SEARCH

SEARCH BY CITATION

Keywords:

  • molecular imaging;
  • neuroimaging;
  • receptor distribution;
  • rodents;
  • signal transduction

Abstract

  1. Top of page
  2. Abstract
  3. Multimodal neuroimaging in rodents: general aspects
  4. Molecular and cellular imaging: the tools
  5. Visualizing receptor expression and receptor occupancy
  6. Imaging receptor function
  7. Challenges in (molecular) neuroimaging of rodents
  8. Conclusions
  9. Acknowledgements
  10. References

Multimodal non-invasive neuroimaging in rodents constitutes an attractive tool for studying neurobiological processes in vivo. At present, imaging studies of brain anatomy and function as well as the investigation of structure–function relationships belong to the standard repertoire of neuroscientists. Molecular imaging adds a new perspective. The mapping of the receptor distribution and receptor occupancy can nowadays be complemented by specific readouts of receptor function either by visualizing the activity of signaling pathways or mapping the physiological consequences of receptor stimulation. Molecular information is obtained through the use of imaging probes that combine a target-specific ligand with a reporter moiety that generates a signal that can be detected from outside the body. For imaging probes targeting the central nervous system, penetration of the intact blood–brain barrier constitutes a major hurdle. Molecular imaging generates specific information and therefore has a large potential for disease phenotyping (diagnostics), therapy development and monitoring of treatment response. Molecular imaging is still in its infancy and major developments in imaging technology, probe design and data analysis are required in order to make an impact. Rodent molecular neuroimaging will play an important role in the development of these tools.


Multimodal neuroimaging in rodents: general aspects

  1. Top of page
  2. Abstract
  3. Multimodal neuroimaging in rodents: general aspects
  4. Molecular and cellular imaging: the tools
  5. Visualizing receptor expression and receptor occupancy
  6. Imaging receptor function
  7. Challenges in (molecular) neuroimaging of rodents
  8. Conclusions
  9. Acknowledgements
  10. References

Neuroimaging plays a central role in the study of brain anatomy and function and is of particular interest for analyzing structural and functional connectivities under both normal and pathologic conditions. The methods provide valuable diagnostic information and have become indispensable tools for assessing the efficacy of therapeutic interventions. Given the relevance of animal models of human disease in biomedical research, it is not surprising that neuroimaging techniques are also used intensively in rodents. Non-invasive imaging is suited for monitoring disease progression and therapy response in individuals, which is of particular interest for chronic pathologies such as neurodegenerative diseases. Longitudinal studies in individuals are attractive from both scientific and ethical points of view. Under conditions when the variability in experimental data is dominated by the biological variability, the assessment of relative changes with regard to a reference state recorded previously in the same individual will enhance the statistical power of the results.

Modern multimodal non-invasive (neuro)imaging allows annotating readouts on tissue morphology, physiology and metabolism with information on cellular and molecular processes (Fig. 1). In this article we will discuss the potential of these so-called target-specific or molecular imaging approaches in experimental neuroscience and describe some relevant imaging modalities as well as the nature and design of molecular reporter molecules for probing molecular processes in vivo. Selected studies of central nervous system (CNS) receptor distribution, ligand–receptor interactions and visualization and quantification of the functional consequences induced by activation or inhibition of these receptors will illustrate the usefulness of molecular neuroimaging.

image

Figure 1.  Multimodal imaging of the mouse brain as illustrated by high-resolution structural MRI (ex-vivo image, only right hemisphere), measurement of hemodynamic changes following infusion of GABAA antagonist bicuculline using fMRI and assessment of cerebral β-amyloid plaque load in a transgenic model of cerebral amyloidosis using fluorescence molecular tomography (FMT) in combination with a plaque-specific fluorescent ligand. The inserts highlight the difference in spatial resolution for the three modalities (pixel dimensions indicated by the square). Scale bar, 1 mm.

Download figure to PowerPoint

Molecular and cellular imaging: the tools

  1. Top of page
  2. Abstract
  3. Multimodal neuroimaging in rodents: general aspects
  4. Molecular and cellular imaging: the tools
  5. Visualizing receptor expression and receptor occupancy
  6. Imaging receptor function
  7. Challenges in (molecular) neuroimaging of rodents
  8. Conclusions
  9. Acknowledgements
  10. References

Structural and functional phenotypes are the result of molecular processes. It is therefore a plausible hypothesis that non-invasive mapping of molecular and cellular events in a temporospatially resolved manner should translate into diagnostic tools of improved sensitivity and specificity. Molecular imaging would also be highly relevant for the evaluation of novel therapies (Rudin & Weissleder, 2003). Molecular processes can be studied at various levels, i.e. visualization of (i) potential drug target molecules such as receptors or enzymes, (ii) the molecular consequences of the target activation by assessing the activation of signaling pathways or (iii) the functional/physiological consequences of ligand–target interactions. Although the imaging of receptor distribution, ligand–receptor interaction (receptor occupancy) and the metabolic and physiological consequences thereof are established tools, imaging assays to assess signal transduction in intact animals have been introduced only recently.

Imaging modalities

Non-invasive imaging of biological samples involves photon propagation through tissue. (This holds with the exception of ultrasound-based methods, which measure the propagation of pressure waves and do not involve electromagnetic radiation.) Photons of different energy can be used for the purpose, provided that tissue absorption is not prohibitively high in the respective frequency domain. For biological tissue, windows of low absorption exist in the high energy spectral domain (X-ray and gamma photons, > 1018 Hz), the near infrared domain of the optical spectrum (0.5 × 1015–3.3 × 1015 Hz, corresponding to wavelengths of 900–600 nm) and the radiofrequency domain (1 × 106–500 × 106 Hz).

The most relevant molecular imaging modality for (clinical) CNS studies is positron emission tomography (PET), which probes the distribution of positron-emitting nuclei (Phelps, 2004). These meta-stable nuclides relax by emission of a positron (the antimatter particle of the electron), which is scattered upon passage through the tissue. Scattering involves loss of excessive energy until the positron energy is low enough to allow capturing by an electron. The antimatter–matter annihilation process generates to gamma photons of 511 keV (and a neutrino), which propagate in opposite directions and are eventually detected by scintillation crystals arranged as a ring surrounding the sample. The simultaneous detection of two gamma photons within a so-called coincidence time window provides spatial information through electronic collimation; the annihilation process must have occurred on the line connecting the two responsive detectors (‘line of response’). Collecting a sufficient number of such events allows the reconstruction of the three-dimensional distribution of the radionuclides (or more precisely the sites of positron–electron annihilation). The coincidence detection principle significantly enhances the sensitivity of the method, as random signals due to stray photons are largely suppressed. At present, PET is the most sensitive molecular imaging modality; local concentrations of the order of 10−9–10−12 m can be detected. Modern PET scanners are designed as hybrid devices; they include a structural imaging modality such as X-ray computerized tomography providing an anatomical reference for the molecular PET data. In addition, structural information can be used to improve the PET image reconstruction.

Light photons are strongly scattered in turbid media and photon transport in biological tissue is commonly described as propagation of a diffuse wave (Ishimaru, 1997; Ntziachristos et al., 2005; Rudin, 2005). Light generated by a source located at a depth d within tissue is detected as a diffuse intensity distribution at the surface, rendering the accurate spatial localization of the source position difficult. Scattering and absorption limit the depth from which signals can be detected to a few centimeters. Nevertheless, fluorescence and bioluminescence imaging are attractive modalities for experimental research as they (i) provide excellent sensitivity, (ii) allow the use of stable reporter molecules and (iii) are relatively economic. In fluorescence imaging, a fluorophor within the tissue is excited using an external light source. In bioluminescent imaging, light is produced through an enzyme (e.g. luciferase)-catalysed reaction. The intrinsic problem that has to be addressed in optical imaging in order to derive quantitative data is the accurate reconstruction of both the location and spatial extent of the photon source within the tissue from the light intensity distribution measured at the tissue surface. A promising solution is fluorescence molecular tomography, which combines the information from surface images obtained for different laser excitation patterns (Ntziachristos et al., 2005; Rudin, 2005). More recent approaches combine fluorescence molecular tomography with computerized tomography. The use of prior (structural) information confines the fluorescence molecular tomography reconstruction problem and hence improves the quality of the reconstructed data (Hyde et al., 2009). The sensitivity of the method is slightly inferior to that of PET and allows detection of concentrations of the order of 10−7–10−10 m. In contrast to PET, quantification in optical imaging remains an issue and awaits the availability of truly tomographic techniques.

Magnetic resonance imaging (MRI) probes the magnetic properties of hydrogen nuclei in the tissue, the major signal contributors being water and adipose tissue. For a description of the underlying principles the reader is referred to the literature (Rudin, 2005). The signal in MRI is governed by multiple parameters such as the proton density, relaxation rates (R1, R2 and R2*; the relaxation rate is the inverse of the relaxation time), diffusion properties of water and water exchange dynamics. This multivariate signal dependence is the source of the high soft tissue contrast provided by the method, which can be tuned by choosing appropriate acquisition parameters. Exogenous contrast agents, which affect relaxation rates, are used to enhance contrast. These paramagnetic or superparamagnetic compounds are based on the strong magnetic moment of electrons, which exceeds that of the proton by a factor of 650. The low quantum energy of the magnetic resonance process is inevitably linked to low intrinsic sensitivity. Typically, the method enables the detection of signal changes induced by local concentrations of the contrast agents of 10−5–10−3 m.

Design of target-specific imaging probes

The imaging of molecular players requires target-specific imaging probes that report on the distribution and concentration of their target molecules. As the concentration of receptor molecules is low there are high demands regarding the sensitivity of the imaging method; 103–106 receptor molecules per cell translate to local concentrations in the range of 10−9–10−6 m. The imaging method should therefore have the sensitivity to detect <10−9 to <10−6 m concentration of a tracer (the tracer occupies only a fraction of the total receptor concentration; hence its concentration will be one or two orders of magnitude lower than the actual receptor concentration) with sufficient signal-to-noise ratio provided that the signal is exclusively due to the bound imaging probe. Any unbound ligand in the system would cause the contrast to deteriorate. The critical parameter to discriminate a target structure from its environment is the contrast-to-noise ratio (or signal-to-background ratio), which should be maximized. Correspondingly, the pharmacokinetic properties of an imaging probe should enable the specific binding to its molecular target, whereas any unbound or non-specifically bound tracer fraction should be rapidly eliminated. When considering molecular targets within the CNS, probe delivery through the intact blood–brain barrier will constitute a major obstacle. In this regard, the structures of well-characterized low molecular weight CNS drugs that are known to cross the blood–brain barrier are rewarding templates for probe design. The isotopic labeling of such receptor ligands or derivatives thereof using radionuclides has led to a number of highly specific imaging ligands. This explains the prominent role of PET for molecular imaging studies of the brain. Nevertheless, other imaging modalities such as fluorescence imaging may also provide relevant molecular and cellular information, in particular when focusing on preclinical applications.

The prerequisites for a target-specific imaging probe are almost identical to those of a therapeutic compound. The probe should have a high affinity to its molecular target (e.g. a receptor), significantly less affinity to other potential binding partners, suitable pharmacokinetic properties (which differ from those of a therapeutic) and must be safe. A diagnostic agent should not prompt any pharmacological response but instead yield a signal that can be detected non-invasively. This may require a signal amplification step, a single ligand–target interaction prompting a signaling response of more than one reporter group. This can be achieved by increasing the payload of reporters, enzymatic activation of reporter systems or trapping of reporter molecules (Rudin, 2005). Alternatively, physiological coupling might be used as the amplifier, i.e. a receptor-mediated physiological response would be mapped as a pseudo-molecular readout. Pharmacological functional MRI (fMRI) experiments, which map the hemodynamic response elicited by ligand–receptor interactions, fall in this last category (Chen et al., 1997;Rudin, 2005).

A key element of a molecular imaging probe is the reporter moiety, which emits photons that can be detected from outside the body. Nuclear imaging reporters are meta-stable nuclei that either directly emit gamma photons or generate gamma photons in a secondary process in the case of positron-emitting nuclei. The most relevant PET isotopes for imaging the CNS are 11C with a half-life of 20 min and 18F with a half-life of 110 min. The longer half-life of 18F is attractive as labeled compounds do not have to be synthesized on site, i.e. there is no need for a cyclotron next to the PET scanner. This advantage is counterbalanced by the fact that most CNS ligands do not contain fluorine; hence fluorination will affect the properties of the compound such as the target affinity or pharmacokinetic. Labeling with 11C in any case needs an on-site cyclotron.

Fluorescence signals are generated by (i) organic dyes characterized by an aromatic system (π-electron system) that is delocalized over several atomic centers, (ii) quantum dots that consist of a quantum confined semiconductor core encapsulated by a semiconductor shell to enhance the quantum confinement (Yoffe, 2001) and that are coated by an organic matrix to achieve biocompatibility (Watson et al., 2003), and (iii) fluorescent proteins such as green fluorescent protein and its longer wavelength analogues (Shaner et al., 2005). In all cases, the fluorescence wavelength is largely governed by the same parameter(s), the dimension(s) of the confinement of the electron involved in the absorption–emission process. For dyes and fluorescent proteins this is the dimension of the delocalized π-electron system of the fluorophor [first described as the ‘particle-in-a-box’ model (Kuhn, 1948)] and for quantum dots the dimensions of the crystal core until the diameter has reached the so-called Bohr radius for the electron–hole pair generated in the excited state (Yoffe, 2001). For bioluminescence, the underlying principle is the same except that the excited state is achieved through an enzyme-catalysed chemical reaction and not by absorption of a light photon. An attractive feature of fluorescence is that it can be modulated by altering the physicochemical environment of the fluorophor. This can lead to effects such as fluorescence resonance energy transfer (Foerster, 1948), alteration of the fluorescence quantum yield (quenching/dequenching) or a shift in fluorescence wavelengths. The susceptibility of the fluorescence parameter to the molecular environment constitutes the basis of so-called activity-based or activatable probes (Tung et al., 2000). Fluorescence is a sensitive method; depending on the depth of the fluorescent source in the tissue, concentrations of the order of 10−7–10−10 m can be detected.

The large majority of MRI probes are based on paramagnetic or superparamagnetic moieties. These contrast agents contain one or several unpaired electrons, the magnetic moment of which is 650 times larger than that of protons. The presence of these compounds in tissue locally affects the magnetic field and causes alterations in the relaxation properties of the water protons that are observed in an MRI experiment. Paramagnetic probes are typically complexes of gadolinium(III) (Gd) containing seven unpaired electrons, in which the metal ion is coordinated by a chelating ligand in order to reduce its toxicity (e.g. as GdDTPA or GdDOTA). Superparamagnetic probes are iron oxide nanoparticles, containing hundreds or thousands of iron atoms, exerting a high magnetic moment. Biocompatibility is achieved by decorating the iron core with an organic coating, e.g. dextrane. The magnitude of the contrast agent-induced changes in proton relaxation times depends on the molar relaxivity of the contrast agents (a measure of their efficiency) and their concentration in tissue. Detection ranges are from 10−6 to 10−4 m. Major efforts are being made to increase the molar relaxivities of contrast agents or to use amplification strategies to increase the probe signal.

Visualizing receptor expression and receptor occupancy

  1. Top of page
  2. Abstract
  3. Multimodal neuroimaging in rodents: general aspects
  4. Molecular and cellular imaging: the tools
  5. Visualizing receptor expression and receptor occupancy
  6. Imaging receptor function
  7. Challenges in (molecular) neuroimaging of rodents
  8. Conclusions
  9. Acknowledgements
  10. References

In view of the potential role of CNS receptors as drug targets, the availability of reporter systems suited for non-invasive quantitative mapping of both receptor distribution and binding capacity would be of great value for diagnosis and therapy evaluation. At present, PET ligands have been developed for a considerable number of CNS receptors (for reviews see Silverman & Melega, 2004; Rudin, 2005). We will illustrate the design principles by discussing PET tracers for metabotropic glutamate receptors (mGluRs).

The mGluRs are glycoprotein-coupled receptors modulating glutamatergic neurotransmission in the CNS. At present, eight mGluR subtypes have been identified. We will focus our discussion on mGluR5, which exerts its function through signal transduction via coupling to phospholipase C. Excessive activation of mGluR5 is associated with a number of psychiatric disorders such as anxiety, depression and schizophrenia, as well as with neurodegenerative diseases such as Parkinson’s disease. Although there is no clinically validated PET tracer targeting mGluR5 available as yet, several potential ligands for mGluR5 are currently under evaluation (Hamill et al., 2005; Patel et al., 2005; Yu et al., 2005a; Ametamey et al., 2006). Their design is based on molecular scaffolds derived from drug candidates that are known to interact with mGluR5 with medium to high affinity. All of these compounds contain the ethynyl core fragment connected to two aromatic rings. An example of this series is 2-methyl-6-(phenylethynyl)-pyridine (MPEP), which has been used as a template for the compound [11C]-ABP688 (Fig. 2). High accumulation of this PET tracer was observed in mGluR5-rich brain regions such as the cerebral cortex, hippocampus and striatum of mice, whereas the radioactivity uptake in areas largely devoid of the target receptor such as the cerebellum was low (Fig. 2; Ametamey et al., 2006). Ligand specificity was demonstrated in autoradiography experiments using mGluR5 knockout mice, which displayed no retention of the tracer in CNS structures. In line with this, target-specific ligand binding was significantly reduced by pretreatment of the animals with the unlabeled (‘cold’) mGluR5 antagonist ligand 2-methyl-6-(3-methoxyphenyl)ethynyl-pyridine (M-MPEP). The extent of the reduction of the local activity allows estimation of the receptor occupancy by the ‘cold’ drug candidate. [The receptor occupancy of a drug is derived from a competition experiment of the PET tracer with the cold (‘cold’ in this context means non-radioactive) drug. Analysis of the tissue activity curve for a radio tracer as a function of the concentration allows estimation of the maximum binding capacity Bmax of the receptor R, Bmax corresponding to the total concentration of accessible receptor molecules. Following pretreatment with a ‘cold’ drug L, the maximum binding capacity for the radio tracer compound is reduced to inline image, where [RL] indicates the concentration of the receptor–drug complex. The receptor occupancy RO, i.e. the fraction of receptors occupied by the drug, is obtained as inline image.] The quantitative assessment of receptor distribution and drug–receptor interactions in the brain in vivo is an attractive tool for the development of CNS drugs. A frequently quoted example demonstrating the utility of receptor occupancy studies relates to the development of antagonists of the dopamine D2 receptor. Dopamine D2 receptor density is increased in a majority of schizophrenia patients and, hence, inhibition of dopamine D2 receptor-induced signaling constitutes an attractive pharmacological strategy. However, it was observed that there is a narrow range between doses required for achieving therapeutic efficacy and doses leading to extrapyramidal side-effects. PET studies using the dopamine D2 receptor tracer [11C]-raclopride revealed that a minimal striatal receptor occupancy of 60% is required for an antipsychotic response, whereas occupancies exceeding 80% lead to significant side-effects (Farde et al., 1988; Tauscher & Kapur, 2001). Hence, quantitative PET analysis was found to be essential for optimizing the dosing of these patients.

image

Figure 2.  Imaging receptor distribution in mouse brain. Images represent the distribution of the radiolabeled ligand [11C]-ABP688 targeting the mGluR5 in rodent brain. (a) Structure of the mGluR5 specific ligand 2-methyl-6-(phenylethynyl)-pyridine (MPEP) and the PET ligand [11C]-ABP688 derived from this template. (b) Autoradiography of rat brain indicates high activity in the prefrontal cortex, hippocampus and striatum. (c) Mouse brain autoradiography indicates ligand specificity. No uptake of [11C]-ABP688 is observed in mGluR5 knockout (lower panel) as compared with wild-type (upper panel) mice. (d) In-vivo mouse PET images reveal high cortical uptake but also significant extracranial uptake (upper panel). Coadministration of the cold 2-methyl-6-(3-methoxyphenyl)ethynyl-pyridine (M-MPEP) diminishes the cerebral activity of the radiotracer, whereas the extracranial signal contributions are still observed, which is indicative of their unspecific nature (lower panel). The in-vivo images illustrate two inherent problems in target-specific imaging: the spatial resolution may not allow unambiguous assignment of signals to anatomical structures and there may be significant signal contributions due to unbound ligands, which will confound information on ligand–receptor interactions. [Images kindly provided by Drs M. Honer and S. Ametamey, ETH Zurich, Switzerland, see also Ametamey et al. (2006).]

Download figure to PowerPoint

Similar target-specific imaging approaches are currently developed as diagnostic tools for the visualization of disease-specific molecular moieties such as protein aggregates associated with neurodegenerative (Alzheimer’s and Huntington’s diseases) or prion (Soto & Estrada, 2008) disease. This will be illustrated for Alzheimer’s disease. The formation of parenchymal β-amyloid peptide aggregates (senile plaques) is a pathological hallmark of Alzheimer’s disease. Demonstration of plaque deposition and quantitative assessment of the plaque burden in patients would be of high diagnostic relevance and critical for patient management and for evaluating the efficacy of disease-modifying therapy (e.g. plaque-reducing therapy). It is therefore not surprising that major efforts are ongoing in developing imaging approaches for assessing this characteristic feature of the pathology using imaging agents that specifically target aggregated β-amyloid peptides. From a clinical perspective, PET approaches are most promising. A variety of potential PET tracers for β-amyloid are currently in clinical evaluation and a number of PET studies have demonstrated the value of such tracers (Klunk et al., 2004; Agdeppa et al., 2003;Shoghi-Jadid et al., 2002; Verhoeff et al., 2004; Kudo, 2006). For rodent models of the disease, e.g. for studies in genetically engineered mice, fluorescent imaging in combination with stable fluorescent probes targeting β-amyloid peptides has been developed as an alternative (see Fig. 1; Bacskai et al., 2003; Hintersteiner et al., 2005; Sigurdson et al., 2007).

Imaging receptor function

  1. Top of page
  2. Abstract
  3. Multimodal neuroimaging in rodents: general aspects
  4. Molecular and cellular imaging: the tools
  5. Visualizing receptor expression and receptor occupancy
  6. Imaging receptor function
  7. Challenges in (molecular) neuroimaging of rodents
  8. Conclusions
  9. Acknowledgements
  10. References

Quantitative assessment of expression levels of the molecular target (target validation) and of the drug–target interaction is essential but not sufficient for demonstration of disease relevance and therapeutic efficacy. It is critical to also evaluate the functional implications of the drug–target interaction.

Signaling pathways

The development of imaging assays to visualize intracellular signaling is attractive for multiple reasons (Rudin, 2008). The analysis of signal transduction in an intact individual with all regulatory processes in place will provide relevant mechanistic information. Multiple membrane receptors/receptor systems use the same signaling pathways. Hence, assessing flux through signaling cascades might be used as a generic tool. It is conceivable that aberrations in cellular signaling would constitute a more sensitive marker of pathology than altered expression levels of receptors.

The activation of molecular pathways can be imaged by assessing either (i) key molecules involved in signal transduction or (ii) protein–protein interactions (Fig. 3). Specific molecular players of the signaling cascade can be probed using a target-specific ligand as outlined in the previous section. This has been demonstrated in vitro for ligands targeting caspases (death proteases) as central constituents of the apoptotic pathway (Berger et al., 2006). The inherent problem in pathway imaging is the fact that imaging probes have to be delivered intracellularly. Alternatively, intracellular targets can be assessed using reporter gene strategies; in this case, the molecular target is commonly expressed as a fusion protein comprising a reporter moiety such as a fluorescent or bioluminescent protein. For example, the oxygen-dependent degradation domain of hypoxia-inducible factor 1α, which is responsible for the oxygen-dependent regulation of the protein, has been fused to firefly luciferase. Stabilization of the oxygen-dependent degradation domain-luciferase fusion protein under hypoxic conditions can thus be monitored in vivo in mice as a bioluminescent signal that is generated following the administration of the enzyme substrate d-luciferin (Safran et al., 2006; Harada et al., 2007), yielding information on hypoxic signaling. More recently, hypoxia signaling has been monitored longitudinally in murine colon tumor xenografts at various levels of the transduction cascade, PET for assessing tumor hypoxia and bioluminescence imaging for monitoring both the stabilization of full-length hypoxia-inducible factor 1α and the induction of hypoxia-inducible factor downstream genes (Lehmann et al., 2009).

image

Figure 3.  Imaging signal transduction in vivo. (a) Two strategies have been developed in cellular systems and have been made available for in vivo imaging in mice: (i) imaging of specific pathway molecules using reporter genes (expression of target–reporter fusion proteins or coexpression using internal ribosomal entry sites (IRES)) or targeting endogenous levels of pathway molecules using low molecular mass (LMM) ligands and (ii) targeting of protein–protein interactions using established cellular assays. (b) Imaging the interaction of two proteins A and B using fluorescence resonance energy transfer (FRET) between a fluorescence donor and an acceptor (Acc) molecule (left panel), the two-hybrid system in which the protein interaction leads to the constitution of a transcription factor consisting of a DNA binding domain (DBD) and an activator domain (AD) that drives the expression of a reporter gene (middle panel), and the protein complementation assay with reconstitution of a split reporter (N-Rep and C-Rep) upon interaction of proteins A and B (right panel).

Download figure to PowerPoint

Protein–protein interactions are central elements of information propagation along signaling cascades. A number of assays have been developed to study these molecular events in cellular systems. They essentially sense the proximity of the two (or more) interacting partners. All of these assays depend on genetically modified cells or organisms and thus constitute research tools. Several of these assays such as the two-hybrid system (Fields & Song, 1989), protein fragment complementation assay (Michnick, 2001) and protein splicing have been adapted for in vivo imaging in mice. The two-hybrid system for studying the interaction between two proteins A and B involves two fusion proteins: protein A fused to the DNA binding domain of a transcription factor and protein B linked to an activator domain. Upon interaction of the proteins A and B, the DNA binding domain and activator domain are brought in sufficient proximity to execute their function as transcription factor, activating the expression of a reporter gene, which can then be detected by imaging. The reporter systems used in in vivo imaging studies were Herpes simplex virus-1 thymidine kinase or Renilla luciferase. PET in combination with 18F-labeled thymidine analogues as enzyme substrates was used for assessing Herpes simplex virus-1 thymidine kinase activity (Luker et al., 2002), whereas bioluminescence imaging revealed the activity of the luciferase (Ray et al., 2002).

The protein fragment complementation assay is based on a split reporter protein, which is reconstituted upon interaction of the target proteins A and B to form a functional reporter system (Michnick, 2001). Again two fusion proteins have to be engineered: protein A with the C-terminal domain of the reporter and protein B with the respective N-terminal domain. The protein fragment complementation assay strategy using firefly luciferase as split reporter has been applied in vivo in mice (Paulmurugan et al., 2002). A slightly modified version of the protein fragment complementation assay is protein splicing. In this case, the reconstituted reporter contains an intein sequence, which is eliminated through protein splicing to form the reporter protein in its proper conformation (Paulmurugan & Gambhir, 2003).

At present, the in vivo imaging of signaling pathways is still in its infancy and the reports published to date are proof-of-feasibility studies (e.g. using genetically engineered cells implanted in mice). Assays based on genetically encoded reporter systems will remain confined to experimental research in cells and animals. Targeting of intracellular targets with low molecular mass compounds is potentially translatable into the clinic but the need for intracellular delivery of such probes constitutes a major hurdle. Nevertheless, even as a preclinical tool, pathway imaging will provide valuable information for basic researchers and drug developers. For example, it has been demonstrated that drug-induced protein–protein interaction can be visualized in the living mouse (Paulmurugan et al., 2004).

Receptor-mediated physiological response

The underlying hypothesis of pharmacological fMRI is that the ligand–receptor interaction will prompt a neuronal response, which through neurovascular coupling translates into a hemodynamic response. As a consequence, cerebral blood flow and cerebral blood volume (CBV) will change. As the oxygen extraction efficiency decreases with increasing cerebral blood flow, venous blood will be better oxygenated giving rise to the so-called blood oxygen level-dependent (BOLD) contrast (Ogawa et al., 1990). For a detailed discussion of the neurophysiological correlates underlying the BOLD signal (or more generally the metabolic and hemodynamic changes associated with neuronal activity), the reader is referred to a review by Logothetis (2008). All of these parameters (cerebral blood flow, CBV and BOLD contrast) can be assessed using fMRI techniques. CBV changes are commonly measured using intravascular superparamagnetic contrast agents such as iron oxide nanoparticles, which increase the relaxation rate R2* (and R2) due to increased susceptibility gradients at the vessel–tissue interface. Due to the long blood half-life of the nanoparticles (one to several hours depending on their size and surface coating), a steady-state concentration will be reached a few minutes after the bolus administration. Under these conditions, any changes in the local relaxation rate inline image can then be attributed to changes in local CBV. [The change in the relaxation rates in a voxel at location inline image and time t following the administration of a contrast agent is proportional to the amount of contrast agent in that voxel, i.e. to the product of its concentration times the local blood volume, inline image, the proportionality factor r2 being the molar relaxivity of the contrast agent (in mmol/s). Assuming the contrast agent concentration to be in steady state, the change in the relaxation rate becomes proportional to changes in inline image, inline image.] Measurements of CBV changes following pharmacological stimulation offer advantages as changes in relative CBV exceed changes in the BOLD intensity, i.e. CBV measurements yield better sensitivity. Pharmacological fMRI has been applied to characterize the functional response elicited by receptor stimulation for a number of CNS receptors (Table 1). In some instances, the CBV changes in response to drug administration have been shown to correlate with the local densities of CNS receptors as illustrated for GABAA receptor stimulation by bicuculline (Fig. 4, adapted from Reese et al., 2000).

Table 1.   Selected pharmacological fMRI studies in rodent brain
Neuro-transmitterMolecular target*Ligandf MRI methodSpeciesReferences
  1. BOLD, blood oxygen level-dependent; CBF, cerebral blood flow; CBV, cerebral blood volume; MEMRI, manganese-enhanced MRI.

  2. *Receptor/enzyme.

Acetyl-cholineAcetylcholine esteraseRivastigmineCBVRatRausch et al. (2005)
DopamineDopamine release from vesiclesAmphetamineBOLDRatChen et al. (1997)
CBVRatSchwarz et al. (2007)
MEMRIRatGozzi et al. (2008b)
DopamineD2 receptorQuinpiroleCBVRatChen et al. (2005)
EticloprideCBFRatChen et al. (2005)
DopamineDopamine transporter2β-carbomethoxy-3 beta-(4-fluoro-pheny)tropaneBOLDRatChen et al. (1997)
GlutamateNMDA (N-methyl-D-aspartate) receptorPhencylidineCBVRatGozzi et al. (2008a)
Serotonin5HT1A8-hydroxy-2-(di-n-propylamino) tetralineCBVRatScanley et al. (2001)
GABAGABAABicucullineCBVRatReese et al. (2000)
CBVMouseMueggler et al. (2001)
image

Figure 4.  Analysis of the physiological response caused by ligand–receptor interaction using pharmacological fMRI. Administration of the GABAA antagonist bicuculline causes increases in CBV in rat brain due to neurovascular coupling (a). Local CBV changes correlate with density of GABAA receptors in respective brain areas (b) (adapted from Reese et al., 2000).

Download figure to PowerPoint

The pharmacological fMRI signal can be measured throughout the brain with a typical temporal resolution of 1 min. This allows the correlation of fMRI signal profiles measured in different brain regions and thereby the elucidation of functional networks as demonstrated for the limbic system in assessing the effects of dopaminergic drugs (Schwarz et al., 2007). Pharmacological fMRI has also been applied to characterize pathological conditions. Examples comprise GABAergic stimulation in models of neurodegeneration (APP23 transgenic mice as models of cerebral amyloidosis) (Mueggler et al., 2001) and transient focal cerebral ischemia (Reese et al., 2002) or dopaminergic stimulation in models of Parkinson’s disease (Chen et al., 1997). It has been demonstrated that functional measures may constitute a sensitive indicator of pathology or stress to the tissue. Although brain structural and static CBV appeared normal in a transient focal cerebral ischemia model in the rat following successful reperfusion, the dynamic CBV response in the affected territory remained compromised following functional activation by the GABAA antagonist bicuculline (Reese et al., 2002).

A number of potential pitfalls associated with pharmacological fMRI studies must be mentioned. (i) The overwhelming majority of fMRI studies are carried out in anesthetized animals. The anesthetics potentially interfere with the ligand–receptor interaction to be studied. Careful experimental design allows the interference of anesthesia to be minimized but it cannot be completely excluded. Alternatives are experiments in awake, immobilized animals, where activation of the stress axis would constitute a potential confound, or the use of manganese-enhanced MRI (MEMRI) (Yu et al., 2005b). (ii) Pharmacological fMRI is based on the integrity of the neurovascular coupling, which might not be the case when studying pathologic conditions. (iii) A drug might interact with a vascular receptor(s) and prompt a cerebrovascular response that is not mediated through neuronal activity. The regional pattern of the observed hemodynamic response might help to discriminate pure vascular from neuron-mediated signals. (iv) Neurons may project to distant regions in the brain, i.e. pharmacologically induced fMRI signals may not necessarily reflect receptor distributions. In the last case, the result would still qualify as a pharmacodynamic readout but not as a molecular image reflecting receptor distribution. Some of these issues were addressed in a study of Schwarz et al. (2004), who compared cocaine-induced CBV changes with the local concentration of cocaine and dopamine derived from microdialysis, which measures extracellular compound levels. In the prefrontal cortex and striatum, cocaine and dopamine levels were found to correlate but the CBV response was distinctly slower. This indicates different temporal characteristics of the receptor-bound cocaine fraction prompting the functional response, whereas the extracellular concentrations essentially mimic the drug’s pharmacokinetic properties. In the motor cortex, dopamine levels remained unaltered, whereas prominent changes in cocaine levels and CBV were observed, again showing different temporal behavior. Obviously the motor cortex CBV response is not mediated by local dopamine levels and can be attributed to either other neurotransmitters or afferent projections from dopaminergic subcortical structures.

At present, fMRI using pharmacological stimulation is a widely used approach to evaluate potential CNS drugs. The method demonstrates (i) the CNS activity of a drug candidate (pharmacodynamic readout), (ii) the time resolution of the fMRI experiment allows the establishment of temporal correlations of regional responses and thus contributes to the elucidation of networks, and (iii) pharmacological fMRI is potentially translatable into the clinic (Wise & Tracey, 2006). In clinical studies the drug-induced hemodynamic responses are commonly measured using the BOLD fMRI as intravascular superparamagnetic contrast agents are not approved for this application.

Challenges in (molecular) neuroimaging of rodents

  1. Top of page
  2. Abstract
  3. Multimodal neuroimaging in rodents: general aspects
  4. Molecular and cellular imaging: the tools
  5. Visualizing receptor expression and receptor occupancy
  6. Imaging receptor function
  7. Challenges in (molecular) neuroimaging of rodents
  8. Conclusions
  9. Acknowledgements
  10. References

Non-invasive neuroimaging in rodents is highly attractive for neuroscience researchers. The combination of various modalities allows the study of structural, physiological, metabolic, cellular and molecular processes in the normal brain and under pathological conditions in a temporospatially resolved manner in the intact animal. The non-invasive nature of the approach enables longitudinal studies, which is of relevance, e.g. for monitoring disease progression in individuals, e.g. in models of neurodegenerative diseases or models thereof, or to monitor the efficacy of therapeutic interventions in an individual allowing comparisons with the pretreatment condition, which is attractive from both a scientific and ethical point of view. Finally, related or even identical methods can be used in experimental and clinical situations, which should facilitate the translation of research data into clinical practice.

However, there are a number of technological challenges that need to be addressed. Small animal imaging demands high spatial resolution. Structural imaging methods provide resolutions of the order of 50 μm, which appears reasonable in view of the physiological motion in an anesthetized, immobilized living subject. The relatively high temporal resolution required for fMRI degrades spatial resolution significantly and typical voxel dimensions are 200–500 μm. Clearly, technological improvements are required in rodent fMRI. Molecular imaging methods such as PET and fluorescence imaging, characterized by high sensitivity, suffer from even more inferior resolution. Frequently the anatomical origin of the signal is not obvious from the molecular imaging data set alone. This has prompted the development of hybrid technologies, such as computerized tomography-PET, MRI-PET (Judenhofer et al., 2008) or computerized tomography-fluorescence molecular tomography (Hyde et al., 2009) combining anatomical and molecular information. The benefit is twofold as hybrid imaging enables better allocation of molecular information to organ structures and improves the reconstruction of molecular data by providing prior knowledge. Additional hybrid methods are currently under development.

The development of sensitive target-specific imaging probes that penetrate the CNS is probably the most challenging problem to be addressed in molecular neuroimaging. Although low molecular weight PET probes will certainly maintain their dominant role in the field, the delivery of fluorescent or MRI agents, which are in bulky and frequently polar compounds, remains an issue. However, these agents offer attractive features. For example, the biophysical properties of such probes can be modulated through the interaction with their molecular target allowing the discrimination of the unprocessed (unbound) from the processed (e.g. bound) state. This greatly improves the signal-to-background ratio. An approach to circumvent delivery issues is to transiently open the blood–brain barrier for drug delivery, e.g. using high intensity focused ultrasound (Hynynen, 2008), or to use reporter gene strategies.

The extraction of quantitative biological information from imaging data sets includes a number of challenges. The biophysical processes underlying the imaging experiment must be understood at a quantitative level for proper data analysis. For example, the intensity derived from MRI data sets depends on the specific measurement conditions and does not directly reflect the concentrations of analytes/contrast agents. Similarly, the fluorescence signal intensity measured at the surface depends on the depth of the fluorescent source embedded in tissue. Quantitative analysis requires knowledge of the three-dimensional distribution of the fluorophor, its fluorescent quantum yield and the scattering and absorption properties of the tissue. A second challenge arises from the fact that a volume element (voxel) of an image contains multiple compartments, i.e. multiple cell types, intracellular and extracellular spaces, vasculature, etc., which all may interact with an imaging agent through either specific or non-specific processes. Moreover, unbound probes may still remain in the circulation. The signal observed is the weighted contribution of all of these compartments. Extensive modeling is required to differentiate individual compartments and to extract quantitative information, e.g. on the interaction of the probe with a specific receptor expressed by a specific cell type. Quantitative PET makes extensive use of multi-compartment models (Slifstein & Laruelle, 2001), whereas for other imaging modalities such analysis approaches are far less established. (In multi-compartment modeling a compartment is considered to be a physiological or biological space, in which the concentration of the tracer is considered to be homogeneous at any time. Tracer transfer between compartments is described by first-order kinetic exchange processes. By fitting model predictions to the experimental data, concentration profiles for the individual compartments can be estimated.) However, even in PET, the proper quantitative assessment of local tracer concentration remains an issue. PET measures local activities, which constitute the weighted sum of the various tissue compartments (plasma fraction, unbound fraction and bound fraction in the various compartments). The art of modeling is to develop a multi-compartment model that is sufficiently comprehensive to capture the essentials of the biological system but simple enough to be mathematically treatable in view of the limited amount of experimental data available (Gambhir, 2004). Additional independent information may be gathered, e.g. from plasma samples, which allows the determination of the plasma activity as a function of time. However, for obvious reasons, plasma sampling is limited in small rodents. Similarly, activity data from tissue (brain) regions devoid of the targeted receptor, which therefore do not comprise a bound tracer fraction, may be used to better confine the quantification problem. In the latter, it is frequently assumed that parameters not involved in receptor binding are identical for the various brain substructures. In view of the inherent complexity of tissue and the limited amount of experimental data available, the modeling problem is inherently underdetermined and approximations have to be made in order to come up with quantitative data. Hence, major efforts are required to advance tissue models, which are key for the accurate determination of local tracer concentrations. A third challenge is to translate the local concentration of an imaging agent into meaningful biological information. For example, the extravasation of contrast agent reflects vascular permeability, which by itself may be regarded an indicator of angiogenesis, as the permeability of newly formed vessels is greater than that of mature vessels. Modeling may help to link imaging information to the molecular and physiological process underlying angiogenesis (Lloyd et al., 2008).

Conclusions

  1. Top of page
  2. Abstract
  3. Multimodal neuroimaging in rodents: general aspects
  4. Molecular and cellular imaging: the tools
  5. Visualizing receptor expression and receptor occupancy
  6. Imaging receptor function
  7. Challenges in (molecular) neuroimaging of rodents
  8. Conclusions
  9. Acknowledgements
  10. References

Non-invasive molecular imaging techniques are rapidly evolving and will lead to novel diagnostic assays of improved sensitivity and specificity as well as to the development of novel readouts to assess the efficacy of therapeutic interventions. Using specific readouts that are based on target-specific reporter probes or genetically encoded reporter molecules, molecular imaging allows the study of gene expression as well as the function of gene products in the living organism. Applications of molecular imaging extend essentially in all fields of biomedical research.

With regard to molecular (neuro)imaging, the passage of the intact blood–brain barrier by target-specific probes constitutes a critical hurdle. At present, brain penetration is predominantly achieved using low molecular weight probes labeled with PET nuclides that are derived from CNS-active compounds. Receptor distribution and occupancy studies based on PET are established tools both clinically and in experimental animal research (Silverman & Melega, 2004). Shuttling large reporter systems such as MRI contrast agents or fluorescent dyes into the CNS is still a major issue and many of the probes developed so far target receptors on the endovascular side such as selectins, adhesion molecules and integrins that are expressed by the activated endothelium frequently encountered in the context of brain pathologies (tumors and neuroinflammation). The use of genetically encoded reporters circumvents the probe delivery problem but is obviously confined to experimental systems and not translatable into the clinic. Reporter gene assays have been extensively used in basic neuroscience. Molecular imaging allows such studies to be carried out non-invasively in a longitudinal manner. The functional consequences of receptor modulation can also be assessed using a metabolic or physiological readout as a surrogate. Pharmacological fMRI provides this kind of information and constitutes a ‘molecular’ imaging strategy that can be readily translated to a clinical setting.

In order to benefit from the potential of molecular imaging, substantial methodological developments are required. The two most important areas are (i) imaging probes that penetrate the CNS and generate a signal with high sensitivity and (ii) tools for quantitative data analysis that translate imaging data into relevant biological information. Major developments are to be expected in both fields.

Abbreviations
BOLD

blood oxygen level-dependent

CBV

cerebral blood volume

CNS

central nervous system

fMRI

functional magnetic resonance imaging

Gd

gadolinium(III)

mGluR

metabotropic glutamate receptor

MRI

magnetic resonance imaging

PET

positron emission tomography

References

  1. Top of page
  2. Abstract
  3. Multimodal neuroimaging in rodents: general aspects
  4. Molecular and cellular imaging: the tools
  5. Visualizing receptor expression and receptor occupancy
  6. Imaging receptor function
  7. Challenges in (molecular) neuroimaging of rodents
  8. Conclusions
  9. Acknowledgements
  10. References
  • Agdeppa, E.D., Kepe, V., Liu, J., Small, G.W., Huang, S.C., Petric, A., Satyamurthy, N. & Barrio, J.R. (2003) 2-Dialkylamino-6-acylmalononitrile substituted naphthalenes (DDNP analogs): novel diagnostic and therapeutic tools in Alzheimer’s disease. Mol. Imaging Biol., 5, 404417.
  • Ametamey, S.M., Kessler, L.J., Honer, M., Wyss, M.T., Buck, A., Hintermann, S., Auberson, Y.P., Gasparini, F. & Schubiger, P.A. (2006) Radiosynthesis and preclinical evaluation of 11C-ABP688 as a probe for imaging the metabotropic glutamate receptor subtype 5. J. Nucl. Med., 47, 698705.
  • Bacskai, B.J., Hickey, G.A., Skoch, J., Kajdasz, S.T., Wang, Y., Huang, G.F., Mathis, C.A., Klunk, W.E. & Hyman, B.T. (2003) Four-dimensional multiphoton imaging of brain entry, amyloid binding, and clearance of an amyloid-beta ligand in transgenic mice. Proc. Natl Acad. Sci. USA, 100, 1246212467.
  • Berger, A.B., Witte, M.D., Denault, J.B., Sadaghiani, A.M., Sexton, K.M., Salvesen, G.S. & Bogyo, M. (2006) Identification of early intermediates of caspase activation using selective inhibitors and activity-based probes. Mol. Cell, 23, 509521.
  • Chen, Y.C.I., Galpern, W.R., Brownell, A.L., Matthews, R.T., Bogdanov, M., Isacson, O., Keltner, J.R., Beal, M.F., Rosen, B.R. & Jenkins, B.G. (1997) Detection of dopaminergic neurotransmitter activity using pharmacologic MRI: correlation with PET, microdialysis and behavioural data. Magn. Reson. Med., 38, 389398.
  • Chen, Y.C.I., Choi, J.K., Andersen, S.L., Rosen, B.R. & Jenkins, B.G. (2005) Mapping dopamine D2/D3 receptor function using pharmacological magnetic resonance imaging. Psychopharmacology (Berl.), 180, 705715.
  • Farde, L., Wiesel, F.A., Halldin, C. & Sedvall, G. (1988) Central D2-dopamine receptor occupancy in schizophrenic patients treated with antipsychotic drugs. Arch. Gen. Psychiatry, 45, 7176.
  • Fields, S. & Song, O. (1989) A novel genetic system to detect protein-protein interactions. Nature, 340, 245246.
  • Foerster, T. (1948) Intermolecular energy migration and fluorescence. Ann. Phys. Leipzig, 2, 5575.
  • Gambhir, S.S. (2004) Quantitative assay development for PET. In Phelps, M.E. (ed.), PET – Molecular Imaging and Its Biological Applications. Springer-Verlag, New York, pp. 125216.
  • Gozzi, A., Large, C.H., Schwarz, A., Bertani, S., Crestan, V. & Bifone, A. (2008a) Differential effects of antipsychotic and glutamatergic agents on the phMRI response to phencyclidine. Neuropsychopharmacology, 33, 16901703.
  • Gozzi, A., Schwarz, A., Crestan, V. & Bifone, A. (2008b) Pharmacological manganese-enhanced MRI (MEMRI) without osmotic breakdown of the blood brain barrier. Proc. Intl. Soc. Mag. Reson. Med., 16, 2314.
  • Hamill, T.G., Krause, S., Ryan, C., Bonnefous, C., Govek, S., Seiders, T.J., Cosford, N.D., Roppe, J., Kamenecka, T., Patel, S., Gibson, R.E., Sanabria, S., Riffel, K., Eng, W., King, C., Yang, X., Green, M.D., O’Malley, S.S., Hargreaves, R. & Burns, H.D. (2005) Synthesis, characterization, and first successful monkey imaging studies of metabotropic glutamate receptor subtype 5 (mGluR5) PET radiotracers. Synapse, 56, 205216.
  • Harada, H., Kizaka-Kondoh, S., Itasaka, S., Shibuya, K., Morinibu, A., Shinomiya, K. & Hiraoka, M. (2007) The combination of hypoxia-response enhancers and an oxygen-dependent proteolytic motif enables real-time imaging of absolute HIF-1 activity in tumor xenografts. Biochem. Biophys. Res. Commun., 360, 791796.
  • Hintersteiner, M., Frey, P., Gremlich, H.U., Kinzy, W., Kneuer, R., Neumann, U., Rudin, M., Staufenbiel, M. & Wiederhold, K.H. (2005) In vivo detection of amyloid deposits by near-infrared fluorescence imaging using a novel oxazine derivative as contrast agent. Nat. Biotechnol., 23, 577583.
  • Hyde, D., DeKleine, R., MacLaurin, S.A., Miller, E., Brooks, D.H., Krucker, T. & Ntziachristos, V. (2009) Hybrid FMT-CT method for in vivo imaging of amyloid-β plaques in a murine model forAlzheimer’s disease. Neuroimage, 44, 13041311.
  • Hynynen, K. (2008) Ultrasound for drug and gene delivery to the brain. Adv. Drug Deliv. Rev., 60, 12091217.
  • Ishimaru, A. (1997) Wave Propagation and Scattering in Random Media. IEEE Press, New York.
  • Judenhofer, M.S., Wehrl, H.F., Newport, D.F., Catana, C., Siegel, S.B., Becker, M., Thielscher, A., Kneilling, M., Lichy, M.P., Eichner, M., Klingel, K., Reischl, G., Widmaier, S., Röcken, M., Nutt, R.E., Machulla, H.J., Uludag, K., Cherry, S.R., Claussen, C.D. & Pichler, B.J. (2008) Simultaneous PET-MRI: a new approach for functional and morphological imaging. Nat. Med., 14, 459465.
  • Klunk, W.E., Engler, H., Nordberg, A., Wang, Y., Blomqvist, G., Holt, D.P., Bergström, M., Savitcheva, I., Huang, G.F., Estrada, S., Ausén, B., Debnath, M.L., Barletta, J., Price, J.C., Sandell, J., Lopresti, B.J., Wall, A., Koivisto, P., Antoni, G., Mathis, C.A. & Långström, B. (2004) Imaging brain amyloid in Alzheimer’s disease with Pittsburgh Compound-B. Ann. Neurol., 55, 306319.
  • Kudo, Y. (2006) Development of amyloid imaging PET probes for an early diagnosis of Alzheimer’s disease. Minim. Invasive Ther. Allied Technol., 15, 209213.
  • Kuhn, H. (1948) Qunatum mechanical theory of light absorption of organic dyes and similar compounds. J. Chem. Phys., 17, 11981212.
  • Lehmann, S., Stiehl, D.P., Honer, M., Dominietto, M., Keist, R., Kotevic, I., Wollenick, K., Ametamey, S., Wenger, R.H. & Rudin, M. (2009) Longitudinal and multimodal in vivo imaging of tumor hypoxia and its downstream molecular events. Proc. Natl Acad. Sci. USA, 106, 1400414009.
  • Lloyd, B., Sczcerba, D., Rudin, M. & Szekely, G. (2008) A computational framework for modelling solid tumour growth. Philos. Transact. A Math. Phys. Eng. Sci., 366, 33013318.
  • Logothetis, N.K. (2008) What we can do and what we cannot do with fMRI. Nature, 453, 869878.
  • Luker, G.D., Sharma, V., Pica, C.M., Dahlheimer, J.L., Li, W., Ochesky, J., Ryan, C.E., Piwnica-Worms, H. & Piwnica-Worms, D.L. (2002) Noninvasive imaging of protein-protein interactions in living animals. Proc. Natl Acad. Sci. USA, 99, 69616966.
  • Michnick, S.W. (2001) Exploring protein interactions by interaction-induced folding of proteins from complementary peptide fragments. Curr. Opin. Struct. Biol., 11, 472477.
  • Mueggler, T., Baumann, D., Rausch, M. & Rudin, M. (2001) Bicuculline-induced brain activation in mice detected by functional magnetic resonance imaging. Magn. Reson. Med., 46, 292298.
  • Ntziachristos, V., Ripoll, J., Wang, L.V. & Weissleder, R. (2005) Looking and listening to light: the evolution of whole-body photonic imaging. Nat. Biotechnol., 23, 313320.
  • Ogawa, S., Lee, T.M., Kay, A.R. & Tank, D.W. (1990) Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc. Natl Acad. Sci. USA, 87, 98689872.
  • Patel, S., Ndubizu, O., Hamill, T., Chaudhary, A., Burns, H.D., Hargreaves, R. & Gibson, R.E. (2005) Screening cascade and development of potential Positron Emission Tomography radiotracers for mGluR5: in vitro and in vivo characterization. Mol. Imaging Biol., 7, 314323.
  • Paulmurugan, R. & Gambhir, S.S. (2003) Monitoring protein-protein interactions using split synthetic Renilla luciferase protein-fragment-assisted complementation. Anal. Chem., 75, 15841589.
  • Paulmurugan, R., Umezawa, Y. & Gambhir, S.S. (2002) Noninvasive imaging of protein-protein interactions in living subjects by using reporter protein complementation and reconstitution strategies. Proc. Natl Acad. Sci. USA, 99, 1560815613.
  • Paulmurugan, R., Massoud, T., Huang, J. & Gambhir, S.S. (2004) Molecular imaging of drug-modulated protein-protein interactions in living subjects. Cancer Res., 64, 21132119.
  • Phelps, M.E. (2004) PET: Molecular Imaging and its Biological Applications. Springer, New York.
  • Rausch, M., Gentsch, C., Enz, A., Baumann, D. & Rudin, M. (2005) Characterization of CNS drugs by mapping changes in regional CBV in the rat: a study paradigm allowing for automated comparison of three-dimensional data obtained multiple imaging sessions. NMR Biomed., 18, 260268.
  • Ray, P., Pimenta, H., Paulmurugan, R., Berger, F., Phelps, M.E., Iyer, M. & Gambhir, S.S. (2002) Noninvasive quantitative imaging of protein-protein interactions in living subjects. Proc. Natl Acad. Sci. USA, 99, 31053110.
  • Reese, T., Bjelke, B., Porszasz, R., Baumann, D., Bochelen, D., Sauter, A. & Rudin, M. (2000) Regional brain activation by bicuculline visualized by functional magnetic resonance imaging: time-resolved assessment of bicuculline-induced changes in local cerebral blood volume using an intravascular contrast agent. NMR Biomed., 13, 4349.
  • Reese, T., Bochelen, D., Baumann, D., Rausch, M., Sauter, A. & Rudin, M. (2002) Impaired functionality of reperfused brain tissue following short transient focal ischemia in rats. Magn. Reson. Imaging, 20, 447454.
  • Rudin, M. (2005) Molecular imaging – Basic Principles and Applications in Biomedical Research. Imperial College Press, London.
  • Rudin, M. (2008) Noninvasive imaging of receptor function: signal transduction pathways and physiological readouts. Curr. Opin. Drug Discov. Devel., 11, 606615.
  • Rudin, M. & Weissleder, R. (2003) Molecular imaging in drug discovery and development. Nat. Rev. Drug Discov, 3, 123131.
  • Safran, M., Kim, W.Y., O’Connell, F., Flippin, L., Günzler, V., Horner, J.W., Depinho, R.A. & Kaelin, W.G. Jr (2006) Mouse model for noninvasive imaging of HIF prolyl hydroxylase activity: assessment of an oral agent that stimulates erythropoietin production. Proc. Natl Acad. Sci. USA, 103, 105110.
  • Scanley, B.E., Kennan, R.P. & Gore, J.C. (2001) Changes in rat cerebral blood volume due to modulation of the 5-HT(1A) receptor measured with susceptibility enhanced contrast MRI. Brain Res., 913, 149155.
  • Schwarz, A.J., Zocchi, A., Reese, T., Gozzi, A., Garzotti, M., Varnier, G., Curcuruto, O., Sartori, I., Girlanda, E., Biscaro, B., Crestan, V., Bertani, S., Heidbreder, C. & Bifone, A. (2004) Concurrent pharmacological MRI and in situ microdialysis of cocaine reveal a complex relationship between the central hemodynamic response and local dopamine concentration. Neuroimage, 23, 296304.
  • Schwarz, A.J., Gozzi, A., Reese, T. & Bifone, A. (2007) In vivo mapping of functional connectivity in neurotransmitter systems using pharmacological MRI. Neuroimage, 34, 16271636.
  • Shaner, N.C., Steinbach, P.A. & Tsien, R.Y. (2005) A guide to choosing fluorescent proteins. Nat. Methods, 2, 905909.
  • Shoghi-Jadid, K., Small, G.W., Agdeppa, E.D., Kepe, V., Ercoli, L.M., Siddarth, P., Read, S., Satyamurthy, N., Petric, A., Huang, S.C. & Barrio, J.R. (2002) Localization of neurofibrillary tangles and beta-amyloid plaques in the brains of living patients with Alzheimer disease. Am. J. Geriatr. Psychiatry, 10, 2435.
  • Sigurdson, C.J., Nilsson, K.P., Hornemann, S., Manco, G., Polymenidou, M., Schwarz, P., Leclerc, M., Hammarström, P., Wüthrich, K. & Aguzzi, A. (2007) Prion strain discrimination using luminescent conjugated polymers. Nat. Methods, 4, 10231030.
  • Silverman, D.H.S. & Melega, W.P. (2004) Molecular imaging of biological processes with PET: evaluating biological bases of cerebral function. In Phelps, M.E. (ed.), PET – Molecular Imaging and Its Biological Applications. Springer-Verlag, New York, pp. 509583.
  • Slifstein, M. & Laruelle, M. (2001) Models and methods for derivation of in vivo neuroreceptor parameters with PET and SPECT reversible radiotracers. Nucl. Med. Biol., 28, 595608.
  • Soto, C. & Estrada, L.D. (2008) Protein misfolding and neurodegeneration. Arch. Neurol., 65, 184189.
  • Tauscher, J. & Kapur, S. (2001) Choosing the right dose of antipsychotics in schizophrenia. CNS Drugs, 15, 671677.
  • Tung, C.H., Mahmood, U., Bredow, S. & Weissleder, R. (2000) In vivo imaging of proteolytic enzyme activity using a novel molecular reporter. Cancer Res., 60, 49534958.
  • Verhoeff, N.P., Wilson, A.A., Takeshita, S., Trop, L., Hussey, D., Singh, K., Kung, H.F., Kung, M.P. & Houle, S. (2004) In-vivo imaging of Alzheimer disease beta-amyloid with [11C]SB-13 PET. Am. J. Geriatr. Psychiatry, 12, 584595.
  • Watson, A., Wu, X. & Bruchez, M. (2003) Lighting up cells with qunatum dots. BioTechniques, 34, 296303.
  • Wise, R.G. & Tracey, I. (2006) The role of fMRI in drug discovery. J. Magn. Reson. Imaging, 23, 862876.
  • Yoffe, A.D. (2001) Semiconductor quantum dots and related systems: electronic, optical, luminescence and related properties of low dimensional systems. Adv. Phys., 50, 1208.
  • Yu, M., Tueckmantel, W., Wang, X., Zhu, A., Kozikowski, A.P. & Brownell, A.L. (2005a) Methoxyphenylethynyl, methoxypyridylethynyl and phenylethynyl derivatives of pyridine: synthesis, radiolabeling and evaluation of new PET ligands for metabotropic glutamate subtype 5 receptors. Nucl. Med. Biol., 32, 631640.
  • Yu, X., Wadghiri, Y.Z., Sanes, D.H. & Turnbull, D.H. (2005b) In vivo auditory brain mapping in mice with Mn-enhanced MRI. Nat. Neurosci., 8, 961968.