Photodynamic therapy (PDT) of tumors can create hypoxia when oxygen is depleted by photochemical consumption or the oxygen supply is compromised by microvascular damage. However, oxygen is a requirement for PDT, and hypoxia during illumination can lead to poorer tumor response. As such, sensitive methods of quantifying tumor oxygen and evaluating its distribution may help in the development and optimization of treatment protocols. In this study, the hypoxia marker EF3[2-(2-nitroimidazol-1[H]-yl)-N-(3,3,3-trifluoropropyl)acetamide]was used to evaluate the oxygenation of PDT-treated radiation-induced fibrosarcoma tumors. Tumor-bearing mice were administered Photofrin (5 mg/kg) 24 h before PDT illumination at 75 mW/cm2, 135 J/cm2 (30 min). EF3 (52 mg/kg) was injected either within 3 min before PDT illumination, with tumor excision at the conclusion of illumination, or within 3 min after illumination, with tumor excision 30 min later. Control animals received EF3 alone, EF3 plus Photofrin,or EF3 plus illumination. After tumor disaggregation, staining with a fluorochrome-conjugated monoclonal antibody, and flow cytometric analysis, control tumors demonstrated an averaged median fluorescence intensity (± SE) of 17.1 ± 2.8. EF3 binding significantly (P = 0.007) increased during PDT to a median fluorescence intensity of 48.9 ± 8.3. In the 30 min after PDT, EF3 binding returned to control levels(median, 18.3 ± 3.3). To evaluate the oxygen concentrations corresponding to these fluorescence intensities, an in vitro standard curve was created based on the in vivo exposure conditions. From this curve, the oxygen tensions of tumors exposed to EF3 under control conditions, during PDT,or after PDT were calculated to be 3.1–5.3, 1.2–2.4, and 3.0–5.2 mm Hg, respectively. Detection of EF3 binding using a monoclonal antibody correlated well with direct detection of binding using a radioactive assay. EF3 binding was linear with drug incubation for times from 1.5 to 60 min. Overall, this work demonstrates that hypoxia during PDT illumination of radiation-induced fibrosarcoma tumors can be detected by the hypoxia marker EF3. Hypoxia during illumination can be labeled separately from that found before or after PDT. Tissue oxygen tensions corresponding to EF3 binding levels can be calculated.

PDT4uses a photosensitizer, light, and oxygen to create tissue damage. Because PDT photochemistry depends on the presence of oxygen, tissue oxygen will be consumed through photochemical reactions. If oxygen consumption outpaces tissue re-oxygenation via the blood vessels, tumor hypoxia may develop during illumination (1, 2). Tissue oxygen may also become depleted during or after illumination if damage to the vascular endothelium reduces perfusion(3).

The depletion of oxygen during PDT has been measured with oxygen-sensitive electrodes during illumination of tumor spheroids in vitro(4) and animal tumors in vivo(1, 5, 6). Decreases in microvessel oxygen tensions during illumination have been measured by optical spectroscopy of phosphorescent probes (7). Treatment conditions creating oxygen depletion are associated with poorer tumor responses,whereas those favoring oxygen maintenance increase tumor cell kill(8, 9, 10), lead to more PDT-associated vascular damage(5), and ultimately, improve the cure rate of mouse tumors(11, 12, 13). The benefits of tumor oxygen maintenance during PDT and the dependence of oxygen consumption on the treatment protocol highlight the need for investigation of PDT-created hypoxia. Sensitive methods of quantifying and evaluating the distribution of tumor oxygen could help in the development and optimization of PDT protocols.

Hypoxia markers, drugs that are bioreduced under hypoxic conditions,are one method of studying tissue oxygenation. Such markers have been used by others to describe the presence of hypoxia in tumors after PDT treatment (14, 15). In these studies, increased tissue hypoxia was detected in the hours after PDT, in association with decreases in vascular perfusion. In contrast to hypoxia that develops during illumination, hypoxia after PDT may improve tumor response because damage to vasculature can contribute to tumor cures(16). Thus, the distinction between tumor hypoxia developing during illumination from that created after treatment is important in the investigation of PDT effects on tumor oxygenation.

This report describes the use of the 2-nitroimidazole hypoxia marker EF3 for detecting tumor hypoxia during or shortly after PDT. EF3 is closely related to the well-characterized marker EF5. It has been shown that EF5 covalently binds to hypoxic cells, forming adducts at a rate that varies inversely with oxygen concentration (17). The cell-bound hypoxia marker is detected with a fluorochrome-conjugated monoclonal antibody and quantified by flow cytometry or visualized by fluorescence microscopy (18). Tumor hypoxia, as labeled by EF5, directly correlates with radiobiologically detected tumor hypoxia(19) and predicts radiation resistance (20). In this study, we demonstrate that EF3 binding detects increases in tumor hypoxia during PDT illumination. Tumor oxygenation during illumination is studied separately from that found before or after treatment.

Animals and Cell Lines.

All experiments were performed using the RIF murine tumor line (kindly provided by Dr. Barbara Henderson, Roswell Park Cancer Institute,Buffalo, NY). Cells were maintained in vitro through regular passage in minimum essential α medium (Life Technologies,Inc., Grand Island, NY) supplemented with 10% FCS, 1%penicillin and streptomycin, and 300 μml-glutamine. For in vitro studies,cells were passaged no more than six times after cell isolation from a tumor. For in vivo experiments, the RIF tumor was propagated on the shoulders of female C3H mice (Taconic, Germantown, NY) by intradermal injection of 3 × 105cells. Tumors were treated ∼1 week after injection, at a size of∼5–7 mm in diameter. At this size, tumors were free of visible necrosis.

Photosensitizer and Hypoxia Marker.

Photofrin was purchased from QLT Phototherapeutics Inc. (Vancouver, BC,Canada), reconstituted in saline to 2.5 mg/ml, and then frozen in aliquots. For injection, the drug was further diluted to 0.5 mg/ml, and mice received 5 mg/kg (via tail vein) 20–24 h before PDT illumination.

The hypoxia marker EF3 was chosen for these studies because it is a more soluble drug than EF5. These drugs demonstrate a similar oxygen dependence to binding, but the increased solubility of EF3 allows the use of higher injected drug doses for the short incubation times of this study. EF3 was synthesized by Dr. M. Tracy and colleagues (SRI International, Menlo Park, CA) and dissolved in saline (20 mm) for mouse injections (via tail vein) at 10 ml/kg (52 mg/kg). Earlier studies with EF5 demonstrated relatively even drug distribution among body tissues (21); therefore, an injection of 10 ml/kg of 20 mm EF3 produced a whole body concentration of 200 μm. 14C-labeled EF3 was synthesized by coupling 2-14C-labeled azomycin acetate (NEN DuPont,Boston, MA) to 3,3,3-trifluoropropylamine.

Light Treatment.

Tumor-bearing, photosensitized (or control) animals were treated with 135 J/cm2, delivered at 75 mW/cm2. These conditions were chosen because a fluence rate of 75 mW/cm2 was shown to deplete RIF tumor oxygenation, based on needle electrode studies(5), and the rate permitted the delivery of a curative fluence (135 J/cm2) over a 30-min period(11).

Illumination was performed using a KTP YAG pumped dye module(Laserscope, San Jose, CA) tuned to produce 630 nm light. Light was delivered through micRolens-tipped fibers (Rare Earth Medical,West Yarmouth, MA) for illumination of a 1-cm diameter treatment area at 75 mW/cm2. Light intensity was measured with a power meter (Coherent, Auburn, CA). Mice were treated with one of the following illumination protocols: (a) Control animals received EF3 with tumor excision performed under anesthesia (ketamine at 175 mg/kg plus xylazine at 10 mg/kg i.p.) 30 min later. Control conditions included EF3 plus Photofrin (no light), EF3 plus illumination (no Photofrin), EF3 alone (neither Photofrin nor illumination), and tumor alone (no EF3, Photofrin, or illumination). For the light controls, illumination was performed during the 30-min EF3 incubation. (b) Tumor hypoxia during PDT was studied in animals receiving EF3 within 3 min before illumination, with tumor excision performed immediately after treatment. Anesthesia was administered during the last 5 min of treatment without interrupting illumination. (c) Tumor hypoxia after PDT was studied in mice receiving EF3 within 3 min after illumination for tumor exposure to drug over the ensuing 30 min. In each treatment group, some animals also received (via orbital plexus) bisbenzamide solution (30 mg/kg in saline; Hoechst 33342; Sigma, St. Louis, MO) at ∼1.5 min before tumor excision. The resulting fluorescence allowed visualization of perfused vasculature in frozen sections cut from these tumors. Mice to receive Hoechst injection during the last 1.5 min of PDT were anesthetized before illumination was begun. EF3 binding in mice anesthetized for the entire 30-min incubation was indistinguishable from that in mice anesthetized only for tumor removal; therefore, mice receiving the same EF3 treatment were pooled regardless of a short or 30-min anesthesia time.

Antibody Staining.

Excised tumors were cut in half perpendicular to the skin surface. One tumor half was coated with Tissue-Tek OCT compound, placed on saline-moistened filter paper, and then frozen on dry ice. These samples were sectioned at 14 μm (Zeiss Microm HM 505 N cryostat) for immunohistochemistry and fluorescence microscopy. The other tumor half was cooled to inhibit further EF3 metabolism and enzymatically digested (167 units/ml collagenase XI, 250 units/ml DNase I, 0.25 mg/ml Pronase E; all from Sigma) to produce a single cell suspension. Both the cell suspensions and the tissue sections were stained for EF3 binding using a previously described protocol (17). Briefly, samples were fixed with 4% PF,rinsed in Dulbecco’s PBS (Sigma), and blocked in PBS containing 0.3%Tween 20 and 1.5% albumin, plus 20% nonfat milk and 5% normal mouse serum. Antibody staining was for 4.5–5 h using a monoclonal antibody(ELK5-A8) conjugated to the fluorochrome Cy5 (Amersham Life Sciences,Arlington Heights, IL) for study by flow cytometry, or to Cy3(Amersham) for study by fluorescence microscopy. Samples were rinsed in PBS containing 0.3% Tween 20 and then PBS with no Tween 20, and stored in 1% PF until flow cytometry or fluorescence microscopy <1 week later. Staining controls included the evaluation of fluorescence in untreated cells and nonspecific antibody binding in tumors from animals not treated with EF3.

In Vitro EF3 Incubations.

RIF cells were exposed to EF3 under controlled oxygen concentrations using a previously described procedure (17). Cells(1 × 106) were plated in Ex-Cell 610-HSF medium (JRH Bioscience, Lenexa, KS) containing 25 mm HEPES (Life Technologies), 10% FCS, and 1% antibiotics in the center of 60-mm glass dishes that had been treated with alkaline media (15% 0.5 m carbonate, 15% newborn calf serum, and 70% water) and 0.1% gelatin to promote cell attachment. After overnight incubation, cell medium was replaced with medium containing 30 μm EF3 or 20 μm EF3 plus 10μ m14C-EF3. Dishes were placed in aluminum O-ring-sealed chambers and connected to a manifold to allow the evacuation of precise partial pressures of air from each chamber and its replacement with N2. After ∼30 min of gas exchanges to reach the desired oxygen concentration (0.005–10%O2), the chambers were removed from the manifold,brought to 37°C over 15 min, and gently shaken at this temperature for 3 h of drug incubation. Cells incubated in nonradioactive EF3 were collected with trypsin, fixed, and stained for flow cytometry as described above. Cells incubated in 14C-EF3 were lysed with 5% trichloroacetic acid, and the radioactivity in the acid-soluble and acid-insoluble components was counted using standard liquid scintillation techniques and a Packard 1900 TR counter(22). EF3 binding is reported in the acid insoluble fraction, corresponding to EF3 detected by antibody staining(17).

For studies of the time dependence of EF3 binding, cells in suspension were gently stirred in a spinner flask under a constant flow of N2 gas. EF3 at a final concentration of 175μ m was added to the flask (using a syringe inserted in an opening in a ceramic stopper), and at given time points after EF3 injection, cell aliquots were removed for antibody staining and flow cytometry. The relative oxygen concentration in the flask was monitored with a Clark-style ceramic electrode (23).

Flow Cytometry.

Flow cytometric analysis was performed with a FACSCalibur (Becton Dickinson, San Jose, CA) maintained by the Flow Cytometry Facility at the University of Pennsylvania Cancer Center. Cells to be analyzed were suspended at ∼0.5–1 × 106cells/ml in 1% PF and strained (22 μm filter) to eliminate cellular aggregates. Before each session, machine settings were adjusted such that a cellular standard produced the same absolute fluorescence. This standard was created by treating V79 cells in vitro with a predetermined exposure of hypoxia marker at a known oxygen concentration (100 μm for 4 h; <0.005%O2). The standard was stained for the hypoxia marker in parallel with the experimental samples, and its Cy5 fluorescence was set at 1000. All samples were read on the FL4 channel(λex = 635 nm,λ em = 661 nm) with a threshold forward scatter of 20. Data were plotted as histograms (cell count versus fluorescence intensity) in Cell Quest (Becton Dickinson) and read into Excel 5.0 (Microsoft) using FCS Assistant v1.3.1a β(shareware).5Cumulative frequency data were calculated in Excel 5.0.

Fluorescence Microscopy.

Fluorescence microscopy was performed on a Nikon Lab-Phot microscope equipped with a 100 W high-pressure mercury arc lamp, cooled (−25°C)CCD camera (“Quantix,” KF1400, Grade I defects;Photometrics), and automatic stage advancement (99S00 stage with 0.1 μm step size; Ludl Electronic Products). Appropriate filter cubes for each fluorochrome were purchased from Omega Optical(Brattleboro, VT); residual infrared light from the filter cubes was eliminated by two serial XF86 filters (Omega Optical)at the camera base. The camera and stage were controlled by a Macintosh 9600 Power PC computer, and adjacent microscopic fields of a section were photographed for later assembly into a single image. Imaging software (IPLab Spectrum; Scanalytics, Inc., Fairfax, VA) precisely recorded the position of each photograph, enabling a section to be rephotographed for different fluorochromes at the same position. Images of Hoechst-labeled blood vessels were made after fixation, and then sections were stained for EF3 and rephotographed. A third photograph was taken after the sections were flooded with Hoechst solution and rinsed; this technique labeled nuclei within the section, indicating section placement within the image and allowing identification of tissue edges. To permit accurate comparisons between sections photographed on different days, an image of hemocytometer-loaded calibration dye (Cy3 fluorochrome) was taken with each photography session and used to assess day-to-day variations in the fluorescence microscopy light source. Photography and image reconstruction were performed in IPLab Spectrum. Adobe PhotoShop 3.0 (Adobe Systems, Inc.,Mountain View, CA) was used to colorize and overlay the three photographs of each section. All images were taken in grayscale, and colors were produced by converting the image to red/blue/green color,and then turning off the green and blue for EF3 versus the red and blue for Hoechst-labeled vessels. Thus, EF3 appears red and Hoechst-labeled vessels appear green to aid in visual discrimination in photos. Color intensity was calculated to account for differences in the exposure time used for each section. Because images were always photographed for optimal exposure (not over- or underexposed), sections demonstrating more EF3 binding were represented by shorter exposure times, not necessarily a brighter image. Quantitative comparison of the images used the observed distribution of pixel intensities, modified by the exposure time and adjusted for the dye standard.

Statistics.

The statistical significance of differences among the flow cytometric data were calculated by the Wilcoxon test performed by JMP (SAS Institute Inc., Cary, NC). P ≤ 0.05 was considered significant.

EF3 Binding to RIF Cells in Vitro

The effects of PDT on tumor oxygenation were studied using the hypoxia marker EF3, where drug incubation was for a short time, i.e., 30 min, during or after PDT illumination. Therefore,the linearity of binding to severely hypoxic (<0.005%O2) RIF cells was investigated for a 175μ m EF3 exposure for times ranging from 1.5 to 60 min. A concentration of 175 μm EF3 was chosen based on exposures to be used in vivo. Mice received injections of 52 mg/kg EF3 to produce a whole body concentration of∼200 μm EF3. At the conclusion of a 30-min incubation, plasma drug levels of EF3 were ∼150μ m (data not shown) as a result of drug metabolism; therefore, an average in vivo exposure was 175μ m EF3 for 30 min. For in vitroincubation of severely hypoxic RIF cells in 175μ m EF3, median fluorescence intensity was linear for exposure times of 1.5–60 min (Fig. 1). The averaged median fluorescence intensity of severely hypoxic RIF cells exposed to 175 μm EF3 for 30 min was 240.

To determine the oxygen dependence of EF3 binding, RIF cells were incubated at oxygen concentrations of 0.005–10% for 3 h in 30μ m EF3. Six-fold less drug was used than for the in vivo studies (30 versus 175μ m), for an exposure six times longer (3 h versus 30 min) because gas changes to establish the lower oxygen concentrations took 30 min to complete. Fig. 2 shows the median fluorescence intensity of the EF3-incubated RIF cells plotted versus the O2 concentration of drug incubation. The curve was calculated based on the fluorescence intensity at each oxygen concentration, corrected for the ratio of maximum (at <0.005% O2) binding at an actual EF3 concentration of 175 μm and a 30-min exposure (see Fig. 1) to that from a 30 μm, 3-h exposure. On the basis of this line, there was an ∼40-fold decrease in fluorescence intensity as the oxygen concentration of drug incubation increased from 0.005 to 10%. Most of this change occurred between 0.1 and 1% O2. At oxygen concentrations of 0.005, 0.1, 1, and 10%, Fig. 2 also depicts binding as found after incubation with 14C-labeled EF3. At oxygen concentrations from 0.005 to 1%, radioactive detection of EF3 closely matched the indirect detection of EF3 with a fluorescent monoclonal antibody. At 10% O2 the radioactive assay detected ∼3-fold less EF3 binding than detection with antibody. This occurs predominantly because of a lack of sensitivity in the fluorescence assay.

EF3 Binding in Control and PDT-treated RIF Tumors.

EF3 binding during incubation in vivo was studied in RIF tumors exposed to control conditions or treated with PDT. Control EF3 incubations were performed in mice receiving EF3 alone (no Photofrin or illumination), EF3 plus Photofrin (no illumination), or EF3 plus illumination (no Photofrin). Flow cytometrically determined median fluorescence intensities in all controls ranged from 7.5 to 19.8, with the exception of a single light control with a median of 36.5. The averaged median fluorescence intensities (± SE) were 13.2 ± 3.6 for mice (n = 3) receiving EF3 without Photofrin or illumination, 15.9 ± 1.4 for mice(n = 2) receiving EF3 and Photofrin, and 22.6 ± 7.2 for mice (n = 3)receiving EF3 and illumination (including the one high value). No significant differences in fluorescence intensities were found between any of these groups; therefore, data were subsequently pooled.

EF3 binding in PDT-treated tumors was studied during the 30-min PDT illumination (75 mW/cm2, 135 J/cm2) or for the 30 min immediately after PDT completion. Fig. 3 depicts representative flow cytometric histograms of a control tumor(EF3 alone) and of a tumor incubated with EF3 during illumination. Compared with the control, the PDT-treated tumor displays a marked shift in the histogram toward higher fluorescence intensities,indicating more cell-bound EF3 and lower oxygen tensions.

The effects of PDT on EF3 binding in RIF tumors are summarized in Fig. 4 as cumulative frequency histograms. Groups containing four to eight mice each were treated with EF3 during PDT, treated with EF3 after PDT,or treated as controls. During PDT, tumors demonstrated a significant increase (P = 0.007) in tumor hypoxia, as indicated by an increase in the averaged median fluorescence intensity(±SE) from 17.1 ± 2.8 in control tumors to 48.9 ± 8.3 in tumors receiving EF3 during illumination. However, this increase in hypoxia was found only during and not shortly after the PDT illumination. Tumors labeled for hypoxia in the 30 min after PDT demonstrated an average median fluorescence intensity of 18.3 ± 3.3, similar to the control median of 17.1 ± 2.8. Low levels of nonspecific antibody binding were detected (median, 3.1 ± 0.3).

Oxygen Tensions in Control and PDT-treated Tumors.

Oxygen concentrations corresponding to EF3-dependent fluorescence intensities in tumors were calculated from the curve in Fig. 2. Table 1 lists the range of fluorescence intensities found at the 50th and 90th percentiles in control and PDT-treated tumors, as well as the oxygen concentration (in percentage and mm Hg) corresponding to these fluorescence intensities. On the basis of median binding, tumors exhibited oxygen tensions of 3.1–5.3 mm Hg under control conditions,1.2–2.4 mm Hg during illumination, and 3.0–5.2 mm Hg after illumination. When tumor oxygenation was calculated from the 90th percentile of binding, oxygen tensions were≤0.76 mm Hg during illumination but ≥0.91 and 1.2 mm Hg in control tumors and tumor studied post-PDT, respectively.

Imaging of EF3 Binding in Control and PDT-treated Tumors.

To observe the spatial distribution of hypoxia in the control and treated tumors, immunohistochemistry was performed on frozen tumor sections. Figs. 5,6 are images of frozen sections from a control tumor (Fig. 5) and from a tumor incubated with EF3 during PDT illumination (Fig. 6). EF3 binding is shown in red, where the intensity of the color indicates of the severity of tumor hypoxia detected. Perfused tumor vasculature is shown in green, whereas tumor regions demonstrating neither Hoechst (perfusion) nor EF3 antibody fluorescence appear black, to be distinguished from regions of the image not containing tissue (shown in dark blue). A typical control tumor (Fig. 5) demonstrated limited, low levels of EF3 binding, with most of this binding occurring in regions lacking Hoechst-stained vessels, i.e., poorly perfused regions. PDT treatment of RIF tumors created more extensive hypoxia, which varied in severity between tumor and section examined. Fig. 6 depicts two sections from one PDT-treated tumor. Both sections demonstrated more severe and widespread hypoxia than that found in controls. However, a noticeably greater area of the section was hypoxic in Fig. 6,B than in Fig. 6 A.

This work describes the use of the hypoxia marker EF3 for studying changes in tumor oxygenation associated with PDT. Although limited time was allowed for EF3 incubation during the illumination period of PDT,significant increases in EF3 binding could be detected as a result of PDT. This increase in EF3 labeling during illumination could be distinguished from that found shortly after illumination, when binding returned to control levels. Furthermore, using an in vitrocurve, we could relate the EF3 binding intensities of treated tumors to a tissue oxygen concentration or oxygen tension.

Tissue oxygen tensions, as determined by EF3 binding, agree well with published measurements of tissue pO2 in tumors of the same RIF strain. Tumor oxygen tensions have been measured with the Eppendorf pO2 Histograph in Photofrin-sensitized RIF tumors before PDT illumination and at specific times during the delivery of 75 mW/cm2; the median tumor pO2 was 3.9 mm Hg in controls, 0.7 mm Hg during the delivery of 0–5 J/cm2, and 2.8 mm Hg during the delivery of 20–50 J/cm2(5). On the basis of the median EF3 binding in this study, similar oxygen tensions of 3.1–5.3 mm Hg in controls and 1.2–2.4 mm Hg in tumors during illumination (75 mW/cm2, 135 J/cm2) were calculated. EF3 detection of hypoxia is an average over the illumination time, whereas the Eppendorf pO2 Histograph measures oxygenation only during needle tracking; therefore, EF3 binding reflects both the 0–5 J/cm2 and 20–50 J/cm2measurements made with the pO2 Histograph. Furthermore, because EF3 binding is averaged over 30 min, instantaneous local variations in pO2 (such as may occur during PDT) may be even larger than measurements allow. In tumors exposed to EF3 after PDT, tumor oxygenation calculated from median binding recovered to 3.0–5.2 mm Hg. Although tumor oxygenation after PDT with these conditions has not been reported using the Eppendorf pO2 Histograph, oxygen measurements for other conditions do demonstrate a similar trend, i.e., recovery of oxygen to control levels or higher immediately following illumination(5).

Using EF3-dependent fluorescence intensities, we calculated tumor oxygen tensions for binding at the 90thpercentile. As expected, oxygen tensions at this binding level were substantially lower. It was found, however, that the majority of cells in control or posttreatment tumors exhibited oxygen tensions >1 mm Hg. In contrast, tumor oxygen tensions as low as 0.05 mm Hg were calculated based on 90th percentile EF3 binding during PDT illumination.

The use of hypoxia markers to study PDT-created changes in tumor oxygenation offers several benefits. Because hypoxia marker binding increases as tumor oxygenation decreases, this technique provides an increasing signal with decreases in oxygen concentration. In the RIF line, for the EF3 incubation conditions used, oxygen concentrations from 0.005 to 1% O2 could be discriminated by antibody binding and from 0.005 to 10% could be discriminated by radioactive assay. This range of O2concentrations may be most important for investigating the consequences of oxygen depletion during PDT. Others have found that at oxygen concentrations from ≤0.5–1%, cell inactivation by PDT is reduced to half of its value under normoxia (24, 25). Cells under anoxic conditions (the equivalent to <0.005% for this study)demonstrate no response to PDT. Thus, the identification of populations of hypoxic tumor cells during PDT illumination could be valuable for understanding tumor responses to a treatment protocol.

Another benefit of hypoxia markers is that these drugs enable the investigation of spatial distributions of oxygen within a tumor. The imaging of tumor hypoxia after PDT has been studied by other investigators using different hypoxia-labeling drugs. Immunohistochemical analysis of binding of the hypoxia marker 7-(4′-(2-nitroimidazol-l-yl)-butyl)-theophylline by van Geel et al.(14) demonstrated increased tumor hypoxia within 2 h after interstitial PDT of RIF tumors. Using Hoechst to label patent blood vessels, van Geel et al.(14) detected decreases in tumor perfusion after PDT and showed that areas of 7-(4′-(2-nitroimidazol-l-yl)-butyl)-theophylline binding did not overlap with the remaining perfused areas. Moore et al.(15) used the marker[123I]iodoazomycin arabinoside to monitor tumor hypoxia after PDT by noninvasive nuclear scintigraphy; a significantly higher relative retention of [123I]iodoazomycin arabinoside could be detected in tumors 24 h after PDT than in corresponding controls. In the present work, we demonstrate the ability to label tumor hypoxia and perfused blood vessels in control and PDT-treated tumors. The PDT-treated tumors clearly demonstrated more intense EF3 binding, i.e., more severe hypoxia, as well as binding over a larger tumor area. Future studies will evaluate these and associated images more closely in an attempt to quantify the findings.

A consideration in using hypoxia markers for study of oxygen tensions in PDT-treated tumors is the possible effect of PDT-created reductions in vascular perfusion on EF3 delivery to tumors. Significant reductions in vascular perfusion during or after PDT could limit EF3 availability to the tumor, resulting in lower levels of EF3 binding than expected. To assess vessel patency during or after PDT, markers of perfusion can be used in conjunction with EF3 to simultaneously study tumor perfusion and hypoxia (14). In this study, qualitative comparison of Hoechst-stained blood vessels in control and PDT-treated tumors suggested that most tumors retained significant perfusion at the conclusion of treatment, indicating that although some changes in tumor perfusion may have occurred, widespread shutdown of tumor vasculature was not visible during PDT. Images of tumors exposed to EF3 after PDT(data not shown) reveal that although low levels of EF3 binding were found after PDT, this binding was well distributed throughout the tumor, indicating good perfusion of EF3 throughout the tumor. Furthermore, using the same treatment conditions of 5 mg/kg Photofrin,135 J/cm2 at 75 mW/cm2,Fingar et al.(3) found no PDT-associated constriction or increased permeability in the vasculature of rat cremaster muscle up to 1 h after illumination. Therefore, the changes in EF3 binding found in this study are most likely associated with the depletion of oxygen by photochemical consumption, with less contribution from PDT-induced changes in vascular perfusion. Image analysis of Hoechst fluorescence will be used to quantify vascular perfusion in these tumors.

The use of hypoxia markers to study tumor oxygenation during PDT illumination required altering typical hypoxia marker protocols. Previously, 2-nitroimidazole exposure times in rodents were at least 2–3 h to allow adequate time for drug binding (20, 26);in this study a 30-min EF3 exposure was chosen to correspond to a 30-min PDT illumination time. This short incubation time was further complicated by the fact that PDT can create changing oxygen concentrations over the course of illumination (1, 5). To confirm that tumor hypoxia during this 30-min incubation would be accurately represented by EF3 labeling, binding was studied as a function of time for exposures ranging from 1.5 to 60 min; drug binding was linear over this range. To confirm that EF3 binding, as detected indirectly through a fluorochrome-labeled antibody, corresponded to the actual EF3-cellular adducts formed, the fluorescence assay was compared to direct detection of EF3 using a radioactive assay. Good agreement between the two assays was found, although radioactivity permitted better detection of low EF3 binding at 10% O2. For the exposure conditions used, the fluorescence assay could detect a maximum oxygen concentration of 1% because of limitations in the sensitivity of antibody staining and flow cytometry.

An in vitro curve was used to determine oxygen concentrations corresponding to EF3 binding levels. In creating this curve, precision gas changes in valve-controlled airtight chambers and sustenance of cells in a minimum volume of medium helped to ensure that the plated cells were equilibrated to the intended oxygen concentration(22). Hypoxia marker binding resulting from this technique is highly reproducible as shown previously (17) and demonstrated within this study. However, EF3 binding in plated cells at<0.005% O2 in these chambers does not perfectly correlate with EF3 binding in cell suspensions (in spinner flasks) at<0.005% O2 for the time course experiments. In particular, 30 min of 175 μm EF3 exposure in the time course experiments produced a median fluorescence intensity of 240, whereas exposure to 30 μm EF3 for 180 min in the chambers resulted in median EF3 binding of 497. The most likely explanation for this discrepancy is that more EF3 binding occurred during the chamber incubations because of binding during the 30-min gas exchanges at room temperature and the subsequent 15 min to bring the chambers to 37°C, which were not considered part of the 180-min incubation time (at 37°C). In addition, although EF3 binding is linear with time of incubation, binding is likely not as linear with drug dose; therefore, a 6-fold decrease in drug dose may have been slightly overcompensated for by a 6-fold increase in incubation time. Because times of EF3 incubation could be more finely controlled for exposures in cell suspension, as in the time course experiments, these experiments were used to estimate the maximum EF3 binding (at<0.005%) expected for the in vivo exposure conditions (30 min at 175 μm). Intermediate oxygen concentrations could not as accurately be studied in cell suspension,however, so relative changes in binding from exposure at different oxygen concentrations were studied using the chambers. Thus, the final curve of fluorescence intensity versus oxygen concentration incorporated data from both types of studies to produce the most accurate representation of absolute EF3 binding at different oxygen concentrations that is possible with our current methods.

In summary, this work demonstrates that EF3 binding can be used to detect and discriminate between tumor hypoxia developing during PDT illumination and that found before or after PDT. A method for calculating tissue oxygen tensions from EF3 binding levels is presented and shown to produce values in good agreement with published data.

We thank Dr. E. Paul Wileyto for statistical advice, Dr. Alex Kachur for assistance with measuring EF3 plasma concentrations, W. Timothy Jenkins for helpful conversations on experimental design,and Carmen Rodriguez for assistance with the laser. We thank Aimee Torres, Patricia Oprysko, Deirdre Pook, and Timothy Nanni for technical assistance. We also thank the Department of Radiation Oncology, Dr. W. Gillies McKenna, Chairman, and Dr. Eli Glatstein, Vice-Chairman, for support.

Fig. 1.

EF3 binding as a function of drug incubation time for RIF cells exposed to 175 μm EF3 at an O2concentration of <0.005%. Cells were stained for EF3 binding with a fluorochrome (Cy5)-labeled monoclonal antibody (ELK5-A8) and read by flow cytometry. Data are the means ± SE (bars) of median fluorescence intensity from duplicate experiments (error bars are too small to be visible around most points). The Pearson correlation coefficient (r) for the linear fit is 0.997.

Fig. 1.

EF3 binding as a function of drug incubation time for RIF cells exposed to 175 μm EF3 at an O2concentration of <0.005%. Cells were stained for EF3 binding with a fluorochrome (Cy5)-labeled monoclonal antibody (ELK5-A8) and read by flow cytometry. Data are the means ± SE (bars) of median fluorescence intensity from duplicate experiments (error bars are too small to be visible around most points). The Pearson correlation coefficient (r) for the linear fit is 0.997.

Close modal
Fig. 2.

EF3 binding as a function of the oxygen concentration for RIF cell incubation in 30 μm EF3 for 3 h. Binding was detected fluorescently (□) by antibody staining and flow cytometry. 14C-EF3 binding was detected by measuring radioactivity (♦). To permit comparison of relative levels of EF3 binding as detected by radioactivity versusfluorescence, the radioactive data were normalized to the EF3 binding at 0.005% and then plotted such that relative binding at 0.005%, i.e., 1, lined up with the fluorescence intensity detected at 0.005%. All data points represent the mean ± SE (bars) of two to three separate experiments. The curve was drawn by correcting each fluorescence intensity for the ratio of maximum binding possible at a 175 μm, 30-min exposure (from Fig. 1) to that found for a 30 μm, 3-h exposure. Fluorescence (∼4.8) resulting from nonspecific antibody binding in vitro was not corrected because it produced an absolute background for all samples, regardless of oxygen concentration.

Fig. 2.

EF3 binding as a function of the oxygen concentration for RIF cell incubation in 30 μm EF3 for 3 h. Binding was detected fluorescently (□) by antibody staining and flow cytometry. 14C-EF3 binding was detected by measuring radioactivity (♦). To permit comparison of relative levels of EF3 binding as detected by radioactivity versusfluorescence, the radioactive data were normalized to the EF3 binding at 0.005% and then plotted such that relative binding at 0.005%, i.e., 1, lined up with the fluorescence intensity detected at 0.005%. All data points represent the mean ± SE (bars) of two to three separate experiments. The curve was drawn by correcting each fluorescence intensity for the ratio of maximum binding possible at a 175 μm, 30-min exposure (from Fig. 1) to that found for a 30 μm, 3-h exposure. Fluorescence (∼4.8) resulting from nonspecific antibody binding in vitro was not corrected because it produced an absolute background for all samples, regardless of oxygen concentration.

Close modal
Fig. 3.

Representative flow cytometry histograms of EF3 binding in a control RIF tumor (○) and a tumor treated with PDT (▪). EF3 (52 mg/kg) incubation was for 30 min; tumors were then enzymatically digested, stained with a monoclonal antibody, and read by flow cytometry. In the PDT-treated (5 mg/kg Photofrin) tumor, EF3 incubation was carried out during the illumination (75 mW/cm2, 135 J/cm2).

Fig. 3.

Representative flow cytometry histograms of EF3 binding in a control RIF tumor (○) and a tumor treated with PDT (▪). EF3 (52 mg/kg) incubation was for 30 min; tumors were then enzymatically digested, stained with a monoclonal antibody, and read by flow cytometry. In the PDT-treated (5 mg/kg Photofrin) tumor, EF3 incubation was carried out during the illumination (75 mW/cm2, 135 J/cm2).

Close modal
Fig. 4.

Cumulative frequency plots of EF3 binding in RIF tumors incubated with EF3 (52 mg/kg) under control conditions (○), during PDT illumination (▪), or shortly after PDT illumination (▴). Nonspecific antibody binding (⋄) was measured in RIF tumors not exposed to EF3 but stained with antibody. All EF3 exposures were for 30 min, followed by tumor excision, digestion, antibody staining, and flow cytometry. In PDT-treated (5 mg/kg Photofrin, 75 mW/cm2,135 J/cm2) tumors, EF3 was administered within 3 min before illumination, with drug incubation during illumination, or within 3 min after illumination, with drug incubation for the ensuing 30 min. Control conditions included exposure to EF3 alone, EF3 plus illumination (no Photofrin), or EF3 plus Photofrin (no illumination);these controls all produced similar levels of EF3 binding and were averaged for graphing. Data points represent the mean ± SE (bars); four to eight animals per group.

Fig. 4.

Cumulative frequency plots of EF3 binding in RIF tumors incubated with EF3 (52 mg/kg) under control conditions (○), during PDT illumination (▪), or shortly after PDT illumination (▴). Nonspecific antibody binding (⋄) was measured in RIF tumors not exposed to EF3 but stained with antibody. All EF3 exposures were for 30 min, followed by tumor excision, digestion, antibody staining, and flow cytometry. In PDT-treated (5 mg/kg Photofrin, 75 mW/cm2,135 J/cm2) tumors, EF3 was administered within 3 min before illumination, with drug incubation during illumination, or within 3 min after illumination, with drug incubation for the ensuing 30 min. Control conditions included exposure to EF3 alone, EF3 plus illumination (no Photofrin), or EF3 plus Photofrin (no illumination);these controls all produced similar levels of EF3 binding and were averaged for graphing. Data points represent the mean ± SE (bars); four to eight animals per group.

Close modal
Fig. 5.

EF3 binding (red) and perfused vasculature(green) in a representative section from a control RIF tumor. Brighter shades of red indicate more hypoxia, with the intensity adjusted for constant camera exposure. EF3 (52 mg/kg) incubation was for 30 min, with Hoechst 33342 (30 mg/kg) injected 1.5 min before tumor excision. At excision, tumors were frozen, sectioned at 14 μm,photographed for Hoechst fluorescence (perfused vessels), stained for EF3, and then rephotographed. Tissue edges were located by flooding the section with Hoechst solution, and non-tissue-containing areas are displayed in blue.

Fig. 5.

EF3 binding (red) and perfused vasculature(green) in a representative section from a control RIF tumor. Brighter shades of red indicate more hypoxia, with the intensity adjusted for constant camera exposure. EF3 (52 mg/kg) incubation was for 30 min, with Hoechst 33342 (30 mg/kg) injected 1.5 min before tumor excision. At excision, tumors were frozen, sectioned at 14 μm,photographed for Hoechst fluorescence (perfused vessels), stained for EF3, and then rephotographed. Tissue edges were located by flooding the section with Hoechst solution, and non-tissue-containing areas are displayed in blue.

Close modal
Fig. 6.

EF3 binding (red) and perfused vasculature(green) in sections from a RIF tumor exposed to EF3 during PDT illumination. Brighter shades of red indicate more hypoxia,with the intensity adjusted for constant camera exposure. This post-photography adjustment resulted in areas of the section(especially noticeable in B), exhibiting maximum brightness because of the substantial EF3 binding in these sections. EF3 (52 mg/kg) incubation was for 30 min, during which PDT (75 mW/cm2, 135 J/cm2) was performed; Hoechst 33342(30 mg/kg) was injected 1.5 min before PDT conclusion and tumor excision. At excision, tumors were frozen, sectioned at 14 μm,photographed for Hoechst fluorescence (perfused vessels), stained for EF3, and then rephotographed. Tissue edges were located by flooding the section with Hoechst solution, and non-tissue-containing areas are displayed in blue.

Fig. 6.

EF3 binding (red) and perfused vasculature(green) in sections from a RIF tumor exposed to EF3 during PDT illumination. Brighter shades of red indicate more hypoxia,with the intensity adjusted for constant camera exposure. This post-photography adjustment resulted in areas of the section(especially noticeable in B), exhibiting maximum brightness because of the substantial EF3 binding in these sections. EF3 (52 mg/kg) incubation was for 30 min, during which PDT (75 mW/cm2, 135 J/cm2) was performed; Hoechst 33342(30 mg/kg) was injected 1.5 min before PDT conclusion and tumor excision. At excision, tumors were frozen, sectioned at 14 μm,photographed for Hoechst fluorescence (perfused vessels), stained for EF3, and then rephotographed. Tissue edges were located by flooding the section with Hoechst solution, and non-tissue-containing areas are displayed in blue.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Salary support for T. M. Busch was provided by NIH Training Grant CA 09677.

4

The abbreviations used are: PDT, photodynamic therapy; EF3,2-(2-nitroimidazol-1[H]-yl)-N-(3,3,3-trifluoropropyl)acetamide;EF5,2-(2-nitroimidazol-1[H]-yl)-N-(2,2,3,3,3-pentafluoropropyl)acetamide;RIF, radiation-induced fibrosarcoma; PF, paraformaldehyde.

5

Available at www.bio.umass.edu.

Table 1

Calculated oxygen concentration (percentage of O2) and partial pressure (mm Hg) corresponding to fluorescence intensities of EF3 binding in control and PDT-treated tumors

Treatment conditionEF3 binding in cells at the 50th percentileEF3 binding in cells at the 90th percentile
FIaO2b (%)Partial pressurec (mm Hg)FIaO2b (%)Partial pressurec (mm Hg)
Control 11.7–22.5 0.41–0.70 3.1–5.3 54.6–89.8 0.12–0.2 0.91–1.5 
EF3 during PDT 32.7–65.2 0.16–0.32 1.2–2.4 106–225 0.006–0.1 0.05–0.76 
EF3 after PDT 11.7–24.8 0.40–0.69 3.0–5.2 47.2–67.3 0.16–0.23 1.2–1.7 
Treatment conditionEF3 binding in cells at the 50th percentileEF3 binding in cells at the 90th percentile
FIaO2b (%)Partial pressurec (mm Hg)FIaO2b (%)Partial pressurec (mm Hg)
Control 11.7–22.5 0.41–0.70 3.1–5.3 54.6–89.8 0.12–0.2 0.91–1.5 
EF3 during PDT 32.7–65.2 0.16–0.32 1.2–2.4 106–225 0.006–0.1 0.05–0.76 
EF3 after PDT 11.7–24.8 0.40–0.69 3.0–5.2 47.2–67.3 0.16–0.23 1.2–1.7 
a

95% confidence interval for the fluorescence intensity (FI) of EF3 binding in cells at the given percentile.

b

O2 concentrations (%) corresponding to the fluorescence intensities were determined from the curve of Fig. 2.

c

Tissue oxygen tensions were calculated from the percentage of O2, given that 1% O2 = 7.6 mm Hg.

1
Tromberg B. J., Orenstein A., Kimel S., Barker S. J., Hyatt J., Nelson J. S., Berns M. W. In vivo tumor oxygen tension measurements for the evaluation of the efficiency of photodynamic therapy.
Photochem Photobiol.
,
52
:
375
-385,  
1990
.
2
Foster T. H., Murant R. S., Bryant R. G., Knox R. S., Gibson S. L., Hilf R. Oxygen consumption and diffusion effects in photodynamic therapy.
Radiat. Res.
,
126
:
296
-303,  
1991
.
3
Fingar V. H., Wieman T. J., Wiehle S. A., Cerrito P. B. The role of microvascular damage in photodynamic therapy: the effect of treatment on vessel constriction, permeability, and leukocyte adhesion.
Cancer Res.
,
52
:
4914
-4921,  
1992
.
4
Nichols M. G., Foster T. H. Oxygen diffusion and reaction kinetics in the photodynamic therapy of multicell tumour spheroids.
Phys. Med. Biol.
,
39
:
2161
-2181,  
1994
.
5
Sitnik T. M., Hampton J. A., Henderson B. W. Reduction of tumour oxygenation during and after photodynamic therapy in vivo: effects of fluence rate.
Br. J. Cancer
,
77
:
1386
-1394,  
1998
.
6
Zilberstein J., Bromberg A., Frantz A., Rosenbach-Belkin V., Kritzmann A., Pfefermann R., Salomon Y., Scherz A. Light-dependent oxygen consumption in bacteriochlorophyll-serine-treated melanoma tumors: on-line determination using a tissue-inserted oxygen microsensor.
Photochem. Photobiol.
,
65
:
1012
-1019,  
1997
.
7
McIlroy B. W., Curnow A., Buonaccorsi G., Scott M. A., Bown S. G., MacRobert A. J. Spatial measurement of oxygen levels during photodynamic therapy using time-resolved optical spectroscopy.
J. Photochem. Photobiol. B
,
43
:
47
-55,  
1998
.
8
Sitnik, T. M., and Henderson, B. W. Effects of fluence rate on cytotoxicity during photodynamic therapy. In: Proceedings of SPIE-The International Society for Optical Engineering, Vol. 2972, pp. 95–102. Bellingham, WA: SPIE, 1997.
9
Iinuma, S., Wagnieres, G., Schomacker, K. T., Bamberg, M., and Hasan, T. The importance of fluence rate in photodynamic therapy with ALA-induced PpIX and BPD-MA in a rat bladder tumor model. In: Proceedings of SPIE-The International Society for Optical Engineering, Vol. 2392, pp. 136–140. Bellingham, WA: SPIE, 1995.
10
Foster T. H., Hartley D. F., Nichols M. G., Hilf R. Fluence rate effects in photodynamic therapy of multicell tumor spheroids.
Cancer Res.
,
53
:
1249
-1254,  
1993
.
11
Sitnik T. M., Henderson B. W. The effect of fluence rate on tumor and normal tissue responses to photodynamic therapy.
Photochem. Photobiol.
,
67
:
462
-466,  
1998
.
12
van Geel I. P., Oppelaar H., Marijnissen J. P., Stewart F. A. Influence of fractionation and fluence rate in photodynamic therapy with Photofrin or mTHPC, Radiat.
Res.
,
145
:
602
-609,  
1996
.
13
Gibson S. L., Foster T. H., Feins R. H., Raubertas R. F., Fallon M. A., Hilf R. Effects of photodynamic therapy on xenografts of human mesothelioma and rat mammary carcinoma in nude mice.
Br. J. Cancer
,
69
:
473
-481,  
1994
.
14
van Geel I. P., Oppelaar H., Rijken P. F., Bernsen H. J., Hagemeier N. E., van der Kogel A. J., Hodgkiss R. J., Stewart F. A. Vascular perfusion and hypoxic areas in RIF-1 tumours after photodynamic therapy.
Br. J. Cancer
,
73
:
288
-293,  
1996
.
15
Moore R. B., Chapman J. D., Mercer J. R., Mannan R. H., Wiebe L. I., McEwan A. J., McPhee M. S. Measurement of PDT-induced hypoxia in Dunning prostate tumors by iodine-123-iodoazomycin arabinoside [see comments].
J. Nucl. Med.
,
34
:
405
-411,  
1993
.
16
Henderson B. W., Fingar V. H. Oxygen limitation of direct tumor cell kill during photodynamic treatment of a murine tumor model.
Photochem. Photobiol.
,
49
:
299
-304,  
1989
.
17
Koch C. J., Evans S. M., Lord E. M. Oxygen dependence of cellular uptake of EF5 [2-(2-nitro-1H-imidazol-1-yl)-N-(2,2,3,3,3-pentafluoropropyl)acetamide]: analysis of drug adducts by fluorescent antibodies vs bound radioactivity.
Br. J. Cancer
,
72
:
869
-874,  
1995
.
18
Evans S. M., Joiner B., Jenkins W. T., Laughlin K. M., Lord E. M., Koch C. J. Identification of hypoxia in cells and tissues of epigastric 9L rat glioma using EF5 [2-(2-nitro-1H-imidazol-1-yl)-N-(2,2,3,3,3-pentafluoropropyl)acetamide].
Br. J. Cancer
,
72
:
875
-882,  
1995
.
19
Lee J., Siemann D. W., Koch C. J., Lord E. M. Direct relationship between radiobiological hypoxia in tumors and monoclonal antibody detection of EF5 cellular adducts.
Int. J. Cancer
,
67
:
372
-378,  
1996
.
20
Evans S. M., Jenkins W. T., Joiner B., Lord E. M., Koch C. J. 2-Nitroimidazole (EF5) binding predicts radiation resistance in individual 9L s.
c. tumors. Cancer Res.
,
56
:
405
-411,  
1996
.
21
Laughlin K. M., Evans S. M., Jenkins W. T., Tracy M., Chan C. Y., Lord E. M., Koch C. J. Biodistribution of the nitroimidazole EF5 [2-[2-nitro-1H-imidazol-1-yl]-N-(2,2,3,3,3-pentafluoropropyl)acetamide] in mice bearing subcutaneous EMT6 tumors.
J. Pharmacol. Exp. Ther.
,
277
:
1049
-1057,  
1996
.
22
Koch C. J., Stobbe C. C., Baer K. A. Metabolism induced binding of 14C-misonidazole to hypoxic cells: kinetic dependence on oxygen concentration and misonidazole concentration.
Int. J. Radiat. Oncol. Biol. Phys.
,
10
:
1327
-1331,  
1984
.
23
Koch C. J., Giandomenico A. R., Iyengar C. W. Bioreductive metabolism of AF-2[2-(2-furyl)-3-(5-nitro-2-furyl)acrylamide] combined with 2-nitroimidazoles. Implications for use as hypoxic cell markers.
Biochem. Pharmacol.
,
46
:
1029
-1036,  
1993
.
24
Chapman J. D., Stobbe C. C., Arnfield M. R., Santus R., Lee J., McPhee M. S. Oxygen dependency of tumor cell killing in vitro by light-activated Photofrin II.
Radiat. Res.
,
126
:
73
-79,  
1991
.
25
Moan J., Sommer S. Oxygen dependence of the photosensitizing effect of hematoporphyrin derivative in NHIK 3025 cells.
Cancer Res.
,
45
:
1608
-1610,  
1985
.
26
Evans S. M., Koch C. J., Laughlin K. M., Jenkins W. T., Van Winkle T., Wilson D. F. Tamoxifen induces hypoxia in MCF-7 xenografts.
Cancer Res.
,
57
:
5155
-5161,  
1997
.