Recent advances in light sources, detectors and other optical imaging technologies coupled with the development of novel optical contrast agents have enabled real-time, high resolution, in vivo monitoring of molecular targets. Noninvasive monitoring of molecular targets can help optimize photodynamic therapy (PDT) by providing the capabilities to monitor the efficacy of treatment. Our lab has developed optical imaging technologies to investigate a wide range of molecular, physiological and morphological responses to photodynamic therapy (PDT). With the idea that drug delivery to the different compartments in the tumor is an important determinant of the treatment effect, we studied drug delivery in vitro and in vivo using optical imaging tools. A molecular specific contrast agent that targets the vascular endothelial growth factor (VEGF) was developed to monitor the changes in the protein expression following PDT. We also studied the PDT-induced physiological changes in vascular permeability and metastasis with in vivo imaging.

Electron paramagnetic resonance imaging (EPRI) is one of the recent functional imaging modalities that can provide valuable in vivo physiological information on its own merit and aids as a complimentary imaging technique to MRI and PET of tissues especially with respect to in vivo pO₂ (oxygen partial pressure), redox status and pharmacology. EPR imaging mainly deals with the measurement of distribution and in vivo dynamics and redox changes using special nontoxic paramagnetic spin probes that can be infused into the object of investigation. These spin probes should be characterized by simple EPR spectra, preferably with narrow EPR lines. The line width should be reversibly sensitive to the concentration of in vivo pO₂ with a linear dependence. Several non-toxic paramagnetic probes, some particulate and insoluble and others water-soluble and infusible (by intravenous or intramuscular injection) have been developed which can be effectively used to quantitatively assess tissue redox status, and tumor hypoxia. Quantitative assessment of the redox status of tissue in vivo is important in investigating oxidative stress, and that of tissue pO₂ is very important in radiation oncology. Other areas in which EPR imaging and oxymetry may help are in the investigation of tumorangiogenesis, wound healing, oxygenation of tumor tissue by the ingestion of oxygen-rich gases, etc. The correct choice of the spin probe will depend on the modality of measurement (whether by CW or time-domain EPR imaging) and the particular physiology interrogated. Examples of the available spin probes and some EPR imaging applications employing them are presented.

Photodynamic therapy (PDT) involves the combined action of light, oxygen and a photosensitizer (PS). It offers unique control in the PS's action because the key cytotoxic agent, singlet oxygen (¹O₂), is only produced in situ upon irradiation. The ¹O₂ production can be controlled in three levels. The first level involves the judicious use of fiber optics to selectively deliver light to disease tissues. The second level is to exert control over the PS's localization by selectively delivering PS to cancer cells. The third level is to exert control of the PS's ability to generate ¹O₂ in responding to specific cancer biomarkers. Here, we present two PDT agents based on the latter two levels of ¹O₂ control. The first PDT agent "PPF" contains a PS (Pyro) and a tumor homing molecule (folate) and a peptide linker. PPF was found to be selectively accumulated in cancer cells via folate receptor (FR) pathway. The second PDT agent "PP_MMP7B" is a matrix metalloproteinase-7 (MMP7)-triggered photodynamic molecular beacon (PMB) containing a PS (Pyro), a ¹O₂ quencher (BHQ3) and a MMP7-cleavable peptide linker. Thus, the ¹O₂ production of PP_MMP7B is highly sequence-specific and its photodynamic cytotoxicity is MMP7-dependent. Since these agents are designed to share functional modules (PS and peptide linker) and common cancer cell model (KB cells overexpress both FR and MMP7), it forms the basis for rational design of receptor-targeted PMB for achieving a multi-level control of ¹O₂ production in cancer cells, which in term, could provide a much higher level of PDT selectivity.

Near-Infrared (NIR) absorbing chromophores have been used in analytical and bioanalytical chemistry extensively, including for determination of properties of biomolecules, DNA sequencing, immunoassays, capillary electrophoresis (CE) separations, etc. The major analytical advantages of these dyes are low background interference and high molar absorptivities. NIR dyes have additional advantages due to their sensitivity to microenvironmental changes. Spectral changes induced by the microenvironment are not desirable if the labels are used as a simple reporting group, e.g., during a biorecognition reaction. For these applications upconverting phosphors seem to be a better choice. There are several difficulties in utilizing upconverting phosphors as reporting labels. These are: large physical size, no reactive groups and insolubility in aqueous systems. This presentation will discuss how these difficulties can be overcome for bioanalytical and forensic applications. During these studies we also have investigated how to reduce physical size of the phosphor by simple grinding without losing activity and how to attach reactive moiety to the phosphor to covalently bind to the biomolecule of interest. It has to be emphasized that the described approach is not suitable for medical applications and the results of this research are not applicable in medical applications. For bioanalytical and forensic applications upconverting phosphors used as reporting labels have several advantages. They are excited with lasers that are red shifted respective to phosphorescence, resulting in no light scatter issues during detection. Also some phosphors are excited using eye safe lasers. In addition energy transfer to NIR dyes is possible, allowing detection schemes using donor-acceptor pairs. Data is presented to illustrate the feasibility of this phenomenon. If microenvironmental sensitivity is required, then specially designed NIR dyes can be used as acceptor labels. Several novel dyes have been synthesized in our laboratories for that purpose.

Indocyanine Green (ICG) is a photosensitive dye frequently used in clinical settings for retinal imaging and measuring cardiac output function. Its low toxicity, high absorption in the near infrared region, and vascular retention have generated much interest in its utility for several therapeutic and diagnostic applications. Its potential, however, is limited by factors such as rapid circulation kinetics and lack of target specificity. Additionally, the optical properties of ICG vary significantly with a wide range of influences such as concentration, solvent, pH, and temperature. In an effort to overcome these limitations and improve ICG's effectiveness as a chromophore for photothermal treatment, we have developed a novel nanometer-sized capsule system as a delivery vehicle for ICG. In our capsules, the ICG molecules are encapsulated within a polymer-salt aggregate core coated with dextran. In this study, we report how the optical properties of ICG are affected by its entrapment within the nano-assembled complexes (NACs). The absorption profiles of both freely dissolved ICG and encapsulated ICG were recorded under various conditions to determine the effects of encapsulation on ICG's light sensitivity and stability at different temperatures. Dilute preparations of freely dissolved ICG demonstrated more rapid degradation kinetics in response to intense broadband light (360 - 1100 nm, 56 mW/ cm²) than NACs containing ICG. Encapsulated ICG also demonstrated greater stability at room and elevated temperatures than freely dissolved ICG. The results suggest that the entrapment of ICG within the capsules influences its ability to undergo physicochemical transformations, including oligomerization and degradation to leucoforms.

For biomedical applications, various optical techniques have been explored, all of which are inherently subject to the limited penetration depth of observation. Nanoparticles have thus been introduced as a potential agent to overcome the limit and to reveal more information on biomedical samples. Optical properties of nanoparticles in a turbid medium, however, have remained difficult to measure precisely. In this paper, we investigate the scattering and absorbing properties of gold nanoparticles of varying diameter and different concentration in absorbing liquid phantoms of silica colloid, based on diffuse photometry. Diffuse reflectance patterns have been collected for each phantom with nanoparticles of different size and concentrations relative to base media. The results imply that particle concentration and size perturbs the medium properties, and the effect is dependent on the scattering properties of the particle. For the case where the particle size stays within a range, the scattering property is dominated by the concentration.

Metal/dendrimer nanocomposites (DNCs) uniquely combine the properties of metallic clusters and the biofriendly polymer host in a nanosized hybrid particle. DNCs can biochemically target tissues and locally reduce femtosecond optical breakdown thresholds, making highly precise and selective photodisruption possible. In this study, we have used high-frequency acoustic monitoring of bubble production dynamics to investigate how DNC properties, solution concentration, and optical parameters affect threshold reduction, actual waiting time, and mechanical characteristics of breakdown. Breakdown is defined here as bubble production with an onset of less than 20 seconds after laser exposure. DNC properties varied include metal content (silver, gold) and terminal group (amino-NH₂, glycidol-OH, and carboxyl- COOH) which determine pH values. Results indicate that DNC metal content markedly influences solution threshold reduction, while DNC terminal group (and thus net surface charge) and solution concentration influence the details of breakdown at these reduced threshold fluences. {Ag(0)} DNCs reduce breakdown threshold fluence 1-2 orders of magnitude more than {Au(0)} DNCs. Furthermore, concentrated DNC solutions and DNCs carrying a net negative charge (carboxyl terminal groups) increase bubble production up to four times and shorten waiting time for breakdown from seconds to milliseconds. Increasing laser fluence for a given DNC solution concentration also shortens breakdown waiting time. Lastly, utilizing the fluorescence properties of silver nanocomposites, we use confocal microscopy to examine KB cell uptake of folate targeted silver DNCs. Cells incubated with folate targeted silver DNCs exhibit a measurable increase of intracellular fluorescence compared to control cells (no DNC incubation). However, while we observe a threshold reduction in KB cells incubated with 500nM folate-targeted DNC solution, there is no threshold reduction in cells incubated with 50nM folate-targeted DNC solution. This suggests that a specific minimum DNC concentration may be required for localized reduced-threshold breakdown to occur.

Luminescent nanoparticles are increasingly used as emitting labels in the design of fluorescence imaging probes, because of their outstanding optical properties, such as in the case of quantum dots, or their role in nano-assembling different functionalities, such as imaging, drug delivery, and therapy. In this work, the potentiality of up-converting nano-crystals for non invasive fluorescence imaging of small animal is explored. These up-converting nano-crystals are lanthanide doped oxide matrices, such as Y₂O₃:Er,Yb compositions. They are produced by a bottom-up process and a thermal treatment. They are functionalized by the coating of a thin polysiloxane shell layer, which can be further derivatized in order to graft biomolecules such as antibodies, peptides, or DNA. Contrary to classical luminescent particles for which light emission occurs at a wavelength superior to that of excitation, these materials emit at 564 and 661 nm upon excitation at 980 or 815 nm. These unique emission properties, due to a multi-photonic process, can allow imaging without any auto-fluorescence from the tissues in the wavelength detection window. Experiments in phantoms mimicking the optical absorption and diffusion of tissues show that these crystals can be detected 4 mm deep at a 10 mg/mL concentration. Luminescence measurements in mouse and the potentiality of these nano-crystals for in vivo imaging are discussed.

The physio-pathological roles of sulfide biomolecules in cellular environments involves redox processes and radical reactions that alter or protect the functional properties of enzymatic systems, proteins and nucleic acids repair. We focus on micromolar monitoring of sulfur-centered radical anions produced by direct electron attachment, using sulfide molecules (a thioether and a disulfide biomolecule) and two complementary spectroscopic approaches: low energy radiation femtochemistry (1-8 eV) and high energy radiation femtochemistry (2.5-15 MeV). The early step of a disulfide bond making RS∴SR from thiol molecules involves a very-short lived odd-electron bonded intermediate for which an excess electron is transiently localized by a preexisting two sulfide monomers complex. The reactive center of oxidized glutathione (cystamine), a major cytoplasmic disulfide biomolecule, is also used as sensor for the real-time IR investigation of effective reaction radius r_eff in homogenous aqueous environments and interfacial water of biomimetic systems. Femtosecond high-energy electrons beams, typically in the 2.5 - 15 MeV range, may conjecture the picosecond observation of primary radical events in nanometric radiation spurs. The real-time investigation of sulfide and disulfide molecules opens exciting opportunities for sensitisation of confined environments (aqueous groove of DNA, protein pockets, sub-cellular systems) to ionizing radiation. Low and high-energy femtoradical probing foreshadow the development of new applications in radiobiology (low dose effect at the nanometric scale) and anticancer radiotherapy (pro-drogue activation).

Bioluminescence Imaging (BLI) has been employed as an imaging modality to identify and characterize fundamental processes related to cancer development and response at cellular and molecular levels. This technique is based on the reaction of luciferin with oxygen in the presence of luciferase and ATP. A major concern in this technique is that tumors are generally hypoxic, either constitutively and/or as a result of treatment, therefore the oxygen available for the bioluminescence reaction could possibly be reduced to limiting levels, and thus leading to underestimation of the actual number of luciferase-labeled cells during in vivo procedures. In this report, we present the initial in vitro results of the oxygen dependence of the bioluminescence signal in rat gliosarcoma 9L cells tagged with the luciferase gene (9L^luc cells). Bioluminescence photon emission from cells exposed to different oxygen tensions was detected by a sensitive CCD camera upon exposure to luciferin. The results showed that bioluminescence signal decreased at administered pO₂ levels below about 5%, falling by approximately 50% at 0.2% pO₂. Additional experiments showed that changes in BLI was due to the cell inability to maintain normal levels of ATP during the hypoxic period reducing the ATP concentration to limiting levels for BLI.

This paper presents a novel fiber optic Fabry-Perot (FP) structure for direct detection of deoxyribonucleic acid (DNA) sequences. The capture DNA is immobilized onto the surface of a silica optical fiber tip by means of the layer-by-layer electrostatic self-assembly (L-b-L ESA) technique. With the FP cavity inside the fiber, the change in optical path length difference (OPD) caused by the immobilization and hybridization can be demodulated. Experimental results demonstrate short (~5 min) response time. In addition, the sensor fabrication is simple and low cost. The signal is stable with high visibility.

Fluorescent techniques developed for probing molecular events in vitro have been translated to small animal in-vivo fluorescence imaging using narrow-band, high-power laser sources, fast photon detection electronics and sophisticated image reconstruction algorithms. Targeted near-infrared fluorescent probes are being developed to increase image contrast and functional information collected by the in-vivo optical images. Besides fluorescence intensity, fluorescence lifetime can also be utilized to probe tissue physiology. Fluorescence lifetime, the average time that a fluorophore remains in the excited state, is specific to the molecule and its environment, but not necessarily to its concentration. Therefore, fluorescence lifetime imaging can be used to probe microenvironments within tissues to differentiate, for example, cancer from surrounding tissue by physiological differences. In the current study we demonstrate the ability to distinguish between two molecular probes in mice using fluorescence lifetime while using single excitation and emission wavelengths. Fluorescence detection was accomplished by measuring diffuse emission at 3 mm distance from excitation and raster-scanning whole-body regions of interest using time-correlated single photon counting technique. The resulting temporal point-spread function data was collected for grid points 1.5 mm apart. Fluorescence decays were de-convolved to determine measured fluorescence lifetimes by least-squares fit with associated chi-square error values. Relative signal from each fluorophore was then determined for specific tissues including liver, kidney and tumor areas. The results demonstrate that relative biodistributions of individual fluorophores with similar photonic characteristics can be simultaneously monitored in vivo using fluorescence lifetime imaging.

The framework of fluorescent targeting probes for optical imaging is similar to that of contrast agents for other modalities. They generally include a biological ligand, specific of the biological process to image, and a label, which confers the probe its optical properties. Moreover, more sophisticated labeling functions, termed "activatable" can be designed. An "activatable probe" will be initially non fluorescent. Only a specific molecular process, such as an enzymatic reaction or cell internalization, is able to activate the probe fluorescence. Such probes are particularly easy to design for the optical imaging modality, because of the easily triggered and well known fluorescence inhibition processes. The optical properties of different commercially available organic dyes and their fluorescence inhibition are therefore examined. Three classes of activatable probes have been listed: (i) activatable probes which rely on the selfquenching of the label; (ii) activatable probes which use RET (Resonance Energy Transfer) between the label acting as a donor, and an organic non-emissive acceptor; (iii) activatable probes which use an inorganic nanostructure as the inhibitor, such as a gold nano-particle. Whereas activatable probes of class (i) can lead to higher fluorescence levels after activation, their initial fluorescence inhibition can be hindered by structural constraints. Probes of class (ii) can therefore be more interesting according to the probe design. The efficiency of probes of class (iii) using nanometer gold particles is reduced because of their plasmon band lying in the visible and not near-infrared domain.

Highly luminescent nanoparticles, such as, trivalent erbium-doped yttrium-oxide, Er³⁺:Y₂O₃, are expected to have a wide range of applications, including imaging, range finding, flash lidar, and other remote-sensing possibilities as well as medical applications. These particles are synthesized by the precipitation from a homogeneous solution of the metal ions and urea at elevated temperatures. The morphology of the calcinated materials, revealed through SEM, shows uniformly spherical aggregates 200 nm or less depending on the ratio of the metal ions in the initial solution. Room temperature optical absorption and emission spectra show that the trivalent erbium ions in Er³⁺:Y₂O₃ nanocrystals possess sharp absorption lines and strong emission in near infrared region that are characteristic of Er ³⁺:Y₂O₃ grown as large single crystals. Low temperature (8 K) spectra obtained from these particles were analyzed in detail for the crystal-field splitting of the ^2S+1L_J multiplet manifolds of Er³⁺(4f¹¹) including the ground-state manifold ⁴I_15/2, and excited manifolds ⁴I_9/2, ⁴F_9/2, ⁴S_3/2, ²H_11/2, ⁴F_7/2, ⁴F_5/2, and ⁴F_3/2. Fluorescence lifetimes and results from an analysis of the intensities of manifold-to-manifold transitions are also reported. Similarity of the nanocrystalline and large single crystal Er³⁺:Y₂O₃, we propose that the simple, inexpensive method described in this study will lead to further investigation of these nanocrystals for their optical properties, especially in the near infrared region of the spectrum.

For investigatoin of intracellular protein interactions, quantum dots are widely used for fluorescent live cell imaging such as total internal reflection fluorescence (TIRF) microscopy and confocal microscopy. In this paper, we performed a quantitative analysis based on fluorescent intensity. For the measurement, A431 cell lines are imaged live with quantum dots using TIRF microscopy. The distribution of quantum dots is affected by a TATHA2 peptide sequence in live cells. This paper also presents the location change of quantum dots due to a nuclear localization signal in A431 cell lines. Confocal microscopy was used to confirm the relation with fluorescent intensity and quantum dot concentration in live cells.

The steady state absorption and fluorescence spectroscopy of 2-(2'-hydroxyphenyl)benzoxazole (HBO) and (2,2'-bipyridine)-3,3'-diol (BP(OH)₂) were studied here free in solution and in human serum albumin (HSA) in order to test their applicability as new biological probes. HBO and BP(OH)₂ are known to undergo intramolecular proton transfers in the excited state. Their absorption and fluorescence spectra are sensitive to environmental change from hydrophilic to hydrophobic, thus allowing the opportunity to use them as environment-sensitive probes. The effect of water on the steady state spectra of the two molecules also shows unique features which may position them as water sensors in biological systems. For HBO in buffer, fluorescence is only due to the syn-keto tautomer, whereas in HSA the fluorescence is due to four species in equilibrium in the excited state (the syn-keto tautomer, the anti-enol tautomer, the solvated syn-enol tautomer, and the anion species of HBO). Analysis of the fluorescence spectra of HBO in HSA indicates that HBO is exposed to less water in the HBO:HSA complex. For the BP(OH)₂ molecule, unique absorption due to water was observed in the spectral region of 400-450 nm. This absorption decreases in the presence of HSA due to less accessibility to water as a result of binding to HSA. Fluorescence of BP(OH)₂ is due solely to the di-keto tautomer after double proton transfer in the excited state. The fluorescence peak of BP(OH)₂ shows a red-shift upon HSA recognition which is attributed to the hydrophobic environment inside the binding site of HSA. We discuss also the effect of probe-inclusion inside well-defined hydrophobic cavities of cyclodextrins.

We address the problem of quantitative evaluation of the absorption S₀-S₁ peaks &lgr; _max of biological chromophores in vacuo by using the state-of-art computational methods of quantum theory. In particularly, we rely on the second order multiconfigurational quasidegenerate perturbation theory (MCQDPT2) following the complete active space selfconsistent field (CASSCF) calculations. The use of augmented effective Hamiltonian operators in the MCQDPT2 framework allows us to correct deficiencies of the standard multistate approaches and to obtain stable saturated solutions for the target low-lying excited states. A high accuracy of the proposed methodology is illustrated for several photoactive protein chromophores in the gas phase including all-trans retinal, green fluorescent protein type chromophores and photoactive yellow protein chromophores. It is shown that our approach provides correct ordering of states and predicts maxima of absorption bands for the S₀-S₁ transitions within only a few nanometers from experimental data.

We present the results of modeling properties of the chromophore, 2-acetyl-4-(p-hydroxybenzylidene)-1-methyl-5- imidazolone (AHBMI), from the newly discovered fluorescent protein asFP595 inside the protein environment by using the combined quantum mechanical - molecular mechanical (QM/MM) method. In this approach, the chromophore unit and the side chains of the nearest amino acid residues are assigned to the quantum subsystem. The starting coordinates of heavy atoms were taken from the relevant crystal structures of the protein. Hydrogen atoms were added manually, and the structure of the model protein system was optimized by using QM/MM energy minimization for the trans-form of the chromophore. The Hartree-Fock/6-31G quantum chemical approximation and the AMBER force field parameters were employed in geometry optimization. The points on potential energy surfaces of the ground and first and second excited electronic states were computed with the complete active space self-consistent field approximation in the quantum subsystem under different choices of the QM/MM partitioning. Possible pathways for the trans-cis photo isomerization presumably responsible for the kindling properties of asFP595 as well as other mechanisms of photo excitation are discussed.

GFP is extensively used in molecular imaging applications. Its fluorescence is due to a autocatalytically formed chromophore located in the center of the protein. Yellow and blue fluorescent mutants of GFP have been created. The protein matrix that surrounds the chromophore influences both the intensity and the wavelength of the fluorescence. We have used conformational searching methods and molecular dynamics simulations to examine the &tgr; and &fgr; dihedral space available to a freely rotating and pyramidalizing chromophore. The calculations have shown that there seems to be a relationship between the quantum yield of the fluorescent protein and the dihedral &tgr; and &fgr; space available to the chromophore. The bright YFP has less rotational space available to a freely rotating chromophore than does wild-type GFP or BFP, which has the most rotational freedom.

Binding proteins suitable for expression and high affinity molecular recognition in the cytoplasm or nucleus of live cells have numerous applications in the biological sciences. In an effort to add a new minimal motif to the growing repertoire of validated non-immunoglobulin binding proteins, we have undertaken the development of a generic protein scaffold based on a single &bgr;-hairpin that can fold efficiently in the cytoplasm. We have developed a method, based on the measurement of fluorescence resonance energy transfer (FRET) between a genetically fused cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP), that allows the structural stability of recombinant &bgr;-hairpin peptides to be rapidly assessed both in vitro and in vivo. We have previously reported the validation of this method when applied to a 16mer tryptophan zipper &bgr;-hairpin. We now describe the use of this method to evaluate the potential of a designed 20mer &bgr;-hairpin peptide with a 3rd Trp/Trp cross-strand pair to function as a generic protein scaffold. Quantitative analysis of the FRET efficiency, resistance to proteolysis (assayed by loss of FRET), and circular dichroism spectra revealed that the 20mer peptide is significantly more tolerant of destabilizing mutations than the 16mer peptide. Furthermore, we experimentally demonstrate that the in vitro determined &bgr;-hairpin stabilities are well correlated with in vivo &bgr;-hairpin stabilities as determined by FRET measurements of colonies of live bacteria expressing the recombinant peptides flanked by CFP and YFP. Finally, we report on our progress to develop highly folded 24mer and 28mer &bgr;-hairpin peptides through the use of fluorescence-based library screening.

Bioluminescent photoproteins, such as aequorin and obelin, are proteins that emit light upon binding calcium. Aequorin and obelin contain four EF-hand domains arranged into a globular structure. The loop region of these EF-hand domains binds calcium by coordinating it in a pentagonal bipyramidal structure with oxygen atoms. The binding of calcium to these EF-hands causes a slight conformational change in the protein, which leads to the oxidation of the internally sequestered chromophore, coelenterazine, producing coelenteramide and CO₂. The excited coelenteramide then relaxes radiatively, emitting bioluminescence at 471 nm in aequorin or 491 nm in obelin. Although calcium is the traditional, and generally the most powerful, triggering ligand in this bioluminescence reaction, alternative di- and trivalent cations can also bind to the EF-hand loops and stimulate luminescence. Species capable of this cross-reactivity include: Cd²⁺, Ba²⁺, Mn²⁺, Sr²⁺, Mg²⁺, and several lanthanides. Magnesium is also known to modulate the bioluminescence of wild-type aequorin, increase its stability, and decrease its aggregation tendency. Both wild-type aequorin and wild-type obelin contain several cysteine residues, aequorin has three and obelin has five. It is believed that these cysteine residues play an important, but as of yet unknown, role in the bioluminescence of these proteins, since mutating most of these residues causes significant loss in bioluminescent activity. In order to explore whether or not these cysteine residues contributed to the specificity of the EF-hand domains for cations we generated four aequorin and obelin mutants and observed their luminescent intensity and decay kinetics by stimulation with calcium, barium, and magnesium. It was found that the cysteine mutations do appear to alter the effects that alternative divalent cations have on the bioluminescence of both aequorin and obelin.

An important goal for in vivo imaging is to be able to non-invasively image single cells. The Olympus IV100 Laser Scanning Microscope, with ultra-thin microscope objectives ("stick objectives"), was used for three-color whole-body imaging of individual two-color cancer cells interacting with the GFP-expressing stromal cells. Cellular dynamics were non-invasively imaged including mitotic and apoptotic tumor cells, stromal cells interacting with the tumor cells, tumor vasculature, and tumor blood flow. This imageable model should lead to a new paradigm of in vivo cancer cell biology and to new visible real-time targets for cancer drug discovery.

Here we describe our cumulative experience with the development and preclinical application of several highly fluorescent, clinically-relevant, metastatic orthotopic mouse models of pancreatic cancer. These models utilize the human pancreatic cancer cell lines which have been genetically engineered to selectively express high levels of the bioluminescent green fluorescent (GFP) or red fluorescent protein (RFP). Fluorescent tumors are established subcutaneously in nude mice, and tumor fragments are then surgically transplanted onto the pancreas. Locoregional tumor growth and distant metastasis of these orthotopic implants occurs spontaneously and rapidly throughout the abdomen in a manner consistent with clinical human disease. Highly specific, high-resolution, real-time visualization of tumor growth and metastasis may be achieved in vivo without the need for contrast agents, invasive techniques, or expensive imaging equipment. We have shown a high correlation between florescent optical imaging and magnetic resonance imaging in these models. Alternatively, transplantation of RFP-expressing tumor fragments onto the pancreas of GFP-expressing transgenic mice may be used to facilitate visualization of tumor-host interaction between the pancreatic tumor fragments and host-derived stroma and vasculature. Such in vivo models have enabled us to serially visualize and acquire images of the progression of pancreatic cancer in the live animal, and to demonstrate the real-time antitumor and antimetastatic effects of several novel therapeutic strategies on pancreatic malignancy. These fluorescent models are therefore powerful and reliable tools with which to investigate human pancreatic cancer and therapeutic strategies directed against it.

More malignant tumors contain more hypoxic regions. In hypoxic tumor cells, expression of a series of hypoxiaresponsive genes related to malignant phenotype such as angiogenesis and metastasis are induced. Hypoxia-inducible factor-1 (HIF-1) is a master transcriptional activator of such genes, and thus imaging of hypoxic tumor cells where HIF-1 is active, is important in cancer therapy. We have been developing PTD-ODD fusion proteins, which contain protein transduction domain (PTD) and the VHL-mediated protein destruction motif in oxygen-dependent degradation (ODD) domain of HIF-1 alpha subunit (HIF-1&agr;). Thus PTD-ODD fusion proteins can be delivered to any tissue in vivo through PTD function and specifically stabilized in hypoxic cells through ODD function. To investigate if PTD-ODD fusion protein can be applied to construct hypoxia-specific imaging probes, we first constructed a fluorescent probe because optical imaging enable us to evaluate a probe easily, quickly and economically in a small animal. We first construct a model fusion porein PTD-ODD-EGFP-Cy5.5 named POEC, which is PTD-ODD protein fused with EGFP for in vitro imaging and stabilization of fusion protein, and conjugated with a near-infrared dye Cy5.5. This probe is designed to be degraded in normoxic cells through the function of ODD domain and followed by quick clearance of free fluorescent dye. On the other hand, this prove is stabilized in hypoxic tumor cells and thus the dye is stayed in the cells. Between normoxic and hypoxic conditions, the difference in the clearance rate of the dye will reveals suited contrast for tumor-hypoxia imaging. The optical imaging probe has not been optimized yet but the results presented here exhibit a potential of PTD-ODD fusion protein as a hypoxia-specific imaging probe.

We have developed an in vivo imaging system consisting of GFP- and DsRed-tagged human colonic cancer cell line, which has peritoneal and lymph node metastatic potential and show high sensitivity to EGFR targeting drugs, and convenient detection devices for GFP and DsRed. The latter includes a small handy fluorescence detection device for external monitoring of the therapeutic effect of the drug and a convenient stereo fluorescent microscope for internal visualization of micrometastases. We applied this imaging system to investigate anti-metastatic effects of EGFR targeting drugs such as gefitinib (Iressa). This system allowed sensitive detection of the development of peritoneal and lymph node metastases from the micrometastasis stage at the cellular level and also permited noninvasive, non-anesthetic monitoring of anti-metastatic effect of the drug in an animal facility without any pretreatment. Significant decreases in the intraabdominal metastatic tumor growth and prevention of inguinal lymph node metastasis by gefitinib treatment could be clearly monitored. These results suggest that convenient, low-cost, true real-time monitoring of therapeutic effect using such a fluorescence-mediated whole body imaging system seems to enhance the speed of preclinical study for novel anti-cancer agents and will allow us to understand the action mechanism of molecular targeting drugs.

Early detection of primary tumors is the key for effective therapeutic intervention and successful patient survival. Small animal models emulating human diseases are powerful tools for our comprehensive understanding of the pathophysiology of tumor formation and metastasis to distant sites. Our long-term goal is to develop a non-invasive, multiphoton-fluorescence lifetime imaging (MP-FLIM) modality that can precisely quantify these steps in animal tumor models at a very early stage. The specific hypothesis is that fluorescence lifetime can be employed as reliable contrast parameter for providing higher detection sensitivity as compared with conventional intensity-based tumor imaging approaches and therefore it is possible to detect smaller tumor volumes (early detection) than those achieved by other prevailing methods. We base this hypothesis on our recent observations that (1) fluorescence lifetime is "intrinsic" to the fluorophore and its measurement is not affected by concentration and/or spectral artifacts as in intensity-based methods, (2) multiphoton excitation can enable increased tissue penetrability and reduced phototoxicity and (3) MP-FLIM approach can discriminate background autofluorescence from the fluorescent proteins in thick tissues thereby achieving a ten-fold increase in signal-to-background ratio over the intensity-based approaches. We present our preliminary data to support this hypothesis in primary tumor detection in nu/nu athymic mouse models.

Target-specific contrast agents are being developed for the molecular imaging of cancer. Optically detectable target-specific agents are promising for clinical applications because of their high sensitivity and specificity. Pre clinical testing is needed, however, to validate the actual sensitivity and specificity of these agents in animal models, and involves both conventional histology and immunohistochemistry, which requires large numbers of animals and samples with costly handling. However, a superior validation tool takes advantage of genetic engineering technology whereby cell lines are transfected with genes that induce the target cell to produce fluorescent proteins with characteristic emission spectra thus, identifying them as cancer cells. Multicolor fluorescence imaging of these genetically engineered probes can provide rapid validation of newly developed exogenous probes that fluoresce at different wavelengths. For example, the plasmid containing the gene encoding red fluorescent protein (RFP) was transfected into cell lines previously developed to either express or not-express specific cell surface receptors. Various antibody-based or receptor ligand-based optical contrast agents with either green or near infrared fluorophores were developed to concurrently target and validate cancer cells and their positive and negative controls, such as &bgr;-D-galactose receptor, HER1 and HER2 in a single animal/organ. Spectrally resolved fluorescence multicolor imaging was used to detect separate fluorescent emission spectra from the exogenous agents and RFP. Therefore, using this in vivo imaging technique, we were able to demonstrate the sensitivity and specificity of the target-specific optical contrast agents, thus reducing the number of animals needed to conduct these experiments.

Conventional optical imaging is restricted with tumor size due to high tissue scattering. Labeling of tumors by fluorescent markers improves sensitivity of tumor detection thus increasing the value of optical imaging dramatically. Creation of tumor cell lines transfected with fluorescent proteins gives the possibility not only to detect tumor, but also to conduct the intravital monitoring studies. Cell lines of human melanomas Mel-P, Mel-Kor and human embryonic kidney HEK-293 Phoenix were transfected with DsRed-Express and TurboRFP genes. Emission of RFP in the long-wave optical range permits detection of the deeply located tumors, which is essential for whole-body imaging. Only special tools for turbid media imaging, such as fluorescent diffusion tomography (FDT), enable noninvasive investigation of the internal structure of biological tissue. FDT setup for monitoring of tumor growth in small animals has been created. An animal is scanned in the transilluminative configuration by low-frequency modulated light (1 kHz) from Nd:YAG laser with second harmonic generation at the 532 nm wavelength. In vivo experiments were conducted immediately after the subcutaneously injection of fluorescing cells into small animals. It was shown that FDT method allows to detect the presence of fluorescent cells in small animals and can be used for monitoring of tumor growth and anticancer drug responce.

The main reason green fluorescent protein (GFP) is so useful in molecular imaging is the fact that its chromophore is formed autocatalytically. We have been using molecular mechanics to examine chromophore formation since 1995 that is well before the first crystal structures of GFP were solved. Our calculations have resulted in a number of predictions that have been borne out by subsequent experiments and a number of them that haven't. Recently we have been supplementing these calculations with calculations based on the crystal structures of immature GFP mutants (i.e. the precyclized form). Preliminary results from these calculations have shown that immature GFP does form a tight-turn in the chromophore forming region, and that chromophore cyclization is probably catalyzed in the manner proposed by Getzoff et al (Biochemistry 44: 1960-1970, 2005).

At the present time the phenomenon of "fluorescence kindling," which is typical of asCP595 protein and some of its mutants, is of great interest for the studies of intracellular traffic. The range of changing of intracellular pH may be rather wide, however, the effect of pH on the fluorescent state of this protein is poorly investigated. Our studies have revealed that the fluorescence intensity of asCP595 increases significantly in alkaline conditions. In addition, the observed change in the position of excitation, emission, and absorption maxima indicates the appearance of new spectral forms of the protein. These conformers are characteristics of new fluorescent states. The changes in the absorption spectrum are indicative of a new dark form of the protein at alkaline pH.

Degradation of the extracellular matrix (ECM) by matrix metalloproteinases (MMPs) enhances tumor invasion and metastasis. To monitor MMP activity, we constructed plasmid that encoded a fluorescent sensor DC, in which an MMP substrate site (MSS) is sandwiched between DsRed2 and ECFP. MMPs are secretory proteins, only acting on the outside of cells; hence, an expressing vector was used that displayed the fluorescent sensor on the cellular surface. The DC was expressed in cells with high secretory MMP, so MSS was cleaved by MMP. Also, GM6001, an MMP inhibitor, causes DsRed2 signals to increase in living cells and on the chick embryo chorioallantoic membrane (CAM). Thus, this fluorescent sensor was able to sensitively monitor MMP activation in vivo. Potential applications for this sensor include high-throughput screening for MMP inhibitors for anti-cancer research, and detailed analysis of the effects of MMP inhibitors.