JP2012255810A - Biochip - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a biochip suitable for optically retrieving for a target cell, allowing a quick and effective collection of the target cell in combination with a target cell retrieval device.SOLUTION: A biochip has a channel in which blood specimen of either collected blood for examination or blood obtained by condensing the collected blood is injected. The channel includes: a target cell pass filter with an aperture allowing a target cell optically labeled for examination in the blood specimen to pass through while not allowing any cell larger than the target cell to pass through; a window part for optical observation disposed on the downstream side of the target cell pass filter which allows the target cell to be observed from the outside of the biochip; a branch channel for collection of the target cell branching from the channel on the downstream side of the window part for optical observation channel; and a valve disposed on the branch part of the branch channel for collection for guiding the target cell to the branch channel for collection.

Description

  The present invention relates to a biochip that is combined with a target cell search device used in a method for examining a specific target cell such as a fetus-derived cell using blood collected for prenatal diagnosis.

  The amniotic fluid test, known as prenatal diagnosis of the fetus, not only has a large invasion to the mother's body, but can also damage the uterus and fetus, leading to miscarriage and stillbirth.

  It is known that about 1 fetal cell, for example, nucleated red blood cells migrates in 1 ml of maternal blood. Therefore, in order to examine the clinical usefulness as a non-invasive prenatal diagnosis, a large-scale clinical study on fetal prenatal diagnosis using maternal blood in the United States and the like from 1995 to about five years (NIFTY study). This clinical study uses flow cytometry to measure fluorescence intensity, target cell size, etc. at high speed with a laser light source while the target cells are specifically fluorescently labeled using an antigen-antibody method and then placed in a liquid flow. Fluorescent activated cell sorting (FACS method) and magnetic microbeads are specifically bound to their target cells via antibodies that react with cell surface antigens and antigens, 2744 cases were also carried out by a technique such as nuclear magnetic cell sorting (MACS method) to collect cells.

  After performing fetal cell separation using these FACS and MACS methods, the obtained cells are diagnosed as XY cells by fluorescence in situ hybridization (FISH method: Fluorescence in situ hybridization) labeled with the DNA. It has been reported that only 41% of the cells were produced and that the false positive rate reached 11%.

  As described above, it is extremely difficult to reliably recover fetal-derived cells from maternal blood without loss by such methods as the FACS method and the MACS method.

In order to solve these problems, Patent Document 1 describes a slide glass with target cells placed on a sample stage using an optical microscope and a CCD camera provided with an XY stage. A method of effectively searching for a target cell by performing image processing based on color, shape, positional relationship, area ratio, and the like divided into a nucleus region, a cell membrane region, and a cytoplasm region is disclosed. However, specimens are spread on a glass slide, such as fetal nucleated erythrocytes, which are about 1 in 1 ml of maternal blood, that is, only about 1 in about 10 ⁸ cells in the blood. In order to search for and find target cells from maternal whole blood, it is necessary to perform a whole blood examination and search, and it takes a considerable amount of time to find them. Even if a search can be made, it is not easy to separate and collect a very small amount of target cells such as fetal nucleated red blood cells.

  In Patent Document 2, a microscope equipped with an objective lens and an imaging device, a sample stage having an XY movement mechanism, on which a slide glass on which a target object-containing sample is placed can be moved to a position where it can be observed with a microscope, and a sample are searched for. An image processing unit having a nozzle that collects the target object, an analysis unit that analyzes image data of the target object and the nozzle acquired by the imaging apparatus, and a storage unit that stores the data and the position thereof, and focusing and A target object automatic search and recovery apparatus including a sample plate in a XY plane and a control unit for controlling the relative position of a nozzle is disclosed. In this apparatus, at the time of focusing, the process of acquiring images is repeated while moving the Z position of the microscope, the contrast intensity of the obtained images is calculated, and the image with the highest contrast intensity is focused. It is said. However, the calculation of contrast intensity in a plurality of images takes time because of a large amount of calculation, and as a result, it takes time to search for a target object.

JP 2004-248619 A JP-A-2005-207986

  The present invention has been made to solve the above-described problems, and is suitable for optical search of target cells. A biochip capable of recovering target cells quickly and efficiently when used in combination with a target cell search device. The primary purpose is to provide it. Furthermore, in order to perform non-invasive and safe prenatal diagnosis of a fetus quickly and accurately, fetal-derived target cells that are present in a very small amount in the collected maternal blood are appropriately concentrated or present in the blood of infants or adults. Abnormal target cells are concentrated appropriately, searched on the flat plate or in the flow path of the biochip with high accuracy in a short time, and after identifying the position, the target cells are reliably recovered, It is a second object of the present invention to provide a simple and efficient method for examining target cells in blood that accurately analyzes chromosomes, DNA, and RNA in cell nuclei. Furthermore, it is a third object to provide a simple and high-performance target cell search apparatus that is suitable for use in the method and can quickly search a large number of blood samples.

The biochip according to claim 1, which has been made to achieve the above object, is a flow into which any blood specimen of blood collected for testing and blood concentrated therein is injected. A biochip having a path,
The flow path includes a target cell passage filter having a gap through which optically labeled target cells to be examined in the blood sample can pass and cells larger than the target cells cannot pass, and the target cell passage An optical observation window disposed downstream of the filter and optically observing the target cell from the outside of the biochip, and recovery of the target cell branched from the flow path downstream of the optical observation window And a valve for guiding the target cell to the branch for recovery, which is arranged in a branch portion of the branch for recovery.

  The biochip according to claim 2 is the biochip according to claim 1, wherein a plurality of the target cell passage filters are arranged in the flow path.

  The biochip according to claim 3 is the biochip according to any one of claims 1 to 2, wherein the valve is a thermally controllable thermal control valve, an electrically controllable electrical control valve, or It is a magnetically controllable magnetic control valve.

  The biochip according to claim 4 is the biochip according to any one of claims 1 to 3, wherein the flow path in the optical observation window is 1 along the flow path without overlapping the target cells. It is formed with the flow path width which can pass in a row | line | column.

  The biochip according to claim 5 is the biochip according to any one of claims 1 to 4, wherein the target cell is optically labeled in the flow path upstream of the optical observation window. The chemical solution injection path for injecting the liquid from the outside is merged.

The test method of target cells in blood is
Optically label cells in blood containing reference cells collected as samples for dictionary creation, irradiate light corresponding to the optical label, color image, and extract the reference cells from the captured color image data A plurality of selected cell regions are cut out with a predetermined image size as target dictionary image data, and the extracted target dictionary image data is binarized based on a predetermined color corresponding to the reference cell. A target dictionary creating step in which the orthonormal basis obtained from the eigenvector is obtained by performing principal component analysis on the plurality of target binary image data obtained and processed in advance,
A plurality of cells other than the reference cells are selected from the color image data, and each selected cell region is cut out as the non-target dictionary image data at the predetermined image size, and each cut out non-target dictionary is used. The image data is binarized based on the predetermined color, and a plurality of non-target binarized image data obtained are subjected to principal component analysis, and orthonormal bases obtained from the eigenvectors are pre-determined into a non-target dictionary. Non-target dictionary creation process,
Collecting a blood sample for testing or concentrating the blood to obtain a blood sample containing target cells to be tested;
A sample labeling step for optically labeling the target cells in the same manner as the optical labeling, after the blood sample is formed into a film on a flat plate and fixed, or before or after being injected into the flow path of a biochip,
An optical label expression step of irradiating the specimen of the flat plate or the biochip with optical label expression light corresponding to the optical label;
The blood sample is irradiated with focusing light at an irradiation angle that can be totally reflected by the flat plate or the biochip, and the blood sample is based on the reflection position of the focusing light on the flat plate or the biochip. A focus control process for focusing the imaging camera on
A color image acquisition step of obtaining color image data by enlarging and imaging the blood sample with the imaging camera;
A binarization processing step of binarizing the color image data captured in the color imaging acquisition step based on the predetermined color;
A contraction / expansion process for contracting / expanding the binarized image data;
A primary candidate search step of assigning a label to each connected pixel region of the image data subjected to the contraction and expansion processing
A secondary candidate search step of re-assigning the same label to connected pixel regions of different labels that are close to each other within the predetermined image size among the labels,
The inspection image data of the predetermined image size including the connected image region indicated by each label after the secondary candidate search step is sequentially cut out, and the inspection image data and each of the target dictionary and the non-target dictionary A third candidate search step of calculating a vector distance, sequentially evaluating both vector distances, and selecting inspection image data that is similar to the target dictionary within a predetermined evaluation value rather than the non-target dictionary;
A target cell specifying step of determining the connected pixel region of the selected image data for examination as the target cell and specifying the position thereof;
A target cell recovery step of recovering the target cell based on the position information;
It has the target cell test | inspection process which analyzes the chromosome, DNA, and / or RNA of the said target cell.

  The method for examining target cells in blood is characterized in that the target cells are fetal nucleated red blood cells.

  The method for examining target cells in blood is characterized in that, during the specimen obtaining step, the concentration is performed by a specific gravity centrifugation method, and then separated and obtained by specific gravity.

  In the method for examining target cells in blood, the blood sample is formed into a film on a flat plate and fixed during the sample labeling step, and the target cells are sucked out while being selected during the target cell recovery step. , Scraping, or scooping out and collecting.

  In the method for examining target cells in blood, the blood sample is injected into the flow path of the biochip during the sample labeling process, and the position is specified in the flow path during the target cell recovery process. It is characterized by.

  The method for examining target cells in blood is characterized in that, during the target cell recovery step, the selected target cells are guided from the flow channel to the flow channel path for recovery.

  The target cell test method in blood comprises a single-stranded complementary sequence that binds the chromosome, DNA and / or RNA to the centromere of the chromosome to be detected or the gene sequence to be detected during the target cell test step. The analysis is carried out by detecting a hybridization reaction with a fluorescent probe for diagnosis combined with a fluorescent dye.

  The method for examining target cells in blood is characterized in that the diagnostic probe is labeled with fluorescence or magnetism.

  In the method for examining target cells in blood, a sample comprising a suspension of the target cells and / or tissues derived therefrom is used as a flow path, compartment, groove, capillary, fiber and / or bead during the target cell examination step. The sample is injected into a biochip reaction field, and the sample is reacted with a labeling reagent that reacts with chromosomes, DNA and / or RNA of the cells in the reaction field, and then the label of the labeling reagent is detected. And

Target cell search device
Optically label cells in blood containing reference cells collected as samples for dictionary creation, irradiate light corresponding to the optical label, color image, and extract the reference cells from the captured color image data A plurality of selected cell regions are cut out with a predetermined image size as target dictionary image data, and the extracted target dictionary image data is binarized based on a predetermined color corresponding to the reference cell. A target dictionary storage unit in which orthonormal bases obtained from the eigenvectors obtained by performing principal component analysis of the obtained binary image data for the target for processing are stored in advance as a target dictionary;
A plurality of cells other than the reference cells are selected from the color image data, and each selected cell region is cut out as the non-target dictionary image data at the predetermined image size, and each cut out non-target dictionary is used. The image data is binarized based on the predetermined color, and a plurality of obtained non-target binarized image data is subjected to principal component analysis, and an orthonormal basis obtained from the eigenvector is used as a non-target dictionary. A pre-stored non-target dictionary storage unit;
A blood chip in which blood is collected for testing or concentrated, and a blood sample on which target cells to be tested are optically labeled is placed on a flat plate or a biochip having a flow channel into which the blood sample is injected XYZ stage that can be held and can be moved and controlled in the XYZ axial directions,
An imaging camera capable of enlarging the blood sample of the planar plate held by the XYZ stage or the biochip and capturing a color image thereof;
An imaging light source that irradiates the planar plate or the biochip with optical label expression light corresponding to the optical label;
A light source for focus control that irradiates light for focus control at an irradiation angle that can be totally reflected by the planar plate or the biochip;
A reflected light detector for detecting a reflection position of the focus control light on the planar plate or the biochip;
Focus control means for moving the XYZ stage in the Z-axis direction based on the reflection position detected by the reflected light detector to focus the imaging camera on the blood sample;
An imaging position control means for controlling the imaging position of the imaging camera by moving the XYZ stage in the XY axis direction;
Color image acquisition means for acquiring color image data of the blood sample by causing the imaging camera to capture an image;
Binarization processing means for binarizing the color image data based on the predetermined color;
Contraction / expansion processing means for contracting / expanding the binarized image data;
Primary candidate search means for assigning a label corresponding to each connected pixel region of the image data subjected to the contraction and expansion processing;
Secondary candidate search means for re-assigning the same label to connected pixel regions of different labels that are close together within the predetermined image size among the labels stored in the label storage unit;
The inspection image data of the predetermined image size including the connected image region indicated by each label after the secondary candidate search step is sequentially cut out, and the inspection image data and each of the target dictionary and the non-target dictionary Tertiary candidate search means for calculating vector distance, sequentially evaluating both vector distances, and selecting inspection image data that is similar to the target dictionary within a predetermined evaluation value rather than the non-target dictionary;
Target cell specifying means for determining that the connected pixel region of the selected image data for examination is the target cell and specifying its position;
It is characterized by providing.

  In the target cell search device, the tertiary candidate search means shifts the position where the inspection image data is cut out one pixel at a time, cuts out a plurality of times at the predetermined image size, and each of the inspection image data cut out at the shifted position The evaluation process is performed every time.

  In the target cell search device, the target cell specifying means has a size and / or roundness of the connected image area of the image data for examination selected by the tertiary candidate search means within a predetermined specified value. The connected image area is determined to be the target cell, and the position of the target cell is specified.

  The target cell search apparatus is characterized in that the focus control light source is also used as the imaging light source.

  The target cell search device combines the examination image data identified as the target cell by the target cell identification means and the plurality of target binarized image data used for creating the target dictionary. And further comprising target dictionary learning means for performing principal component analysis and updating and storing the orthonormal basis obtained from the eigenvector as the target dictionary in the target dictionary storage unit.

  The target cell search device includes the test image data not selected by the tertiary candidate search means, or the test image data not determined to be the target cells by the target cell specifying means, and the non-target A principal component analysis is performed on a plurality of the non-target binarized image data used for creating the dictionary, and the orthonormal basis obtained from the eigenvector is updated to the non-target dictionary storage unit as a non-target dictionary. And non-target dictionary learning means for storing the information.

  The target cell search device further includes a display, and the target cell specifying means specifies the similarity of the target cell, the image of the test image data, and the color cut out in the same range as the test image data At least one selected from an image of image data and an image of the entire color image data can be displayed on the display.

  The target cell search device includes a stocker capable of accommodating a plurality of the planar plates after the target cell identification and the planar plates before the identification, and the planar plate between the stocker and the XYZ stage. It further has a conveyance mechanism for conveying.

  According to the method for examining target cells in blood using the present invention, it is non-invasive and safe for the mother and the fetus, and does not induce miscarriage or stillbirth for the purpose of performing the examination. Diagnosis can be made quickly and accurately. In this test method, fetal-derived target cells, which are present in a very small amount in the collected maternal blood, are appropriately concentrated by centrifugation or the like to reduce the volume per sample and improve efficiency. Also, according to this test method, it is possible to quickly and accurately diagnose the presence or absence of abnormal target cells in the blood regardless of age or age from children such as infants after birth to adults.

  In this test method, fetal-derived target cells that are appropriately concentrated from the maternal blood and abnormal target cells that are appropriately concentrated from blood of infants or adults can be accurately obtained in a short time on a flat plate or in the flow path of a biochip. Since it is possible to instantly search and optically determine the position and specify it, and then reliably capture and collect the few target cells, it is highly reliable.

  In addition, specific fluorescence can be obtained by amplifying genes such as chromosomes, DNA and RNA in the cell nucleus by polymerase chain reaction (PCR) or the like, or by performing the FISH method on the nuclear cells without amplification. Since it can detect optically with a probe or magnetically with a magnetic probe, perform chromosome diagnosis / DNA diagnosis, accurately analyze and perform highly reliable genetic diagnosis, fetal diagnosis can be performed easily and efficiently. Contributes to advanced medical diagnosis before birth and precise and advanced medical diagnosis based on blood component tests of infants and adults.

  Furthermore, according to the method for examining target cells in blood and the target cell search device using the present invention, a target dictionary is created by principal component analysis of a cell image serving as a reference for target cells, and other cell A non-target dictionary is created by principal component analysis of the image, and the vector distance between the blood image collected for testing and the dictionary and the testing image is calculated, and the similarity to the dictionary is evaluated. Since a target cell can be searched with a small amount of calculation, a high-speed and highly efficient search is possible. In addition, the color image itself is not searched by image processing, but the color image is binarized with a predetermined color corresponding to the target cell, and the binarized image is processed and searched. The amount of calculation is smaller than that of image processing, and a higher-speed search is possible. Also, by focusing the light for focus control on a flat plate etc. and performing focusing based on the reflection position, the amount of calculation is far less than when calculating the contrast intensity of the image, and focusing can be performed at high speed. Therefore, high-speed search becomes possible. In addition, when the imaging light is also used as the focus control light, a simple configuration can be obtained.

  In addition, the target cell search apparatus using the present invention can process a blood sample quickly and search for a target cell at high speed and with high efficiency, so that the target cell is separated, reliably captured and recovered, and its cell nucleus It is a simple and high-performance target cell search apparatus suitable for a non-invasive diagnostic system that accurately analyzes genes such as chromosomes, DNA, and RNA, and performs highly reliable genetic diagnosis.

  In addition, the biochip of the present invention is suitable for searching for target cells from blood samples such as blood and its concentrated / separated components, and can immediately separate and collect the searched target cells.

It is a flowchart which shows the test | inspection method of the target cell in blood. It is a flowchart explaining the principal component analysis in a target dictionary preparation process in detail. It is a block diagram of a target cell search device. It is the side view (a) which shows the outline | summary of the imaging part in FIG. 3, and the side view (b) which shows the modification. FIG. 5 is a schematic diagram illustrating an overview of a focusing operation of the imaging unit in FIG. 4. It is a flowchart explaining the tertiary candidate search process in FIG. 1 in detail. It is a side view which shows the outline | summary of the imaging part in another target cell search device. A schematic plan view (a) of a biochip to which the present invention is applied, an enlarged plan view (b) of a target cell passage filter part 72, an enlarged plan view (c) of an optical observation window part 73, and an optical observation window part 73 It is an expanded side sectional view (d) of channel 76. It is a side view which shows typically the bonding method of the biochip to which this invention is applied. These are the target cell image (a), red blood cell image (b), and white blood cell image (c) after staining. It is the normal cell image example (a) after FISH, the chromosome 21 trisomy cell fluorescence observation image (b). 1 is a schematic plan view of a biochip for chromosome testing and genetic testing to which the present invention is applied. It is a cell observation image after staining (a) and a fluorescence observation image after FISH (b) on a biochip for chromosome inspection and genetic inspection to which the present invention is applied. FIG. 5 shows DNA microarray fluorescence observation images for genetic testing on a biochip for chromosome testing and genetic testing to which the present invention is applied, showing normal cell DNA microarray results (a) and 21 trisomy cell DNA microarray results (b), respectively. ing. It is a color image example (a) cut out as target dictionary image data, and a color image example (b) cut out as non-target dictionary image data. The color image (a) which imaged the concentrated blood sample (a), the binarized image (b) after the binarization process, and the binarized image (c) after the contraction and expansion process.

  Hereinafter, although embodiment of this invention is described in detail, the scope of the present invention is not limited to these forms.

  In the method for examining target cells in blood using the present invention, the target dictionary creation step S1 and the non-target dictionary creation step S2 of FIG. 1 are performed in advance. Thereafter, the collected blood sampled is subjected to the sample acquisition step S11 and the sample labeling step S12 in the figure to optically label the target cells in the concentrated blood sample, and the optical label expression step S13 and the focus control step S14. , Color image acquisition step S15, binarization processing step S16, contraction and expansion step S17, primary candidate search step S18, secondary candidate search step S19, tertiary candidate search step S20, and target cell specifying step S21 The target cell in the blood is examined by performing the target cell recovery step S23 and the target cell test step S24. When the concentrated blood sample is fixed to the flat plate, the area that can be imaged at one time is narrower than that of the flat plate. Therefore, the imaging area of the flat plate is moved to perform processing from the optical labeling step S13 to the target cell specifying step S21. Is repeated to search for target cells for the entire flat plate (step S22). If the concentrated blood sample on the flat plate can be imaged at one time, it is not necessary to repeat step S22.

  In the following, the case where the target cells in the blood to be examined are fetal nucleated red blood cells (NRBC) will be described in detail. However, the target cells are not limited to this, and not only undifferentiated NRBC but also non-nucleated red blood cells. Without being limited, various blood cells may be used, and specifically, white blood cells, platelets, and the like may be examined. For example, as erythrocytes other than NRBC, abnormal erythrocytes such as erythrocytes of patients infected with malaria, sickle erythrocytes (sickle cell disease), large erythrocytes (vitamin B12 deficiency, etc.), small erythrocytes / hypochromic ( Iron deficiency), spherical red blood cells (such as genetic spherocytosis), elliptical red blood cells (such as genetic elliptical red blood cells), oral red blood cells (such as hemolytic anemia), teardrop red blood cells (such as myelofibrosis), target red blood cells (such as myelofibrosis) Non-nucleated red blood cells such as thalassemia), erythrocytes (such as congenital abeta-lipoproteinemia), and crescent red blood cells (such as disseminated intravascular coagulation) may be examined.

  In the target dictionary creation step S1, cells in maternal blood containing the reference NRBC collected as a dictionary creation sample are optically labeled, and light corresponding to the optical label is irradiated for color imaging to obtain color image data. obtain. Such color image data can be obtained in the same manner as in the specimen acquisition step S11 to the color image acquisition step S15 described later. The operator sorts a plurality of NRBCs from the captured color image data while viewing the images, and cuts out each sorted cell region as target dictionary image data with a predetermined image size. This predetermined image size is the size of an image that allows just one NRBC to be cut out (included), and is appropriately adjusted depending on the imaging magnification. As an example, the predetermined image size is 52 pixels × 52 pixels. Each extracted target dictionary image data is binarized based on a predetermined color corresponding to the reference NRBC. Such binarization processing is performed in the same manner as the binarization processing step S16 described later. A plurality of obtained binarized target image data is subjected to principal component analysis, and orthonormal bases obtained from the eigenvectors are set as target dictionaries in advance.

  FIG. 2 shows a flowchart of principal component analysis. First, in step S31, a plurality of (n) NRBC target binarized image data D (p, q) are vectorized. Here, p represents the X coordinate of the pixel, and q represents the Y coordinate of the pixel. For example, when the target binarized image data D (p, q) is 52 pixels × 52 pixels, p and q are each 1 to Takes a value of 52. If the maximum values of p and q are P and Q, the binarized image data for target D (p, q) is a matrix of P rows and Q columns, so this is a matrix of one column of P × Q dimensions. It can be represented by a target dictionary vector Dv (m). Here, m is an integer value of 1 to P × Q, and is 1 to 2704 in the case of 52 pixels × 52 pixels.

  For example, Dv (1) = D (1,1), Dv (2) = D (1,2), Dv (3) = D (1,3)... Dv (m) = D (P, Q ). Since there are n such target dictionary vectors, each vector is represented by Dv1, Dv2,... Dvn.

Next, in step S32, an average vector of dictionary vectors Dv1, Dv2,... Dvn is calculated by the following equation (1).

Next, in step S33, a covariance matrix Cp with the average vector is calculated by the following equation (2).

  Next, in step S34, the covariance matrix Cp is subjected to eigenvalue decomposition. Subsequently, in step S35, eigenvectors for a plurality of (for example, 40) eigenvalues from the higher order of eigenvalues are obtained. The number of eigenvalues is preferably a number that can represent 90% or more. A matrix in which these eigenvectors are arranged is an orthonormal basis, and this is the target dictionary B.

  In the target cell search device 1 (see FIG. 3) to be described later, the target dictionary creating means 9 creates the target dictionary B as described above and stores it in the target dictionary storage unit 31.

  In the non-target dictionary creation step S2, a plurality of cells other than the reference NRBC are selected from the color image data used in the target dictionary creation step S1 or the color image data obtained in the same manner while viewing the image. The sorted cell regions are cut out as non-target dictionary image data with a predetermined image size. Each of the extracted non-target dictionary image data is binarized based on a predetermined color, and a plurality of obtained non-target binary image data are subjected to principal component analysis and orthonormal obtained from the eigenvectors. Let the base be a non-target dictionary C in advance.

  When creating a non-target dictionary, it is preferable to select cells such as erythroblasts that are not NRBC, as well as leukocytes such as neutrophils, as cells other than NRBC.

  The image clipping method, binarization processing, principal component analysis method, and the like in the non-target dictionary creation step S2 can be performed in the same manner as the target dictionary creation step S1 except that the target cells are different. The detailed explanation is omitted. In the target cell search device 1 (see FIG. 3) described later, the non-target dictionary creating means 10 creates the non-target dictionary C as described above and stores it in the non-target dictionary storage unit 32. Further, the target dictionary B and the non-target dictionary C may be created by the target cell search device 1, for example, the manufacturer of the target cell search device 1 creates in advance, and it is data separately from the device. May be sold at.

  Returning to FIG. 1, the specimen acquisition step S11 will be described.

  In the sample acquisition step S11, the maternal blood collected for testing is concentrated to a desired component by centrifugation or the like, and a concentrated blood sample containing fetal-derived target cells to be tested is acquired. For example, a specific gravity adjusting liquid that is a suspension of immiscible suspensions of two types having specific gravity larger than that of red blood cells or white blood cells, specifically, a Percoll liquid that is a fine suspension of polyvinylpyrrolidone-coated silica gel (GE Healthcare Japan) Maternal blood is put into a centrifuge tube together with the product made by Co., Ltd. and centrifuged. More specifically, maternal blood collected from heparin is injected into a centrifuge tube containing two layers of Percoll solution with different specific gravities, and is centrifuged at a high speed with a centrifuge and centrifuged from the difference in specific gravities. Correspondingly, erythrocyte and leukocyte-containing layers containing NRBC are collected from the maternal blood divided into a plurality of layers with a pipette to obtain a concentrated blood sample. Concentrated blood samples may be collected at high speed in parallel from different maternal blood by simultaneously centrifuging using a plurality of centrifuge tubes. In addition to the specific gravity centrifugation method using such a specific gravity adjusting solution, nylon wool method, lectin method using sugar chain lectin, leukocyte and NRBC capture method using a porous filter, plasma removal method using centrifugation, dehydration method, A concentrated blood sample may be collected by various known methods such as an external filtration method and a blood adsorption method.

  In the sample labeling step S12, the concentrated blood sample is formed into a film on the flat plate 8 (see FIG. 3) and fixed, and then the entire target cell is optically labeled. The flat plate 8 is a sample stage suitable for optical microscope observation such as a slide glass. Examples of optical labels include fluorescent and staining labels such as Giemsa staining, hematoxylin staining, and light staining; antibodies using anti-CD71 antibody, anti-CD36 antibody, anti-glycophorin A antibody, and green fluorescent dye Fluorescent labeling using labeled fluorescent dyes such as certain fluorescein isothiocyanate (FITC) labeled monoclonal antibodies, orange fluorescent dyes phycoerythrin (PE) labeled monoclonal antibodies, Texas red labeled monoclonal antibodies, allophycocyanin labeled monoclonal antibodies Can be mentioned. When fluorescently labeling NRBC, the concentrated blood sample obtained in the sample acquisition step S11 is dropped on the flat plate 8, spread uniformly and dried, and then applied to a fluorescent staining solution such as Giemsa staining solution and dried.

  The planar plate 8 optically labeled by staining the whole cell as described above is repeatedly subjected to the optical labeling expression step S13 to the target cell specifying step S21, and the position of the target cell is specified by image processing. Steps S13 to S21 will be described after the configuration of the target cell search device 1 in the present invention suitable for performing these steps S13 to S21 is described.

  As illustrated in FIG. 3, the target cell search device 1 includes an imaging unit 2, a processing unit 3, a display 4, a stocker 5, and a transport mechanism 6 as an example.

  The imaging unit 2 operates under the control of the processing unit 3 to capture a color image obtained by enlarging the concentrated blood sample, and includes an XYZ stage 11, an imaging camera 15, an imaging light source 14, a laser light source 12, and a focal point. The camera 13 is provided. A side view of the imaging unit 2 is shown in FIG.

  As shown in the figure, the XYZ stage 11 can position and hold the flat plate 8 and can control movement in the XYZ axial directions. The XYZ stage 11 is a flat end, and the flat plate 8 is placed on the top (upper side in the figure), and the flat plate 8 is positioned and fixed by a positioning mechanism (not shown). The XYZ stage 11 has an XYZ moving mechanism (not shown), is connected to the imaging position control means 23 (see FIG. 3) of the processing unit 3, and is in the X-axis direction (left-right direction in the figure) and Y-axis direction. The movement in the direction (front and back in the figure) is controlled, connected to the focus control means 22 (see FIG. 3) of the processing unit 3, and the movement in the Z-axis direction (vertical direction in the figure) is controlled. As such an XYZ stage 11, a known one can be adopted.

  The imaging camera 15 is a color CCD camera as an example, and has a magnifying lens 16 such as an optical microscope in the imaging direction. This imaging camera 15 is arranged at a position facing the flat plate 8 held by the XYZ stage 11, and expands the target cells contained in the concentrated blood sample on the flat plate 8 at a magnification that can be identified, The color image is taken.

Imaging light source 14 is for irradiating the light for imaging the plane plate 8, corresponding to the optical labeling of concentrated blood sample, irradiating the optical indicator expressing light L _T of the wavelength of expressing optical label. For example, if the optical label is a fluorescent label, an optical indicator expressing light L _T is the light that can excite the fluorescent substance wavelength. In this case, a light emitting diode that oscillates at a specific wavelength such as ultraviolet or near ultraviolet corresponding to a fluorescent material is used as the imaging light source 14. Also, if the optical label is a dyed label visible, optical indicator expressing light L _T is visible light. In this case, a white light emitting diode or a halogen lamp is used as the imaging light source 14.

The imaging light source 14 may be configured to be disposed on the same side as the imaging camera 15 with respect to the flat plate 8 as shown in FIG. 4A (configuration of epi-illumination), or as shown in FIG. Alternatively, the flat plate 8 may be arranged on the opposite side of the imaging camera 15 (transmission illumination configuration). In the case of epi-illumination shown in FIG. 4A, the imaging camera 15 images the reflected light of the concentrated blood sample, and in the case of the transmitted illumination shown in FIG. 4B, the imaging camera 15 images the transmitted light of the concentrated blood sample. If the transmitted illumination is the XYZ stage 11, and holds the edge portions of the flat plate 8 so as not to interfere with the irradiation of the optical indicator expressing light L _T. It is preferable that the target cell search apparatus 1 has both epi-illumination and transmitted illumination as the imaging light source 14 and can select either one. Which illumination is used to select an image is selected as appropriate based on the color of the optical marker, the sharpness, and the like.

  The laser light source 12 is a focus control light source that emits focus control light. The laser light source 12 is disposed at a position where the flat plate 8 on the XYZ stage 11 is irradiated with the laser light Le at an angle θ. This angle θ is an angle at which the laser beam Le is totally reflected by the flat plate 8.

  The focus camera 13 corresponds to the reflected light detector in the present invention, and is a monochrome CCD camera, for example. As shown in the figure, the focusing camera 13 tilts the imaging direction at an angle θ with respect to the plane plate 8 so that the reflected light Le ′ of the laser beam Le reflected by the plane plate 8 can be detected. Are disposed at positions facing the laser light source 12.

  Returning to FIG. 3, the processing unit 3 of the target cell search device 1 controls the operation of the target cell search device 1 in an integrated manner, and although not shown, a central processing unit (not shown) that performs various calculations and processing ( CPU), ROM, RAM, hard disk, one or a plurality of computers having an input / output interface, etc., and an operation program thereof. The following means are realized in cooperation with the computer and the operation program.

  The processing unit 3 includes a target dictionary creation unit 9, a non-target dictionary creation unit 10, an image acquisition unit 21, a focus control unit 22, an imaging position control unit 23, a transport control unit 24, a target dictionary storage unit 31, and a non-target dictionary storage unit. 32, search storage unit 41, binarization processing means 42, contraction / expansion processing means 43, primary candidate search means 44, secondary candidate search means 45, tertiary candidate search means 46, target cell specifying means 47, target dictionary A learning unit 51 and a non-target dictionary learning unit 52 are included. Each storage unit 31, 32, 41 is a hard disk as an example.

  The stocker 5 accommodates a plurality (for example, 80) of flat plates 8 before and after the search for target cells, and can control the loading and unloading of each flat plate 8. The moving mechanism 6 conveys the flat plate 8 between the stocker 8 and the imaging unit 2.

  Returning to the inspection method and continuing the description, the operator accommodates the plurality of flat plates 8 labeled in the sample labeling step S12 in the stocker 5. When the target cell search device 1 is operated, the transport control unit 24 controls the stocker 5 and the transport mechanism 6 to set one flat plate 8 from the stocker 5 on the XYZ stage 11. Further, the imaging position control means 23 controls the XYZ stage 11 so that a predetermined imaging position of the flat plate 8 enters the imaging area of the imaging camera 15.

In the optical label developing step S13, the image acquisition unit 21 to irradiate the optical indicator expressing light L _T to the imaging light source 14. As a result, the fluorescent label of the concentrated blood sample is excited and the cell nucleus develops fluorescence.

  In the focus control step S14, the focus control means 22 irradiates the laser light source 12 with the laser light Le (see FIG. 4). The focus control means 22 acquires a captured image from the focus camera 13, moves the XYZ stage 11 in the Z-axis direction based on the reflection position of the reflected light Le ′ of the laser light Le reflected by the flat plate 8, By controlling the distance between the flat plate 8 and the magnifying lens 16, the imaging camera 15 (magnifying lens 16) is focused on the concentrated blood sample.

  FIG. 5 shows an observation image 55 of the focus camera 13. The focus camera 13 observes the reflection position of the reflected light Le ′ on the flat plate 8. The distance between the magnifying lens 16 and the plane plate 8 (see FIG. 4) is the distance at which the focal point of the magnifying lens 16 is exactly aligned with the plane plate 8, and the reflected light Le ′ is the observation image 55 as shown in FIG. The imaging area of the focusing camera 13 is adjusted in advance so that the image is observed in alignment with the prescribed position 56 at the center of the screen. In this way, when the flat plate 8 is close to the magnifying lens 14, the reflection position of the reflected light Le ′ is shifted as shown in FIG. Observed by shifting to the left side. On the contrary, when it is far away, the reflection position of the reflected light Le ′ is observed to be shifted from the specified position 56 to the right as shown in FIG. The focus control unit 22 performs focusing by moving the XYZ stage 11 in the Z-axis direction so that the reflected light Le ′ is observed just overlapping the specified position 56.

  When focusing is performed in this manner, the XYZ stage 11 is only moved so that the reflected light Le ′ is aligned with the specified position 56. For example, the imaging camera 15 performs imaging while gradually moving the XYZ stage 11 in the Z-axis direction. Since the amount of calculation is much smaller than that in the case of calculating the contrast of the image and controlling to find the position where the contrast becomes the highest, focusing can be performed at high speed.

  In the color image acquisition step S15, the color image acquisition unit 21 causes the imaging camera 15 to capture the image and acquire the color image data X (i, j), along with the X and Y information (position information) on the plane plate 8, for search. The data is stored in the storage unit 41. Since the area of the plane plate 8 that can be imaged at once by the imaging camera 15 is usually a narrow area, the captured color image data X (i, j) is a part of the plane plate 8 image data. Here, i represents the position in the X-axis direction in the image data, and j represents the position in the Y-axis direction. For example, if the color image data X (i, j) is 2000 pixels × 1000 pixels in size, i is 1 to 1 An integer value of 2000, j is an integer value of 1-1000. Since the flat plate 8 is positioned on the XYZ stage 11, the position i, j in the obtained color image data X (i, j) is specified, and the flat plate of the color image data X (i, j) obtained is captured. By referring to the position at 8, the position on the plane plate 8 can be specified.

  In the binarization processing step S16, the binarization processing means 42 reads the color image data X (i, j) stored in the search storage unit 41, and based on a predetermined color corresponding to the reference cell. The binarization process is performed and the search storage unit 41 stores the binarization process. In this case, the predetermined color is a color of a cell nucleus that is fluorescently labeled with the target NRBC, and is obtained in advance by averaging the colors of a plurality of NRBCs as a reference obtained from the sample. Here, the red (R) component, the green (G) component, and the blue (B) component of the predetermined color are used for the calculation, and the maximum value of the gradation of each color (for example, in the case of 8-bit gradation) It is normalized so that the maximum value is 1 by dividing by 255). The normalized R component of the predetermined color is Xr, the G component is Xg, and the B component is Xb. Xr, Xg, and Xb are stored in advance in the search storage unit 41.

In the binarization processing, the binarization processing unit 42 normalizes the RGB component of each pixel of the color image data X (i, j) with the maximum value of the gradation of each color, and then performs the following (3 ) Formula is calculated. Here, the normalized R component of the pixel is r, the G component is g, and the B component is b.

  Pixels for which C> 0.99, that is, pixels having a color approximating a predetermined color are set to 1, for example, and other pixels are set to 0, for example, to binarize all pixels. The binarization processing unit 42 stores the binarized binarized image data IR (i, j) in the search storage unit 41 separately from the color image data X (i, j).

  If the binarization process is performed based on the color information of the NRBC cell nucleus in this way, the data amount is smaller than that of the color image, and therefore the search can be performed at high speed in the following steps.

  In the contraction / expansion step S17, the contraction / expansion processing unit 43 reads the binarized image data IR (i, j) from the search storage unit 41, and if there is 0 among the eight pixels surrounding the pixel of interest. After performing the contraction process for setting the target pixel to 0 for all the pixels, if there is 1 in 8 pixels surrounding the target pixel, the expansion process for setting the target pixel to 1 is performed for all the pixels. The binarized pixel data IR (i, j) is replaced and stored in the search storage unit 41. Noise is removed by the contraction and expansion processing. The contraction / expansion process may be performed a plurality of times as necessary.

  In the primary candidate search step S18, the primary candidate search means 44 applies to a connected pixel area in which pixels 1 are adjacent to each other in the binarized image data IR (i, j). Then, a labeling process for assigning a label is performed so that each connected pixel region can be identified, and the assigned label is stored in the search storage unit 41 in association with the connected pixel region. The labeling process can be performed by a known method.

  In the secondary candidate search step S19, the secondary candidate search means 45 reads the label stored in the search storage unit 41, and re-applies the same label to the connected pixel regions of different labels that are close within a predetermined image size. It is reassigned and stored in the search storage unit 41. That is, when the inspection image data is cut out in the next tertiary candidate search step S20, the same label is re-assigned to the connected pixel regions of different labels that fall within a predetermined image size and can be cut out simultaneously. This predetermined image size is the same image size as the target dictionary image data and the non-target dictionary image data when the target dictionary B and the non-target dictionary C are created. For example, when the predetermined image size to be cut out is 52 pixels × 52 pixels, when different connected pixel regions are included in the 52 pixels × 52 pixels, the same label is given again.

  By assigning the same label to the adjacent connected pixel regions in this way, the calculation amount to be processed in the next tertiary candidate search step S20 is reduced, so that the search can be performed at high speed.

  Further, the secondary candidate search means 45 averages the X coordinate and Y coordinate of the connected pixel area of the same label, obtains the barycentric coordinates Lx and Ly of the connected pixel area, and associates them with the label for search. The data is stored in the storage unit 41.

  In the tertiary candidate search step S20, the tertiary candidate search means 46 sequentially cuts out inspection image data having a predetermined image size including the connected image region indicated by each label after the secondary candidate search step, and the inspection image data and Then, the vector distance between each of the target dictionary B and the non-target dictionary C is calculated, and both vector distances are sequentially evaluated to obtain test image data that is similar to the target dictionary within the predetermined evaluation value than the non-target dictionary. Select. This tertiary candidate search step S20 will be described in detail with reference to the flowchart of FIG.

  First, in step S41, the tertiary candidate search means 46 reads the barycentric coordinates Lx and Ly of one label from the search storage unit 41, and determines the cut-out position of the inspection image data having a predetermined image size. When the target cell is distorted, the center of gravity is biased. Therefore, when the inspection image data is cut out around the center of gravity coordinates Lx and Ly, the target cell is biased out and protrudes from the inspection image data. May end up. For this reason, not only the inspection image data is cut out and evaluated with the center-of-gravity coordinates Lx and Ly as the center coordinates, but the inspection image data is cut out by shifting the center coordinates to be cut out from the center-of-gravity coordinates Lx and Ly one pixel at a time. It is preferable. The maximum range for shifting the pixels is preferably about 1/8 to 1/20 in the vertical and horizontal directions of a predetermined image size. Here, as an example, it is assumed that the evaluation is performed by shifting from the center-of-gravity coordinates Lx, Ly to ± 4 pixels. In this case, the evaluation process is performed by cutting out a total of 81 times for one label. If the search is performed by shifting one pixel at a time, the target cell can be searched with high accuracy. In addition, when there is no distortion in the shape of the target cell, or when the target cell can be cut out from the inspection image data so as not to protrude, the process of shifting by one pixel may not be performed.

  In this way, in order to cut out the inspection image data by shifting one pixel at a time, in step S41, the tertiary candidate search means 46 sequentially selects one center coordinate to be cut out within the range of ± 4 pixels from the barycentric coordinates Lx and Ly. specify.

  Next, in step S42, the tertiary candidate search means 46 binarizes the inspection image data K (p, q) having a predetermined image size (52 pixels × 52 pixels) centered on the coordinates specified in step S41. The cut-out image data IR (i, j) is cut out and read from the search storage unit 41. Here, p represents a pixel in the X-axis direction, and q represents a pixel in the Y-axis direction. When the predetermined image size is 52 pixels × 52 pixels, p and q are integer values of 1 to 52.

  The tertiary candidate search means 46 vectorizes the inspection image data K (p, q). That is, it is expressed by a P × Q-dimensional inspection vector A, that is, a one-column matrix A (m). This vectorization is performed in the same manner as the vectorization performed in the target dictionary creation step S1.

  For example, the inspection vector A (m) is such that A (1) is K (1,1), A (2) is K (1,2),... A (m) is K (P, Q). Thus, when the inspection image data K (p, q) is 52 pixels × 52 pixels, it is a matrix of one column in which 2704 pixel values are arranged in order.

In step S43, the tertiary candidate search means 46 calculates the addition average value h of each element (each pixel) of the inspection vector A (m) by the following equation (4).

In step S44, the tertiary candidate search means 46 performs norm calculation of the inspection vector A (m). The norm Vr of the inspection vector A (m) is calculated by the following equation (5).

In step S45, the tertiary candidate search means 46 normalizes the inspection vector A (m) with the calculated addition average value h and norm Vr. Specifically, the inspection normalization vector Av (m) is calculated by the following equation (6).

  In step S46, the tertiary candidate search means 46 calculates the distance Gb between the test normalization vector Av and the target dictionary B, and calculates the distance Gc between the test normalization vector Av and the non-target dictionary C.

  As the distance Gb between the inspection normalization vector Av and the target dictionary B, an inner product A · B between the vectors is calculated. Further, the distance Gc between the test normalization vector Av and the non-target dictionary C calculates the inner product A · C between the vectors.

In step S <b> 47, the tertiary candidate search unit 46 calculates the similarity with the evaluation function S represented by Expression (7), and stores it in the search storage unit 41.

  This evaluation function S takes a negative value when the test normalization vector Av (test vector A) is closer to the target dictionary B than the non-target dictionary C, and the value approaches -1 as the target dictionary B is closer. .

  In step S48, the tertiary candidate search unit 46 determines that the inspection vector A is a non-target when the value S calculated in step S46 is negative, that is, in a range smaller than the predetermined evaluation value (value 0). It is selected as being similar to the target dictionary B rather than the dictionary C, and a flag or the like is stored in the search storage unit 41 so that the selection can be understood.

  In step S49, the tertiary candidate search means 46 returns to step S41 when the evaluation has not been completed for the range of ± 4 pixels of the barycentric coordinates Lx, Ly, and cuts out the inspection image data shifted by one pixel and performs S42. -S48 is performed. The tertiary candidate search means 46 updates the center position where the value of the evaluation function S is closest to −1 as the center of gravity coordinates Lx and Ly, while evaluating the center position by shifting one pixel at a time, and the similarity at that time is updated. The search storage unit 41 is overwritten and stored in correspondence with the label.

  In step S50, the tertiary candidate search means 46 determines whether or not the similarity has been calculated for all labels, and when all the calculations have not been completed yet, repeats steps S41 to S49 for another label. When all the labels are finished, the tertiary candidate search step S20 is finished.

  In the target cell specifying step S21, the target cell specifying means 47 determines that the connected pixel region of the test image data selected in the tertiary candidate search step S20 is a target cell, specifies its position, and outputs it. . When there are a plurality of connected pixel areas determined as target cells in the tertiary candidate search step S20, a predetermined number (for example, three) of connected pixel areas are selected from those having a high similarity (S is close to -1). The target cells may be determined and their positions may be specified and output.

  Furthermore, as necessary, the target cell specifying means 47 has the size and / or roundness of the connected image region of the test image data selected in the tertiary candidate search step S20 within predetermined specified values. You may perform the process which determines that a connection image area | region is a cell nucleus of a target cell.

  Specifically, the target cell specifying unit 47 determines the width in the X-axis direction and the width in the Y-axis direction as the size of the connected pixel region, and both widths are within predetermined predetermined values for size determination ( For example, when the number of pixels of a cell is 52 pixels × 52 pixels, a cell within a range of 20 to 40 pixels) is determined as a target cell. The predetermined predetermined value for size determination is obtained by replacing the size range that can be taken by the cell nucleus of the NRBC in the inspection image data with the number of pixels.

Further, the target cell specifying means 47 calculates the roundness of the connected pixel region by the following equation (8), and this roundness is within a predetermined prescribed value for roundness discrimination (for example, 0.5 to 0.5). 1.5) is determined to be a target cell.
Where R = L ÷ π ÷ 2
L is the perimeter of the connected pixel region.

  The target cell specifying means 47 specifies that the size and / or roundness is within a predetermined specified value as the cell nucleus of the target cell, and searches for a flag or the like so that it can be identified. The data is stored in the storage unit 41. It is preferable to determine the target cell from both the size and the roundness.

  The target cell specifying means 47 determines all the labels selected in the tertiary candidate search step S20, and then specifies the positions Lx, Ly and the position on the flat plate 8 specified as NRBC as a cell recovery device (FIG. (Not shown).

  This completes the search for target cells in a partial region of the flat plate 8. Subsequently, the imaging position control unit 11 moves the XYZ stage 11 in the X-axis direction and the Y-axis direction so that the imaging regions overlap each other or slightly, and the target cell specifying step from the optical label expression step S13 described above. By repeating the steps up to S21, the search for one flat plate 8 is completed.

  In the color image acquisition step S15, the XYZ stage 11 is moved in the XY directions for each imaging region to repeatedly perform imaging, and after acquiring the entire color image data of one flat plate 8, binarization processing is performed. You may perform from step 16 to target cell specific process S21.

  The planar plate 8 for which the search has been completed is accommodated in the stocker 5 by the transport mechanism 6, the other plane plate 8 before the search is set from the stocker 5 to the XYZ stage 11, and the search is similarly started. In this way, the target cell search device 1 automatically searches all the flat plates 8 accommodated in the stocker 5.

  The target cell search device 1 includes a position where a target cell is specified, a degree of similarity thereof, an image of test image data, an image of color image data cut out in the same range as the test image data, and an image of the entire color image data. It is preferable that at least one selected from can be displayed on the display 4. When an image of inspection image data or an image of color image data is displayed on the display 4, it is preferable to display an instruction image such as a red circle or an arrow at the specified NRBC position.

  In addition, the target cell search device 1 uses the test image data K (p, q) determined by the target cell specifying unit S21 to be an NRBC cell nucleus and a plurality of binarizations for targets used for creating the target dictionary. A target dictionary learning means 51 (FIG. 5) that performs principal component analysis together with the image data D (p, q) and updates and stores the orthonormal basis obtained from the eigenvector as the target dictionary B in the target dictionary storage unit 31. 3) may be further provided. By learning and updating the target dictionary B in this way, the target dictionary B becomes more accurate, so that the NRBC search accuracy can be improved.

  Further, the target cell search device 1 has the test image data K (p, q) that the tertiary candidate search means 46 did not select, or the test image data K that has not been determined that the target cell specifying means is NRBC. (P, q) and a plurality of non-dictionary image data used to create the non-target dictionary C are subjected to principal component analysis, and the orthonormal basis obtained from the eigenvector is set as the non-target dictionary C. You may further provide the non-target dictionary learning means 10 to update and memorize | store in the target dictionary memory | storage part 32. FIG. By learning and updating the non-target dictionary C in this way, the non-target dictionary C becomes more accurate, so that the NRBC search accuracy can be improved.

  This completes the NRBC cell search using the target cell search device 1.

  In the target cell recovery step S22, the target cell is recovered based on the position of the target cell output from the target cell search device 1. For example, the flat plate 8 is set on the XY stage of an inverted microscope, and the XY stage is moved so as to observe the positions Lx and Ly where the target cells are specified. While confirming the inverted microscope, the operator moves the position roughly with a motor so that the tip of the glass capillary for cell recovery having a slightly larger lumen than the target cell is located at the position Lx, Ly of the target cell. Manipulate a micromanipulator that moves a minute position to collect target cells. Target cells can be collected by various known methods. For example, an enzyme-containing solution such as physiological saline, Ringer's solution, phosphate buffered saline (PBS), or protein kinase K is dropped on a desired target cell with a glass capillary for cell recovery, and left for a while to slide glass. After the target cells fixed to the cell are easily removed, the target cells are selectively sucked out by using the same glass capillary or another cell recovery glass capillary under negative pressure to obtain the target cells. The target cells may be scraped or scooped with a spatula.

  In the target cell inspection step S24, the chromosomes, DNA and / or RNA in the cell nucleus of the collected target cells are analyzed. For this purpose, a highly sensitive analysis method is required. For example, fluorescence in situ hybridization (FISH method: Fluorescence in situ hybridization), DNA, etc., for labeling the chromosomes and RNA in the cell nucleus of the obtained cells. Polymerase chain reaction (PCR method) that amplifies DNA, primer extension preamplification (PEP method) that amplifies genomic DNA using random 15-mer in combination with Taq DNA polymerase to initiate copying through the genome : Primer extension preamplification), comparative genomic hybridization (CGH method: comparative genomic hybridization), in situ PCR method, denatured oligonucleotide to detect genomic DNA excess, deletion, amplification, etc. Amplify DNA etc. using origin polymerase ( An analysis method such as DOP-PCR (Degenerated Oligonucleotide Primer PCR) can be mentioned. Such an inspection process may be performed in a plate well, a microtube, or the like, but a detection biochip may be used in order to reliably inspect a valuable sample without waste. The detection biochip has a reaction field composed of, for example, a flow path, a compartment, a groove, a capillary, a fiber, and / or a bead, and a path through which a labeling reagent can be injected as necessary, an extra reagent, a by-product, It has a route through which unnecessary chemicals can be discharged, and these analysis methods can be carried out in minute amounts in the reaction field.

  For example, by examining the obtained cells by the FISH method using monoclonal antibodies against CD71 and glycophorin A, it is possible to examine abnormalities in chromosomes, abnormal numbers in XY chromosomes, abnormalities in 21 · 18 or 13 trisomy In addition, sex determination, individual identification, and Rh (D) type can be examined by PCR, and a plurality of gene abnormalities can be diagnosed by PEP.

  During the target cell inspection step S24, a chromosome, DNA and / or RNA is used to detect a diagnostic fluorescent probe in which a chromosome-specific gene and a fluorescent dye are bound, for example, a fluorescently labeled probe that specifically binds to a centromere. A hybridization reaction may be detected and analyzed. For example, CEP DNA FISH probes, CEP Chromosome Enumeration DNA FISH Probes, and CEP probe kits, which are CEP series (CEP is a registered trademark) manufactured by Abbott. The labeling material for the diagnostic probe may use fluorescence or magnetism.

  This completes the method for examining target cells in blood.

  As shown in FIG. 7, the optically labeled concentrated blood sample may be injected into the flow channel 71 of the biochip 70 and the target cell search device 1 may search for the NRBC. In the figure, components similar to those already described are assigned the same reference numerals and detailed description thereof is omitted.

  In the figure, a biochip 70 is held on the XYZ stage 11 of the target cell search device 1. This biochip 70 has one flow path 71. A sample container 61 containing a concentrated blood sample is connected to the input side of the flow channel 71 of the biochip 70 via a conduit, and the sample container 61 is connected via a control valve 62 that can be controlled to open and close. The pressure pump 63 is connected. Further, a three-way opening / closing control valve 64 is connected to the output side of the flow path 71, a collection container 65 is connected to one output pipe 67 (recovery flow path) of the three-way opening / closing control valve 64, and the other output pipe. 68 is connected to a disposal container 66. The three-way opening / closing control valve 64 is an opening / closing control valve having one input port and two output ports and capable of opening / closing each output port. In this case, the three-way opening / closing control valve 64 controls opening / closing of the output pipe 67 side and the output pipe 68 side. A separate opening / closing control valve may be provided for each of the output pipe 67 and the output pipe 68. The biochip 70 is made of a transparent material such as a pressure resistant resin, specifically a polyester resin such as polyethylene terephthalate (PET) or an acrylic resin so that the NRBC flowing through the channel 71 from the outside can be optically observed. It is made of plastic resin such as cyclic polyolefin and silicone resin.

  Since the NRBC has a size of about 10 μm, the flow path is formed with a pipe width of 10 μm to 20 μm, preferably 10 μm to 15 μm.

  The output pipe 68 side of the control valve 62 and the three-way opening / closing control valve 64 is opened, the output pipe 67 side is closed, each valve is closed after the concentrated blood sample flows into the flow path 71, and the optical labeling expression step S13 already described. -Target cell specific process S21 is performed. When the NRBC is specified, the control valve 62 and the three-way opening / closing control valve 64 are controlled to guide the NRBC to flow into the recovery container 65, and control is performed so that other than the NRBC flows into the disposal container 66. By repeating this process, NRBC is recovered (target cell recovery step S22).

  Alternatively, the output pipe 68 side of the control valve 62 and the three-way opening / closing control valve 64 is opened, the output pipe 67 side is closed, and the concentrated blood sample is slowly allowed to flow through the flow path 71 while the optical label expression step S13 to target cell identification. When step S21 is repeated and NRBC is specified, the three-way opening / closing control valve 64 is guided to control the three-way opening / closing control valve 64 so that the NRBC flows into the recovery container 65, and the three-way opening / closing control valve is flowed so as to flow into the disposal container 66 other than NRBC. 64 is controlled (target cell recovery step S22). Since the flow path 71 is narrower than the flat plate 8, the size of the image data to be image-processed is small, so that the image processing can be performed at high speed and the NRBC can be specified at high speed. Therefore, the NRBC can be specified before passing the three-way opening / closing control valve 64, and the NRBC can be recovered.

  Using such a biochip 70 eliminates the need to fix the concentrated blood sample on the flat plate 8 and to remove and collect the NRBC from the flat plate 8, thereby improving work efficiency and reducing the NRBC. It can be collected quickly. Therefore, inspection efficiency can be improved.

  Further, the biochip 70 of FIG. 7 may be replaced with the biochip 75 shown in FIG. 8 to search and collect target cells. In this case, the three-way opening / closing control valve 64 shown in FIG. 7 is unnecessary, and the collection container 65 and the disposal container 66 are connected to the biochip 75.

  The biochip 75 shown in FIG. 8 is formed of, for example, a silicone resin (for example, polydimethylsiloxane: PDMS), a plastic resin such as PET, acrylic resin, or cyclic polyolefin, or glass, and blood collected for testing. And a flow path 76 into which the concentrated blood sample is injected. The channel 76 is provided with a target cell passage filter 72, an optical observation window 73, a recovery branch 78, and an open / close control valve 79. In the figure, the positional relationship between the flow path 76 and the target cell passage filter 72 is schematically shown.

  The target cell passage filter 72 is a cell (non-target cell) 92 larger than the target cell 91 and capable of passing the target cell 91 to be examined in the concentrated blood sample as shown in FIG. Has a gap 82 that cannot pass through. For example, the target cell 91 is a red blood cell such as NRBC, and the non-target cell 92 is a white blood cell larger than the red blood cell. The non-target cells 92 are discharged to the outside of the biochip 75 through the disposal path 77 branched from the flow path 76 immediately upstream of the gap 82. When the number of non-target cells 92 is small, a storage tank for non-target cells 92 is formed in the biochip 75, and the non-target cells 92 removed by the target cell passage filter 72 are not discarded outside and stored in the storage tank. It is good also as a structure to accommodate.

  A plurality of such target cell passage filters 72 may be arranged along the flow path 76. By arranging the plurality of target cell passage filters 72, the separation accuracy between the target cells 91 and the non-target cells 92 is improved, so that the search for the target cells 91 is further accelerated.

  An optical observation window 73 is disposed downstream of the target cell passage filter 72. The enlarged plan view is shown in FIG. 2C, and the enlarged side sectional view is shown in FIG. The optical observation window 73 is formed of a transparent material so that optical observation of the target cell 91 flowing through the flow path 76 from the outside of the biochip 75 is possible. The position of the XYZ stage 11 is controlled so that the imaging camera 15 of the target cell search device 1 images the optical observation window 73.

  The channel 76 of the optical observation window 73 is preferably formed with a channel width that allows the NRBC to pass in one row along the channel 76 without overlapping. Specifically, since the NRBC has a substantially disk shape, the cross section that can pass through the disks in a single row and aligned in the direction of the surface is rectangular so that the flat surface can be observed from the optical observation window 73. It is preferable to form the flow path 76 with a hole. For example, the width of the flow path 76 parallel to the optical observation window 73 is formed with a width L1 (for example, 10 μm to 15 μm) such that a disk having a disk-like NRBC shape (diameter of about 10 μm) recessed at the center passes. The width of the flow path 76 perpendicular to the observation window 73 is formed with a width L2 (for example, 3 μm to 4 μm) that allows a disk (thickness of about 3 μm) to pass through. When the flow path 76 is formed in this manner, the flat surface can be reliably observed without overlapping the NRBCs, so that it is possible to accurately distinguish between normal enucleated erythrocytes and NRBCs and further improve the NRBC search efficiency. be able to. In the figure, an example is shown in which the flow path 76 is formed with such a narrow flow path width from the place past the target cell passage filter 72.

  Downstream of the optical observation window 73, a recovery branch path 78 (recovery flow path path) for recovering NRBC from the flow path 76 is branched. An opening / closing control valve 79 capable of branch control is disposed at a branch portion of the recovery branch path 78. This open / close control valve 79 has the same function as the open / close control valve 64 already described, and can open and close the recovery branch 78 and the flow path 76 respectively. The open / close control valve 79 can perform open / close control by inputting a control signal to the control electrode 85.

  The on-off control valve 79 is a thermally controllable thermal control valve, an electrically controllable electrical control valve, a magnetically controllable magnetic control valve, or a pneumatic control valve that can be controlled by air pressure. It is preferable. When the open / close control valve 79 is a thermal control valve, for example, the valve is formed of a bimetallic material in which two metal plates having different thermal expansion coefficients are stacked, a shape memory alloy, or a thermoresponsive polymer such as poly-N-isopropylamide. The valve is activated by applying heat. For example, the heat control valve is irradiated with laser light to generate heat, and the heat control valve is operated with this heat. When the open / close control valve 79 is an electric control valve, for example, the valve is formed of an electromagnetic control valve or a piezoelectric element that operates by energizing an electromagnetic coil. A metal terminal electrically connected to the electric control valve is formed on the surface of the biochip 75, and a control electric signal is applied to this terminal to operate the electric control valve. When the open / close control valve 79 is a magnetic control valve, for example, the valve is formed of a magnetostrictive element. The magnetic control valve is operated by applying a magnetic field by bringing a magnet close to the magnetic control valve. When the open / close control valve 79 is a pneumatic control valve, an air valve made of an elastic material such as PDMS is formed in the biochip, and the pneumatic control valve is operated by pressurization / decompression of air. In the case of a heat control valve or a magnetic control valve, since the valve can be operated without contact with the biochip 75, it is not necessary to consider contact failure, and it is preferable because it can be operated with high reliability.

  The recovery container 65 of FIG. 7 is connected to the recovery branch 78, and the waste container 66 of FIG.

  Although not shown, the target cell search device 1 has a control mechanism for operating the open / close control valve 79 (for example, a laser light source and its control means in the case of a thermal control valve). The open / close control valve 79 is provided in the biochip 75 and the target cell search device 1 performs this control, whereby the specified NRBC can be reliably recovered. Further, there is no need to provide the three-way opening / closing control valve 64 outside, and when testing a plurality of specimens, there is no need to replace the three-way opening / closing control valve 64 for each specimen, and NRBC is quickly collected from a large number of specimens. can do.

  As shown in FIG. 8, a chemical solution injection path 80 for injecting an optical labeling chemical solution for optically labeling NRBC from the outside may be joined to the flow channel 76 upstream of the optical observation window 73. . In this case, NRBC can be optically labeled in the flow channel 76 by injecting a concentrated blood sample that is not optically labeled into the flow channel 76 and injecting an optically labeled chemical solution from the chemical solution injection path 80 (specimen labeling step). S12). An example of the cell image after staining is shown in FIG. FIG. 10 (a) is a typical image of the target cell NRBC. FIG. 10B is a typical image of normal red blood cells (RBC), which is an example of a non-target cell, and FIG. 10C is a typical image of white blood cells (WBC), which is an example of another non-target cell. It is a typical image.

  Instead of the opening / closing control valve 79, a three-way opening / closing control valve may be arranged at a branching point of the recovery branch 78 and the flow path 76. A direction switching valve that can control whether it flows may be provided.

  Instead of the opening / closing control valve 79, a manual valve that can be opened and closed may be provided. In this case, when the NRBC is specified, for example, the NRBC is tracked and imaged by the imaging camera 15, and the operator manually operates the valve while viewing the image to collect the NRBC in the collection container 65 (target Cell recovery step S22).

  As an example, the biochip 75 is formed by bonding a base plate member 86 and a cover plate member 87, which are two plate members, as shown in a side view in FIG. The base plate 86 has a groove that becomes a flow path 76, a target cell passage filter 72, a disposal path 77, and a collection branch path 78, and a groove into which the opening / closing control valve 79 is fitted, and is formed by integral molding. This base plate material 86 is made of transparent resin or glass. When a hard resin is used, it is manufactured by nanoimprinting, transfer molding, cutting using a mold or Si mold, or injection molding using a mold. can do. When a soft resin such as PDMS or elastomer is used, it can be manufactured by nanoimprinting or transfer molding using a mold or Si mold, or injection molding using a mold. When glass is used, it can be manufactured by dry etching, wet etching, or the like. When electrical wiring is required, wiring such as Cr and Au is manufactured by vacuum deposition or sputtering. After the opening / closing control valve 79 is incorporated into the base plate 86, the cover plate 87, which is a transparent resin flat plate, is bonded by bonding, thermal bonding, or physical adsorption, thereby completing the biochip 75. For example, a frame-shaped instruction display indicating the position may be attached to the optical observation window 73 by printing, engraving, or the like. Moreover, it is also possible to manufacture a flow path in the cover plate material 86 as necessary, or to manufacture a biochip by superposing three or more layers.

  The target cells (S23 in FIG. 1) collected by the biochip 75 are subjected to processing such as the target cell inspection step S24.

In the target cell search device 1, the focus control laser light source 12 may also be used as the imaging light source 14. More specifically, without placed the imaging light source 14 in FIGS. 4 (a) and 7, the laser light source 12 to irradiate the optical indicator expressing light L _T as light Le for focusing. If it does in this way, it will become a simple structure. Since the irradiation range (irradiation area) of the laser beam Le is narrow, this configuration can be particularly preferably employed when imaging the channel 71 in FIG. 7 and the channel 76 in FIG. 8 where the imaging range is narrow.

  An example in which the method for examining cells in blood using the present invention is performed and used for prenatal diagnosis using fetal nucleated red blood cells (NRBC) in maternal blood is shown below.

Example 1
The test method for fetal cells is as follows: [1] Specimen acquisition step for acquiring a concentrated blood sample by concentration of NRBC by density gradient centrifugation of maternal blood; [2] Blood is formed by smearing the concentrated blood sample on a flat plate. Specimen labeling step of preparing a smear, optically labeling with Giemsa staining, [3] Optical labeling step of expressing an optically labeled sample and capturing the color image, focus control step, color image acquisition step, [4 A binarization process for binarizing the color image to remove noise, a contraction / expansion process, [5] a primary candidate search process for searching and specifying the position of the NRBC, a secondary candidate search process, a tertiary candidate search process, Target cell identification step, [6] Target cell recovery step of capturing and recovering NRBC on the plate based on the position information, [7] Gene information in the nucleus of the recovered target cell The target cells inspection step of analyzing the biological technique is chromosomal tests are performed.

  Below, the basic operation procedure of each process [1]-[7] is demonstrated in detail.

[1] Specimen acquisition process A stock solution of polyvinylpyrrolidone-coated silica gel fine powder (GE Healthcare Japan Co., Ltd .: Percoll) is diluted with physiological saline (0.15M NaCl aqueous solution: 0.9% NaCl aqueous solution) to obtain a specific gravity. Percoll dilutions of 1.085 g / ml and 1.075 g / ml were prepared.

  After injecting 5 ml of Percoll diluted solution with a specific gravity of 1.085 g / ml into a 15 ml centrifuge tube, 5 ml of Percoll diluted solution with a specific gravity of 1.075 g / ml was slowly injected and laminated. Maternal blood collected from a pregnant woman with heparin was diluted 2-fold with physiological saline. 5 ml of diluted maternal blood was slowly injected into the 15 ml centrifuge tube and laminated. If a larger amount of maternal blood is used, the number of centrifuge tubes may be increased or a 50 ml centrifuge tube may be used.

  The centrifuge tube was centrifuged for 30 minutes. In order from the bottom, the erythrocyte layer, the Percoll diluent layer having a specific gravity of 1.085 g / ml, the NRBC and leukocyte-containing layer, the specific gravity of 1.075 g / ml Percoll diluent, the lymphocyte layer, and the plasma layer were separated. The plasma layer was removed with a disposable pipette. Subsequently, the lymphocyte layer and the specific gravity 1.075 g / ml Percoll dilution layer were also removed.

  The target layer, in which NRBC and leukocytes are present, appears on the surface layer and was carefully pipetted.

  The collected recovery solution of the target layer is placed in a 25 ml centrifuge tube, and then physiological saline or phosphate buffered saline (PBS) pH 7.2 is added thereto, followed by centrifugation for 10 minutes, and the supernatant is removed three times. Washing was performed by repeating.

  Thereto, PBS was added and diluted to 10 μl to obtain a cell suspension containing NRBC.

[2] Sample labeling step A Giemsa stock solution containing methylene blue, eosin and Azure II was diluted with PBS to give a 1% Giemsa solution dilution.

  After 2 μl of the NRBC-containing cell suspension obtained in the above-described specimen acquisition step was dropped on a slide glass, the suspension was uniformly smeared on the slide glass using a cover glass to form a membrane. If necessary, cold air was applied to the smear to dry the smear. The slide glass together with the dried smear film was immersed in methanol in a pad for 10 minutes to immobilize the smear film. The smear film was dried while applying cold air as necessary.

  Next, the slide glass together with the dried smear was immersed in a 2% Giemsa solution diluted solution in another pad for 10 minutes and stained. The slide glass together with the stained smear was washed twice with ultrapure water in another pad. The smear film was dried while applying cold air as needed to prepare a specimen-labeled blood smear sample.

The target cell search device described in the embodiment was prototyped. Target dictionaries and non-target dictionaries were created in advance. One of the target dictionary image data used for the creation of the target dictionary is enlarged and shown in FIG. FIG. 15B is an enlarged view of one of the non-target dictionary image data used for creating the non-target dictionary. These images are 52 pixels × 52 pixels in size.
[3] Optical labeling step, focus control step, color image acquisition step The slide glass prepared in the specimen labeling step was set in the stocker of the prototype target cell search device. When the target cell search device 1 is operated, the target cell search device 1 transports the slide glass to the XYZ stage, irradiates the blood smear sample with the imaging light source, and automatically focuses and color images. . An example of the captured color image is shown in FIG. This color image has a size of 2448 pixels × 2050 pixels.

[4] Binarization processing step and contraction / expansion step Subsequently, the target cell search device binarizes the color image data and performs contraction / expansion processing. FIG. 16B shows a binarized image obtained by binarizing the color image of FIG. 16A, and FIG. 16C shows a binarized image obtained by contracting and expanding the color image. In each image, the position of the NRBC specified in the target cell specifying step described later is indicated by an arrow.

[5] Primary candidate search step, secondary candidate search step, tertiary candidate search step, target cell specifying step Next, the target cell search device performs a primary search step to a target cell specifying step, and FIG. One target cell was automatically identified from the image of. In the target cell specifying step, the size of the connected image region of the test image data selected in the tertiary candidate search step is in the range of 20 to 40 pixels, and the roundness is in the range of 0.5 to 1.5. The connected image area inside was determined to be the nucleus of the target cell. The identified NRBC image was visually confirmed by the operator to be NRBC.

In the tertiary candidate search step, the NRBC shown in FIG.
Distance Gb = 44.004847 with target dictionary B
Distance Gc = 29.512506 with non-target dictionary C
Evaluation function S = −0.197128166
In the target cell identification step,
X-axis direction width (maximum value) = 25 pixels Y-axis direction width (maximum value) = 24 pixels Area = 463
Perimeter L = 82
Radius R = Perimeter length ÷ π ÷ 2 = 13.050705
Roundness = 0.865293
Met.

[6] Target cell recovery step A blood smear sample containing NRBC found by a target cell search apparatus was set on the stage of an inverted microscope equipped with a micromanipulator. The stage was moved to the position specified in the target cell identification step, and manually adjusted to the position where NRBC was present while looking at the measure attached to the microscope.

  Next, using a micromanipulator, a 100 μg / ml proteinase K solution (Wako Pure Chemical Industries, Ltd., 164-14004 dissolved in TE buffer) is sucked into a glass capillary for cell recovery, and the proteinase is recovered from the recovery capillary. The solution K was discharged toward the target NRBC and allowed to stand at room temperature for 5 minutes. When the vicinity of the target NRBC was easily peeled off by the proteinase K solution for a while, the target NRBC was peeled off with the cell peeling glass capillary, and the NRBC was sucked with the cell collecting glass capillary, captured and collected. The recovered NRBC was put in a 600 μl microtube, washed twice with PBS, and finally adjusted to a volume of about 2 μl to obtain a recovered NRBC suspension. The NRBC collected by the micromanipulator can be directly transferred onto a slide glass for chromosome examination and fixed to Carnoy instead of being collected in a microtube.

[7] Target cell test process (chromosome test: FISH method)
First, a solution of ethanol: acetic acid = 3: 1 was prepared as a Carnoy liquid. As an aging solution, a 2 × SSC (standard citric acid-added physiological saline) /0.1% NP-40 (Nonidet-40) solution was prepared. As the ethanol washing solution, 70% ethanol aqueous solution, 85% ethanol aqueous solution, and 100% ethanol were prepared. A 70% formamide / 2 × SSC solution was prepared as a denaturing solution. A 50% formamide / 2 × SSC solution was prepared as a formamide washing solution.

  Next, 7 μl of hybridization buffer (Abbott, attached to the FISH probe) and FISH probe (Abbott 32-190002, 32--190001, 32-192018, 32) -130018, 32-131018, 32-1332023, 32-131024, etc.) 1 μl and 2 μl of ultrapure water were prepared in a microtube. The mixing ratio varies depending on the number of FISH probes used. The above example is an example in the case of using one type of FISH probe. This was centrifuged for several seconds after stirring, and then heated with a block heater at 73 ° C. for 5 minutes to denature the probe. This probe solution was kept at 45 ° C. until just before use.

  To the recovered NRBC suspension, 5 μl of Carnoy's solution was added, and after centrifugation at 1500 rpm for 5 minutes, the supernatant was removed, and 2 μl of Carnoy's solution was further added. The above operations of centrifugation, removal of supernatant, and addition of 2 μl of Carnoy's solution are repeated three times.

  This Carnoy's solution containing NRBC was dropped onto the center of the frontier (FRC-05, Matsunami Glass Industrial Co., Ltd.), which is a cell-immobilized coated slide glass, and placed in a humidified tapper and left at 67 ° C. for 10 minutes. Then, it dried with cold air and obtained the NRBC fixed slide.

  In the staining pad, the aging solution (2 × SSC / 0.1% NP-40) was heated to 37 ° C., and the NRBC-immobilized slide was immersed therein for 30 minutes. Next, the NRBC-immobilized slide was washed and dehydrated in a separate staining pad in the order of 70% ethanol aqueous solution, 85% ethanol aqueous solution, and 100% ethanol in order, followed by washing and dehydration. And dried.

  Depending on the state of the sample cells, the FISH success rate is improved by performing lithium dodecyl sulfate (LIS) solution treatment and pepsin solution treatment as necessary.

NRBC-immobilized slides were immersed in a room temperature LIS solution (10-100 mM LIS / 0.1 M Tris (tris hydroxymethylaminomethane), pH 7.4) for 30 minutes in a staining pad. The NRBC-immobilized slide is then washed three times in a separate staining pad for 1 minute immersion in 2 × SSC solution at room temperature. Next, it was immersed in a pepsin solution (10-100 μg / ml pepsin / 0.01 M HCl, pH 2) heated to 37 ° C. in a staining pad for 10 minutes. The room temperature formalin refix solution (1% formaldehyde / 50 mM MgCl ₂ ) in another staining pad was then added to the room temperature enzyme quenching solution (50 mM MgCl ₂ / PBS solution, pH 7.4) in another staining pad for 5 minutes. / PBS solution, pH 7.4) for 10 minutes. Further, the sample was immersed in PBS at room temperature for 5 minutes, and washed and dehydrated in the order of 70% ethanol aqueous solution, 85% ethanol aqueous solution, and 100% ethanol as an ethanol cleaning solution, followed by drying with cold air.

  In the staining pad, the denaturing solution (70% formamide / 2 × SSC solution) was heated to 73 ° C., and the NRBC-immobilized slide was immersed therein for 5 minutes. Next, the NRBC-immobilized slide was washed and dehydrated in a separate staining pad in the order of 70% ethanol aqueous solution, 85% ethanol aqueous solution, and 100% ethanol in order, followed by washing and dehydration. And dried.

  After the NRBC-immobilized slide was heated to 45 ° C. with a heater, 10 μl of the probe solution stored at 45 ° C. was dropped onto the NRBC-immobilized slide, covered with a cover glass, and air bubbles were expelled. Sealed with KOKUYO Co., Ltd. Alternatively, a hybridization mask seal (manufactured by SC World Co., Ltd.) is attached to the cell immobilization position on the NRBC-immobilized slide, heated to 45 ° C. with a heater, and 2 μl of the probe solution stored at 45 ° C. is then added. The solution was dropped on a well formed like a bank with a hybridization mask seal on an NRBC-immobilized slide, and a sealing seal was affixed to prevent bubbles from entering.

  The NRBC-immobilized slide to which the probe solution was dropped was placed in a humidified tapper in an incubator adjusted to 37 ° C., and hybridization was performed for 4 to 18 hours.

  The cover glass or the mask mask for hybridization was removed, and the NRBC-immobilized slide after hybridization was washed with 2 × SSC solution at room temperature.

  The NRBC-fixed slide was washed by soaking in a 2 × SSC / 0.3% NP-40 solution at room temperature for 5 minutes in a staining pad, and 2 × SSC / 0.3% at 73 ° C. in another staining pad. Washed by soaking in NP-40 solution for 2 minutes, then soaked in 2 × SSC for 1 minute at room temperature in a separate staining pad.

  On this NRBC-immobilized slide, 150 μg / ml DAPI (4 ′, 4%, 4%, DABCO (1,4-diazabicyclo [2,2,2] octan, manufactured by Wako Pure Chemical Industries, Ltd .; 049-25712)) is contained. 6-diamidino-2-phenylindole, manufactured by Dojindo Laboratories Co., Ltd .; 340-07971) or DAPI solution containing a fluorescent fading inhibitor (manufactured by Invitrogen; SlowFade (registered trademark) Gold Antifade Reagent with DAPI S36938, etc.) The cover glass was covered, and the cover glass was sealed with non-fluorescent and colorless nail polish to prevent the cover glass from slipping and the DAPI solution from drying.

  The NRBC-immobilized slide was observed and imaged using an appropriate filter with a fluorescence microscope. The NRBC nucleus on the NRBC-immobilized slide was in a state where a number of fluorescent spots corresponding to the number of chromosomes were observed by hybridization with a fluorescently labeled FISH probe. FIG. 11 shows examples of images of normal cells and chromosome 21 trisomy cells. FIG. 11 (a) is a photomicrograph of a fluorescence image after FISH of a normal cell, FIG. 11 (b) is a photomicrograph of a fluorescence image after FISH of a 21-chromosome trisomy cell, where X is the X chromosome, and Y is Y chromosome.

  Data analysis on the number of chromosomes based on the number of fluorescent spots was performed using the captured images, and it was possible to determine a congenital abnormality due to an abnormal number of chromosomes.

(Example 2)
If necessary, it is possible to collect nuclei after chromosome examination (FISH) and perform genetic examination using a DNA microarray. By using a DNA microarray, a single cell can be used to test for many diseases. The basic operation procedure will be described in detail.

  NRBC-immobilized slides after chromosome examination (FISH) as in [1] to [7] of Example 1 were used.

[7 '] Target cell test process (gene test: DNA microarray method)
From the NRBC-immobilized slide after FISH, NRBC or NRBC nuclei were collected in a microtube using a micromanipulator attached to an inverted microscope.

  In a microtube from which NRBC was collected, a random primer (SP180-1 manufactured by Operon Biotechnology Co., Ltd.), PCR enzyme (EX Taq DNA Polymerase RR001A manufactured by Takara Bio Inc.), buffer (attached to the PCR enzyme), dNTPPmix Etc. (attached to the PCR enzyme) were added to carry out a PCR amplification reaction.

  From the solution after the PCR reaction, the amplified gene is purified using a PCR product purification kit (Qiagen 28104), and the PCR reaction is performed using UULISYS Nucleic Acid Labeling Kits, Alexa Flour 546 (Molecular Probe U-21652). The gene amplified in step 1 was fluorescently labeled and purified with Microcon YM-30 (Millipore 42409) to obtain a fluorescently labeled PCR product. As another method for obtaining fluorescent-labeled PCR products, Cy3-dUTP (GE Healthcare PA55021) is added to the reaction solution during PCR reaction, directly labeled during PCR reaction, and purified with PCR product purification kit. There is also a method.

  This fluorescently labeled PCR product and 2x hybridization buffer (12 x SSC, 0.4% SDS, 10 x Denhardt's solution, 0.2 mg / ml denatured salmon sperm DNA) are mixed in equal amounts and heat-denatured at 94 ° C for 1 minute. And then cooled rapidly on ice.

  A DNA microarray (also called a DNA chip slide glass Takara-Hubble Slide Glass TX720 manufactured by Takara Bio Inc., Oligo manufactured by Sumitomo Bakelite Co., Ltd.), which is also called a DNA chip in which partial sequences of DNA related to congenital anomalies are arranged and immobilized. 10 μl was dropped on a DNA immobilization substrate BS-11607, which was immobilized on probe DNA by SC World Co., Ltd., and covered with a cover glass. Alternatively, a hybridization mask seal (manufactured by SC World Co., Ltd.) was attached on the DNA microarray, and 2 μl of the hybridization solution was dropped into the well and sealed.

  The DNA microarray was placed in a humidified tapper in an incubator set to a hybridization temperature (45 to 60 ° C.), and hybridization was performed for 4 to 18 hours.

  Remove the cover glass and seal from the DNA microarray and immerse in 5 minutes each in 3 different washing solutions: 2 x SSC solution at room temperature, 2 x SSC / 0.2% SDS solution at 65 ° C and 0.05 x SSC solution at room temperature. Washing was performed.

  Fluorescence was observed and imaged with a DNA chip scanner or a fluorescence microscope. When DNA corresponding to the DNA spotted on the DNA microarray is present in the hybridization solution, a DNA hybridization reaction occurs at that location, and thus fluorescence labeled on the DNA in the hybridization solution is observed.

  When data analysis was performed using the captured image, when there was a gene sequence abnormality related to a congenital abnormality, fluorescence was observed at the spot where the relevant probe was spotted.

  Depending on the method of genetic testing, direct labeling during PCR reaction using fluorescently labeled nucleic acids such as Cy3-dUTP (GE Healthcare PA55021) during PCR amplification, and purification with a PCR product purification kit, In some cases, fluorescent labeling operation after PCR is not necessary.

(Example 3)
(3-1) Chromosome inspection by FISH on biochip

  A 0.15M KCl low expansion solution was added to the concentrated blood sample obtained by the Percoll method, and erythrocytes without nuclei were destroyed at room temperature for 10 minutes (some anucleated erythrocytes remained unbroken). ). Next, the supernatant was removed by centrifugation at 1500 rpm for 5 minutes. Furthermore, the operation of adding PBS, centrifuging, and removing the supernatant was repeated three times to obtain a concentrated blood sample in which the amount of nucleated red blood cells was reduced.

  Target cells were collected and NRBC suspension was obtained in the same manner as in [2] to [6] of Example 1 except that this concentrated blood sample with a reduced amount of nucleated red blood cells was used. PBS was added thereto, passed through a leukocyte removal filter, or a concentrated blood sample from which leukocytes were removed using CD45-labeled magnetic beads was injected into the biochip on the biochip reaction detection apparatus shown in FIG. In the figure, the liquid flow path 101 is indicated by a solid line, and the control pattern wiring of the open / close control valve 102 disposed in the middle of each liquid flow path 101 is indicated by a broken line. As shown in the figure, the biochip reaction detection apparatus has a FISH chamber 103 containing a cell immobilization filter 104, a PCR chamber 105, and a DNA microarray chamber (gene testing chamber) 106.

  Each reaction solution prepared in the biochip reaction detection apparatus was set (the details of the composition and preparation of each solution are the same as those of the FISH and DNA microarray described above).

  5-10 μl of the concentrated blood sample solution that has passed through the leukocyte removal filter or 5 μl of the NRBC suspension collected from the blood smear sample slide is sent to the FISH chamber 103 in the biochip, and the cells are placed on the cell immobilization filter 104. Immobilized.

  5 μl of aging solution (2 × SSC / 0.1% NP-40) was fed to the FISH chamber 103 in the biochip, and the biochip was heated to 37 ° C. with the apparatus and placed for 30 minutes. Next, washing was performed by slowly feeding 10 μl of 70% ethanol aqueous solution, 10 μl of 85% ethanol aqueous solution, and 10 μl of 100% ethanol at room temperature in this order, and air was sent to the FISH chamber 103 and air dried.

  5 μl (2 × SSC / 0.1% NP-40) of a denaturing solution (70% formamide / 2 × SSC solution) is fed to the FISH chamber 103 in the biochip, and the biochip is heated to 73 ° C. with the apparatus. Let sit for 5 minutes. Next, the biochip was cooled to 4 ° C. with an apparatus, and then washed with 10 μl of a 70% ethanol aqueous solution cooled to 4 ° C. Next, the apparatus was returned to room temperature and slowly washed in the order of 10 μl of 85% ethanol aqueous solution and 10 μl of 100% ethanol for washing, and air was sent to the FISH chamber and air dried.

  Place 2 μl of the prepared hybridization probe solution in the FISH chamber 103 of the flow path chip, heat the biochip to 73 ° C. with the apparatus, heat denature for 5 minutes, and then place at 37 ° C. for 4 to 16 hours. A hybridization reaction was carried out (vibration was applied as necessary to promote the hybridization reaction).

  At room temperature, 10 μl of 2 × SSC / 0.3% NP-40 was poured into the FISH chamber and left for 5 minutes.

  10 μl of 2 × SSC / 0.3% NP-40 was poured into the FISH chamber, and the biochip was heated to 73 ° C. with the apparatus and placed for 2 minutes.

  At room temperature, 20 μl of 2 × SSC was poured into the FISH chamber and washed.

  2. 2 μl of 150 μg / ml DAPI contrasting staining solution containing 3% DABCO was placed in the FISH chamber.

  The chamber-immobilized cells were observed and imaged using an appropriate filter mounted on the apparatus.

  NRBC was determined from the cell shape of the captured image, and a congenital abnormality due to an abnormal number of chromosomes was determined from the FISH result of NRBC. An example of normal cell observation results is shown in FIG. FIG. 13 (a) is a photomicrograph of NRBC, which is a target cell. FIG. 13 (b) is a micrograph of NRBC, which is a target cell after FISH, 21 is 21 chromosomes, X is X chromosome, and Y is Y chromosome. As is clear from FIG. 13, the evaluated target NRBC was normal with no chromosome number abnormality.

(3-2) Genetic test using DNA microarray on biochip A genetic test was performed in the biochip directly from the NRBC in the FISH chamber 103 of the biochip after the chromosome test or from the separately collected NRBC. When using the NRBC in the FISH chamber 103 in the biochip after the chromosome test, 5 μl of hypotonic solution (75 mM KCl) is placed in the FISH chamber 103 for 5 minutes at room temperature. Next, 5 μl of cell disruption solution (100 μg / ml proteinase K / 0.01 M EDTA / 0.5% SDS / TE buffer solution) was injected, and the cells were heated to 37 ° C. with an apparatus for 10 minutes to disrupt the cells.

  The cell disruption solution is transferred from the FISH chamber 103 of the biochip to the PCR reaction chamber 105, 5 μl of PCR mix solution (primer / PCR enzyme / PCR buffer / Cy3-dUTP / dNTPmix) is added, and the temperature of the biochip is controlled by the device. (The cycle of 94 ° C. for 1 minute → 37 ° C. for 2 minutes → 72 ° C. for 1 minute was repeated) to carry out the PCR reaction.

  The reaction solution after the PCR reaction is sent to the genetic testing chamber 106 containing the DNA microarray in which the probe DNA to be tested is immobilized, 2 μl of 2 × hybridization buffer is added, and the biochip is heated to 94 ° C. with the apparatus. After heat denaturation for 2 minutes, the mixture was rapidly cooled to 37 ° C. and allowed to stand at 37 ° C. for 4 to 16 hours to allow hybridization reaction (vibration was applied as necessary to promote the hybridization reaction).

  After completion of the hybridization reaction, the gene testing chamber 106 is washed with 10 μl of 2 × SSC at room temperature, and then 10 μl of 2 × SSC / 0.2% SDS solution is injected and heated to 65 ° C. with the apparatus for 5 minutes. placed. Next, washing was performed with 20 μl of 0.05 × SSC at room temperature.

  Using an appropriate filter mounted on the apparatus, the DNA microarray in the gene testing chamber 106 was observed with fluorescence, and images were taken.

  The arrangement of the DNA microarray is shown in Table 1. In Table 1, a 30 base synthetic oligo DNA (custom product) that does not match the human gene sequence at each concentration is used as a positive control, and X, Y, 13, 18, 21 manufactured by GSP Laboratory Co., Ltd. are used as test probes. A DNA microarray probe based on the FISH probe near the centromere of each chromosome (concentration: 10 μM: custom-made product). Trisomy 13 is a so-called Patou syndrome, Trisomy 18 is a so-called Edward syndrome, and Trisomy 21 is a test for chromosomal abnormalities that cause so-called Down syndrome.

  Judgment of congenital anomalies due to gene anomalies was made from the fluorescence locations and the fluorescence values of the captured images. As is apparent from Tables 1 and 2 showing the normal cell observation results and the 21 trisomy observation results and FIG. 14, when the number of chromosomes is abnormal, a difference appears in the detected fluorescence value. In comparison, it was found that the fluorescence value was observed about 1.5 times brighter.

  Thus, in the trisomy 21 cell, the fluorescence value of the chromosome 21 probe was about 1.5 times higher than that of the normal cell, and it was possible to determine the abnormal number of chromosomes. In this way, the determination of the number of chromosomes is quantitatively compared with the magnitude of the fluorescence value, and the determination of genetic abnormality is based on the hybridization results with the probe. It was qualitatively shown to be used for diagnosis.

  The method for examining target cells in blood using the present invention is a non-invasive and safe method for examining pre-natal diagnosis of specific target cells such as fetal cells using maternal blood. Useful for prenatal diagnosis of the fetus. In addition, it is useful for diagnosis and treatment by examining abnormal target cells present in the blood of infants and adults. The target cell search apparatus using the present invention and the biochip attached thereto are used to accurately and quickly perform the method for examining target cells in blood.

1 is a target cell search device, 2 is an imaging unit, 3 is a processing unit, 4 is a display, 5 is a stocker, 6 is a transport mechanism, 8 is a flat plate, 9 is a target dictionary creation means, 10 is a non-target dictionary creation means, 11 is an XYZ stage, 12 is a laser light source, 13 is a focusing camera, 14 is an imaging light source, 15 is an imaging camera, 16 is a magnifying lens, 21 is an image acquisition means, 22 is a focus control means, and 23 is an imaging position control means. , 24 is a transport control unit, 31 is a target dictionary storage unit, 32 is a non-target dictionary storage unit, 41 is a search storage unit, 42 is a binarization processing unit, 43 is a contraction / expansion processing unit, and 44 is a primary candidate search Means 45, secondary candidate search means 46, tertiary candidate search means 47, target cell specifying means 47, target dictionary learning means 51, non-target dictionary learning means 52, observation image 55, observation position 56, 61 is a sample container, 2 is a control valve, 63 is a pressure pump, 64 is a three-way opening / closing control valve, 65 is a collection container, 66 is a disposal container, 67 is one output pipe, 68 is the other output pipe, and 70 and 75 are biochips. , 71 and 76, a flow path, 72 a target cell passage filter, 73 an optical observation window, 77 a discard path, 78 a recovery branch path, 79 a three-way open / close control valve, 80 a chemical solution injection path, and 82 a gap 85 is a control electrode, 86 is a base plate material, 87 is a cover plate material, 91 is a target cell, 92 is a non-target cell, 101 is a liquid flow path, 102 is an open / close control valve, 103 is a FISH chamber, and 104 is cell immobilization filter, 105 PCR chamber, the DNA microarray chamber 106, L1 · L2 width, Le laser beam for focus control, Le' reflected light, _{L T} is an optical indicator expressing light, theta is the angle, S1 is the target dictionary Product Step S11 is a non-target dictionary creation step, S11 is a sample acquisition step, S12 is a sample labeling step, S13 is an optical label expression step, S14 is a focus control step, S15 is a color image acquisition step, and S16 is a binarization processing step. , S17 is a contraction / expansion step, S18 is a primary candidate search step, S19 is a secondary candidate search step, S20 is a tertiary candidate search step, S21 is a target cell specifying step, S22 is a step, S23 is a target cell recovery step, S24 Is a target cell inspection step, and S22, S31 to S35 and S41 to S50 are steps in the flowchart.

Claims (5)

A biochip having a flow path into which a blood sample of blood collected for testing and blood concentrated from the blood is injected,
The flow path includes a target cell passage filter having a gap through which optically labeled target cells to be examined in the blood sample can pass and cells larger than the target cells cannot pass, and the target cell passage An optical observation window disposed downstream of the filter and optically observing the target cell from the outside of the biochip, and recovery of the target cell branched from the flow path downstream of the optical observation window A biochip comprising: a branch path; and a valve arranged at a branch portion of the recovery branch path to guide the target cell to the recovery branch path.   The biochip according to claim 1, wherein a plurality of the target cell passage filters are arranged in the flow path.   3. The valve according to claim 1, wherein the valve is a thermally controllable thermal control valve, an electrically controllable electric control valve, or a magnetically controllable magnetic control valve. The biochip described.   The flow path in the optical observation window is formed with a flow path width that allows the target cells to pass in one row along the flow path without overlapping. The biochip according to crab.   5. A chemical solution injection path for injecting an optical labeling chemical solution for optically labeling the target cell from the outside into the flow channel upstream of the optical observation window portion. The biochip according to any one of the above.