Chapter 3 Materials and Methods
3.5 Setting of microscope and operation procedure
The images of the samples are scanned by the system described in 3.1. The setting of the
microscope is in scan mode large FOV(340μm x 340μm scanning plane), scanning speed 20kHz, resolution of the image (512 x 512), and scanning along z-direction for 30 layers that the distance between each layer is 1μm. The highest intensity appears at the middle layer. After the z-scan for the area is finished, we can move the stage in x-y plane for the scan of other areas of the samples.
The laser intensity can be adjusted by rotate the half wave plate and its value is measured by a portable power meter placed after the polarizer. The excitation intensity for the z-scan is about 300mW.
Figure 3.4.1 the method to stain DAPI
Figure 3.4.2 complete sample
Chapter 4 Result and discussion
4.1 Image processing
The signals acquired by the microscope are separated to different channel by the wave bands, 610/75(nm), 525/50(nm), 460/50(nm). We take the images of DAPI in channel 525/50 for analysis since it shows the highest intensity in our system (signal of DAPI in 525/50 is a little stronger than in 460/50, and there is no DAPI signal in 610/75). The images are processed with a public domain program, ImageJ. To quantify DNA content of cells, we need to segment nucleus from the image and acquired the integrated intensity of each nucleus (the sum of the intensity value of each pixel in the nucleus). The following is the procedure to process the image and is all done with ImageJ.
With the 30 stacks of images, first sum their intensity into one image (image/stacks/z project/sum slices). To segment the nucleus, the image needs to be adjusted the threshold and transformed to
binary form. There are several methods which have different outcomes and may affect the result of the nucleus segmentation. The method used in this thesis is in image/adjust/auto threshold/method Huang. After autothreshold, use the function, watershed (process/binary/watershed), which can
separate two connected nuclei and analyze particles (analyze/analyze particles) for nucleus
segmentation. In the function, analyze particles, one can set the limit of minimum size (pixel^2) of the nucleus and exclude all small noise (set the minimum size limit, 150). The selection information is then stored in the ROI manger. To acquire the data, use measure in ROI manager on the original image.
Figure 4.1.1 Image processing and data acquisition
4.2 z-stack summation image
Considering the nucleus in the cell attached on the cover slip is not a completely flat object, the z-scan is used to quantify the total DNA content of the nucleus. The Figure (4.2.1) shows the integrated intensity of each cell from two image that one is single layer which has the highest intensity in the z-stacks and the other is the image of which the intensity is the summation of the z-stacks.
The spatial distributions along the z-axis of the nuclei on the cover slip are similar that most of them have the highest intensity in the same layer and the effect of z-summation simply is the linear intensity enhancement. The advantage of the enhancement is that the cell segmentation becomes more precise and it reflects the total DNA content. However it takes time which is proportional to the number of the scanning stacks. If the analysis only needs DNA content, the scan of z-stacks is unnecessary.
y = 1.8391x + 751.78 R² = 0.9799
0 20000 40000 60000 80000 100000 120000 140000
0 10000 20000 30000 40000 50000 60000 70000 80000
z-stacks summation
single layer
integrated intensity of each nucleus
Figure 4.2.1 integrated intensity of cells in z-stack summation image – in highest intensity single layer image
There is also a problem that should be noticed in this experiment. The enhancement ratio which defined as the slope in the Figure (4.2.1) is not a stable value. With the condition that the same sample and the successive scans, the enhancement ratios of the different z-stacks summation vary approximately from 2 to 9. The difference may be caused by the different overlapping volume of the two-photon excitation volume from different layers which is related to the laser intensity.
4.3 Comparison of CL1-0 and CL1-5
According to some measurements of DNA content by flow cytometry, the DNA content distribution of a group of the same species cells can be divide into three divisions which represent the different phases of the cycle, G1 phase, S phase and G2/M phase[13]. G1 phase is from the end of the mitosis until the DNA beginning to synthesis for cell duplication. When DNA is synthesizing until double content, the cell cycle is in the S phases. In G2/M phase, DNA in the cells remains double in G2 phase then entering the mitosis phase. Generally, most of the cells are in G1 phase that shows the largest peak in the distribution histogram. G2/M phase also has a smaller peak. However the nucleus in the late mitosis stage may be recognized as two objectives by the cell segmentation algorithm, some of the G2/M population will be classified into G1 phase.
As mention in the discussion 4.2 that laser intensity may not be stable, the fluorescence signal should be normalized before combining all the data together. To normalize, all integrated intensities of the cell in same image are divided by the median integrated intensity which has high probability in G1 phase. Then the peak of the normalized distribution will appear at the integrated intensity which has value equal to 1. There are 1324 CL1-0 nuclei and 984 CL1-5 nuclei for the analysis. The integrated intensity distribution histogram of CL1-0 and CL1-5 are shown below.
0 (normalized with the median intensity)
CL1-5
The two distributions seem similar that both of them have the G1 peaks and without the G2/M peak.
Trying to differentiate two cell lines by the integrated intensity data, the average value and the standard deviation are calculated. The mean and standard deviation of CL1-0 and CL1-5 are (1.085, 0.514) and (1.163, 0.715). Though the difference looks small, that the both larger values of CL1-5 seems to suggest that CL1-5 has higher DNA content abnormality. Because of DNA
synthesis, the signals of normal and abnormal DNA content are merged in the distribution
histogram. To avoid the problem, another attempt to find the difference of cell lines is to count the nuclei of which the integrated intensity is larger than 2.6. There are 1.1% nuclei of CL1-0 and 4.5%
nuclei of CL1-5 satisfying the condition that also suggest the higher DNA abnormality of CL1-5.
Use the same method on the analysis of nuclei area. We can see that CL1-0 nuclei are larger than the CL1-5 nuclei and the dispersion of the CL1-0 area is also larger. From the Figure 4.3.3, it shows that the integrated intensity increase with the area. However the distribution histograms of
integrated intensity seem to be different from the distributions of the area. We also wonder how the distribution of average intensity is.
Figure 4.3.1 Integrated intensity distribution histogram of CL1-0 and CL1-5
0
To get the average intensity, integrated intensity of each nucleus is divided by its area. The origin average intensity is also normalized by multiplying the same value which makes the mean of average intensity equal to the mean of integrated intensity. The distribution of the average intensity shows that most of the nuclei of CL1-0 have the value 1 average intensity. But the nuclei of CL1-5 have the similar counts from average intensity 1 to 1.4 that may imply the disperse DNA density in the CL1-5 nuclei.
Figure 4.3.2 area distribution histogram of CL1-0 and CL1-5
Figure 4.3.3 integrated intensity – area diagram of CL1-0 and CL1-5
0
Table 4.4.1 Statistical value of some quantities of CL1-0 and CL1-5 Conclusion
Though the difference of the two cell lines can be recognize from the histograms, it is more convenient to have a value as the index of two cell lines furthermore the invasive ability. As the table showed below, the percentage difference between the standard deviation of the integrated density is the largest. And it may have the potential to measure the DNA content abnormality and the invasive ability of cells.
CL1-0 CL1-5
Figure 4.3.4 Nucleus average density distribution histogram of CL1-0 and CL1-5
References
[1] 國家衛生福利部, “民國 103 年主要死因分析,” (2016)
[2] Siegel, R., et al. (2013). "Cancer statistics, 2013." CA: A Cancer Journal for Clinicians 63(1):
11-30.
[3] American Cancer Society
[4] Team, T. N. L. S. T. R. (2011). "Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening." New England Journal of Medicine 365(5): 395-409.
[5] Dyszkiewicz, W., et al. (2000). "Prognostic significance of DNA ploidy in squamous cell lung carcinoma: is it really worth it?" The Annals of Thoracic Surgery 70(5): 1629-1633.
[6] Volm, M., et al. (1985). "DNA distribution in non-small-cell lung carcinomas and its relationship to clinical behavior." Cytometry 6(4): 348-356.
[7] Bunn PA, Krasnow St, Makuch RW, Schlam ML, Schechter GP: Flow cytometry analysis of DNA content of bone marrow cells in patients with plasma cell myeloma: Clinical implications.
Blood 59528-535, 1982.
[8] Choma, D., et al. (2001). "Aneuploidy and prognosis of non-small-cell lung cancer: a meta-analysis of published data." Br J Cancer 85(1): 14-22.
[9] Boyd, R. W., Nonlinear optics, 3rd ed. (Academic, London, 2008)
[10] DeVoe, R. J., et al. (2003). Voxel shapes in two-photon microfabrication.
[11] Born, M & Wolf, E., Principles of optics : electromagnetic theory of propagation, interference and diffraction of light, 7th(expanded) ed. (Cambridge University Press, Cambridge, UK; New York;, 1999)
[12] Kapuscinski, J. (1995). "DAPI: a DNA-specific fluorescent probe." Biotech Histochem 70(5):
220-233.
[13] Roukos, V., et al. "Cell cycle staging of individual cells by fluorescence microscopy."
(1750-2799 (Electronic)).
[14] Chu, Y.-W., et al. (1997). "Selection of Invasive and Metastatic Subpopulations from a Human Lung Adenocarcinoma Cell Line." American Journal of Respiratory Cell and Molecular Biology 17(3): 353-360.