imaging, sensitivity can be enhanced with higher strength magnetic fields (up to 20 T from 3T of the human device) and dedicated coils. The design and construction of specialized mouse RF coils greatly increase sensitivity to the MR signal.50

As with other imaging modalities, important information can be obtained with injectable contrast agents that alter tissue response to the RF pulse. For example, contrast agents can be used to change local Tl and T2 relaxation. In practice, the interaction between contrast agents and adjacent water molecules leads to a local reduction in T1 or T2 relaxation times. Two general types of contrast agents exist: those that primarily decrease T1, thus augmenting the signal on T1-weighted images, and those that predominantly decrease T2, thus diminishing the signal on T2- weighted images. Paramagnetic metal cations51 such as chelated gadolinium come under the category of contrast agents affecting T1, while SPIO and nanoparticles (magnetic particles often the size of nanometers), such as monocrystalline iron oxide nanoparticles, are contrast agents affecting T2. SPIOs can provide targeted molecules or cells with a large magnetic moment that creates substantial disturbances in the local magnetic field, leading to a rapid dephasing of nuclei. However, the major disadvantage of SPIO particles for molecular imaging is their relatively large size (bigger than that of nanoparticles) and opsonization by plasma proteins, resulting in rapid clearance by phagocytic cells and reduced transendothelial passage and tissue penetration. Despite its high spatial resolution, MRI has low molecular sensitivity (10^sup - 3^-10^sup -5^ mol/L). With the development of targeted contrast agents directed at specific molecules, the range of MR applications has dramatically expanded by combining MRI's non invasiveness and high spatial resolution with the specific localization of molecular targets. In the rapidly growing field of MR contrast agents, several protocols have been developed to improve the grade of achievable contrast.52

Magnetic susceptibility contrast induced by iron oxide particles has been successfully used for high-resolution single-cell imaging53 of T lymphocytes54 and stem cells.55

Kircher et al.56 developed a novel quantitative, noninvasive, high-resolution imaging approach to follow the recruitment of antigen specific CD8+ T cells to target tumors. To label the lymphocytes they used improved superparamagnetic particles, i.e., highly derivatized cross-linked iron oxide nanoparticles.

Engineered natural killer (NK) cells directed against HER-2/neu receptors were studied by Daldrup-Link et al. using MR imaging.57 The cells were genetically modified to express a receptor for the ErbB^sub 2^ antigen (HER^sub 2^-neu) and labeled with ferumoxides and ferucarbotran using lipofection and electroporation. The modified cells were injected into the tail vein of mice carrying HER^sub 2^-neu positive tumors. NK cell accumulation was followed by MR over time, demonstrating that the modified cells targeted the HER^sub 2^-neu positive tumor, whereas the unmodified cells did not accumulate within the tumor.

Maki et al.58 described the feasibility of tracking a stem cell line of bone and cartilage cells with fluorescein isothiocyanate (FITC)-labeled poly-l-lysine-CF(3) (PLK-CF[3]) using mouse ATDC5 cells. By adding it into culture medium, FITC-labeled PLK-CF(3) was easily internalized by ATDC5 cells. No acute or long-term toxicities were seen at less than 160 mug/mL. Labeled cells transplanted into the cranial bone of mice were detected for up to 7 days by MRI, demonstrating that FITC-labeled PLK-CF(3) is a useful positive contrast agent for MR tracking in bone and cartilage.

Cohen et al.59 described the use of the heavy chain of murine ferritin, an iron storage molecule with ferroxidase activity, as a novel endogenous reporter for the detection of gene expression by MR imaging. C6 glioma cells stably expressing a TET-EGFP-HA-ferritin construct permitted the dynamic detection of TET-regulated gene expression by MRI, which was then independently validated by FL microscopy and histology.

The application of such a novel MR reporter gene that generates significant contrast in the absence of exogenously administered substrates opens new possibilities for non-invasive molecular imaging of gene expression by MR imaging.

A comprehensive review of MRI contrast agents has been recently published.60

MRI can be applied to functional studies based on hemodynamic response, including measurement of perfusion, blood flow and blood oxygenation. For example, functional MR imaging (fMRI) with very rapid low-resolution sequences (20-30 per minute) is employed to assess regional changes in hemodynamic variables related to neuronal functional activity. Increased neuronal activity leads to a regional increase in the amount of oxyhemoglobin in brain blood vessels with respect to deoxyhemoglobin; this results in an increase of signal due to an excess of oxyhemoglobin, which, in turn, is related to neural activity. This type of signal is known as the blood oxygen level-dependent effect (BOLD).61

Besides fMRI, another important MR-based modality is MR spectroscopy (MRS), which allows to detect changes in the local distribution of some molecules other than water. The most common MR detectable nuclei exploited to assess biological processes by MRS are: ^sup 1^H, ^sup 19^F, ^sup 31^P, ^sup 13^C and ^sup 23^Na. Hydrogen, the most widely used nucleus, permits MR spectroscopy based on the presence of methyl groups in amino acids, lactate, N- acetylaspartate, creatine, choline, and methyl and methylene groups in lipids. Fluorine-labeled compounds introduced into cells can be used as an indicator of a cellular process. ^sup 31^P MRS was the first nucleus to be used for the study of pH, energy and phospholipid metabolism of living systems in vivo.62, 63 Most molecules in living systems consist of a carbon backbone and hydrogen atoms, which forms the basis of ^sup 1^H and ^sup 13^C MRS. ^sup 13^C MRS has relatively low sensitivity owing to the low natural abundance of ^sup 13^C (1.1%). However, measurement of ^sup 13^C MRS does not interfere with the water signal as ^sup 1^H MRS. ^sup 31^P MRS has intermediate MR sensitivity.64

MRS imaging (MRSI) combines MRI with MRS. With MRSI a particular region within an organism or a sample can be studied, obtaining a wealth of chemical information about that region, as would be achievable by an NMR spectroscopy.65, 66 Compared with single voxel methods, MRSI of the brain offers metabolic information with improved anatomical coverage and spectral resolution, but may be difficult to perform in the clinical setting.

Computed tomography and ultrasound imaging

These two modalities will be reviewed briefly as they are employed mainly in morphologic and vascular imaging.

In CT, an X-ray tube is the radiation source used to reveal inner organ structure based on the assessment of X-ray attenuation by tissues. The different molecular composition of various body structures determines the differences in X-ray absorption, so the differential X-ray attenuation by bone, fat, air and water results in high-contrast images of anatomical structures.

Micro-CT scanners scaled-down from clinical CT scanners for preclinical applications provide 3D images of small animals with an image resolution (50-100 [mu]??) scaled proportionally to that of a human CT. Images of intact rodent organs at spatial resolutions from cellular (20 [mu]m) down to subcellular dimensions (e.g., 1 [mu]m) can also be acquired, thus filling the resolution gap between microscope imaging, which resolves individual cells in thin tissue sections, and mini-CT imaging of intact volumes.67

However, the contrast between different soft tissues or blood vessels is intrinsically poor in X-ray CT images. To improve the visualization of abdominal organs and vasculature, highly attenuating exogenous contrast agents are used in clinical and preclinical imaging. These nonionic extracellular water-soluble agents, such as the iodinated agents typically used for clinical evaluations, are cleared from the vessels in a few minutes via the kidneys, making them useful for studying renal structure and function.68 But imaging the vasculature of animal models with these contrast agents is not currently possible because the majority of commercially available micro-CT scanners are too slow to capture the images before the contrast media is cleared from the blood. Recent advances in microCT technology have resulted in scanners that can produce volumetric images in 1-8 s;69 however, the use of rapidly clearing agents still poses problems because the first-pass phase of an extracellular intravenous contrast agent persists in the vessel for only a few heartbeats. New blood pool contrast agents for prolonged vascular residence time have been recently developed for small animal imaging.70, 71 These blood pool agents are also important for tumor characterization, particularly for detecting, characterizing, and assessing tumor aggressiveness, neovascularization, and response to therapy.72

X-ray CT imaging is an excellent tool for investigating the skeleton, some inner organs, the lungs and fat distribution.73 Investigators are working on the possibility to study functional parameters such as blood flow and vascular permeability.74 Because of their high anatomic resolution, micro-CT scanners can also be used for functional phenotyping of the rodent heart.75 Although CT imaging is of limited relevance for molecular and cellular imaging when used alone, it plays an essential role in obtaining an anatomic reference frame for functional imaging interpretation, as will be discussed in the section on multimodality at the end of this review.

Basic US imaging is founded on the use of ultrasound waves (not radiation) to produce either static or dynamic images of body organs and to measure the velocity of blood flow. Ultrasound refers to sound waves emitted at a frequency higher than 20 kHz. Clinical diagnostic US scanners use frequencies in the range of 1 to 20 MHz. US biomicroscopy typically operates at frequencies of 40 to 200 MHz, while scanning acoustic microscopy uses frequencies higher than 200 MHz. The choice of US frequency is a trade-off between resolution (resolution improves with frequency) and penetration depth (penetration drops with frequency). US probes contain one or more acoustic transducers emitting wave pulses that are reflected as echoes bouncing off the interfaces between materials of different acoustic impedance (product of the density of the material and the speed of sound in that material). The echo signal encodes information about the interface depth and the differences between the impedances of the materials at the interface. The echoes received by the probe are converted into a single dimensional signal, but 2D and 3D images can be obtained by combining echoes collected from different angles around an organ under examination.76

Based on the Doppler effect (the frequency shift related to the velocity between the structure and the probe), ultrasonography can also reveal whether organs (usually blood) are moving towards or away from the probe and the velocity of such motion. This is a particularly useful feature in cardiovascular studies. Ultrasonography is being increasingly used for research in small animals.77 With dedicated high and ultra-high frequency instruments, dynamic processes, such as the heart cycle, can be visualized even in small animals with a heart rate above 400 beats/min78 and neoplastic lesions detected in a transgenic prostate cancer mouse model.79

US contrast agents alter wave absorption and reflection, thus enhancing the differences in the signals coming from different tissues. The majority of US contrast agents consist of intravenous injections of very small bubbles (microbubbles) of air or perfluorocarbon gas stabilized by a protein, lipid, or polymer shell.14 Microbubbles work well not only because they provide a strongly reflective blood/gas interface, but also because they resonate in the ultrasound beam, making them several orders of magnitude more reflective than normal blood.

The use of air microbubbles in molecular and cellular imaging exploits the chemical or electrostatic properties of the microbubble shell in order to increase the impedance or the echogenic signal in specific areas and tissues where they non-specifically bind. In the "specific targeting" method, microbubbles can be used for molecular and cellular imaging, following the attachment of specific antibodies or other ligands to their surface.80 This leads to the accumulation of targeted contrast agents at a specific site through the use of adhesion molecules that recognize disease antigens. Potential ligands include antibodies, peptides, polysaccharides. With encapsulation strategies and the advent of second harmonic ultrasonic imaging,81 recent attempts at microbubble targeted imaging have proven successful in imaging the molecular features of various disease states, including inflammation, angiogenesis and thrombosis.82

Most targeted ultrasound contrast agents are microbubbles, but other vehicles include acoustically active liposomes agents83 and perfluorocarbon emulsions nanoparticles.84

Multimodal imaging strategies

Molecular and cellular imaging modalities have different temporal and spatial resolutions, probe mass, tissue penetration depth and sensitivity for measuring properties related to morphology or function (Table I). Each imaging modality has drawbacks and limitations. With a multimodal approach, the advantages of the different techniques can be accrued and the weakness of an individual modality reduced, thus offering the prospect of improved diagnostics, therapeutic monitoring, and preclinical research.85

Imaging modalities can be roughly divided into two groups: those that excel primarily at providing structural information (i.e., CT, MR, and ultrasound) and those that offer functional or molecular information (i.e., PET, SPECT and optical imaging). Therefore, it is not surprising that much work has gone and is ongoing in the integration of PET and SPECT with CT or MR to permit functional and anatomical assessment in the same subject at the same time.

There are different approaches to combining functional and anatomical information: the historical method is by software integration of images obtained at different acquisition times; the second, more recent approach is based on integrated hardware systems (e.g., PET/CT and SPECT/CT) that obtain double information on morphology and function in a single study.

Software integration is performed in two steps: the first is to generate a geometrical transformation of the two 3D datasets (registration) and the second is to combine the images by generating a single image (Figure 1) from the two registered originals (fusion). Image registration must reckon with animal positioning in different scans and imaging parameters such as section thickness, image matrix size and scan length. Various methods have been developed for human studies to determine the geometrical transformation of different imaging modalities under study,86 including head holder alignment, external fiducial markers, stereotactic frames, internal landmark matching and automated fusion methods. Since the first three methods use external markers attached to the object's skin or face, they cannot be used in a retrospective way if no markers were used in the studies. The stereotaxic frame method is the most accurate means of localization, but its use is limited by frame invasiveness. Combining the use of the imaging features of each modality seems to be a more flexible way to integrate the different techniques. Several groups have been working to improve the performance in small animal studies of image registration methods that are effective in human studies. A method using an imaging chamber that can be rigidly and reproducibly mounted on separate small animal PET and CT scanners was developed at UCLA.87 The imaging chamber works synergistically in combination with a registration phantom to achieve accurate image registration. Rowland et al. employed a voxel-based method to register small animal PET emission images and MR data; the method requires no manual identification of image features and makes no use of surgically implanted or external fiducial markers or stereotactic apparatus.88 Another 3D automated fusion method to integrate PET, CT and SPECT small animal images was developed by the Jan et al.89

Figure 1.-Multimodality imaging in a preclinical oncological study. A) CT, B) PET and C) PET/CT fused axial images of a transgenic mouse naturally developing breast cancers. The PET/CT fused image shows [^sup 18^F]FDG pathological accumulation in the tumor (color map) superimposed on the CT anatomic image (grey map). Courtesy of Nuclear Medicine, Policlinico S.Orsola, Bologna, Italy.

Figure 2.-Hardware integration of CT and PET images: A) acquisition scheme by two separate scanners positioned next to each other on a bed that moves through the system: morphology and function are assessed sequentially; B) acquisition by one hybrid system in which the two scanning systems are integrated in one single gantry: morphology and function are assessed simultaneously.

The alternative multirnodality approach to building an imaging system in which two or more modalities are integrated into a single unit was developed to overcome the limited success of the integration approach in human extra-cranial studies.90 The fact that the brain is well approximated by a rigid body contained within the skull effectively allows for accurate registration, whereas in other parts of the body, organ shape and location depend critically on how a subject is positioned in the scanner and on organ changes over time. In molecular and cellular imaging, another major disadvantage of acquiring sequential scans on two different scanners is that two different parameters cannot be simultaneously measured or time- dependent changes for those parameters correlated.

Hardware integration of different modalities has been undertaken in two ways: in one the scanners are placed side by side and are integrated by a bed that moves through the system (Figure 2A). The other involves the development of a fully integrated system with a single set of detectors or sensors that can detect the radiation from the different modalities (Figure 2B).

Six years ago, the first side-by-side integrated system prototype was developed at the University of Pittsburgh, in which a clinical human CT scanner was combined with a clinical PET device.91 Combined PET and CT scanner systems are now part of clinical applications.92 Research is being conducted on combining high-resolution animal PET with dedicated CT systems to integrate PET and CT information in preclinical studies.93

In the early 1990s, Hasegawa et al. from the University of California were the first group to develop a simultaneous emission/ transmission system based on a linear array of high purity germanium detectors that were used to simultaneously detect both the 100-200 keV emission gamma rays from an injected radiotracer and the transmitted beam from an X-ray tube.94 This is the first fully integrated system with a single set of detectors for emission and transmission radiation.

CT, however, has low soft-tissue contrast and uses relatively high doses of ionizing radiation, which may have biological effects on the animal models being studied.95 On the other hand, MR imaging can provide high spatial resolution and excellent soft-tissue contrast for anatomic imaging, but suffers from poor signal strength, resulting in low sensitivity for functional imaging. Thus, PET/MR holds promise as a valuable integrated system that affords the advantage of higher soft tissue contrast in MR anatomical imaging and less radiation exposure than PET/CT. An additional advantage to performing PET in a high magnetic field is good resolution enhancement due to reduced positron range, even though very high magnetic fields are needed to improve resolution for commonly used positron emitters such as ^sup 18^F or ^sup 11^C.96 Simon Cherry's group started developing MR compatible PET detectors at UCLA.97 The detector principle is based on the use of long optical fibers that guide the light from scintillation crystals positioned within the magnetic field to position sensitive photomultiplier tubes outside, where the fringe field drops below 10 mT. The light guide is 3-4 m long. Based on this technology, the first simultaneous PET and MR imaging studies of phantoms at 1.5 T were performed with a single-layer lutetium oxyorthosilicate (LSO) ring measuring 54 mm in diameter.98

The feasibility of simultaneous PET and MR imaging, as well as MR spectroscopy, for small animals has been shown using a similar prototype.

Nevertheless, the PET device performed poorly owing to limited axial extent, low sensitivity and light guide length. For larger axial coverage of the PET insert, it may be advantageous to use a different detector concept. Avalanche photodiodes (APDs) have been used in a prototype small animal PET system99 to read out the scintillation light from bismuth germanate or Cerium-doped LSO crystals.100 The main advantage here is that the APDs operate in high magnetic fields without performance degradation. Moreover, these semiconductor detectors are very compact, thus offering the opportunity to build very compact modules with potentially minimum interference with MR imaging.101

In addition to combining functional and anatomical information, it could be very useful to analyze the same molecular process with different imaging modalities to obtain functional-functional information. To this end, several approaches can be envisaged: a) vectors containing a reporter gene whose expression can be detected by two different functional imaging modalities; b) vectors expressing fusion proteins that can be visualized by two different imaging techniques; c) bicistronic vectors containing two different reporter genes controlled by the same unique promoter (using internal ribosome entry sites, i.e. 1RES sequences).

Combined emission tomography-optical imaging seems to be the most effective in gaining further advantages (Figure 3): A) BL permits rapid, cost-effective real-time imaging in animal models, which is especially useful for the study of molecular process kinetics; B) emission tomography imaging provides tomographic, fully quantitative images and, because this imaging modality has been well developed for clinical use, it constitutes a link between preclinical and clinical studies.

Research is under way to couple optical imaging with PET and SPECT to allow functional-functional multimodal imaging. Alexandrakis et al have recently investigated different detector designs that could be utilized for optical and PET imaging102 allowing, for example, the expression of two reporter genes to be monitored at the same time or other novel approaches to dissecting molecular pathways and interactions.

MR and emission tomography imaging have also been combined in a recent study. De Vries et al.103 labeled immature phagocytic dendritic cells taken from the blood of melanoma patients by using dextran-coated SPIO and [111In]oxine to monitor cellular therapy (of the mature cells) by both MR and scintigraphic imaging. SPIO combined with radionuclide labeling enabled them to quantify (by scintigraphy) and to image in detail (by MR) the anatomical localization of the migrated cells.

Figure 3. - Example of a functional-functional multimodal imaging strategy for reporter gene expression analysis. Stably transfected MCF7 (breast cancer) cells expressing Luciferase and a mutant form of the D2 receptor controlled by the same Estrogen inducible promoter were injected into CDl mice subcutaneously. Reporters expression was then evaluated by PET and optical analysis. For PET imaging, 30-min acquisition was performed the same day as the CCD scan after intraperitoneal injection of (18F]FESP (7.4 MBq or 200 mCi) and 15-min uptake. For bioluminescence imaging, 10-min CCD acquisition was performed 20 min after intraperitoneal Luciferin administration.

Moreover, a triple fusion reporter vector harboring a BL synthetic Renilla luciferase (hrl) reporter gene, a reporter gene encoding the monomelic red FL protein (mrfpl) and HSVlI-sr39tk as PET gene, was found to preserve the most activity for each protein component and was, therefore, investigated in detail. Metastases of a human melanoma cell line (A375M) stably expressing the triple fusion were imaged by PET and optical technologies over a 40-50-day time period in living mice.104

In conclusion, this review gives an overview of the highly complex world of molecular and cellular imaging, where several variables have to be examined in the study of a specific molecular process to obtain the most information: imaging technique, probe, contrast agent, reporter gene, single versus multimodality approach. Multimodality imaging seems to be the choice of the future because of the numerous opportunities it may provide for improving on currently existing multimodality imaging systems and for the development of new combinations of modalities. As multimodality imaging progresses, the focus of research will shift (and in some cases this has already happened) to work out how to make the best use of two complementary datasets.

References

1. Koo V, Hamilton PW, Williamson K. Non-invasive in vivo imaging in small animal research. Cell Oncol 2006;28:127-39.

2. Massoud TF, Gambhir SS. Molecular imaging in living subjects: seeing fundamental biological processes in a new light. Genes Dev 2003;17:545-80.

3. Maggi A, Ciana P. Reporter mice and drug discovery and development. Nat Rev Drug Discov 2005;4:249-55.

4. Lewis JS, Achilefu S, Garbow JR, Laforest R, Welch MJ. Small animal imaging: current technology and perspectives for oncological imaging. Eur J Cancer 2002;38:2173-88.

5. Cherry SR. In vivo molecular and genomic imaging: new challenges for imaging physics. Phys Med Biol 2004;49:R13-48.

6. Ntziachristos V, Ripoll J, Wang LV, Weissleder R. Looking and listening to light: the evolution of whole-body photonic imaging. Nat Biotechnol 2005;23:313-20.

7. Levin CS. Primer on molecular imaging technology. Eur J Nucl Med Mol Imaging 20O5;32 Suppl 2:S325-45.

8. Ottobrini L, Ciana P, Biserni A, Lucignani G, Maggi A. Molecular imaging: a new way to study molecular processes in vivo. MoI Cell Endocrinol 2006;246:69-75.

9. Contag CH, Bachmann MH. Advances in vivo bioluminescence imaging of gene expression. Annu Rev Biomed Eng 2002;4:235-60.

10. Zhao H, Doyle TC, Coquoz O, Kalish F, Rice BW, Contag CH. Emission spectra of bioluminescent reporters and interaction with mammalian tissue determine the sensidvity of detecdon in vivo. J Biomed Opt 2005;10:41210.

1 1 . Morawski AM, Lanza GA, Wickline SA. Targeted contrast agents for magnetic resonance imaging and ultrasound. Curr Opin Biotechnol 2005;16:89-92.

12. Artemov D, Bhujwalla ZM, Bulte JW. Magnetic resonance imaging of cell surface receptors using targeted contrast agents. Curr Pharm Biotechnol 2004;5:485-94.

13. Bulte JW, Kraitchman DL. Iron oxide MR contrast agents for molecular and cellular imaging. NMR Biomed 2004;17:484-99.

14. Klibanov AL. Microbubble contrast agents: targeted ultrasound imaging and ultrasound-assisted drug-delivery applications. Invest Radiol 2006;41:354-62.

15. Pomper MG. Molecular imaging: an overview. Acad Radiol 2001;8:1141-53.

16. Chatziioannou AF. Instrumentation for molecular imaging in preclinical research: micro-PET and micro-SPECT. Proc Am Thorac Soc 2005:2:533-6,510-11.

17. Turkington TG. Introduction to PET instruments. J Nucl Med Technol 2001;29:4-11.

18. Schafers KP, Reader AJ, Kriens M, Knoess C, Schober O, Schafers M. Performance evaluation of the microPET focus: a third- generation microPET scanner dedicated to animal imaging. J Nucl Med 2005;46:455-63.

19. Phelps ME. Positron emission tomography provides molecular imaging of biological processes. Proc Natl Acad Sci U S A 2000;97:9226-33.

20. Couturier O, Luxen A, Chatal JF, Vuillez JP, Rigo P, Hustinx R. Fluorinated tracers for imaging cancer with positron emission tomography. Eur J Nucl Med MoI Imaging 2004;31:1182-206.

21. Aliberti S, Grignani G, Agiietta M. Positron-emission tomography in cancer therapy. N Engl J Med 2006;354:1958-60.

22. Peremans K, Cornelissen B, Van Den Bossche B, Audenaert K, Van de Wiele C. A review of small animal imaging planar and pinhole SPECT gamma camera imaging. Vet Radiol Ultrasound 2005;46:162-70.

23. Beekman FJ, McElroy DP, Berger F, Gambhir SS, Hoffman EJ, Cherry SR. Towards in fit o nuclear microscopy: iodine-125 imaging in mice using micro-pinholes. Eur J Nucl Med MoI Imaging 2002;29:933- 8.

24. Leyton J, Latigo JR, Perumal M, Dhaliwal H, He Q, Aboagye EO. Early detection of tumor response to chemotherapy by 3'-deoxy3'- [18F]fluorothymidine positron emission tomography: the effect of cisplatin on a fibrosarcoma tumor model in vivo. Cancer Res 2005;65:4202-10.

25. Solomon B, Binns D, Roselt P, Weibe U, McArthur GA, CuIIinane C et al. Modulation of intratumoral hypoxia by the epidermal growth factor receptor inhibitor gefitinib detected using small animal PET imaging. MoI Cancer Ther 2005;4:1417-22.

26. Yagle KJ, EaryJF, Tait JF, Grierson JR, Link JM, Lewellen B etal. Evaluation of F-annexin V as a PET imaging agent in an animal model of apoptosis. J Nucl Med 2005;46:658-667

27. Gheysens O, Gambhir SS. Studying molecular and cellular processes in the intact organism. Prog Drug Res 2005;62:117-50.

28. Gross S, Piwnica- Worms D. Spying on cancer: molecular imaging in vivo with genetically encoded reporters. Cancer Cell 2005;7:5-15.

29. Doubrovin M, Serganova I, Mayer-Kuckuk P, Ponomarev V, Blasberg RG. Multimodality in vivo molecular-genetic imaging. Bioconjug Chem 2004;15:1376-88.

30. Dubey P, Su H, Adonai N, Du S, Rosato A, Braun J etal Quantitative imaging of the T cell antitumor response by positron- emission tomography. Proc Natl Acad Sei U S A 2003;100:1232-7. 31. Su H, Forbes A, Gambhir SS, Braun J. Quantitation of cell number by a positron emission tomography reporter gene strategy. MoI Imaging Biol 2004;6:139-48.

32. Xiong Z, Cheng Z, Zhang X, Patel M, Wu JC, Gambhir SS et al. Imaging chemically modified adenovirus for targeting tumors expressing integrin Ialpha|v(betal3 in living mice with mutant herpes simplex virus type 1 thymidine kinase PET reporter gene. J Nucl Med 2006;47:130-9.

33. Welsh DK, Kay SA. Bioluminescence imaging in living organisms. Curr Opin Biotechnol 2005;16:73-8.

34. Rice BW, Cable MD, Nelson MB. In vivo imaging of light- emitting probes. J Biomed Opt 2001;6:432-40.

35. Ntziachristos V, Ripoll J, Wang LV, Weissleder R. Looking and listening to light: the evolution of whole-body photonic imaging. Nat Biotechnol 2005;23:313-20.

36. Zacharakis G, Kambara H, Shih H, Ripoll J, Grimm J, Saeki Y etal. Volumetric tomography of fluorescent proteins through small animals in vivo. Proc Nad Acad Sei U S A 2005;102:18252-7.

37. Shah K, Weissleder R. Molecular optical imaging: applications leading to the development of present day therapeutics. NeuroRx 2005;2:215-25.

38. Hassan M, Klaunberg BA. Biomedical applications of fluorescence imaging in vivo. Comp Med 2004;54:635-44.

39. Zhao H, Doyle TC, Coquoz O, Kalish F, Rice BW, Contag CH. Emission spectra of bioluminescent reporters and interaction with mammalian tissues determine the sensitivity of detection in vivo. J Biomed Opt 2005:10:41210.

40. Cao YA, Wagers AJ, Beilhack A, Dusich J, Bachmann MH, Negrin RS et al. Shirting foci of hematopoiesis during reconstitution from single stem cells. Proc Nad Acad Sei U S A 2004;101:221-6.

41 . Schimmelpfennig CH, Schulz S, Arber C, Baker J, Tarner I, McBride J et al. Ex vivo expanded dendritic cells home to T-cell zones of lymphoid organs and survive in vivo after allogeneic bone marrow transplantation. AmJ Pathol 2005;167:1321-31.

42. Ciana P, Di Luccio G, Belcredito S, Pollio G, Vegeto E, Tatangelo L et al. Engineering of a mouse for the in vivo profiling of estrogen receptor activity. Mol Endocrinol 2001;15:1104-13.

43. Ciana P, Raviscioni M, Mussi P, Vegeto E, Que I, Parker MG etal. In vivo imaging of transcriptionally active estrogen receptors. Nat Med 2003;9:82-6.

44. Klotz DM, Hewitt SC, Ciana P, Raviscioni M, Lindzey JK, Foley J et al. Requirement of estrogen receptor-alpha in insulin-like growth factor- 1 (IGF-l)-induced uterine responses and in vivo evidence for IGF-1/estrogen receptor cross-talk. J Biol Chem 2002;277:8531-7.

45. Di Lorenzo D, Villa R, Biasiotto G, Belloli S, Ruggeri G, Albertini A et al. Isomer-specific activity of dichlorodyphenyltrichloroethane with estrogen receptor in adult and suckling estrogen reporter mice. Endocrinology 2002;143:4544-51.

46. Van der Pluijm G, Que I, Sijmons B, Buijs JT, Lowik CW, Wetterwald A et al. Interference with the microenvironmental support impairs the de novo formation of bone metastases in vivo. Cancer Res 2005;65:7682-90.

47. Citrin D, Scott T, Sproull M, Menard C, Tofilon PJ, Camphausen K. In vivo tumor imaging using a near-irifrared-labelled endostatin molecule. Int J Radiat Oncol Biol Phys 2004;58:536-41.

48. Ray P, Pimenta H, Paulmurugan R, Berger F, Phelps ME, Iyer M et al. Noninvasive quantitative imaging of protein-protein interactioas in living subjects.Proc Natl Acad Sci U S A 2002;99:3105-10.

49. Michalet X, Pinaud FF, Bentolila LA, TsayJM, Doose S, LiJJ et al. Quantum dots for live cells, in vivo imaging, and diagnostics. Science 2005;307:538-44.

50. Nieman BJ, Bock NA, Bishop J, Chen XJ, Sled JG, Rossant J et al. Magnetic resonance imaging for detection and analysis of mouse phenotypes. NMR Biomed 2005;18:447-68.

51. Bhorade R, Weissleder R, Nakakoshi T, Moore A, Tung CH. Macrocyclic chelators with paramagnetic cations are internalized into mammalian cells via a HlV-tat derived membrane translocation peptide. Bioconjug Chem 2000;11:301-5.

52. Artemov D, Bhujwalla ZM, Bulte JW. Magnetic resonance imaging of cell surface receptors using targeted contrast agents. Curr Pharm Biotechnol 2004;5:485-94.

53. Hinds KA, HiIlJM, Shapiro EM, Laukkanen MO, Silva AC, Combs CA et al Highly efficient endosomal labeling of progenitor and stem cells with large magnetic particles allows magnetic resonance imaging of single cells. Blood 2003;102:867-72.

54. Dodd CH, Hsu HC, Chu WJ, Yang P, Zhang HG, MountzJD Jr et al. Normal T-cell response and in vivo magnetic resonance imaging of T cells loaded with HIV transactivator-peptide-derived superparamagnetic nanoparticles. J Immunol Methods 2001;256:89-105.

55. Bulte JW, Douglas T, Witwer B, Zhang SC, Strahle E, Lewis BK et al Magnetodendrimers allow endosomal magnetic labeling and in vivo tracking of stem cells. Nat Biotechnol 2001;19:1141-7.

56. Kircher MF, Allport JR, Graves EE, Love V, Josephson L, Lichtman AH et al. In vivo high resolution three-dimensional imaging of antigen-specific cytotoxic T-lymphocyte trafficking to tumors. Cancer Res 2003;63:6838-46.

57. Daldrup-Link HE, Meier R, Rudelius M, Piontek G, Piert M, Metz S et al. In vivo tracking of genetically engineered, anti-HER2/ neu directed natural killer cells to HER2/neu positive mammary tumors with magnetic resonance imaging. Eur Radiol 2005;15:4-13.

58. Maki J, Masuda C, Morikawa S, Morita M, Inubushi T, Matsusue Y et al. The MR tracking of transplanted ATDC5 cells using fluorinated poly-1-lysine-CF(3). Biomaterials 2007;28:434-40.

59. Cohen B, Dafni H, Meir G, Harmelin A, Neeman M. Ferritin as an endogenous MRI reporter for noninvasive imaging of gene expression in Co glioma tumors. Neoplasia 2005;7:109-17.

60. Delikatny EJ, Poptani H. MR techniques for in vivo molecular and cellular imaging. Radiol Clin North Am 2005;43:205-20.

61. Price RR, Allison J, Massoth RJ, Clarke GD, Drost DJ; NMR Task Group #8. Practical aspects of functional MRI (NMR Task Group #8). Med Phys 2002;29:1892-912.

62. Moon RB, R