Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Fluorescence microscopy enables the spatial assignment of fluorophores within cells. The limitations of current fluorescence cellular imaging, which mainly uses filter-based systems, include endogenous autofluorescence, exogenous dye contributions, and spectral overlap between probes and the studied fluorophore. Autofluorescence artifacts occur during biological imaging; these may distort experimental findings. The major endogenous fluorophores responsible for native fluorescence include aromatic amino acids (tryptophan, 340 nm), pyridoxine (390 nm), collagen (390 nm), elastin (410 nm), reduced nicotinamide adenine dinucleotides (470 nm), flavins (550 nm), and porphyrins (630 and 690 nm), all of which may perturb single point or filter detection systems (Richards-Kortum and Sevick-Muraca, 1996). Exogenous fluorophores also may interfere with data interpretation; an example is phenol red, which serves as a pH indicator in some media or fluorescence diagnostics such as organelle probes. Additionally, the cellular microenvironment not only affects fluorescence quantum yields but also the chromophore's emission wavelength maxima and bandwidth. Fluorophores have different photodynamic activities or photobleaching properties that may lead to shifts in emissions signals. There is a need for complete fluorescence spectra acquisition in cellular environments, which overcomes these limitations.

Fluorescence spectral bioimaging, using Fourier spectroscopy, allows spatial and spectral measurements that enable characterization of the fluorescence emission profile at every acquired pixel. A recent technique, it has been used successfully in fluorescence in situ hybridization applications to diagnose genetic defects by fluorescent staining, termed spectral karyotyping, of chromosomes (Schröck et al., 1996). Malik et al. (1996) have used spectral bioimaging microscopically to monitor protoporphyrin IX production after exogenous 5-aminolevulinic acid administration. Subsequent illumination of the cells caused photobleaching and photoproduct formation, with a distinctly different spectral peak appearing in the acquired spectrum.

In this study, a similar system was used to correctly identify the fluorophore motexafin gadolinium (MGd) in intracellular sites after MGd treatment. The biolocalization of MGd was confirmed by co-staining with organelle-specific probes and spectral overlay. MGd is a radiation enhancer that targets tumors and is now in human clinical trials (Rowinsky, 1999). To better understand the biochemical mechanisms that underlie these activities cell culture localization studies were performed. Here, wavelength-dependent microenvironmental changes, photobleaching, and subcellular fluorophore concentrations after MGd treatment were determined in both the murine Rif-1 fibrosarcoma and EMT6 mammary sarcoma cell lines.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. The synthesis and chemical characterization of MGd (also known as gadolinium texaphyrin, NSC #695238) used in this study has been described previously (Sessler et al., 1999). The compound was formulated in 5% aqueous mannitol, and the pH was adjusted to 5.5 with acetic acid to yield a final MGd concentration of 2.2 mM. Fluorescent probes for staining mitochondria (MitoTracker Green FM; MTG), lysosomes (LysoTracker red DND-99; LTR), Golgi apparatus (BODIPY FL C₅-ceramide, C₅-ceramide), endoplasmic reticulum (Rhodamine B, hexyl ester, perchlorate; R6), and the nucleus (Hoechst 33342; HO342) were purchased from Molecular Probes (Eugene, OR).

Cells. Murine EMT6 mammary sarcoma and Rif-1 radiation-induced fibrosarcoma cell lines were maintained through established in vivo/in vitro propagation procedures (Rockwell et al., 1972; Young et al., 1996). Cells were grown in Waymouth's medium (MB752/1; Life Technologies, Grand Island, NY) supplemented with 14% fetal bovine serum (FBS; JRH Biosciences, Woodland, CA) and penicillin/streptomycin (Sigma, St. Louis, MO).

Colony Survival Assay. EMT6 cells were cultured and harvested into 25-cm² plastic tissue culture flasks at a density of 200 cells/flask. Twenty-four hours was allowed for cell attachment to the flask at 37°C and 5% CO₂. MGd was added to yield a range of final concentrations from 0 to 150 µM. After 2 or 24 h MGd exposure the cells were washed three times with growth media without serum. The flasks were replenished with serum-supplemented media and returned to the incubator for an additional 6 to 10 days. The colonies were fixed with 10% buffered formalin and stained with crystal violet. Each determination was performed in triplicate and repeated twice for each data point.

Fluorescence Spectral Bioimaging. Female BALB/c mice, 7 to 8 weeks old, were purchased from Simonsen Laboratories, Gilroy, CA. Animals received care in compliance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication, 1996). The right flanks of the mice were shaved and depilated with Nair (Carter-Wallace, Inc., New York, NY) the day prior to tumor inoculation. EMT-6 cells (5-7 × 10⁵) were subcutaneously implanted into the right hind flanks of recipient mice. The tumor-bearing animals were studied 7 to 10 days after implantation (tumor size approximately 70 mm³).

Fluorescence Microscopy. The cells were seeded (4 × 10⁴) into Lab-Tek 10-cm² Flaskette glass chamber slides (Nunc, Inc. Naperville, IL) containing 4 ml of media. The cells were allowed to incubate for 24 h to enable attachment to the flask. MGd was added to the media to yield a final concentration of 25 µg/ml and incubated for predetermined amounts of time. After the incubation, the cells were washed with serum-free media, coverslipped, and sealed with nail polish. The specimens were viewed using a 63× oil-immersion objective via a Zeiss Axioplan-2 microscope. The SD200 SpectraCube system was coupled to the microscope via a C-mount. Both fluorescence and phase contrast micrographs were captured. All cell preparations were performed in triplicate.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

In Vivo Fluorescence Imaging. The chemical structure, absorption, and fluorescence profile of MGd is shown in Fig. 1. MGd has a large Soret band at 473 nm (log  = 5.06). Due to its extended porphyrin aromatic conjugation system, the texaphyrin macrocycle has a low-energy band occurring at 740 nm (log  = 4.59). The signature fluorescence emission band of MGd is centered at 758 nm and was used to monitor real-time MGd biolocalization in EMT6 mammary sarcoma-bearing mice. The animals were intravenously injected with MGd (40 µmol/kg) and then the tumor and surrounding normal skin were illuminated with 470-nm light, corresponding to the Soret band, 5 h postinjection. The resultant fluorescence emission signals were captured with a CCD camera coupled to an interferometer, and then the signal was Fourier-transformed, allowing spectral identification at each pixel (Fig. 2). Integration of the signal over the emission captured range (720-780 nm), for four animals, and subtraction of the endogenous fluorescence revealed a 6-fold enhancement in the tumor relative to surrounding normal tissue.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Chemical structure of MGd and an absorption spectrum of 25 µg/ml MGd in phosphate-buffered saline with 10% FBS.

View larger version (38K): [in this window] [in a new window]   Fig. 2.   Fluorescence spectral bioimaging of MGd in an EMT-6 mammary sarcoma-bearing mouse. A black-and-white photo illustrating the tumor-bearing mouse in the prone position is shown in a. The mouse received 40 µmol/kg MGd and was spectral-imaged 5 h after intravenous administration. A fluorescence spectral typing map, using a pixel classification of 720 to 780 nm, was obtained and overlaid on the pre-imaged black-and-white photo (b). Two pixel areas, averaged from four animals, showing tumor (orange) and nontumor (green) regions from animals receiving MGd (full lines) and receiving no MGd (dashed lines), were processed to yield fluorescence emission profiles (c). The curves represent average values ± 20%.

MGd Photobleaching. The fluorescence quantum yield of MGd is low. However, it can be used for detection and quantitation due to the large extinction coefficient at the excitation wavelength of 473 nm (log  = 5.06), a large Stokes shift (758 nm), and a detection system that exhibits optimal quantum efficiencies in the desired wavelengths. Because irreversible destruction or photobleaching of the fluorophore limits the illumination intensity and time of exposure, studies were undertaken to determine the effect of illumination on MGd photobleaching. An illumination intensity of 63 mW/cm² caused no significant photobleaching, whereas using an intensity of 209 mW/cm² did (Fig. 3). Interestingly, repeating an illumination at 63 mW/cm² did not significantly change the fluorescence profile and emission (data not shown).

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Fluorescence spectra of MGd-treated Rif-1 cells. Cells were incubated with 25 µg/ml MGd and then imaged at 24 h with an illumination intensity of 63 mW/cm2 () followed by another acquisition using 209 mW/cm2 (). The data reflect fluorescence pixels (3 × 3) within a punctate fluorescent area compared with background (background, open symbols).

Uptake Kinetics. The localization of MGd in EMT6 and Rif-1 cells was evaluated for incubation periods up to 72 h in medium supplemented with 14% FBS. No inherent cytotoxicity, measured using a clonogenic assay, was observed using MGd concentrations up to 150 µM in serum supplemented medium (data not shown). Using a 72-h incubation time in serum-free medium the IC₅₀ was 25 µM so for serum-free media experiments incubation periods were confined to 24 h or less.

View larger version (38K): [in this window] [in a new window]   Fig. 4.   MGd localization within EMT6 cells. EMT6 cells were treated for 4 h with 25 µg/ml MGd. The transmission and fluorescence micrographs are illustrated in a and b, respectively. The fluorescence emission spectra, analyzed at various regions, are shown in c.

View larger version (35K): [in this window] [in a new window]   Fig. 5.   MGd localization in the nucleus of EMT6 cells after an incubation of 48 h. Transmission and fluorescence micrographs are depicted in a and b, respectively. Emission profiles from the nuclear, extranuclear punctate, cytoplasmic, and extracellular regions are shown in c.

View larger version (45K): [in this window] [in a new window]   Fig. 6.   Time-dependent localization of MGd in punctate regions of EMT6 and Rif-1 cells. Temporal profiles of EMT6 and Rif-1 cells in serum-supplemented medium are shown in a and b, whereas those in medium lacking serum are shown in c and d, respectively. Data were averaged from 10 cells.

Staining with Organelle Probes. The subcellular localization of each of the studied organelle probes was confirmed based on the unique staining patterns of the Rif-1 and EMT6 murine cell lines. Typical localization patterns for the murine radiation-induced fibrosarcoma cell line (Rif-1) are illustrated in Fig. 7. In Fig. 7b, the cells were sequentially stained with LTR, C5-ceramide, and HO342. The red fluorescent LTR stains the lysosomes, the green C5-ceramide targets the Golgi apparatus, and the blue fluorescent HO342 accumulates in the nuclei. Three exposures, with specific filter cubes matched to each probe, were acquired and then overlaid to yield the depicted image. Figure 7c denotes the distinct green curlicue staining characteristic of the mitochondria, and Fig. 7d shows the red fluorescent R6 accumulating in the endoplasmic reticulum.

View larger version (92K): [in this window] [in a new window]   Fig. 7.   Representative staining of Rif-1 cells with fluorescent organelle-specific probes. In the transmission micrograph in a, cells were stained with fluorescence probes (b) specific for the nuclei (blue), Golgi apparatus (green), and lysosomes (red). The fluorescence micrograph in c depicts the mitochondria as green and the nuclei as blue. The endoplasmic reticulum is illustrated by the red staining in d, with the nuclei appearing blue.

View larger version (26K): [in this window] [in a new window]   Fig. 8.   In vitro normalized fluorescence emission spectra of organelle-specific fluoroprobes. The emission maxima for HO342, C5-ceramide, and LTR are shown in a. C5-ceramide, at high concentrations in the Golgi apparatus, forms longer wavelength excimers that result in a wavelength shift from 522 nm, C5-ceramide (a), to 637 nm, C5-ceramide (b). The normalized fluorescence emission spectra of MTG and R6 are shown in b. R6 fluorescence was observed with two filter cubes, the custom filter cube, R6 (a), and the MGd/C5-ceramide filter cube, R6 (b).

Colabeling with Organelle Probes. After exposure of the cells to MGd for varying times, organelle-specific fluoroprobes were added for colocalization studies. It was necessary to wash the cells with serum-free medium before incubation with the organelle-specific probes, because most probes, and also the MGd, are planar aromatic systems and may interact noncovalently and cause unwanted artifacts that change biolocalization. Dichroic and emission filters were used because, for multispectral labeling, it is ideal that the dye undergoing analysis and the organelle-specific fluorophore have the same coincident wavelength custom excitation. Special attention was given to the long-pass filters, so the different fluorophores could be simultaneously spectrally resolved at every pixel within the CCD array study area.

View larger version (37K): [in this window] [in a new window]   Fig. 9.   Colocalization of MGd within lysosomes. Rif-1 cells were incubated with MGd for 6 h in serum-supplemented medium and then incubated with LTR and HO342. A transmission micrograph is shown in a, and a multiple exposure image obtained with filter sets appropriate for HO3423 (blue), LTR (red), and MGd (yellow) is shown in b. Spectral analysis of four extranuclear pixel areas is depicted in c.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Motexafin gadolinium, or MGd, is a paramagnetic, aqueous-soluble expanded pentadentate porphyrinoid that is selective for neoplastic cells (Young et al., 1996). MGd alone, using the conditions described here, has no observed therapeutic impact. However, radiation enhancement with MGd in combination with radiation therapy has been seen in single and multifraction studies in experimental tumor models (Miller et al., 1999; Minamikawa et al., 1999). MGd is presently in a phase III clinical trial for patients undergoing whole brain radiation therapy for brain metastases (Rowinsky, 1999) and is also being evaluated in nonclinical models as a chemosensitizer with doxorubicin and bleomycin.

Intracellular localization studies may provide mechanistic insights into MGd's mode of action as both a radiation enhancer and a chemosensitizer. Some discordance has been observed when assessing in vitro radiation responses as MGd displayed minimal radiation enhancement in in vitro clonogenic assays, whereas good tumor responses were recorded in murine models (Miller et al., 1999). It is hoped that interferometric Fourier fluorescence microscopy may aid in defining the importance of tumor-host interactions versus intracellular interactions.

Noninvasive spectral bioimaging of EMT6 tumor-bearing mice showed enhanced localization of MGd in the sarcoma compared with the surrounding normal tissue, resulting in an enhancement ratio of 6 to 1. The latter fluorescent result was in accord with previous studies using magnetic resonance imaging to detect the paramagnetic gadolinium coordinated into the central core of the macrocycle and ¹⁴C-labeled MGd (Miller et al., 1999; Minamikawa et al., 1999). The latter radiolabeled biodistribution study performed in SMT-F mammary tumor-bearing mice yielded a tumor-to-muscle ratio of 8.5 to 1 at 5 h after intravenous administration of 9.9 µmol MGd/kg (Miller et al., 1999). It may also be possible to study the intracellular localization of MGd using subcellular fractionation with radiolabeled MGd. However, cultured cells are difficult to fractionate, they are more problematic than tissues due to differences in cytoskeletal organization, and additionally, with subfractionation it is difficult to cleanly resolve many organelles into distinct fractions (Howell et al., 1989; Pasquali et al., 1999), so alternative more facile approaches to discerning intracellular localization sites of drugs are sought.

The uptake of MGd within EMT6 and Rif-1 cells, assessed using interferometric Fourier fluorescence microscopy was immediate. Retention increased significantly with time and was more pronounced in serum-free media. MGd partitioned within the lysosomes, endoplasmic reticulum, and, to a lesser extent, the Golgi apparatus and mitochondria in serum-supplemented medium. Only in 15% of EMT6 cells, at 48 h, was MGd observed in the nucleus. In serum-free medium, more MGd was partitioned within the mitochondria.

Interferometric Fourier fluorescence microscopy is a potentially valuable technique that will resolve subcellular localization of fluorophores using known organelle-specific markers. Fluorescence is often wavelength-dependent, and varies with alterations in kinetic and physiological factors with different cellular compartments. The latter cannot be analyzed from single-point determinations as is obtained using fluorescence filter technology. Monitoring the fluorescence emission range not only aids in subcellular localization studies but is also valuable in assessing cellular changes using different biochemical markers. Experiments of this kind are currently in progress in this laboratory.