Differences in light scattering properties of a tumorigenic and a non-tumorigenic model for tissue were demonstrated using a variety of light scattering techniques, the majority of which are in vivo compatible. In addition to determining that light scattering differences exist, models for the microarchitectural changes responsible for the light scattering differences were developed.
© 2007 Optical Society of America
Measurements and analysis of light transport through tissue can provide morphological and biochemical information to aid in the detection of abnormal pathologies such as cancer. This work focuses on measurements of morphological changes. Multiple techniques are used to measure light scattering differences and to constrain a physical model for microarchitectural changes. The objective is to demonstrate an ability to measure and characterize changes in morphology that are similar to those expected to occur during precancerous changes.
Most cancers originate in the epithelium, which is comprised of cells with very little interstitial structural material. Consequently, one model for carcinogenesis would be to measure cancerous and non-cancerous epithelial cells. The ideal model would be one in which the cancerous cells were derived directly from the non-cancerous cells, as happens in vivo. Such models are not readily available for epithelial cells. There are, however, such analogous models for fibroblast cells. Rat1 and Rat1-T1 cells are one such model. As a model of cancer, we compare exponentially growing Rat1-T1 cell cultures with Rat1 cell cultures that have reached the plateau phase of growth. Rat1 cells are immortal but not tumorigenic and were derived from rat embryo fibroblast cells by an unknown spontaneous event. Tumorigenic Rat1-T1 cells were derived from Rat1 cells, by transfection of a mutant ras oncogene . The reason for measuring cells in different proliferative states is that cell proliferation plays an important role in cancer initiation and progression. Transformed cells have a higher proliferative index than the normal tissue from which they are initiated. Rat1 cells in the plateau phase of growth were chosen to model a normal tissue for two reasons: 1) they are a non-tumorigenic cell line; and 2) most of the cells in a normal tissue are non-proliferating, and are specifically arrested in the G1-phase of the cell cycle. Plateau-phase Rat1 cells are arrested predominantly in the G1-phase of the cell cycle. Rat1-T1 cells in the exponential phase of growth were chosen as a model of malignant tissue for the opposite two reasons: 1) they are a highly tumorigenic cell line; and 2) most of the cells in a nodular malignancy are proliferating, and have a high fraction of S- and G2-phase cells. Exponentially-growing Rat1-T1 cells are all proliferating and have a very high fraction of S- and G2-phase cells. We have used this and a related carcinogenesis model extensively as described in detail in several earlier publications [11, 23–26].
The use of polarized light for imaging tissue and other turbid media has been previously implemented in several configurations. As early as 1978, polarized images were obtained of monkey retina . These images were obtained by illuminating a point with linearly polarized light and examining the surrounding retina through either a polarizer parallel to the incident illumination or perpendicular to the incident illumination. More recently, the effects of scatterer size and concentration on polarized images have been investigated. As the size of the particles increases, the image taken through a polarizer parallel to the incident polarization changes from having two lobes to having four lobes, while the image taken through a perpendicular polarizer remains a cross pattern [3, 4]. Physical models for the observed patterns have been proposed [5–8].
In addition to polarized imaging, wavelength dependent polarized point measurements have been made. The wavelength dependence of polarized backscattered light is dependent on particle size and has been used to estimate the size of nuclei in the epithelium [9,10] using table top set-ups that collected light from locations within the illumination volume.
Fiber optic probes for wavelength dependent measurements of polarized light scattering have been designed and implemented [11–13]. These probes have linearly polarized illumination using 200μm optical fibers and multiple 200μm collection fibers with different polarizations. The simplest of these probes has three fibers all in a row with the center for light delivery and two collection fibers, one for the parallel and one for the perpendicular polarizations. This probe has been reported to provide in vivo measurements of nuclear diameter . A modification of this probe that has angled collection fibers allows for optical sectioning by changing the distance from the probe to the tissue or tissue phantom . Finally, the probe described in section 2.2 has one polarized delivery fiber and three collection fibers, forming a “T”. This probe can provide information on the size of scattering centers and on the number density of scattering centers. It has been used to characterize scattering in epithelial cells where it was found that the average effective scatterer size is slightly greater than 0.5 μm .
In this paper, results of polarized light imaging, unpolarized and polarized wavelength dependent fiber optic measurements, and angular dependent single scattering measurements are all described for the tumorigenic as well as the nontumorigenic model of tissue. Monte Carlo simulations are then used to model light transport in tissue and elucidate the microarchitectural changes that are the underlying causes of changes in light scattering. The model of tissue used in the Monte Carlo simulations assumes a broad size distribution. The choice of a broad distribution of scatterer sizes was motivated by our own work as well as work of others [14–17]. The specific model we are using is described in Table 1. This model was obtained by fitting polarized, angular-dependent, single scattering measurements of cells to the sum of scattering from three log-normal size distributions. We have previously published fits to the data which only used the two smaller distributions . Subsequently, the size distribution of the nuclei was determined and this size distribution was added to the fits. Others have proposed the size density function ρ(d) = Ad -α for describing scatterers in tissue, where d is the diameter of the scatterers and A and α are constants [18–20]. In other words, as the size of the particles increase, the number of them decreases. This is qualitatively true for our model described in Table 1. A decrease in scatter number with size has also been found in a light scattering and electron microscopy study to determine the size of subcellular organelles .
2. Materials and Methods
2.1. Imaging System
The imaging system shown in Figure 1 can be used to collect polarized images of light scattering and to measure the reduced scattering coefficient. To collect polarized light scattering images, the vertically mounted laser is used. This linearly polarized HeNe laser is focused on the center of the sample with a spot size of about 44 μm. The laser power at the sample is 3 μW. Scattered light emitted from the sample is collected out to a radius of ~5 mm from the center of the laser spot. The system uses a 50/50 beam splitter which allows the incident laser light to pass through it and at the same time acts as a mirror for focusing the scattered light onto the camera. The beam-splitter is placed at an angle of ~45° to the vertical. The beam splitter is tilted slightly away from 45° to eliminate specular reflection. The camera contains a nitrogen cooled CCD that is kept at -100 °C. The linear polarizer in front of the camera lens can be rotated to obtain images of light scattered from the sample with polarization parallel and perpendicular to the incident light polarization. A 3.0 OD filter situated just below the linearly polarized laser prevents excess light from entering the CCD camera, thus protecting it from damage. A linear polarizer is situated between the laser and the OD filter. This polarizer allows the incident polarization to be precisely controlled and therefore leads to more accurate measurement of parallel and perpendicular images. Typical integration times were 200 ms for a parallel image and 10 s for a perpendicular image.
The second portion of the optical set-up which measures the reduced scattering coefficient of the cells uses an unpolarized 633 nm laser which is incident at an angle of 25° to the vertical. This set-up also has an OD filter and similar focusing lens in front of the laser. The laser power is 4 μW at the sample. The image is captured by the CCD camera via the reflecting beam-splitter. To obtain an unpolarized image, images are taken with the analyzing polarizer in two positions oriented 90° from each other and added together.
2.2. Fiber-Optic System
The fiber-optic probe system shown in Figure 2 consists of tungsten lamps to provide illumination, a fiber-optic probe to deliver to and collect light from the sample, a spectrograph to disperse the light, and a thermoelectrically cooled CCD to quantitate the collected light intensities. Measurements of both unpolarized and polarized light scattering can be made. There are two light delivery fibers which are illuminated sequentially. One fiber delivers unpolarized light while the other delivers linearly polarized light as described in the caption to Fig. 2. When the unpolarized light delivery fiber is on, light is collected by fiber 2 which is 550 μm from the delivery fiber (center-to-center). When the polarized delivery fiber is on, light is collected by fibers 1, 3, and 4 as shown in Fig. 2. Each of these fibers is 550 μm from the polarized delivery fiber (center-to-center). All fibers are normal to the sample being measured. The collection fibers are lined up along one axis of the CCD camera, while the other axis is wavelength. Data can be collected from 425 to 1042 nm.
Non-tumorigenic Rat1 fibroblast cells harvested in the plateau phase of growth (Rat1P) were used as a model for non-tumorigenic tissue and tumorigenic Rat1-T1 fibroblast cells harvested in the exponential phase of growth (Rat1-T1E) were used as a model for tumor igenic tissue. Rat1 fibroblast cells harvested in the exponential phase of growth (Rat1E) and tumorigenic Rat1-T1 fibroblast cells harvested in the plateau phase of growth (Rat1-T1P) were also measured. To determine the effects due to tumorigenicity, comparisons of Rat1P versus Rat1-T1P, and of Rat1E versus Rat1-T1E were made. To determine the effects due to growth stage, comparisons of Rat1P versus Rat1E, and of Rat1-T1P versus Rat1-T1E were made.
2.4. Preparation of Cells
Monolayer cultures were routinely maintained and subcultured for up to 20 passages (cumulative population doublings 120) as described in detail elsewhere . In brief, cells were cultured as monolayers in standard tissue culture flasks using Dulbeccos Modified Eagles Medium (DMEM, Hyclone) containing 4.5 g/l D-glucose, 5% (v/v) fetal calf serum (Hyclone), 100 IU/ml penicillin, and 100 g/ml streptomycin (Hyclone), referred to as complete medium. Cell suspensions were obtained from monolayer cultures by treating 10 minutes with 0.25% trypsin in a phosphate-buffer (pH 7.4) containing 1 mM EDTA and 25 mM HEPES, followed by the addition of cold complete medium. Cell suspensions for measurements were passed twice through an 18 gauge needle, centrifuged into a pellet, and the medium was removed. The cell pellet was resuspended in PBS and centrifuged again to form a cell pellet for the imaging and fiber-optic experiments.
2.5. Cell Counting and Volume Analysis
An aliquot of each cell suspension was counted using an electronic particle counter equipped with a pulse-height analyzer (Coulter Electronics). A cell volume distribution was obtained and gates were set to select only intact cells, excluding acellular debris. Three counts were taken for each sample and averaged to determine the concentration of cells in the suspension. Cell volume distributions containing ≥104 cells were saved and processed to obtain the mean volume of the cells in the suspension. Absolute volumes were determined through calibration of the particle counter using five different sizes of polystyrene microspheres (Duke Scientific).
2.6. Determination of cell proliferative status
Flow cytometry was used to characterize the proliferative status of the cells. Briefly, cell suspensions (~1 × 106 cells) were fixed in 70% ethanol and stored at 4°C. About 24 hrs prior to flow cytometry analysis, the cells were centrifuged for 10 minutes at 1500 rpm and 4°C, and the fixation solution was removed. The cells were stained using 1ml of propidium iodide (50 μg/ml) in PBS containing CaCl2 and MgCl2 (Hyclone). RNAse (100 μg/ml) was added for the digestion of RNA. The samples were allowed to sit overnight at 4 °C before flow measurements to ensure complete RNA digestion. DNA content histograms were then collected and deconvolved using WinList and ModFit LT software packages (Verity Software House; Topsham, ME) to estimate fractions of cells in the G1-, S- and G2- phases of the cell cycle. These data confirmed that the cells were in the expected growth stages. For plateau phase cultures, the percent of cells in G1 (gap 1) is in the mid to upper 80’s, while 6–10% cells are in S (synthesis) and 4–7% in G2 (Gap 2). For exponential cultures, about 50% of the cell are in G1 and ~35% in S and ~14% in G2.
2.7. Spectroscopic Measurements of Cell Pellets and Data Analysis
The cell pellets were measured in a cylindrical, black delron sample holder of diameter 10 mm and depth 10 mm. Polarized light scattering data were acquired with both an imaging system (Section 2.1) and a fiber optic system (Section 2.2). The imaging system provides spatial information at 633 nm. The fiber optic system generates wavelength dependent data from specific spatial locations of the samples.
2.7.1. Wavelength Dependent Light Scattering Measurements
Three to seven preparations each of Rat1P, Rat1-T1E, Rat1E and Rat1-T1P were measured over several weeks. The cell measurements are made with the fiber-optic probe (Figure 2) held by an optical mount and the sample chamber raised using a jack until the probe is just in contact with the cells. Both polarized measurements with light source P and unpolarized measurements with light source U were made. All measurements were corrected to account for the wavelength dependent properties of the probe, spectrograph and CCD array using spectralon (Labsphere) as a reference. Spectralon has uniform reflectance over the range of wavelengths used.
The unpolarized data were normalized to the area under the curve from 500 nm to 800 nm and the slope from 500 to 800 nm subsequently determined. The polarized data were normalized to the area under the curve from 947 to 997 nm since the polarizers do not polarize light in this wavelength range. This normalization corrects for any differences in light collection efficiency of the detection fibers. Two intensity ratios were computed, I1/I3 and I1/I4, where I1, I3, and I4 are the light intensities from fibers 1, 3, and 4 respectively.
2.7.2. Polarized images
Six preparations of Rat1P, five preparations of Rat1-T1E, four preparations of Rat1E and five preparations of Rat1-T1P were measured as described in Section 2.1 over a time period of several months. Before nearly every experiment, a measurement of a specific sample of polystyrene spheres was measured to assure that the system was functioning propertly. Three to five images each were collected with the collection polarization parallel to and perpendicular to the incident light polarization. The raw images were normalized to 1 sec integration times and corrected for the difference in reflectance properties of the beam splitter for the two polarizations. The beam splitter was found to reflect one polarization a factor of 2.16 more efficiently than the perpendicular direction. Averages of images obtained with the detection polarization parallel to the incident light polarization were then divided by the average of images taken with the detection polarization perpendicular to the incident polarization. These ratios were computed on a pixel by pixel basis. The four quadrants in the image were then averaged in a manner consistent with the expected symmetry of the images if the excitation had been perfectly vertical. Intensity values along vertical and horizontal lines passing through the point of laser incidence (e.g. vertical and horizontal lines of Figure 5a,b) were calculated for the ratio images. The horizontal line provides information analagous to I1/I3 calculated for the fiber optic system except that rather than determining wavelength dependence, the spatial dependence is obtained. Intensity values along horizontal and vertical lines of the parallel images were also determined. A point by point ratio of these lines provides information analogous to I1/I4 measured by the fiber optic system.
2.8. Determining the Reduced Scattering Coefficient
Unpolarized images, obtained with angled illumination, were averaged and analyzed to determine the reduced scattering coefficient. The calculation for reduced scattering coefficient is based on the work of Wang et al. . The formula μ′s = sin(αi)/n(xp-xl) is used, where μ′s is the reduced scattering coefficient, αi is the angle of incidence (with respect to the vertical), n is the refractive index of water at 633 nm, which is 1.3325, xl is the center of the laser beam, and xp is the center of the peripheral circle. The center of the peripheral circle is the center of the contours of intensity that are distant enough from the point of laser incidence that they are circular. An algorithm was written to determine the center of the peripheral circle. Briefly, the intensity is determined at a 3 mm point to the left of the point of laser incidence along the y = 0 axis. The point on the other side of the peripheral circle (also with y = 0) that has the same intensity is then determined. The center of the peripheral circle is the average of these two points. Circularity of contours of log intensity near the peripheral circle demonstrates that the center of the peripheral circle was determined using points far enough from the point of laser incidence. This check was performed with all data.
2.9. Measurement of angular dependent light scattering
The system shown in Fig. 4 was used to make angular dependent measurements of single scattering from dilute cell suspensions. A HeNe laser (633 nm) is incident on the sample and scattered light is detected by an avalanche photodiode (Hamamatsu C546-01 APD module) that is rotated around the sample. Measurements were obtained with unpolarized light, light polarized parallel to the scattering plane and light polarized perpendicular to the scattering plane. The polarizer (Versalight, Meadowlark Optics) in front of the laser and the polarizer in front of the detector can be oriented either both parallel to or both perpendicular to the scattering plane. Measurements were taken every 2° to 3° for scattering angles 4°–10°. For scattering angles 10° to 165°, they are taken every 5°. When both polarizers are parallel to the scattering plane, measurements from 85°–100° are taken every 1° to 2° for better resolution. Measurements are taken in both a forward and backward direction as a system hysteresis check. At angles less than 30° (i.e. near forward scattering) an optical density filter of 1 is needed to reduce the incident laser light.
Two sample cells are used for the measurements, one clear for the small angles under 30° and one with a black backstop for all other measurements so the back reflection does not overload the light scattered at high angles from the cells. Cell suspensions of with concentration between 200,000 to 500,000 cells/ml were measured.
2.10. Monte Carlo Simulations
In previous work we have described a model for the size, shape and refractive index distributions of scattering centers in cells which was developed to agree with the measured polarized, angular dependent single scattering from cells . In the present paper, three log-normal distributions are used to model the scattering centers of the cells. Distribution 1 in Table 1 is meant to model small scattering centers such as protein complexes in water. Distribution 2 is intended to model organelles which are in a medium with a slightly higher index of refraction. (The higher index of refraction is motivated by the idea that there are proteins, sugars and lipid precursors present in the medium which will raise the index of refraction.) Distribution 3 is the measured distribution of nuclear sizes and the index of refraction values are taken from the literature. Light scattering parameters (i.e. scattering coefficients and phase functions, P(θ)) calculated from these distributions by the T-matrix method  were used as inputs to Monte Carlo simulations of light transport. The accuracy of the Monte Carlo code was detemined by comparison of simulation and experimental measurements of polystyrene spheres. The wavelength dependent Monte Carlo and experimental results agreed with in errors for 0.5 μm diameter spheres and for a mix of 0.5 μm 1.5 μm diameter spheres. A small discrepancy was found between the experimental and simulation results for images of log(I∥/I⊥). All simulation results were offset from the experimental results by 0.1. This small consistent discrepancy was irrelevant to the examination of differences performed in this paper. Each simulation was performed at least 3 times in order to obtain a standard deviation for the result. To try and reproduce the measured experimental differences between the tumorigenic and non-tumorigenic models, wavelength dependent and spatially resolved light scattering results from the Monte Carlo simulations were compared with the experimental measurements. The model of table 1 was used with one modification as the non-tumorigenic control model. The number density of the largest size distribution was decreased by 30% in order to significantly speed-up the simulations. The mean size, and relative number density of the distributions of the non-tumorigenic model were varied to generate the tumorigenic model.
3.1. Comparison of scattering from the tumorigenic and non-tumorigenic models.
Images of light scattering from Rat 1-T1 cells in the exponential phase of growth (i.e. the tumorigenic model) are shown in Fig. 5. The log of intensity as a function of position obtained with crossed excitation and analyzer polarizers, log(I⊥), has the expected shape with four nearly symmetric lobes [3,4]. The small deviation from symmetry arises from the tilt of the beam splitter away from 45° that was necessary to avoid specular reflectance. The log of the intensity obtained with parallel excitation and analyzer polarization, log(I∥), has the expected shape for scatterers with a relatively small size parameter (size parameter = 2Nr/λ, where r and N are the radius and refractive index of the particle, respectively, and λ is the wavelength) [3,4]. The ratio of light scattering intensity with parallel and perpendicular collection polarization was computed, log(I∥/I⊥). Results for the non-tumorigenic model (plateau Rat1 cells) and the tumorigenic model (exponential Rat1-T1 cells) are shown in Fig. 6a and b, respectively. The difference of the two images (b-a) is shown in Fig. 6c. The values along horizontal and vertical lines through the center of log(I∥/I⊥) images of plateau Rat1 cells and exponential Rat1-T1 cells were determined and their average and standard deviations plotted for each cell type in Fig. 7. At all locations the average of log(I∥/I⊥) is greater for the exponential Rat1-T1 cells than for the plateau Rat1 cells. The error bars overlap everywhere except for the horizontal lines at a distance of 0.06 mm from the center of the images.
Measurements with the fiber optic probe confirm the result that I∥/I⊥ is greater for the tumorigenic model than for the non-tumorigenic model. Figure 8a shows the ratio of light collected by fibers 1 and 3 as a function of wavelength. The ratio, I1/I3, is analogous to points on Figure 7a at distances of 550 μm from the point of laser incidence. The ratio of I1/I3 for exponential Rat1-T1 cells to I1/I3 for plateau Rat1 cells at 633 nm is 1.08±0.06. The equivalent ratio obtained from the horizontal lines in Figure 7a is 1.18±0.10. Results for I1/I4 are shown in Figure 8b. The values for Rat1-T1 cells in the exponential phase of growth are clearly lower.
Unpolarized data from the fiber optic probe also demonstrates a difference between Rat1-T1 cells in the exponential phase of growth and Rat1 cells in the plateau phase of growth. In Figure 9, there is a significant difference between the slope of the data for the tumorigenic and non-tumorigenic models with the slope being significantly greater for the tumorigenic model (exponential Rat1-T1 cells).
3.2. Differences in single scattering properties
All of the light scattering results presented so far were obtained using techniques that primarily measure multiple scattering. The differences in scattering properties seen with these methods must ultimately result from changes in single scattering properties. To measure the changes in single scattering, polarized angular dependent light scattering was measured. The data in Figure 11 have been normalized to have the same intensity from 30° – 50° to account for any (small) differences in the concentration of cells in the samples. Light scattering at large angles is greater for the exponential Rat1-T1 cells than for the plateau Rat1 cells. Similarly, there is a small increase in light scattering at small angles for exponential Rat1-T1 cells. These results indicate that there are relatively more of the smallest scatterers in the tumorigenic model and that the concentration or size of the very largest scatterers has also increased.
3.3. Reduced scattering coefficients
Reduced scattering coefficients were determined for both the tumorigenic and non-tumorigenic models as well as for Rat-T1 cells in the plateau phase of growth and Rat1 cells in the exponential phase of growth. The results in Fig. 10 show that all reduced scattering coefficients were near 4 cm-1. These data were used to help constrain the model in section 3.5.
3.4. Summary of light scattering differences
In addition to the measurements of Rat1-T1 cells in the exponential phase of growth and Rat1p cells in the plateau phase of growth that have already been presented, light scattering measurements were also made on Rat1-T1 cells in the plateau phase of growth and of Rat1 cells in the exponential phase of growth. Table 2 summarizes the differences that were found in light scattering properties. The first line compares the tumorigenic and non-tumorigenic models. Differences were found for some part of the spectral or spatial ranges for all but one of the metrics. The second and third lines compare exponential and plateau phase cells for each of the two cell types. The results are very consistent for the two cell types and with the changes seen between the tumorigenic and non-tumorigenic models. The final two lines of the table compare tumorigenic and non-tumorigenic cells both when they are when they are in the same growth phase. The only differences is consistent with the difference seen in the tumorigenic and non-tumorigenic models.
3.5. Microarchitectural changes that can result in the observed light scattering changes.
The measured differences in light scattering must result from microarchitectural changes. To understand what these microarchitectural changes might be, the effects on light scattering properties of changing the scattering distributions in Table 1 were determined. Initially, simulations were run to determine the effects of individually changing the mean size or number density of each distribution in Table 1. None of the changes in these six parameters resulted in the same qualitative changes as were seen experimentally. For example a 20% decrease in the mean size of the middle distribution decreased I1/I4, increased I1/I3, but had no effect on slope. However, based on the results of these first six simulations, combinations of parameter changes were chosen that were expected to result in the changes seen experimentally. Two combinations of parameter changes resulted in the same changes as the light scattering differences between the tumorigenic and the non-tumorigenic model. One combination was a 20% decrease in the mean size of the middle size distribution and a 20% decrease in the concentration of the middle size distribution and is referred to as simulation 1. The other combination was a 20% decrease in the mean size of the middle size distribution and a 20% increase in the concentration of the smallest size distribution and is referred to as simulation 2. For both of these simulations, the reduced scattering coefficient was adjusted (by multiplying all concentrations to a constant) to be within 0.5 cm-1 of that for the non-tumorigenic control model.
The differences in log(I∥/I⊥) for the tumorigenic versus non-tumorigenic simulation models are shown in Fig. 12a and b. These images resemble Fig. 6c with the same cross shape. (Note that the intensity scale is slightly different.) The differences seen in the fiber optic measurements were also qualitatively reproduced by the simulations. The slope of the unpolarized data is more negative for the exponential Rat1-T1 cells than for the plateau Rat1 cells as demonstrated by Fig. 12c. In the experiments I1/I3 was found to be greater for the exponential Rat1-T1 cells and I1/I4 was found to be greater for the plateau Rat1 cells. The same results were found in the simulations as shown in Fig. 12d and e. The simulation results for angularly resolved scattering also show changes very similar to those found in the experiments as can be demonstrated by comparisons of Figs. 11 and 13. The angularly resolved scattering results where quite similar for the two simulations and consequently only the results for one simulation are shown.
4. Discussion and Conclusions
Significant differences in light scattering properties have been found between a tumorigenic and a non-tumorigenic model of tissue. The results in Table 2 indicate that the differences in proliferative status are the major contributor to the differences in the tumorigenic and non-tumorigenic models. A Raman and infrared spectroscopy study of these models demonstrated differences in the biochemical composition and, analagous to this study, the majority of the differences were due to changes in proliferative status rather tumorigenicity per se .
The present light scattering results can also be compared with previous work on similar cell lines. Previously, it was found that the slope of similar unpolarized light scattering measurements is steeper and there is relatively more high angle scattering for exponential cells than for plateau phase MR1 rat fibroblast cells . Additionally, the average scatter size appears to be smaller for exponential than plateau phase MR1 cells . These results are identical to what was observed in this work.
The light scattering measurements of model cell systems can also be compared to in vivo measurements of cervical tissue . The slope was steeper and I1/I3 was larger for dysplastic tissue than for non-dysplastic tissue which is consistent with the results in the present paper. The change in I1/I4, however, is opposite of what was observed here. Further work is needed to understand the microarchitectural changes underlying these light scattering changes in cervical tissue.
The many light scattering techniques that were used not only demonstrated a difference in light scattering properties between the tumorigenic and non-tumorigenic models of tissue, but also helped elucidate the microarchitectural changes that cause the light scattering changes. Monte Carlo simulations of light transport were performed for several models of scatter size distributions and relative refractive indices. It was found that a 20% decrease in the size of the scatterer distribution representing the organelles combined with a 20% increase in the number density of the smallest distribution of scatterers (simulation 2) qualitatively reproduced the experimentally measured light scattering changes. Smaller percentage changes would have likely yielded better agreement with experimental results. It was found that a 20% decrease in the size of the scatterer distribution representing the organelles combined with a 20% decrease in the number density of this same distribution of scatterers (simulation 1) also qualitatively reproduced the experimentally measured light scattering changes. However, the change in I1/I3 is much greater than that seen experimentally. No changes in the size of the distribution representing homogeneous nuclei were needed to reproduce the experimental results. The simulations demonstrated that changes in nuclear size or number density had relatively small effects on measured light scattering properties compared to changes in the other size distributions. The effects of index of refraction variations were not studied. Increasing the relative refractive index of the smallest scatterer size distribution is expected to have a very similar effect to increasing the number density of the smallest scatterer size distribution .
Changes in subnuclear microarchitectural features corresponding to cancerous or precancerous changes have been reported in the literature. In both instances, the tissue model assumed that the mass density could be described by M(r) ~ rD. In this model, an increase in D corresponds to a decrease in smaller structures and an increase in larger structures . For ex vivo rat esophageal samples, D appears to increase with dysplasia, indicating that there are relatively more large scatterers in the dysplastic tissue , the opposite of what was seen for our tumorigenic model. In a study of azoxymethane treated rat colons, the fractal dimension appears to be lower for the azoxymethane treated colons than for the controls indicating a relative increase in small scattering centers , consistent with results of this paper. (Azoxymethane induces colon carcinogenesis, however, these animals were sacrificed before any histological changes were present.)
In conclusion, differences in light scattering properties of a tumorigenic and a non-tumorigenic model have been demonstrated using a variety of light scattering techniques, the majority of which are in vivo compatible. The changes in light scattering due to tumorigenicity were smaller than those due to changes in proliferative status possibly because the tumorigenic and non-tumorigenic cells differed only by a ras mutation. In addition to determining that light scattering differences exist, models for the changes in microarchitecture that cause the light scattering differences were developed.
The authors gratefully acknowledge funding through the NIH NCI (grant CA71898) and use of the NIH flow cytometry resource at Los Alamos National Laboratory (grant RR01315). We also thank Anabel Guerra for doing some of the flow cytometric analysis.
References and links
1. L. A. Kunz-Schughart, A. Simm, and W. Mueller-Klieser, “Oncogene-associated transformation of rodent fibroblasts is accompanied by large morphologic and metabolic alterations,” Oncol. Rep. 2,651–661(1995).
3. A.H. Hielscher, J.R. Mourant, and I.J. Bigio, “Influence of particle size and concentration on the diffuse backscattering of polarized light from tissue phantoms and biological cell suspensions,” Appl. Opt. 36,125–135(1997). [CrossRef] [PubMed]
5. M. Dogairu and T. Asakaru, “Polarization-dependent backscattering patterns from weakly scattering media,” J. Opt. (Paris) ,24,271–278 (1993). [CrossRef]
6. M. J. Rakovic and G. W. Kattawar, “Theoretical analysis of polarization patterns from incoherent backscattering of light,” Appl. Opt. 37,3333–3338 (1998). [CrossRef]
7. T. M. Johnson and J. R. Mourant, “Polarized wavelength-dependent measurements of turbid media,” Opt. Express 6,200–216 (1999). [CrossRef]
8. M. R. Ostermeyer, D. V. Stephens, L. Wang, and S. L. Jacques “Nearfiled polarization effects on light propagation in random media,” in: OSA TOPS on Biomedical Optical Spectroscopy and Diagnostics , E. Sevick-Muraca and D. Benaron, Eds., Vol.3. pp.20–25 (1996).
9. K. Sokolov, R. Drezek, K. Gossagee, and R. Richards-Kortum, “Reflectance spectroscopy with polarized light : Is it sensitive to cellular and nuclear morphology,” Opt Express 5,302–317 (1999). [CrossRef] [PubMed]
10. V. Backman, R. Gurjar, K. Badizadegan, I. Itzkan, R.R. Dasari, L.T. Perelman, and M.S. Feld, “Polarized light scattering spectroscopy for quantitative measurement of epithelial cellular structures in situ,” IEEE J. Quantum Electron. 5,1019–1026(1999). [CrossRef]
11. J.R. Mourant, T, M. Johnson, and J.P. Freyer, “Characterizing mammalian cells and cell phantoms by polarized backscattering fiber-optic measurements,” Appl. Opt. 40,5114–5123 (2001). [CrossRef]
12. A. Myakov, L. Nieman, L. Wicky, U. Utzinger, R. Richards-Kortum, and K. Sokolov, “Fiber optic probe for polarized reflectance spectrocopy in vivo : Design and performance,” J. Biomed. Opt. 7,388–397 (2002). [CrossRef] [PubMed]
13. L. Nieman, A. Myakov, J. Aaron, and K. Soklov, “Optical sectioning using a fiber probe with an angled illumination-collection geometry: evaluation in engineered tissue phantoms,” Appl. Opt. 43,1308 –1319 (2004). [CrossRef] [PubMed]
14. J. R. Mourant, T. M. Johnson, S. Carpenter, A. Guerra, and J. P. Freyer, “Polarized angular dependent spectroscopy of epithelial cells and epithelial nuclei to determine the size scale of scattering structures,” J. Biomed. Opt. 7,378–387 (2002). [CrossRef] [PubMed]
16. M. Bartlett, G. Huang, L. Larcom, and H. Jiang, “Measurement of particle size distribution in mammalian cells in vitro by use of polarized light spectroscopy,” Appl. Opt. 43,1296–1307 (2004). [CrossRef] [PubMed]
18. D. Passos, J.C. Hebden, P.N. Pinto, and R. Guerra, “Tissue phantom for optical diagnostics based on a suspension of microspheres with a fractal size distribution,” J. Biomed. Opt. 10,064036-1—11, (2005). [CrossRef]
19. B. Gelebart, E. Tinet, J. M. Tualle, and S. Avriller, “Phase function simulation in tissue phantoms: a fractal approach,” Pure Appl. Opt. 5,377–388 (1996). [CrossRef]
21. H. Fang, M. Ollero, E. Vitkin, L.M. Kimerer, P.B. Cipolloni, M. N. Zaman, S. D. Freedman, I. J. Bigio, I. Itzkan, E. B. Hanlon, and L. T. Perelman, “Noninvasive sizing of subcellular organelles with light scattering spectroscopy,” IEEE J. Sel. Top. Quantum Electron. 9,267–276 (2003). [CrossRef]
23. J.R. Mourant, J.P. Freyer, A.H. Hielscher, A.A. Eick, D. Shen, and T.M. Johnson, “Mechanisms of light scattering from biological cells relevant to noninvasive optical-tissue diagnostics,” Appl Opt. 37,3586–3593 (1998). [CrossRef]
24. J.R. Mourant, M. Canpolat, C. Broker, O. Esponda-Ramos, T. Johnson, A. Matanock, K. Stetter, and J.P. Freyer, “Light scattering from cells: the contribution of the nucleus and the effects of proliferative status,” J. Biomed. Opt. 5,131–137 (2000). [CrossRef] [PubMed]
25. J. R. Mourant, S. Carpenter, K. W. Short, P. Kunapareddy, L. Coburn, and J. P. Freyer “Biochemical differences in tumorigenic and non-tumorigenic cells measured by Raman and infrared spectroscopy,” J. Biomed. Opt. 10,031106-1 to031106-15 (2005). [CrossRef]
26. J. R. Mourant, A. H. Hielscher, A. A. Eick, T. M. Johnson, and J. P. Freyer, “Evidence of intrinsic differences in the light scattering properties of tumorigienic and nontumorigenic cells,” Cancer Cytopath. 84,366–374 (1998).
27. M. I. Mischenko and L. D. Travis, “Capabilities and limitations of a current fortran implementation of the T-matrix method for randomly oriented, rotationally symmetric scatterers,” J. Quantum Spectrosc. Radiat. Transf. 60,309324 (1998).
28. J. R. Mourant, T.J. Bocklage, T. M. Powers, H. M. Greene, K. L. Bullock, L. R. Marr-Lyon, M. H. Dorin, A. G. Waxman, M. M. Zsemlye, and H. O. Smith, “In vivo light scattering measurements for detection of precancerous conditions of the cervix,” accepted Gynecological Oncology 2007.
29. R. Drezek, M. Guillard, T. Collier, I. Biodo, A. Malpica, C. Macaulay, M. Follen, and R. Richards-Kortum, “Light scattering from cervical cells throughout neoplastic progression: influence of nuclear morphology, DNA content, and chromatin texture,” J. Biomed. Opt. 8,7–16 (2003). [CrossRef] [PubMed]
30. A. Wax, C. Yang, M. G. Müller, R. Nines, C. W. Boone, V. E. Steele, G. D. Stoner, R. R. Dasari, and M. S. Feld, “In situ detection of neoplastic transformation and chemopreventive effects in rat esophagus epithelium using angle-resolved log coherence interferometry,” Cancer Research ,63,3556–3559 (2003). [PubMed]
31. Y. L. Kim, Y. Liu, R. K. Wali, H. K. Roy, M. J. Goldberg, A. K. Kromin, K. Chen, and V. Backman, “Simultaneous measurement of angular and spectral properties of light scattering for characterization of tissue microarchitecture and its alteration in early precancer,” IEEE J. Qunatum Electron. 9,243–256 (2003). [CrossRef]