2.1.

Optical Imaging System Design

The schematic diagram of our experimental system is shown in Fig. 1. The whole system includes three parts: (i) the semiconductor laser source for excitation, (ii) optical filters/near-infrared (NIR) area CCD detection system, and (iii) a standard frame-grabber carder NI-1427 as the data acquisition system. A semiconductor laser (Changchun New Industries Optoelectronics Technology Company, Changchun, China) with a wavelength of 690 nm and a power of 100 mw, was used as light source. The laser light was collimated and fell on a beam expander. The expanded beam was reflected, and then an aperture was used to select a spot with a suitable size for both the sample and the detect range of the system. This selected light was used to photoexcite the region of interest in the sample to produce [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence signal. An achromatic lens was used to collect the emission from the sample, which included the [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence signal, and other possible radiations, such as PS fluorescence, backscattered laser light, and autofluorescence from tissues in vivo. The emission feature of singlet oxygen was centered in the wavelength region of 1270 ± 10 nm. Spectral discrimination of the singlet oxygen emission from the interference mentioned above was achieved by using a long-pass filter (LP1150, Thor labs) and a NIR bandpass filter (1270BP20, Omega optical) centered at 1270 nm with a full width at half maximum of 20 nm and an attenuation optical density (OD) better than OD5. The transmitted emission from the filters was coupled by an NIR lens onto an InGaAs area detector (XEVA-CL-FPA-1.7–320-TE1–60Hz, Xenics), whose spectra response ranged from 900 to 1700 nm with a uniform quantum efficiency of 84%. The CCD surface was cooled at –20°C to reduce thermal noise. We used a standard frame-grabber carder NI-1427 for the image data acquisition. Because of the high quantum efficiency of the NIR area CCD, it showed strong capability in weak-signal detection. Throughout our study, singlet oxygen luminescence images could be acquired with an integration time as short as 1 s. In addition, due to the area imaging method, the proposed system could achieve two-dimensional measurement of singlet oxygen without scanning. Thus, the proposed system was able to produce singlet oxygen luminescence images faster than the raster-scanning methods.

Fig. 1

Schematic layout of the NIR area CCD-based [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence imaging system.

2.2.

Experimental Sections

In the study, benzoporphyrin derivative monoacid ring A [(BPD-MA), Novartis Ophthalmics Europe Ltd., Basle, Switzerland] was dissolved in 75% ethanol in different concentrations and then illuminated with light source to produce [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ . BPD-MA has been studied as a PS for photodynamic therapy in age-related macular degeneration and presents a rather high singlet oxygen quantum yield in ethanol (φ_Δ = 0.81).²² All samples were held in a 10 × 10 × 5 mm quartz cuvette mounted on a sample stage, which was able to maintain the source-sample-detector geometry constant during experiments. The space intensity of the exciting light was 60 mw/cm², and images were captured at an exposure time of 1 s throughout these experiments.

In order to show that it was able to measure singlet oxygen luminescence by this system in vivo, a pilot study was conducted on a total of six female Kunming mice (Experimental Animal Centre of San Yat-sen University of Medical Science, Guangzhou, China). In this study, BPD-MA was dissolved in 5% dextrose-in-water solution, as illustrated in the product instruction. All the mice were anesthetized with a 1.4 mg/g intra-peritoneal dose of Urethane (Shanghai Beihe Chemicals Company, Shanghai, China), and then depilated on the zone of dorsal with 8% Na₂S (Guangzhou Chemical Reagent Factory, Guangzhou, China). For the first three mice, BPD-MA solution was subcutaneously injected into the hairless area with a 0.5-μg dosage for photosensitization. After a waiting period of 30 min, which allowed for the diffusion of PS in the tissues, light exposure was performed with a density of 60 mw/cm². The second set of mice was not injected with BPD-MA solution, but was irradiated using the same exciting light. This was the control group for this part of the study.

Further experiment has been done on tumor-bearing nude mice with BPD-MA intravenously injected. Two male nude mice (BALB/C, Shanghai Laboratory Animal Center, CAS, Guangzhou, China) were used to establish transplanted cancer. Tumors were induced by subcutaneous injection of a suspension of 10⁶ HCT116 cells into the dorsal of seven-week-old male nude mice of 24 g weight. When the volume of the tumor reached ∼120 mm³, one of the mice received an intravenous injection of BPD-MA. After a waiting period of 3 h that allowed for the accumulation of PS in the tumor, light irradiation was performed in the anesthetized animals. Treatment irradiation of the tumor was performed with a spatial intensity of 60 mw/cm². The second mouse was not injected with BPD-MA, and the tumor was irradiated by the same exciting light as in the control group.

3.1.

Validation of Singlet Oxygen Luminescence Signal

To demonstrate that the signal detected by our system was indeed derived from singlet oxygen, experiments were performed with three samples of BPD-MA dissolved in 75% ethanol at the concentration of 0.5 μg/ml. The first sample contained no additional reagents, whereas the second sample had 1,4-Diazabicyclo[2,2,2]octane [(DABCO), Guangzhou Aladdin Reagent Company, Guangzhou, China], an acknowledged quencher of singlet oxygen, added to it. The last one was first bubbled with nitrogen gas to displace the dissolved oxygen, and then the bubbling stopped and the sample was exposed to air. Figure 2 shows the singlet oxygen images of samples irradiated by the exciting light of the same fluence rate. In Fig. 2a, the shapes and details of the [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence image were observed from the sample with no additional reagents added. As expected, in Fig. 2b the NIR signal disappeared with the addition of DABCO. Similarly, the signal was absent from the sample equilibrated with nitrogen, as shown in Fig. 2c. When nitrogen bubbling stopped and the sample equilibrated with nitrogen previously was exposed to air, the signal appeared again due to the diffusion of air back into the liquid sample, as shown in Fig. 2d.

Fig. 2

Comparison of near-IR emission centered at wavelength 1270 ± 10 nm, detected from BPD-MA in ethanol: (a) sample with no additional reagents added, (b) sample with addition of singlet oxygen quencher DABCO, (c) sample deoxygenated, and (d) sample in (c) oxygenated environment.

3.2.

Linear Sensing Range and Detection Sensitivity

BPD-MA was dissolved in 75% ethanol with concentrations of 0.125, 0.25, 0.5, 1.0, 2.0, and 4.0 μg/ml. These samples were prepared in order to test the linear sensing range of the system. A background image was recorded after the irradiation of 75% ethanol without PS dissolved. The [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence images were taken immediately after the start of the irradiation. In this series, each pixel represents the background-subtracted signal at 1270 nm. Figure 3a shows the series of [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence images of samples with different concentrations. It is clear that the [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence signal increases with the increasing of sample density for the same fluence rate. From the data in Fig. 3a, the average [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence intensities are 10.3, 17.2, 38.9, 78.4, 134.9, and 283.9 counts, respectively, as plotted in Fig. 3b. These average signal intensities were calculated by integration method, which averaged data between upper and lower limits in the image. The upper limit, which was used to remove the hot pixels in the spectral image, was determined by the mean of pixel value in region of signal area, and the lower limit was determined by the mean of background in the image and the mean of pixel value in region of signal area. The relationship between the average [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence intensity and the concentration of the BPD-MA is shown in Fig. 3b. It was found that the average [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence intensity showed a linear relationship with the concentration of BPD-MA from 0.125 to 4 μg/ml. The linear fitting analysis gave an r ² of 0.993.

Fig. 3

Detection of singlet oxygen luminescence from BPD-MA samples with different concentrations. (a) Images of BPD-MA dissolved in 75% ethanol with the concentration of 0.125, 0.25, 0.5, 1.0, 2.0, and 4.0 μg/ml from the top left to bottom right. The top images use colorbar ranging from 0 to 70 and the bottom images have a color bar ranging from 0 to 400. (b) The dependence of average [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ signal intensity on the concentration of BPD-MA samples derived from (a), the linear fitting give an r ² value of 0.993, and the error bars represent the standard deviation of three samples with the same concentration in each concentration.

We could calculate the detection sensitivity of the system as follos:^{23, 24}

Fig. 4

Effects of spectral smoothing, linear spatial filter used for reducing noise level. (a) original singlet oxygen luminescence image, (b) 3 × 3 pixels, (c) 5 × 5 pixels neighborhood are averaged to replace each pixel. Intensity graphs of dashed row A, across the background, are plotted on the right side of these images. The SNRs achieved from (a), (b), and (c) are 20.0621, 37.7165, and 55.2642, respectively.

3.3.

Resolving Power

The pixel size of the area CCD is 30 × 30 μm, and the magnification of this system is 0.762. The spatial resolution of 39.4 μm can be calculated in theory. To investigate the spatial resolution of the system, a series of filaments with diameters (D) of 200, 150, 100,and 50 μm were set in the center of the cross section of the collimated laser beam, sequentially. Thus, the exciting light was divided to form a light spot with a dark line in the center. Figure 5 shows the [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence images obtained on irradiation of BPD-MA dissolved in 75% ethanol. The dark fringe appeared in each of the images in Fig. 5, and each one could be distinguished. However the sharpness of dark fringes in the images deteriorated with the reduction of the dark line's width in the laser light. Because the dark fringe in the last image can be distinguished, corresponding to the 50-μm case; thus, the spatial resolution of the system is <50 μm. The resolving power, which could further be optimized by improving the magnification of the system, is sufficient and has promise in singlet oxygen monitoring in PDT.

Fig. 5

The spatial resolution of the system: The [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence images of BPD-MA samples excited by the divided laser lights with dark lines of (a) 200, (b) 150, (c) 100, and (d) 50 μm widths, respectively. All the images were obtained with an integration time of 1 s.

3.4.

Singlet Oxygen Luminescence Imaging In Vivo

A pilot study was conducted on a total of six female Kunming mice to demonstrate the system's potential in vivo. One group of three mice was injected with BPD-MA subcutaneously at 0.5 μg dosage, and the other group of three mice was injected without BPD-MA. Figure 6 shows the results of the pilot experiments in vivo. Figure 6a is a white-light image of a Kunming mouse taken using a commercial digital camera (Olympus, E300). The area marked by a circle represents the tissue area irradiated by the laser. Figure 6b illustrates singlet oxygen images obtained from one of the Kunming mice without subcutaneous injection of BPD-MA during irradiation. The CCD integration time was 1 s. This image was not background corrected, thus representing the background emission. Figure 6c shows a [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence image series acquired from one of the mice with subcutaneous injection of BPD-MA. Each image was obtained with an integration time of 1 s. These images illustrated the singlet oxygen luminescence from the area irradiated during light exposure. Compared to the weak light acquired from the mouse without PS, shown in Fig. 6b, the [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence could be measured clearly from the mouse injected with PS in vivo, as shown in Fig. 6c. From the data obtained from the two groups, the average [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence signal intensity as a function of the amount of irradiation can be calculated by the integration method, as shown in Fig. 6d. The curve that represents signal change from Kunming mice with subcutaneous injection of BPD-MA shows a decreasing [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence signal during the long exposure time, compared to the weak light intensity and no intensity depletion in the case without BPD-MA. Because this data set was not background subtracted, the data mainly represent the background light present in the measurements. The causes for the reduction in [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ signal might likely be due to PS photobleaching, the photochemical depletion of molecular oxygen in tissues.^{25, 26} The error bars in Fig. 6d represent the standard deviation between the three animals in each experiment.

Fig. 6

(a) White-light image of a Kunming mouse, the area marked by a circle is for irradiation. (b) 1270-nm emission detection from one of the mice without BPD-MA injected during irradiation. (c) Singlet oxygen luminescence time series-image obtained from one of the mice injected with BPD-MA subcutaneously during light exposure. (d) The dependence of average light intensity on the amount of irradiation, each point represents the mean of the average spatial light intensities obtained from three animals. The error bars represent the standard deviation between the three animals in each experiment. The singlet oxygen images in (b) and (c) were acquired with an integration time of 1 s.

Because the use of tumor-bearing model to study PS accumulation and reactive oxygen species (ROS) production under laser irradiation has high relevance and intravenous injection can be an accurate method of efficient tissue diffusion, we used two tumor-bearing nude mice for further experiments. One of the mice was infused with a BPD-MA concentration of 1.50 mg/kg body weight, as per product instructions. The other mouse in the study was not injected with BPD-MA. Figure 7a is a white-light image of a tumor-bearing nude mouse with the tumor area irradiated with the laser marked by a circle. Figure 7b shows the NIR emission from the tumor-bearing mice without intravenous injection of BPD-MA at the moment when the amount of irradiation was 15 s. The CCD integration time was 1 s. Similar to Fig. 6b, Fig. 7b represents the background emission too. Figure 7c shows an [TeX:] ${}^1{\rm O}_2$ ${}^1{\mathrm{O}}_2$ luminescence image series acquired from the tumor-bearing mouse with intravenous injection of BPD-MA. Each image was obtained with an integration time of 1 s. These images represent measurements of singlet oxygen in the tumor tissues during PDT. We recorded singlet oxygen luminescence images from both the tumor-bearing nude mice to monitor the singlet oxygen changes during long-term light exposure. All images were recorded with an integration time of 1 s. Average singlet oxygen signal intensity was calculated from these images by integration method. Figure 7d shows average singlet oxygen signal intensity's depletion in real time as a function of the amount of irradiation and presents data for 600 s of light exposure. Data analysis of these measurements was performed without any background subtraction. By analyzing the spectral images from Figs. 6d and 7d, the signal intensity from the mice injected with BPD-MA (both subcutaneously and intravenously) was obviously stronger than the signal from the mice without injection of BPD-MA. The signal shows similar depletion trend from both Kunming mouse with subcutaneous injection of BPD-MA and tumor-bearing nude mouse with intravenous injection of BPD-MA, indicating that the singlet oxygen measurement could be achieved in vivo during PDT by this proposed system.

Fig. 7

(a) White-light image of a tumor-bearing nude mouse; the tumor for irradiation is marked by a circle. (b) The measurement of singlet oxygen from the tumor in the mouse without intravenous injection of BPD-MA during irradiation. (c) Singlet oxygen luminescence time series-image obtained from the tumor in the mouse injected with BPD-MA intravenously during light exposure. (d) Real-time average singlet oxygen signal depletion as a function of the light exposure. The singlet oxygen luminescence images in (b) and (c) were acquired with an integration time of 1 s.