KR102655391B1 - Method for counting cells in microdroplet - Google Patents

Abstract

The present invention relates to a method of simply and economically counting cells by dropping a liquid sample containing cells and using an optical microscope image of the droplet, and a method of calculating the cell concentration in the liquid sample using this. In detail, it relates to a method for counting cells in microdroplets (first invention) and a method for tracking changes in the number of cells in microdroplets using the same (second invention).

Description

Method for counting cells in microdroplet

The present invention relates to a droplet-based cell counting method. More specifically, a method of simply and economically counting cells by dropping a liquid sample containing cells and using an optical microscope image of the droplet, and using this to It relates to a method of calculating cell concentration in a liquid sample.

In biological research using cells or in clinical practice, it is very important to know the exact number or concentration of cells, that is, to count cells. In particular, microbial enumeration is an essential part of cell-based research and has been widely used in food and drug safety testing, biomedical testing, and environmental monitoring.

Conventional methods for counting cells include total cell number measurement methods such as microscopic count and electronic particle counter, plate culture method, and membrane filtration method. , viable cell number measurement methods such as most probable number (MPN) are known. However, these methods have the disadvantage of not being able to analyze a large amount of samples at once, making it difficult to obtain quick results, and also making it difficult to obtain accurate and uniform data.

Recently, automated equipment that can quickly count multiple samples or programs that can count cells in thousands of images have been proposed, but the equipment is expensive and requires a process of applying physicochemical modifications such as fixing or staining the cells.

Meanwhile, microfluidic systems based on microdroplets with a diameter of several to hundreds of μm are useful for analysis at the single to multiple cell level. In this case as well, there is a need to detect and count cells within microdroplets, and fluorescence image analysis and laser-induced fluorescence (LIF) analysis are used as conventional methods for detecting cells within droplets. Fluorescence imaging allows accurate cell counting, but has the problem of handling detection reagents (e.g., fluorescently tagged antibodies), and LIF can rapidly count cells with continuous laser scans, but within droplets. The disadvantage is that it is difficult to detect when a single cell is present. Additionally, there is a problem that it is practically impossible to observe cells within microdroplets over time according to these conventional methods.

US Patent 10267726 proposes a method of obtaining a 2D image of the particle diffraction pattern of a droplet, reconstructing a 3D image from it, and then counting the particles in the droplet. However, at this time, the droplets are not microdroplets with a diameter of several mm or ㎛, and expensive equipment and complicated processes are required to count the particles in the droplets.

Korean Patent Publication No. 10-2022-0079727 relates to a cell concentration measurement device that counts cells using a deep learning method from droplet images, but there is no explanation of the structure of artificial intelligence or learning data. Therefore, it is unknown how to specifically measure cells in microdroplets.

US Patent 10267726 Korean published patent 10-2022-0079727

The purpose of the present invention is to provide a method for automatically counting changes in cells in droplets in phase contrast images over time by complementing the automated analysis, which is a drawback of conventional cell counting methods.

The present invention for achieving the above-mentioned object is characterized by a method for counting cells in microdroplets (first invention) and a method for tracking changes in the number of cells in microdroplets using the same (second invention).

The first invention includes the steps of (A) converting a liquid sample into microdroplets; (B) acquiring a phase contrast image from the microdroplets; (C) converting the phase difference image into a grayscale image; (D) obtaining a gray numerical profile from the grayscale image; and (E) calculating the number of peaks from the gray numerical profile to determine the number of cells in the microdroplet. Counting cells in the microdroplet comprising: It's about method . At this time, in step (E), the peak may be a downward peak. At this time, the acquisition of the gray numerical profile, calculation of the number of peaks, etc. may be automatically performed using a conventionally known image analysis program, for example, "Image J" as in the example below, but is not limited thereto.

In order to utilize the present invention in conventionally widely used software, the present invention adds a step (C') of obtaining an inverted image by inverting the grayscale image to black and white between steps (C) and (D), , in step (E), it may also be possible to make the peak an upward peak. At this time, an analysis program such as “Image J” can be used to invert the image into black and white.

In the present invention, the number of microdroplets may be one or multiple. In the case of multiple microdroplets, a plurality of microdroplets may be formed as a single layer adjacent to each other within a grid formed in the microdroplet device.

In the present invention, the diameter of the microdroplets is less than 1 mm, preferably several to hundreds of ㎛, more preferably several to several tens of ㎛.

In the present invention, because the microdroplets are 'fine', they can evaporate quickly when exposed to the atmosphere. Therefore, it is desirable to cover the upper surface of the microdroplet with an evaporation prevention cover.

The second invention relates to a method for tracking changes in the number of cells in microdroplets over time, by applying the cell counting method in microdroplets according to the first invention to one or a plurality of microdroplets at different times.

The second invention relates to a method of counting cells at intervals in the same manner as the first invention and tracking whether the cells increase during that time. Therefore, the receiving structure of the microdroplets, the size of the microdroplets, the cell counting method, etc. It may be the same as in the first invention.

According to the cell number counting method of the present invention, it is possible to count the number of cells in droplets based on images, which was impossible in the prior art, and changes in cell number over time can be easily and accurately confirmed. In addition, if a single cell exists within the droplet, not only can the change in number be observed at the single cell level, but counting errors due to the growth pattern of the strain clumping together can also be resolved, providing fast and accurate results without separate staining or labeling. You can get it.

1 is a conceptual diagram and partial photograph of the microfluidic device used in the present invention.
Figures 2a and 2b are a top view of a liquid droplet fixed in an embodiment of the present invention and a vertical cross-sectional conceptual diagram of a liquid droplet and a cell fixed between a grid and a cover slip, respectively.
Figure 3 shows an optical image of a droplet (left) and a binary black-and-white image converted to a predetermined threshold (right).
Figure 4 shows optical images, phase contrast images, and fluorescence images for the same microdroplet.
Figure 5A is an example of a grayscale image and a gray numerical profile obtained therefrom.
Figure 5b is an example of a black-and-white inverted grayscale inverted image and an inverted gray numerical profile obtained therefrom.
FIG. 6A is an exemplary flowchart for temporal observation of multiple droplets, and FIG. 6B is an example of an algorithm for this.
Figure 6c is an example of an image conversion process and a screen of a related program in the process of the method according to the present invention.
7A and 7B are exemplary images and analysis diagrams, respectively, showing the accuracy of the cell counting method according to the present invention.
Figure 8 is a diagram showing that the cell counting method in microdroplets according to the present invention is stable.

Hereinafter, the present invention will be described in more detail with reference to the attached drawings and examples. However, these drawings and examples are merely examples for easily explaining the content and scope of the technical idea of the present invention, and the technical scope of the present invention is not limited or changed thereby. Based on these examples, it will be obvious to those skilled in the art that various modifications and changes are possible within the scope of the technical idea of the present invention.

In describing the invention, if it is judged that a detailed description of the known technology related to the invention may unnecessarily obscure the gist of the invention, the detailed description will be omitted.

The present invention relates to a method for automatically counting the number of cells in a droplet and a method for tracking changes in the number of cells in a droplet using the same, and is characterized by counting the number of cells based on a phase contrast image. To obtain an image of a droplet, the number of cells can be counted by measuring a phase contrast image while the droplet is fixed to a grid, converting it to the reverse phase, and then analyzing the droplet, but is not limited to this. As explained in detail below, when counting cells according to the present invention was compared with the cell counting results obtained from fluorescence images known to be the most accurate in counting, it was found that stable counting was possible with less error compared to the prior art, making it a cell counting method. You can check its effectiveness.

[Example]

1. Fabrication of microdroplets

To produce microdroplets (hereinafter simply referred to as 'droplets') containing cells, a microfluidic device as shown in Figure 1 was used. A microfluidic device is a device that uses technology to produce a very small amount of fluid using microchannels with a size of tens to hundreds of micrometers. In Figure 1, A is a conceptual diagram of producing microdroplets using a microfluidic device, and B is a partial photograph of the glass-PDMS microchannel used in the present invention. In the drawing, C is a photo of a glass grid produced by applying photoresist to fix droplets. In the example photo, 7×9 rectangular grids are formed. D is an enlarged photo of a grid formed in C. As shown in the example, each grid is preferably marked with a predetermined distinguishing mark (here, "B4").

When injecting immiscible fluids such as oil and aqueous using a device with the same structure as illustrated in Figure 1, shear stress on the continuous phase occurs at the point where the two fluids intersect. The disperse phase breaks down. As a result, droplets with the composition of an aqueous solution injected in a continuous phase are generated through a droplet-based microfluidic device. These droplets have monodisperse properties and can be manufactured from nano-liter to pico-liter through volume control. Additionally, each droplet can become a micro-sized chemical reactor. In the present invention, cells expressing red fluorescent protein were added to the continuous phase to create microdroplets containing cells, thereby creating an environment in which cells could grow.

After dispensing droplets into the grid, mineral oil was dispensed and covered with a cover glass to prevent evaporation. Accordingly, the droplets could be observed for a long time.

The materials and specifications used are shown in the table below.

Planar photographs of the plurality of droplets fixed on the grid and after 100 minutes, and vertical cross-sectional diagrams of the droplets and cells fixed inside the grid and cover glass are shown in Figures 2a and 2b, respectively.

2. Acquisition and analysis of images

(1) Optical image acquisition and analysis

First, general optical images were obtained using a conventional optical microscope to observe droplets and cells. In Figure 3, an optical image (bright field image) taken with an optical microscope and a binary black and white image (binary image) converted to a predetermined threshold value are attached. As can be seen in the figure, it was difficult to distinguish between black and white in a simple optical image, making it impossible to distinguish between cells within the droplet. The converted binary image also had problems. For reference, in this embodiment, general optical, phase contrast, and fluorescence images were obtained by shooting with a black-and-white camera (CCD/sCMOS). Even if shot with a color camera, if they are converted to black-and-white or grayscale, the images are the same as those shot with a black-and-white camera. Therefore, it does not matter whether the camera actually used is black-and-white or color.

In other words, the droplet previously covered and fixed with the cover glass is pressed upwards and downwards between the cover glass and the bottom surface of the grid, as shown in FIG. 2b. However, as can be seen in the binary image photo (right photo in Figure 3), the part of the droplet that is in contact with the grid (the pressed gray part; the inner part marked with a black circle in the left photo of Figure 3) is black and white, and it is difficult to distinguish between black and white. Disadvantages were identified.

(2) Phase contrast image acquisition and analysis

To increase the signal of cells compared to the background, phase contrast images were obtained using a phase contrast microscope to observe droplets and cells. Specifically, the condenser was removed from the optical microscope used in (1) above, and instead, an objective lens containing a phase plate and a condenser were attached to change the microscope to a phase contrast microscope.

The obtained phase contrast image confirmed that the cell border appeared darker and clearer than the general optical image obtained in (1) above for the same droplet. Figure 4 shows the optical image, phase contrast image, and fluorescence image for the same microdroplet. In the figure, in the fluorescence image, cells expressing fluorescence are displayed in white.

As can be seen in Figure 4, compared to the fluorescence image, which accurately represents the cells, the presence of cells in the droplet was displayed much more clearly in the phase contrast image than in the optical image. Accordingly, cell counting was performed automatically on phase contrast images using “Image J,” a conventionally known image analysis program. However, while the location and number of cells were clearly recognized visually, in "Image J", the gray area (noise generated by upward and downward compression of microdroplets) in the phase contrast image acted as noise, so cell counting was not performed.

“Image J” used in the present invention is a Java-based image processing program developed by the National Institutes of Health and the Optical and Computational Instrumentation Institute (LOCI, University of Wisconsin). The program can calculate user-defined signal thresholds and pixel value statistics from image files in a variety of formats, and a built-in editor and Java compiler are provided to develop custom acquisition, analysis, and processing plug-ins.

(3) Conversion and analysis of phase contrast images

Grayscale conveys only luminance information because the value of each pixel represents the amount of light. This luminosity increases the more light there is, and appears as white in grayscale. On the other hand, when the amount of light is small, the luminous intensity is also low and appears black. If we connect this with the previous phase contrast image, the difference in refractive index and thickness is converted into the square of light and dark in the phase contrast image.

In order to improve the problem of "Image J" not being able to analyze the gray area in the phase contrast image, the phase contrast image was converted to grayscale in an image program, and then the gray numerical profile was obtained from the converted grayscale image. Related images and graphs are attached in Figure 5a. In the drawing, A is a grayscale image converted from a phase contrast image of a microdroplet, B is an example of a gray numerical profile obtained from this grayscale image, and C is the number and position of cells in the same microdroplet as in A above. This is a fluorescence image where is clearly displayed.

In the drawing, the profile of B is the gray numerical profile of the horizontal pixel indicated by a red line in image A. Referring to C, which is a fluorescence image, the cell indicated by a black arrow in A is a cell, and the corresponding position pixel in B is black and is located below. It can be seen that it represents a peak. That is, as shown in A of Figure 5A, even if the cells are converted to grayscale, they still appear black because the amount of light transmitted is small, and the microdroplets appear white because they are transparent and also transmit light well. Through this, when the grayscale value is expressed as a profile, the part with cells shows a peak value in the downward direction where the grayscale is close to 0, while the grayscale value in the droplet part shows that the grayscale value increases from the gray part. Through this, it was confirmed that cells within microdroplets could be accurately counted.

Meanwhile, "Image J" has a function to find the upward peak in the profile ('Find Maxima'), but does not have a function to find the downward peak, so it cannot automate the coefficient of the downward peak from the gray numerical profile above. Therefore, in the present invention, it is also possible to select a method of automatically counting cells in microdroplets by inverting the grayscale image to black and white to obtain an inverted image, then obtaining a gray numerical profile from this, and counting the upper peaks in the profile. . Figure 5b shows an example of a grayscale inverted black-and-white image and an inverted gray numerical profile obtained therefrom. As can be seen in the figure, the cell is converted from black to white as the profile is turned upside down by inversion. Therefore, if you use the 'Find Maxima' function of the image software, you can count the number of peak values and automatically count the white (cell part indicated by a white arrow) value.

3. Simultaneous analysis of multiple microdroplets

As described in 1 above, the plurality of microdroplets fixed in the grid according to the method of the present invention maintained their shape without merging or deforming even over time (see Figure 2a). Using these characteristics, each of the plurality of microdroplets produced was separately analyzed. The overall analysis was carried out in accordance with the flowchart in FIG. 6A, and an example of the algorithm for analysis is shown in FIG. 6B. The basic image analysis program used was the above-described “Image J.”

Through this algorithm, optical images undergo a conversion process as illustrated in Figure 6c. In Figure 6c, A is an optical image of a droplet fixed to a grid, B is an image with a ROI (region of interest) set based on a specific threshold of the bright field image, and C and D are microdroplets each assigned a identification number. , An example of a program screen showing threshold (prominence) settings, E and F are images showing the positions of pixels that exceed a certain threshold in each microdroplet, and an enlarged image of one of these exemplary microdroplets. F is a microdroplet numbered '50' and the marked dot is a cell.

First, in the optical image, the RGB image captured by the microscope is converted to grayscale (8 or 16 bit) with a black background. Afterwards, use the program's functions (e.g. "Despeckle", "Watshed", "Analyze Particles" functions, etc.) to remove noise, separate only the boundary surface of the microdroplets, connect the droplet parts that were broken due to noise, and collect the fine particles to be used for analysis. The characteristics of the tax amount were defined. At this time, considering the spherical nature of the droplet, the minimum value of sphericity was set to 0.70. After going through the above command process, the program designates the analysis area that meets the conditions as the region of interest (ROI) to be analyzed. The program automatically numbers each region of the ROI, so that each droplet has a unique number. Uniquely numbered droplets maintain the same area number even when images taken at various times are stacked, so changes in the analysis area over time can be observed equally. Afterwards, the results of the previously designated area remain in the result window and may cause confusion in the next fluorescence image analysis. Therefore, use the program's function (e.g. "Clear results") to delete the results for ROI.

Next, a stack, which is a set of images that stacks several images in one window, was created using the program function (e.g., "Images to Stack") for several fluorescent images taken at each incubation time. The previously specified ROI for the stack was raised using the ROI manager, and a loop was started to repeat the operation for several images. Each image that makes up the previous image stack is called a slice. Therefore, the operation below was performed for the nth slice and repeated when moving to the next slice.

Afterwards, the image is converted to 8 bits and the following coefficient operation is performed on the inside of the previously specified ith ROI area. The user specifies a threshold for the significant fluorescence signal of cells to be distinguished from the background within the ROI area. Accordingly, output=[Count] was inserted in the same command line to identify and count the number of fluorescence signals exceeding the specified threshold using the program's function (e.g., "Find maxima"). Additionally, the output=[Point Selection] command was inserted in the next line so that the user could visually check whether the fluorescence signal counted by the program was correct, so that the program could check the counted signal with the marked points after counting. In addition, "Add Selection" was added so that the user can check even if the counted point is transferred to the droplet with the next number. Finally, “Next Slice [>]” was added at the end of the loop so that the same operation could be performed on other slides in the stack. Additionally, as a way to increase user convenience, you can assign a keyboard shortcut (“Plug-in > Shortcut > Create shortcut”) when running the Image J plugin. Therefore, cell counting can further save time by being a procedure that starts when the user presses a shortcut button designated on the keyboard.

4. Accuracy analysis of the present invention

In order to confirm the reliability of the counting method according to the present invention, it was cross-validated by comparison with fluorescence images.

According to the method mentioned in 1 above, 136 microdroplets were fixed on a grid and cultured in a given environment, and phase contrast images and fluorescence images were taken respectively according to the culture time. Since the strain used in this example expresses a red fluorescent protein, the fluorescence signal increases as the cells grow. In other words, since the fluorescence signal in the fluorescence image comes only from cells, the number of cells in the droplet was accurately measured from the fluorescence signal, and the accuracy/reliability of the number of cells measured in the same droplet was analyzed by the method according to the present invention.

An example of an analysis image of the number of cells in a droplet according to the present invention and an image of an analysis of the number of cells in a fluorescence image of the same droplet is shown in Figure 7a, and an analysis graph showing the accuracy of cell counting according to the present invention over culture time is shown in Figure 7b. It is shown in .

A total of 136 droplets were analyzed to confirm that the point of the pixel value counted for the same number of droplets in the image of the same grid was accurately counted with the naked eye (Figure 7a), and the method presented in 3 above (automated program) The cell count value over time was determined using (Figure 7b).

When the counted points and their number were checked with the naked eye, it was confirmed that the phase contrast and fluorescence images for each droplet matched the number and counted points (Figure 7a). Additionally, as shown in Figure 7b, the results of counting on the same grid at 1-hour intervals showed an error of less than 2 in the average number of cells in the droplet between the two methods at all time points. The standard deviation values of each method were also similar, confirming that the previous method can be used as a counting method. Therefore, the effectiveness and reliability of the counting method according to the present invention were confirmed.

5. Stability analysis of the present invention

The stability of the counting method according to the present invention was analyzed.

To confirm the stability of the counting method, the same number of droplets was counted repeatedly 25 times based on the number of cells counted in the fluorescence image. The counting process utilized the program completed (improved) in the process of the present invention, but the principle is based on the cell counting method according to the present invention.

The correlation between the fluorescence method and the number of fluorescence is shown for a fixed number of cells (2 cells per microdroplet) (A in Figure 8) and by varying the number of cells from 1 to 10 (B in Figure 8). Confirmed.

The upper control line (UCL) and lower control line (LCL) were designated based on the standard deviation obtained through 25 experiments using the same image and code. In general, the upper and lower control limits, which are control limits in a process, are used as indicators to maintain the process in a controlled state or to investigate and determine whether the manufacturing process is in a well-managed state. Therefore, when a process is within the two limit ranges, it means that the process is well managed. When this is applied to the proposed method, both graphs in Figure 8 are within the control limits, meaning that the proposed method is a fairly reliable method.

The number of cells counted through the conventional fluorescence image and phase contrast image was indicated with a box plot and management limit line for the number of cells in the droplet from 1 to 10 (B in Figure 8). Through this, the range of cell values counted using phase contrast images in all cell count ranges appears to be within the management limit, showing that the method presented below is a reliable method.

As discussed above, it was confirmed that the method for counting cells in microdroplets according to the present invention is reliable and safe compared to the fluorescence image analysis method known as an accurate cell counting method. Therefore, according to the present invention, it will be possible to quickly and accurately count actual pathogens and various types of cells as they are, even without biochemical changes such as fluorescent labels.

In addition, it was confirmed that it is possible to track changes in the number of cells in microdroplets over time by applying the cell counting method in microdroplets according to the present invention.

In addition, it will be natural to separate a small portion of the bulk liquid sample and use the cell counting method in microdroplets according to the present invention to quickly and accurately measure the cell concentration/number in the liquid sample.

Claims (10)

(A) converting a liquid sample into microdroplets;
(B) acquiring a phase contrast image from the microdroplets;
(C) converting the phase difference image into a grayscale image;
(D) obtaining a gray numerical profile from the grayscale image: and
(E) determining the number of cells in the microdroplet by calculating the number of peaks from the gray numerical profile;
A method for counting cells in microdroplets, comprising:
In claim 1,
In step (E), the peak is a downward peak. A method for counting cells in microdroplets.
In claim 1,
Between steps (C) and (D),
A step (C') of obtaining a reversed image by inverting the grayscale image to black and white is added,
In step (E), the peak is an upward peak. A method for counting cells in microdroplets.
The method of any one of claims 1 to 3,
A method for counting cells in microdroplets, wherein the microdroplets are formed in a plurality of single layers adjacent to each other within a grid formed in the microdroplet device.
The method of any one of claims 1 to 3,
A method for counting cells in microdroplets, characterized in that the diameter of the microdroplets is less than 1 mm.
The method of any one of claims 1 to 3,
A method for counting cells in microdroplets, characterized in that the upper surface of the microdroplets is covered with a predetermined evaporation prevention cover.
A method for tracking changes in the number of cells in microdroplets over time, characterized in that the cell counting method in microdroplets according to any one of claims 1 to 3 is applied to one or a plurality of microdroplets with time difference.
In claim 7,
A method for tracking changes in the number of cells in a microdroplet over time, characterized in that the microdroplets are formed in a plurality of single layers adjacent to a grid formed in the microdroplet device.
In claim 7,
A method for tracking changes in the number of cells in microdroplets over time, characterized in that the diameter of the microdroplets is less than 1 mm.
In claim 7,
A method for tracking changes in the number of cells in a microdroplet over time, characterized in that the upper surface of the microdroplet is covered with a predetermined evaporation prevention cover.