2.1.

Sample Preparation

Tumors implanted on the CAM of chicken eggs or on the back of female BALB/c mice were used as samples. Ten fertilized hen eggs (Goto Hatchery, Inc., Gifu, Japan) were incubated at 37°C. Here, the first day of incubation is defined as the first day of embryonic development. The EMT6 mouse breast cancer cell (CRL-2755, ATCC) was cultured in Waymouth’s MB 752/1 medium (11220-035, Life technologies) containing 10% fetal bovine serum (S1820, Biowest) and an antibiotic antimycotic solution ($100\times$) (A5955-100ML, Sigma-Aldrich). Cells were prepared at a concentration of $1.25\times {10}^7\text{cells}/\mathrm{mL}$ for transplantation after removing part of the egg shell. A polytetrafluoroethylene ring was placed onto the part of the CAM that included the branch of the blood vessel on the 11th day of embryonic development. Immediately after excluding, $20\mu \mathrm{L}$ of the tumor cell solution was dropped in the ring. The ring was removed on the 13th day of embryonic development. The measurement was conducted 8 days after tumor implantation on the CAM.

Eight syngeneic female BALB/c mice (5 weeks of age) were used. The EMT6 tumor cells were prepared at a concentration of $2\times {10}^6\text{cells}/\mathrm{mL}$ for injection. The mouse received subcutaneous injections of 0.1 mL cell suspension in the dorsal region using a 27-gauge needle. Sixteen days after implantation, the tumor was ready for measurements. To prepare the tissue, the mouse was euthanized by an overdose of anesthesia. The animal experimentation protocol was approved by the Animal Experiments Committee of Osaka University. The animal experiment was performed in accordance with the regulations on animal experiments established by the Animal Experiments Committee of Osaka University. The sample size is the number of individuals.

Tumor tissues were resected and cut into 1-mm thick slices using surgical knives and scissors. Each section was sandwiched between slide glasses. The sample thicknesses of the tumor tissues were fixed at 1 mm using spacers. For a high-accurate calculation of optical properties with inverse Monte Carlo method, the sample thickness was adjusted for the reflectance and transmittance of the samples to be $>0.5\%$, which was the limit of detection sensitivity of the optical setup.

2.2.

Integrating Sphere Measurements

A double-integrating sphere system with an intervening sample was designed to measure the optical properties of biological tissues. This is a convenient tool, because it can measure diffuse reflectance ($R_{\mathrm{d}}$) and total transmittance ($T_{\mathrm{t}}$) simultaneously. Figure 1 schematically diagrams the optical properties measurement system, which uses a xenon light source [L2274(GS) and C8849, Hamamatsu Photonics K.K.]. The spheres equip with a light baffle between the detector port and the sample port. Samples were placed between two 100-mm outer diameter integrating spheres (CSTM-3P-GPS-033SL, Labsphere), which were made of a diffusely reflective material, Spectralon. The entrance port of reflectance sphere and the sample port for the spheres had 10-mm diameters. The beam-illuminated area on the sample had 1-mm diameter. The incident light was diffusely reflected from the sample surface and diffusely or directly transmitted through the sample. Then, the light was scattered in the spheres and transported through an optical fiber (CUSTOM-PATCH-2243142, Ocean Optics) to a spectrophotometer (Maya2000-Pro, Ocean Optics) as $R_{\mathrm{d}}$ and $T_{\mathrm{t}}$. The average measurement integration time was 100 ms. Spectralon standards (Labsphere Inc.) were used to calibrate the diffuse reflectance spectrum. From the experimental data, the optical properties were calculated with the inverse Monte Carlo method as described in Sec. 2.3.

Fig. 1

Schematic of the optical properties measurement system using a double-integrating sphere.

2.3.

Inverse Monte Carlo Method

We employed the inverse Monte Carlo technique to calculate the optical properties of the samples from the measured $R_{\mathrm{d}}$ and $T_{\mathrm{t}}$ values. The tissue’s optical properties were calculated for each wavelength point. The algorithm consisted of the following steps: (a) estimate a set of optical properties; (b) calculate the reflectance and transmittance with the Monte Carlo code developed by Wang et al.³⁰; (c) compare the calculated results with the measured values of the $R_{\mathrm{d}}$ and $T_{\mathrm{t}}$; and (d) reiterate the above steps until the calculated and measured values agree within the specified acceptance margin of 99.5%. This iterative process yields the set of optical properties that most closely match the measured values of reflectance and transmittance of the tissue. The cross talk between the spheres was not taken into account. The optical properties measurement system has been calibrated by gel tissue simulation phantom that was prepared using hemoglobin powder as the absorber and Intralipid as the scatter.³¹ The ${\mu }_{\mathrm{a}}$ of the hemoglobin was determined with a spectrophotometer (model U-3500, Hitachi). Different aliquots of hemoglobin solution were added to yield final ${\mu }_{\mathrm{a}}$ of the gels at 403 nm of 0.3, 0.6, 1.2, and $2.4{\mathrm{mm}}^{-1}$. Final range of the gels ${\mu }_{\mathrm{s}}^{\prime }$ in the wavelength range from 350 to 1000 nm was 0.67 to $4.82{\mathrm{mm}}^{-1}$. To validate the method for measuring the ${\mu }_{\mathrm{s}}^{\prime }$ in combination of inverse Monte Carlo technique, the ${\mu }_{\mathrm{s}}^{\prime }$ of latex sphere solution was measured. The theoretical ${\mu }_{\mathrm{s}}^{\prime }$ of the solution was calculated from Mie theory.

In these calculations, the anisotropy factor was fixed at 0.9 because this is the typical value in many tissues.³² Additionally, because the average refractive index of a single cell is 1.38 at the wavelength of 405 nm (as shown in Ref. 33), the refractive index was fixed at 1.38. To investigate the validity of the above working hypothesis, we evaluated the sensitivity to refractive indices and anisotropy factors. The optical properties were calculated using inverse Monte Carlo with different anisotropy factors and different refractive indices. We calculated at $g=0.7$ and 0.9 for $n=1.38,140$, and 1.431 (as shown in Ref. 34).

When considering a photon and several scattering events, the reduced scattering coefficient (${\mu }_{\mathrm{s}}^{\prime }$) can be defined³⁵ to describe a multiple scattering process as ${\mu }_{\mathrm{s}}^{\prime }={\mu }_{\mathrm{s}}$ ($1-g$).

In the inverse Monte Carlo method, the cross talk between the spheres was not taken into account because our estimation using the equation described by Pickering et al.³⁶ showed that the increment of the signal by the cross talk between the spheres is under 0.1%.

2.4.

Optical Penetration Depth

Two equations were used to calculate the optical penetration depth ($\delta$). When ${\mu }_{\mathrm{a}}\ll 3$ ${\mu }_{\mathrm{s}}^{\prime }$, the $\delta$ can be estimated as³⁷

2.5.

Histological Study

Harvested CAM and mouse tumor tissues were fixed with a 20% buffered formalin solution (Mildform 20NM, Wako Pure Chemical Ind.) for 32 and 26 days, respectively. Then, the tumors were sliced through the plane with the largest tumor diameter, embedded in paraffin, and sectioned at $3\mu \mathrm{m}$. Histology slides were prepared at Genostaff Co. Ltd., Tokyo, Japan. Sections were mounted on glass slides, stained with hematoxylin and eosin (H&E), and scanned with a computerized image analyzer (NanoZoomer 2.0-RS, Hamamatsu Photonics K.K.). Image analysis was performed with NDP.Scan 2.5 software that accompanied the computerized image analyzer. The number of cell nuclei was ascertained from nine randomly selected locations in the H&E selection. Each location had a total area of $\mathrm{90,000}{\mu \mathrm{m}}^2$. Cell diameter at the cut surfaces of the tissue is measured. The major axis of the cells in CAM and mouse tumor tissues was derived from 120 and 113 tumor cells, respectively.

2.6.

Tissue Extraction and High-Performance Liquid Chromatography Analysis of Protoporphyrin IX

CAMs with tumors were used in photosensitizer accumulation studies. Twenty-four samples were analyzed at time intervals ranging from 0 to 24 h following 5-aminolevulinic acid (ALA) administration. A 1-mg/egg i.v. dose of ALA was used in experiments involving the high-performance liquid chromatography (HPLC) analysis of extracted photosensitizer. To extract the protoporphyrin IX (PpIX) from tumors, tumors were homogenized in ice-cold $0.01\mathrm{mol}/\mathrm{L}$ phosphate buffered saline eight times the initial weight of the tissue using a sonicator (Vibra cell VCX130, Sonics & Materials, Inc.) for 30 s. The $100\mu \mathrm{L}$ resulting homogenate was mixed with $10\mu \mathrm{L}$ of 50% acetic acid and $300\mu \mathrm{L}$ of $N,N$-dimethylformamide (DMF)/2-propanol(IPA) (100:1 v/v). This mixture was vigorously shaken for 1 min and the phases were then separated by centrifugation. The supernatant was collected for PpIX analysis by HPLC. The pellet was suspended with $150\mu \mathrm{L}$ of DMF/IPF, shaken for 1 min, and centrifuged again. The supernatant was collected. The supernatants obtained were mixed, and HPLC analysis was carried out using a HPLC system consisted of Alliance e2695 separations module and a model 2475 Multi-Wavelength Fluorescence Detector from Waters. HPLC grade solvents from Wako Pure Chemical Industries were used as the mobile phases. The PpIX extracted from tumor was dissolved in HPLC mobile phase [acetonitrile/10 mmol/L tetrabutylammonium hydroxide solution (7:3 v/v)]. HPLC separation was carried out at a flow rate of $1.0\mathrm{mL}/\min$ on a Capcell Pak C18 UG120 (4.6 mm i.d. $\times 150\mathrm{mm}$; particle size, $5\mu \mathrm{m}$) column (Shiseido Co.) The column temperature was maintained at 40°C. The fluorescence was monitored at 630 nm with excitation set at 400 nm.

2.7.

Statistical Analysis

The data are presented as the mean with the standard deviation. Statistical analyses were performed using the Student’s $t$-test with a significance level of $P<0.05$.

3.1.

Optical Properties of Chorioallantoic Membrane and Mouse Tumor Tissues

The optical properties of CAM and mouse tumor tissues are measured with the double-integrating sphere optical setup and inverse Monte Carlo technique. Figures 2(a) and 2(b) show the calculated ${\mu }_{\mathrm{a}}$ and ${\mu }_{\mathrm{s}}^{\prime }$ spectra of the CAM and mouse tumor tissues, respectively. Hemoglobin absorption peaks occurred around 410 and 545 nm. In the wavelength range from 437 to 515 nm, the ${\mu }_{\mathrm{a}}$ values of the CAM tumor tissues were significantly higher than those of mouse tumor tissues. The difference was most pronounced at the wavelength of 489.3 nm, which was $0.40\pm 0.06$ and $0.30\pm 0.03{\mathrm{mm}}^{-1}$ ($P=0.0008$) for CAM and mouse tumor, respectively. For both models, the ${\mu }_{\mathrm{s}}^{\prime }$ spectrum was greater at shorter wavelengths, had a maximum value of $2.9\pm 0.3{\mathrm{mm}}^{-1}$ at the wavelength of 350 nm, and smoothly decreased over the wavelength range to $0.8\pm 0.1{\mathrm{mm}}^{-1}$ at a wavelength of 1000 nm. Additionally, both spectral curves had similar slopes. Hence, the difference in the values of ${\mu }_{\mathrm{s}}^{\prime }$ in the CAM and mouse tumor tissues was negligible.

Fig. 2

Absorption coefficient (${\mu }_{\mathrm{a}}$) and reduced scattering coefficient (${\mu }_{\mathrm{s}}^{\prime }$) spectra of chorioallantoic membrane (CAM) and mouse tumors in the wavelength range from 350 to 1000 nm: (a) absorption coefficient spectra and (b) reduced scattering coefficient (${\mu }_{\mathrm{s}}^{\prime }$) spectra. The error bars denote the standard deviation.

3.2.

Sensitivity to Anisotropy Factor and Refractive Index of Inverse Monte Carlo Method

Sensitivity to $g$ and $n$ of inverse Monte Carlo method was tested by the calculation with various $g$ and $n$. Figure 3 shows the ${\mu }_{\mathrm{a}}$ spectra of the CAM tumor tissue for $g=0.7$ and 0.9 for $n=1.38,1.40$, and 1.431. Figure 4 shows the ${\mu }_{\mathrm{s}}^{\prime }$ spectra of the CAM tumor tissue for $g=0.7$ and 0.9 for $n=1.38,1.40$, and 1.431. The sensitivity to different anisotropy factors and refractive indices was a little in these anisotropy factor and refractive index ranges.

Fig. 3

Absorption coefficient (${\mu }_{\text{a}}$) spectra of CAM tumors tissues for different refractive indices $n=1.38$ (solid), $n=1.40$ (long dashed), and $n=1.431$ (dashed) in the wavelength range from 350 to 1000 nm: (a) anisotropy factor $(g)=0.7$ and (b) $g=0.9$.

Fig. 4

Reduced scattering coefficient (${\mu }_{\mathrm{s}}^{\prime }$) spectra of CAM tumors tissues for different refractive indices $n=1.38$ (solid), $n=1.40$ (long dashed), and $n=1.431$ (dashed) in the wavelength range from 350 to 1000 nm: (a) anisotropy factor $g=0.7$ and (b) $g=0.9$.

3.3.

Optical Penetration Depth

Figure 5 shows the optical penetration depth ($\delta$) derived from the data of ${\mu }_{\text{a}}$ and ${\mu }_{\mathrm{s}}^{\prime }$ in Figs. 2(a) and 2(b). The CAM tumor tissue had a shorter $\delta$ than that of the mouse tumor tissue in the wavelength range of 451 to 512 nm ($P<0.05$). The $\delta$ values of the CAM and mouse tumor tissues were linearly correlated with a correlation coefficient of 0.99, and the standard deviation of $\delta$ for the CAM tumor was same or less than that of the mouse tumor tissue in the wavelength range from 350 to 1000 nm.

Fig. 5

Optical penetration depth ($\delta$) of CAM and mouse tumors in the wavelength range from 350 to 1000 nm. The error bars denote the standard deviation.

3.4.

Histological Analysis

Figure 6 shows the H&E staining section from the tumors. The numbers of cell nuclei in the CAM and mouse tumor tissues were $246\pm 33$ and $346\pm 38$ per $\mathrm{90,000}\mu {\mathrm{m}}^2$, respectively. The CAM tumor tissue was significantly less dense than that of the mouse tumor tissue ($P<0.00002$). The major axis of the tumor cell in the CAM tissue ranged from 8.96 to $33.5\mu \mathrm{m}$ (mean, $17.8\pm 4.7\mu \mathrm{m}$), whereas that in the mouse tissue ranged from 9.45 to $30.4\mu \mathrm{m}$ (mean, $17.1\pm 3.9\mu \mathrm{m}$). Hence, the major axis of the 2 tumors did not significantly differ. Additionally, the CAM tumor tissue was similar to the mouse tumor tissue in that it has spindle-shaped tumor cells arranged in irregular intertwining bands.

Fig. 6

H&E staining of a section from the tumor: (a and c) CAM tumor, and (b and d) mouse tumor.

3.5.

Protoporphyrin IX Accumulation in Chorioallantoic Membrane Tumor Tissues

Metabolic properties of PpIX were evaluated by HPLC analysis of photosensitizer extracted from tumor tissue. PpIX levels peaked at 4 h after i.v. administration of ALA (Fig. 7).

Fig. 7

Protoporphyrin IX (PpIX) concentrations in the CAM tumor tissues as a function of following injection.