LIGHT SCATTERING MICROSCOPY FOR TUMOR CELL RECOGNITION

Light scattering microscopy is suggested to provide some insight into cell architecture and to give information on cell malignancy or morphological changes upon necrosis or apoptosis. In the present paper a microscope setup for scattering experiments with high angular resolution is reported, and first data obtained from test models (polystyrene beads), cell monolayers and 3-dimensional cell spheroids are reported. Data are compared with simulations based on Mie scattering and discussed in view of their possible use for tumor cell recognition withi n tissue samples. The discussion includes present limitations due to spatial resolution or inhomogeneous illumination and suggests possibilities of improvement.


Introduction
In label-free diagnostics of cells and tissues, intrinsic fluorescence and non-elastic light scattering appear to be promising methods.In particular, autofluorescence was used increasingly for tumor detection of various organs including bladder, lung, larynx, breast or skin [1−7].When excited by near ultraviolet light, fluorescence of the coenzymes nicotinamid adenine dinucleotide (NADH) as well as flavin mono-and dinucleotide (FMN/FAD) seems to play a predominant role, since it reflects their state of oxidation and, consequently, cell physiology [8−10].Previously, isogenic cell lines prior and subsequent to activation of tumor suppressor genes could be distinguished on the basis of their fluorescence spectra and decay times, and this difference was related to various metabolic pathways, in particular anaerobic glycolysis and respiration [11].
For a further distinction of tumor and less malignant cells, the application of Raman spectroscopy was suggested [12][13][14], and previous experiments of isogenic glioblastoma cells revealed slight differences in the vibrational spectra around 970 cm -1 [15].In the literature, also measurements of elastic scattering are reported, and angular dependence of light scattering [16] as well as wavelength dependent backscattering [17,18] were used to distinguish tumorigenic and nontumorigenic cells or cells undergoing apoptosis.Measurements were based on differences in Mie scattering by cells or organelles of various size and were commonly performed with cell suspensions or cells within a flow cytometer [19].Since those measurements did not permit simultaneous visualization of sub-cellular structures, we now modified a conventional microscope for measuring elastic light scattering at high spatial and angular resolution.

Test Models
Polystyrene beads with a diameter of 10.2µm were used to evaluate the setup.For measurements beads were suspended in sterile filtered Earl's Balanced Salt Solution (EBSS) and transferred to a glass capillary.

Cell Cultures
Chinese Hamster Ovary cells (CHO) were routinely grown in RPMI 1640 culture medium supplemented with 10% fetal calf serum (FCS) and 1% Penicillin/Streptomycin at 37°C and 5% CO 2 .MCF-7 human breast carcinoma cells were cultivated in DMEM/Ham's F-12 culture medium supplemented with 10% fetal calf serum (FCS) and 1% Penicillin/Streptomycin at 37°C and 5% CO 2 .For measurement of single cells or small cell clusters, CHO cells were cultivated as monolayers on glass slides.

Cell Spheroids
After seeding 200 cells per well in a 96 well plate coated with agarose to prevent the cells from adhesion at the bottom, multi-cellular spheroids of MCF-7 breast cancer cells were cultivated for one week in DMEM/Ham's F-12 culture medium supplemented with 10% fetal calf serum (FCS) and 1% Penicillin/Streptomycin at 37°C and 5% CO 2 .For measurements spheroids were transferred to a glass capillary of rectangular shape with the dimensions 4.25x1.25x50mm(Hilgenberg GmbH, Malsfeld, Germany).All probes were measured in sterile filtered EBSS or Hepes to reduce the scattering background. .

Microscopy
For measurement of light scattering cell monolayers or cell spheroids located either on a cover slide or within a micro-capillary were illuminated within an inverted microscope (Axiovert 200, Carl Zeiss Jena, Germany).The expanded beam of a 470 nm laser diode (LDH 470 with driver PDL 800-B, PicoQuant, Berlin, Germany) was deflected by an adjustable mirror and focused into the aperture of the microscope objective lens (40×/0.75 or 40x/1.30oil immersion) as described earlier [20].Thus, the samples were illuminated by a parallel beam under variable angles.Scattered light was collected by the same objective lens, selected by a beam splitter and further focused by a Bertrand lens generating an image of the aperture plane.Within this image a central region of ± 1° was selected by a tiny aperture, so that backscattering could be measured in an angular range between slightly above 180° and 225−240°.A scheme of the setup is depicted in Figure 1.Measurements were combined with microscopic imaging of the aperture plane (using the Bertrand lens) or the object plane (omitting the Bertrand lens).

Bead Measurements
To compare simulations and experiments of light scattering with angular resolution, we used polystyrene beads of various diameters.Some agreement was found for beads of various sizes, e.g.10.2 ± 0.02 µm in comparison with theoretical spheres of 10.2 µm diameter and a refractive index of 1.59, as depicted in Figure 3.

Cell Measurements
First results of angular dependent light scattering of living cells were obtained from CHO colonies on glass slides.Figure 4 shows measurements of a healthy colony (red line) in comparison with a colony in an advanced state of decline, exhibiting numerous blebs and increasing amounts of cell fragments (blue line).These degraded cells are characterized by an angular dependence of light scattering which is similar to the phase function of small spheres (with repeated scattering minima every 4−5 degrees, cf. Figure 1).Less pronounced minima appearing at a higher frequency are observed for the healthy cells, since in this case scattering is expected to originate mainly from the comparably large (whole) cells.

Spheroid Measurements
In a first step towards tumor detection by measurements of angular dependent light scattering we analyzed spheroids of MCF-7 breast cancer cells with an average diameter of 200 µm.By an appropriate aperture the field of view could be varied between about 30µm and 250µm to select an area of interest.In addition, focal planes with different distances z from the bottom of the spheroid could be imaged.Typical results are depicted in Figure 5, with the focal plane being located at z = 0 µm (red curve) or z = 60 µm (blue curve).Cellular structures are more evident in the second case and may result in more pronounced maxima and minima in the angular dependence of light scattering (Figure 5).

Discussion
According to our preliminary results (and also to those of other authors) elastic light scattering may offer some information on cell morphology, which is complementary to parameters describing cell function, e.g.autofluorescence or Raman scattering.The advantage of our present setup is that small object fields can be selected in the microscope permitting measurements even of single cells or small tissue (e.g.tumor) sections of specific interest.Object size can be limited by appropriate diaphragms in the microscope (in our case about 30 µm).However, the light spot on the sample is not diffraction limited, since the laser is focused into the aperture plane of the microscope, resulting in a parallel beam in the plane of the sample.
A further question is related to the optimum coherence length of laser light, which should be larger than the scattering objects to profit from interference patterns occurring in Mie scattering.
On the other hand, additional interference may arise from various parts within the microscope, and artefacts may be created, if the coherence length is too long.In the present case a coherence length of the laser diode of 220 µm (determined by a Michelson interferometer) appeared to be appropriate to resolve Mie scattering and to avoid most artefacts.However, some inhomogeneous illumination of the samples could not be totally excluded.
Light scattering by several cells and sub-cellular structures may result in rather complex interference patterns, and angular resolution in two rather than in one direction might give additional information about individual scatterers.Therefore, rotation of the samples is planned in future experiments.In addition, measurements with polarized light [21,22] could provide further information on sub-cellular architecture.
Since Mie scattering is very sensitive to the refractive indices of the scatterers and the surrounding medium, it appears difficult to determine sizes of the scattering particles in absolute units.However, it appears possible to measure differences between tumor cells and less malignant cells, if these cells have different morphologies, e.g.different sizes of their nuclei.It appears also possible to measure swelling or shrinking processes upon necrosis or apoptosis.
Only future experiments may prove, whether mixed cultures of tumor and less malignant cells can be analyzed or whether cancer cells infiltrating healthy tissue can be detected by light scattering experiments.

Figure 1
Figure 1 Microscope setup for backscattering experiments with angular resolution.

Figure 2
Figure 2 shows the simulated phase function for Mie scattering (non-polarized light) by single polystyrene beads with diameters of 4µm, 8µm and 16µm, assuming refractive indices of n b = 1.59 for the beads and n m = 1.334 for the surrounding

Figure 2
Figure 2 Simulated phase function for single polystyrene beads with diameters of 4µm, 8µm and 16µm assuming refractive indices of nb= 1.59 (bead) and nm = 1.334 (surrounding medium).

Figure 3
Figure 3 Angular dependence of light scattering of polystyrene beads in the range of 190°−220°; red line: measurement of a bead with Ø 10.2 ± 0.02 µm; blue line: simulation for a sphere with Ø 10.22 µm (excitation wavelength: 470±5 nm; Plan-Neofluar 40x/1.30microscope objective lens with immersion oil).Inlay: image of the object plane with scattering beads.

Figure 4
Figure 4 Angular dependence of light scattering of CHO cells in the range of 190°-220°; red line: healthy colony; blue line: degraded colony with blebs and cell fragments (excitation wavelength: 470±5 nm; Plan-Neofluar 40x/0.75microscope objective lens); Inlay: transmission microscopy of both colonies.

Figure 5
Figure 5 Angular dependence of light scattering of a MCF-7 spheroid with a focal plane at 0 µm (red line) or 60 µm (blue line) from the glass slide; angular range 195°−230°; illuminated field ~90 µm, as marked by the upper blue light spot (the lower spot corresponds to an artifact) (excitation wavelength: 470±5 nm; Plan-Neofluar 40x/1.3microscope objective lens with oil immersion).