Light scattering from cancerous cells and tissues

Motivation

In this project we focus on measuring light scattering from cells and tissues to develop new diagnostic parameters for early cancer detection. Early cancer is associated with the changes in tissue organization at cellular level, such as changes in inter-cellular organization, cellular shape and density, cell nuclei size, cell nuclei density and intra-nuclear structure and other changes on a small scale not seen under conventional light microscopy [1].  Understanding of sub-cellular origins of light scattering in cells is an important step in developing new diagnostic parameters. We are studying scattering in multi-cell objects, such as cell suspensions (HeLa, HT29, T84), cell monolayers  (HeLa, HT29, T84) and epithelial tissue cancer model (pre-cancer in the rat esophagus).

Light Scattering Spectroscopy: Principle and Experimental Set-up

The angular and spectral distributions of light scattered elastically by a particle are determined by d, the size of the particle relative to the λ, wavelength of the incident light, and by m, the refractive index mismatch between the particle and the surrounding medium. The barebones of our light scattering experiment are shown on Figure 1. The incident polarized plane wave undergoes scattering in the sample at focal plane FP1 of the lens L1. Light scattered at different points of the sample, but at the same angle with respect to the incident beam is collected at a point P in the focal plane FP2. FP2 is also known as Fourier plane, since lens L1 Fourier transforms spatial distribution, I(x,y,z), of scattered light into angular distribution, I(θ,φ). In the case of broadband illumination, each point P will be associated with a spectrum, I(λ,θ=const, φ=const). In our experiment we usually collect angular-wavelength distribution I(λ,θ,φ=const) for fixed azimuthal angle j. We also collect light with the polarization either parallel, I_II(λ,θ,φ=const), or perpendicular, I_⊥(λ,θ,φ=const),  to the incident light.

Figure 1: Light Scattering Experimental Geometry

From the scattering point of view biological cells are weakly scattering with m<1.1. Also relative to the visible range of illumination wavelengths 400-700 nm some sub-cellular structures, such as nuclei or cells (5-20 um) are large, some are of the order of, such as nucleoli and mitochondria (150 nm – 2 um), and some are numerous in quantitaty, but small in size, such as protein complexes (10th of nm). We use Mie theory, which gives an exact numerical solution for scattering of a plane wave on a sphere to optimize azimuthal angles of collection φ for different size particles (Figure 2) [2].

Figure 2: Fourier plane for l=550 nm, 5 um and 50 nm sphere

Images show Fourier planes for two particle sizes of 5 um and 50 nm for various polarization and scattering angle ranges for fixed wavelength of 550 nm. Backscattering and forward scattering is determined with respect to the direction of the incident beam (θ=0 – same direction as incident beam). The grayscale shows variation in scattering intensity. Analyzing absolute scales, we conclude, that forward scattering in angles 0-10 degrees is dominated by 5 um particle with  a very large (~10¹⁰) factor in amplitude. Backscattering I_⊥ has similar ratio of the 5 um to 50 nm signal intensity, but an absolute magnitude of 5 um signal is about 10⁶ lower, then forward scattering. Finally, backscattering I_II has the largest contribution from 50 nm particle (~10^-5). Moreover one can note, relative angular uniformity of 50 nm.

Experimental set-up for collecting light scattering signals is presented on Figure 3. Light from broadband 75W Xe arc lamp is collimated using 4F-system to allow for control of beam diameter and divergence (not shown). The light delivered to the sample is linearly polarized by polarizer P. Scattered light is collected by a Fourier lens, which converts spatial distribution of scattered light intensity at the sample plane into 2D angular distribution of scattered light at Fourier plane (FP). Fourier plane is imaged through coherent fiber bundle onto an entrance slit of the spectrograph, which disperses a slice of the angular distribution I(θ,φ-fixed) in wavelength onto 2D CCD detector (θ,λ,φ-fixed). One end of the fiber bundle is free to rotate, allowing for different azimuthal angles, φ of scattering.

Figure 3: Experimental set-up

Measurements of Light Scattering in Beads, Cells and Tissues

Measurements of Light Scattering in Beads

We use polystyrene micro/nano-spheres and Mie theory to calibrate our experimental set-up for absolute magnitude of the signal and distortions. The advantage of this sample type is a known value of refractive index and size distribution. Scattering maps I(l,q,j=const) for 5 mm and 50 nm bead mixture backscattering (I_II-BS) and forward scattering (I_II-FS) are presented on Figure 4a. Data were normalized to reflectance standard, which provides almost uniform response in angle, wavelength and polarization. Large particle spectra are oscillatory in angle and wavelength, unlike smaller ones. From comparison to Mie theory, it is obvious, that I_II-FS is dominated by large particles, while I_II-BS is dominated by smaller structures (Figure 4b).

Figure 4: Mixture of 5 um and 50 nm beads. (a) Forward/backscattering I(l,q,j=const)  (b) Mie theory fit to wavelength spectra I(λ,θ=const,φ=const)

Measurements of Light Scattering in Cells

We use cell monolayers and cell suspensions to study light scattering at cellular level. Three types of cells are used in our light scattering experiments: HeLa, HT29 and T84.

Figure 5: T84 cell data. (a) Phase contrast and DAPI-fluorescence staining (b) Forward/backscattering I(l,q,j=const)  of T84 cell monolayer (c) Mie theory fit to wavelength spectra I(λ,θ=const, φ=const)

We can get information about overall structure of a cell sample by using phase contrast microscopy, as well as nuclear size distribution from DAPI-stained fluorescence image (Figure 5a). Scattering maps I(λ,θ,φ=const) for backscattering (I_II-BS) and forward scattering (I_II-FS) of the T84 cell monolayer are presented on Figure 5b. Forward scattering I_II-BS data were analyzed with Mie theory, where size distribution of the nuclei from fluorescence images was used to get an average index contrast m=1.055 (Figure 5c). Backscattering intensity I_II-BS followed an inverse power law behavior (I~λ^-γ), which was a straight line on a log-log plot (Figure 5c). This type of behavior is characteristic of the smaller particles and one of the models to describe this behavior was developed in our tissue experiments [5].

References

1. Cotran RS, Kumar V, Collins T and Robbins SL. Robbins pathologic basis of disease, Saunders (1999).

2. Hulst HCvd. Light scattering by small particles, Wiley (1957).

3.Siglin JC, Khare L and Stoner GD. "Evaluation of Dose and Treatment Duration on the Esophageal Tumorigenicity of N-Nitrosomethylbenzylamine in Rats." Carcinogenesis, 16(2):259-265 (1995).

4. V. Backman, R. Gurjar, K. Badizadegan, R. Dasari, I. Itzkan, L. T. Perelman, M. S. Feld, “Polarized Light Scattering Spectroscopy for Quantitative Measurement of Epithelial Cellular Structures In Situ”, IEEE J. Sel. Top. Quant. Elect., 5, 1019-1026 (1999).