Rapid quantification of cell numbers using computer images.
Rapid quantification of cell numbers using computer images.
复制标题
使用计算机图像快速量化细胞数量。
作者：
Thomas Steenstrup;Kari Clase;Kevin Hannon
The quantification of cell numbers is a standard measure for proliferation and is often used to interpret tissue culture experiments. However, counting manually using a microscope or microscopic images is time consuming. In this paper, we present a novel method for rapid quantification of nuclei numbers using fluorescent immunohistochemical staining and conventional imaging software. This technique is easily applicable, reliable and a desirable alternative to manual cell counting. In addition, we show that the technique can be used to analyze the differentiation rate of skeletal muscle cells. To determine the feasibility of using pixel quantification as an alternative to manual nuclei counting, we used proliferating BC3H1 skeletal muscle cells, which proliferate in high serum medium (7). By assaying cell numbers in proliferating cultures, we show that manual counting of nuclei can be replaced by simple pixel quantification of nuclei staining. At various time points after plating, cultures of BC3H1 cells were fixed and stained using Hoechst 33342 dye (Sigma, St. Louis, MO, USA). Images of the cultures were captured and post-processed in Adobe Photoshop. The images were adjusted with the Level function to remove background and create a uniform staining. Using the Threshold function, we converted an image to a two-bit image. Figure 1A displays an image captured by the charge-coupled device (CCD) camera (Sony, New York, NY, USA). Figure 1B displays the same image after level adjustment and conversion to a two-bit format. Using the Select Color function, the nuclei staining was selected and the number of pixels was determined by the Histogram function. By comparing the number of nuclei to the number of pixels in five images, an average number of pixels per nuclei was determined. This conversion factor was used to determine the average number of nuclei at five different time points of the BC3H1 cultures, using five images for each time point. Figure 1C demonstrates that manual counting and pixel numbers show a similar increase in cell numbers as the cultures proliferate from very low density to confluency. Figure 1D establishes a linear correlation between manually counted nuclei numbers and numbers based on converted pixel values. It should be noted that the cells decrease in size as they reach confluency, which introduces a small variation between the two methods of cell counting. Linear regression between the two methods is significant and proves the validity of our method (Figure 1D). Having shown that pixel quantification can be used to determine the number of nuclei in cultures, we investigated whether the technique could be used to analyze the amount of myosin staining and differentiation in muscle cultures. Normally with images of differentiated cell cultures, differentiation is measured by determining the number of myosin-positive cells or nuclei. Here, we show that quantifying the number of myosin-positive pixels is comparable to these techniques. To evaluate this, we Benchmarks
