JP6951463B2 - Methods for Predicting Survival Outcomes in Subjects with Hepatocellular Carcinoma (HCC) or Cholangiocarcinoma (CCA) - Google Patents

Description

The present disclosure relates to a method for generating a predicted survival outcome for a subject suffering from liver cancer. More specifically, the predicted survival outcome is shown in a cancer tissue sample obtained from a subject, two genes or proteins (specifically, polo-like kinase 1; PLK1) and epithelial cell transformation. It is generated based on the expression profile of 2 (epithelial cell kinase 2; ECT2)). The present disclosure also provides a device configured to generate a predicted survival outcome by the disclosed method.

Primary liver cancer has two major histologically distinct subtypes: hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (CCA), the diagnosis and treatment of which is their baseline clinical. It is uniquely determined based on its characteristics. Extensive tumor heterogeneity in HCC or CCA is a complex multifactorial etiology that includes environmental factors such as hepatitis B virus (HBV), hepatitis C virus (HCV), parasite infections, and chemical carcinogens. Due to the existence of. Other risk factors include gender and racial / ethnic differences (liver cancer mainly affects men and is prevalent in the Asian population) (http://globocan.iarc.fr/), plus smoking. , Excessive drinking and unhealthy lifestyle such as ^{dietary factor 3.} HBV and HCV are the major pathogens of HCC and are responsible for up to 90% of liver cancers worldwide. On the other hand, CCA is rare except in Southeast Asia such as northeastern Thailand. For example, in northeastern Thailand, infection by liver fluke (O. viverlini) is endemic, and about 60% of liver cancers are CCA ⁴ Is. These global differences may be due to the presence of different pathogenic factors between different ethnic groups. According to one hypothesis, different causative factors can induce different molecular mechanisms to initiate malignant transformation independently, resulting in genomic heterogeneity between tumors. As with many other solid tumors, the biological and genetic heterogeneity in each type of HCC and CCA described herein provides high resistance to treatment for these cancers worldwide. It is the second most lethal malignant tumor in Japan ^1,2 . Therefore, identify the underlying genes or proteins that are expressed in the driving of cancer development, better stratify patients based on the identified cancer driver, and provide precision medicine treatment. Efforts have been made to make it easier.

See, eg, Chen et al. ^45, in promoting early HCC recurrence was found that the expression of ECT2 are closely related to the activation of the Rho / ERK signaling axis (signaling axis). Also, Sun et al. ^46, another gene that PLK1 is expressed significantly higher in HCC samples, be a independent prognostic factors for HCC, it has demonstrated to be applicable to diagnosis and treatment. Takai et al. ⁴⁷ drew similar conclusions in a paper supporting the findings of Sun et al. ^46. In addition, researchers have devised several diagnostic or prognostic measures that take advantage of the proven association between liver cancer and the expression of PLK1 or ECT2. For example, U.S. Patent Application Publication No. 2011/03119280 discloses a DNA biochip for HCC diagnosis by scanning multiple genetic markers, including PLK1. Most of the above approaches focus on cancer prognosis, which correlates with individual expression of the candidate gene or candidate protein, rather than cancer prognosis, which correlates with detailed information generated based on the correlation between candidate genes or candidate proteins. However, the derivation of diagnostic results based on the interrelationships of multiple expressed candidate genes has been proposed in several publications on other cancer types. Soma et al., US Patent Publication No. 2012/0197540, provides a method for predicting the risk of death in breast cancer patients by measuring the expression levels of multiple genes, including both PLK1 and ECT2, in tissue samples of breast cancer patients. do. The method then uses each of the measured gene expression levels to calculate a prognostic score indicating the subject's risk of death by a predetermined algorithm. However, similar methods or platforms that utilize the interrelationships of co-expressed genes or proteins for the prognosis or diagnosis of liver cancer have not been disclosed. Prognostic outcomes obtained from such methods or platforms may be applicable to improve treatment outcomes or medical resource allocation.

The present disclosure relates to a tissue sample obtained from a subject or a method for measuring or quantifying the expression of PLK1 and ECT2 in a tissue sample. In particular, the expression described herein can refer to at least one of transcriptome expression or protein expression in a sample. Preferably, the tissue sample is HCC or CCA affected.

One of the purposes of the present disclosure is to use measured or quantified expression of PLK1 and ECT2 in tissue samples to generate prognostic or diagnostic results for subjects suffering from HCC or CCA. To provide. Prognostic or diagnostic outcomes are generally associated with a subject's predicted survival outcome or risk of death. Prognostic or diagnostic results may be referenced by a physician to perform precision medicine.

A further object of the present disclosure is preferably configured to fully or semi-automatically quantify the expression of PLK1 and ECT2 and then generate prognostic or diagnostic results regarding the medical condition of a subject suffering from liver cancer. To provide the device.

At least one of the aforementioned objectives can be achieved in whole or in part by the present invention. One of the embodiments of the present invention includes a method for generating a predicted survival outcome of a subject suffering from liver cancer, wherein the method is one of hepatocellular carcinoma or cholangiocarcinoma collected from the subject. The step of preparing a tissue sample containing, or a tissue sample which is one of hepatocellular carcinoma or cholangiocarcinoma, and polo-like kinase 1 (PLK1) and epithelial cell transformation 2 (ECT2) in the prepared sample. To quantify transcriptome expression and / or protein expression in, at least one step of performing a nucleic acid-based assay and / or a proteomik-based assay and expression of ECT2 in a specific region of the prepared sample. In the step of deriving the ratio of the expression of PLK1 to the sample and the step of generating the predicted survival outcome of the subject based on the derived ratio, when the derived ratio is less than a predetermined value, the above A step of setting the predicted survival outcome of the subject as the first predicted survival outcome and setting the predicted survival outcome of the subject as the second predicted survival outcome when the derived ratio is equal to or higher than the predetermined value is included. .. More specifically, the first predicted survival outcome corresponds to an overall survival rate of greater than 40% to 60% over a given period of survival analysis, and the second predicted survival outcome is that of the survival analysis. It corresponds to an overall survival rate of 40% to 60% or less in the above predetermined period. Typically or preferably, the predetermined period is, but is not limited to, 12 to 36 months.

In some embodiments, the nucleic acid-based assay is any one or combination of microarray-based techniques and polymerase chain reaction (including digital PCR) -based techniques.

In some embodiments, the proteomics-based assay comprises a set of tissue microarrays and immunohistochemical staining of human tissue, or a set of tissue microarrays and immunohistochemical staining of human tissue.

In another aspect of the present disclosure, a device for generating a predicted survival outcome for a subject suffering from liver cancer is disclosed. A first device capable of identifying, detecting, or determining transcriptome expression and / or protein expression of polo-like kinase 1 (PLK1) and epithelial cell transformation 2 (ECT2) in the tissue sample of the subject. The module is configured to derive the ratio of PLK1 expression to ECT2 expression in a particular region on the prepared sample and to generate the predicted survival outcome of the subject based on the derived ratio. The second module is derived by setting the predicted survival outcome of the subject as the first predicted survival outcome when the derived ratio is less than a predetermined value. When the ratio is equal to or greater than the predetermined value, the predicted survival outcome of the subject is configured to be the second predicted survival outcome. More specifically, the first predicted survival outcome corresponds to an overall survival rate of greater than 40% over a predetermined period of survival analysis, and the second predicted survival outcome corresponds to the predetermined survival analysis. Corresponds to an overall survival rate of 40% or less over the period. Typically, or preferably, the nucleic acid-based assay refers to any one or combination of techniques based on microarrays (such as microarrays) and polymerase chain reactions (including digital PCR). can. Also, the proteomics-based assay can be tissue microarray and immunohistochemical staining of human tissue.

For a plurality of embodiments, the disclosed apparatus is a signal formed to continuously or periodically emit a signal separable by the first module above with the expressed transcriptome or protein of PLK1 and ECT2. It further comprises a test module capable of performing at least one nucleic acid-based assay or proteomics-based assay for binding to a transduction moiety.

In some embodiments, the subject is preferably of Asian descent in order to obtain more accurate results. In addition, the tissue sample of the subject contains one carcinoma of hepatocellular carcinoma or cholangiocarcinoma, or is one carcinoma of hepatocellular carcinoma or cholangiocarcinoma, and is among hepatocellular carcinoma or cholangiocarcinoma. It is one of.

The results of consensus clustering using a hierarchical clustering algorithm are shown. Left panel: Empirical Cumulative Distribution (CDF) plot corresponding to a consensus matrix in the range of clusters (K) = 2-8. Center panel: Corresponding change in area under the CDF. Right panel: Consensus matrix corresponding to the number K of clusters in the range 2-5. The results for the CCA sample are shown on the top and the results for the HCC sample are shown on the bottom. (A) is a heatmap of CCA subtypes based on consensus clustering and hierarchical clustering of the most variable genes. The x-axis represents the CCA subtype consensus cluster. CCA samples are represented in columns and are categorized into four major clusters by dendrogram. Genes are represented by rows. Expression of 1189 genes (identified by t-test as different expression levels between C1 and C2 (FDR <0.05, multiple change> 2)) is _{shown in log 2} from -3 to 3. .. (B) is a heat map of HCC subtypes (FDR <0.05, multiple change> 2) based on 1020 genes having different expression levels, and is shown in the same manner as in (a). The x-axis represents the HCC subtype consensus cluster. (C) is a subclass mapping of CCA subtypes and HCC subtypes. Significant relationships between clusters are represented by Bonferroni-adjusted p-values. A Venn diagram showing the number of genes defined by Student's T-test (FDR <0.05, multiple change> 2) with different expression levels between C1 and C2 samples of CCA and HCC is shown. The left panel shows the number of genes upregulated in C1 in CCA (n = 578) and HCC (n = 656), and the number of overlapping genes (n = 218) (the probability of overlapping genes is statistically significant. Some (p = 3.20x10-97; hypergeometric test)). The right panel shows similar results for genes that are upregulated in C2. A Kaplan-Meier survival analysis of the CCA subtype (upper panel) or HCC subtype (lower panel) is shown. It shows the significant pathways of the C1 or C2 subtypes of HCC or CCA identified by GSEA analysis. It is represented by a log 10p value of 2 to 0 (p value of 0.01 to 1). (A) Shows a comparison between the Thai CCA subtype and the published signature (shown on the y-axis). Each column represents a tumor sample. Positives for each signature are represented by gray bars, negatives are represented by black bars, and white when FDR> 0.05. (B) Shows a comparison between the Thai HCC subtype and the published signature (shown on the y-axis). (A) Frequency of HCC chromosomal abnormalities (upper panel), or frequency> 30% of HCC chromosomal abnormalities (lower panel), and (b) similar results for CCA. (A) Shows the frequency of chromosomal abnormalities in the C1 subtype of CCA (upper panel) or the C2 subtype of CCA (lower panel). The copy count gain is shown in light gray and the copy count loss is shown in black. (B) Shows the frequency of chromosomal abnormalities in the C1 subtype of HCC (upper panel) or the C2 subtype of HCC (lower panel). The frequency of samples with copy number variation (CNV), as well as the relationship between genes with high concordance and tumor subtypes is shown. All Kaplan-Meier plots of 378 HCC cases. Representative images of CCA and HCC cases at 200x magnification based on immunohistochemical staining for PLK1 (left panel) or ECT2 (right panel) are shown. (A) Clarify the correlation between PLK1 array expression and TMA score (left panel) or ECT2 array expression and TMA score (right panel) in CCA cases (upper panel) or HCC cases (lower panel). , (B) To clarify the correlation between PLK1 array expression and CT2 array expression, or the correlation between PLK1 TMA score and ECT2 TMA score in CCA or HCC cases. The results of Kaplan-Meier survival analysis of CCA cases (left panel) or HCC cases (right panel) based on the median cutoff of PLK or ECT2 are shown. (A) Kaplan-Meier survival analysis of all CCA and HCC cases based on a combination of protein expression ratiometrics (ECT2 / PLK1). For low cutoff, the low cutoff is defined by the ECT2 / PLK1 ratio ≤ 1 for low, and for high, the high cutoff is defined by the ECT2 / PLK1 ratio> 1. (B) Kaplan-Meier survival analysis of HCC cases is shown in the same manner as in (a). (A) Thai HCC subtype vs. Chinese HCC subtype, (b) Asian-American (AsA) subtype, (c) Thai CCA subtype vs. Japanese (Japanase) CCA subtype (D) Subclass mapping and associated Kaplan-Meier survival analysis for Thai HCC subtypes vs. European-American (EA) HCC subtypes. (E) Shows a subclass mapping of Thai CCA subtypes vs. Caucasian CCA subtypes. Significant relationships between clusters are represented by Bonferroni-adjusted p-values from 0 to 1 and subtypes in the cohort. Inconsistent subtypes in the Thai cohort are indicated by an X. Significant associations between clusters are shown in light gray with p <0.05. Box plots of BMI in C1 and C2 subtypes of Thai HCC cases, Asian American HCC cases, and Thai CCA cases are shown. Median and standard deviation are also shown. (A) A heat map showing the correlation between metabolite abundance and gene expression in the CCA sample, and (b) a heat map showing the global correlation between the metabolite abundance and gene expression in the HCC sample are shown. Based on Pearson R-values of -1 to 1, positive correlations are indicated by light gray bars and negative correlations are indicated by black bars. (A) Hierarchical clustering of 81 metabolites distinguishes between HCC-C1 cases (light gray bars) and C2 cases (black bars), and (b) Hierarchical clustering of 77 metabolites. , CCA-C1 cases (light gray bar) and C2 cases (black bar) are shown to be distinguished. Samples are represented by columns, metabolites are represented by rows, and metabolite abundance is represented by log2. (A) Highly matching metabolite / gene network Analysis Pathway Analysis and (b) Highly matching metabolite / gene network Analysis Pathway Analysis are shown. Upregulated metabolites in the C1 subtype are shown in light gray and upregulated metabolites in the C2 subtype are shown in dark gray. (A) A box plot of the abundance of three representative bile acid-related metabolites in HCC samples of C1 (n = 15) and C2 (n = 14), along with Student's T-test p-values, and ( b) A box plot of the abundance of three representative bile acid-related metabolites in CCA samples of C1 (n = 33) and C2 (n = 18), along with Student's T-test p-values. (A) CIBERSORT analysis of C1 subtype of HCC vs. C2 subtype of HCC, and (b) CIBERSORT analysis of C1 subtype of CCA vs. C2 subtype of CCA are shown. High to low associations between cell types are scaled from 1 to -1. The size of the circle indicates the significance of the association, and the larger the circle, the higher the significance. (A) A box plot of the abundance of three types of leukocytes in HCC samples of C1 and C2, along with standard deviation and Student's T-test p-values, and (b) C1 with standard deviation and Student's T-test p-values. And a box plot of the abundance of three types of leukocytes in the CCA sample of C2 is shown.

Hereinafter, the present invention will be described based on a representative embodiment or a preferred embodiment of the present invention with reference to the accompanying description and drawings. It should be understood that the description and drawings corresponding to the embodiments are for the purpose of clarifying and assisting the understanding of the present invention. One of ordinary skill in the art can devise various modifications without departing from the scope of the invention as defined by the appended claims.

As used herein, the term "gene" can refer to a DNA sequence of functional importance. The gene can be a natural nucleic acid sequence or a recombinant nucleic acid sequence derived from a natural source or synthetic construct. The term "gene" can also refer to, but is not limited to, for example, cDNA and / or mRNA encoded by a genomic DNA sequence, or cDNA and / or mRNA derived directly or indirectly from a genomic DNA sequence.

As used herein, the term "transcriptome" refers to a collection of RNA transcripts (whether encoded or not) transcribed in a particular tissue, preferably produced in the tissue. Includes all or substantially all of the RNA transcripts. These transcripts include messenger RNA (mRNA), selectively spliced mRNA, ribosome RNA (rRNA), transfer RNA (tRNA), and many other transcripts that are not translated into protein (eg, for example. Also included are (i) antisense molecules such as nuclear small RNA (snRNA), (ii) small interfering RNA (siRNA) and microRNA, and (iii) other RNA transcripts of unknown function. The transcriptome can also refer to a protein translated from an RNA transcript as an extension of gene transcription.

One aspect of the disclosure relates to a method for generating a predicted survival outcome for a subject suffering from liver cancer. In essence, the method involves preparing a tissue sample of one of the hepatocellular carcinomas or bile duct cancers taken from the subject and the polo-like kinase 1 (PLK1) in the prepared sample. And a step of performing at least one kinase-based assay and / or proteomics-based assay to quantify transcriptome and / or protein expression of epithelial cell transformation 2 (ECT2), and the prepared sample. A step of deriving the ratio of the expression of PLK1 to the expression of ECT2 in the above specific region, and a step of generating the predicted survival outcome of the subject based on the derived ratio, the derived ratio. Is less than a predetermined value, the predicted survival outcome of the subject is the first predicted survival outcome, and when the derived ratio is greater than or equal to the predetermined value, the predicted survival outcome of the subject is the second. Includes steps to determine the predicted survival outcome of. Preferably, the first predicted survival outcome corresponds to an overall survival rate greater than 40% to 60% over a given time period of the survival analysis, and the second predicted survival outcome corresponds to the given predetermined survival analysis. Corresponds to an overall survival rate of 40% to 60% or less over a period of time. The inventors of the present disclosure have a relative expression of the tumor markers PLK1 and ECT2 in affected tissue having or suffering from hepatocellular carcinoma or cholangiocarcinoma, and adversely affect the survival outcome or subject of the test subject. We found that it correlates with acute liver cancer. The resulting predicted survival outcome can be an effective tool for facilitating the planning of strategic treatments and the availability of medical resources for the treatments according to the medical needs of the study subjects.

According to some embodiments, tissue samples suffering from hepatocellular carcinoma or cholangiocarcinoma in the present disclosure can be obtained by any surgical or medical procedure known in the art. Tissue samples are preferably taken by liver biopsy. Liver biopsy may be, but is not limited to, percutaneous, transjungular, or laparoscopic biopsy. In some embodiments, the type of liver biopsy performed depends on the assay performed after the biopsy to quantify transcriptome expression and / or protein expression associated with PLK1 and / or ECT2. For example, tissue samples obtained by laparoscopic examination are more suitable for proteomics-based assays such as tissue microarrays that require sufficient area in the sample for immunochemical staining and signal detection. On the other hand, nucleic acid-based assays in which the mRNA or transcriptome corresponding to the expression of PLK1 and / or ECT2 is extracted from the sample, proportionally amplified and finally quantified are obtained from a percutaneous biopsy. Tissue samples of the subject are preferred.

In some embodiments, the nucleic acid-based assay is either one or a combination of microarray-based techniques and polymerase chain reaction-based techniques (including digital PCR). Nucleic acid-based assays generally extract all of the mRNA or transcriptome from a tissue sample of a given size and then selectively quantify the mRNA or transcriptome associated with the expression of PLK1 and / or ECT2 (PLK1). And / or with or without selective amplification of ECT2 mRNA or transcriptome). Preferably, the Transcriptome of PLK1 and / or ECT2 hybridizes on a substantially complementary probe immobilized on the platform. The hybridization process is configured to emit a signal, preferably an optical signal, that can be detected by one or more sensors located near or within the platform. The probe is tagged with a signaling moiety designed to generate a light signal during hybridization and double-stranded nucleic acid formation. The signal emitted is preferably proportional to the number of hybridization reactions that occur on the platform, specifically quantifying the expression of PLK1 and / or ECT2 by reference to pre-established standards. Is proportional to the number of hybridization reactions allowed.

Similarly, for multiple embodiments, the proteomics-based assay is tissue microarray and immunohistochemical staining of human tissue, whereby the expressed protein or peptide of PLK1 and / or ECT2 signals one or more. Expression of PLK1 and / or ECT2 is quantified by selectively binding to a moiety and then detecting the signal emanating from the attached signaling moiety. For example, in a tissue microarray (TMA), the obtained tissue sample is used to prepare a core (0.5 to 2.0 mm) of a tissue sample fixed with formalin and embedded in paraffin. The prepared tissue sample is immunohistochemically treated with a TMA fragment of 1 to 10 μm in the form of a slide. In addition, slides containing the cut segments of tissue samples are deparaffinized in xylene and rehydrated in alcohol with a concentration gradient. Then, using an acidic buffer, antigen recovery is performed under pressure for 5 to 60 minutes. Anti-PLK1 (mouse monoclonal) is diluted 1: 1000 with a 2% skim milk solution and then contacted with the treated tissue sample at room temperature for 1-4 hours. Anti-ECT2 (mouse monoclonal) is diluted 1: 500 with a 2% skim milk solution and then contacted with the treated tissue sample. In the disclosed method, sensors or systems such as those commercially available by contacting the antibody with a targeted tumor marker or protein (eg, Envision + system (Dako) used in conjunction with 3,3'-diaminobenzidine (DAB)). The result is that an antigen-antibody complex that can be separated using is obtained. Preferably, the slides are counterstained with hematoxylin, dehydrated, clarified and covered with a cover glass. Detection of tagged or stained PLK1 and / or ECT2 can be performed at a magnification of 200-fold or higher for quantification of expression of PLK1 and / or ECT2 in a particular region. It should be noted that the quantification may be performed manually by trained personnel or by using a signal sensor or image analyzer connected to the computing device. Preferably, the treated tissue sample can be scored for various parameters or characteristics of the image (eg, tumor cell percentage, and indicator signal or staining intensity). Scores of the properties of the different images generated can be further processed (eg, multiply by each other or one or more) to obtain more useful test results for subjects with hepatocellular carcinoma or cholangiocarcinoma. Input to the algorithm, etc.). In some embodiments, the expression level of PLK1 or ECT2 in the disclosed method corresponds to the product obtained by multiplying the score obtained from the processed sample or the score obtained from the image of the processed sample. Or as defined in the product. The expression level can be quantified by multiplying the score of the percentage of tumor cells expressing PLK1 or ECT2 by the score of the signal or staining intensity of the expressed PLK1 or ECT2, thereby expressing the expression of each tumor marker. You can get a calculated score that effectively represents your level. In such embodiments, the expression level of the tumor marker under consideration corresponds to the calculated score. The scale of the score for each characteristic of the image can range from 1 to 15, more preferably 1 to 10, and most preferably 1 to 4. Multiple regions of the treated sample are scored independently. In some embodiments, in the disclosed method, this independent score is used to calculate the average in order to better represent the expression of PLK1 and / or ECT2. This may result in more accurate results.

In the disclosed method, after detecting and quantifying the expressed transcriptome and / or protein via a nucleic acid-based or protein-based assay, the PLK1 for expression of ECT2 in a particular region of the prepared sample. Derivation of expression ratio. For example, when the expression level of PLK1 is quantified as 12 and the calculated expression level of ECT2 is 8, the derived ratio is 1.5. In some embodiments, the ratio is calculated or calculated by swapping the numerator and denominator of the aforementioned ratios, such that instead the numerator is the expression level of ECT2 and the expression level of PLK1 is the denominator. Can be done. The aforementioned region preferably has dimensions of 0.1 to 1.0 mm in length, 0.1 to 1.0 mm in length, and 1 to 4 um in thickness, but is not limited thereto. In the disclosed method, to simplify the process of deriving the predicted survival outcome of the subject, if the derived ratio is less than a predetermined value, the predicted survival outcome of the subject is taken as the first predicted survival outcome and derived. When the ratio is equal to or higher than the above-mentioned predetermined value, the predicted survival outcome of the subject is defined as the second predicted survival outcome. In various embodiments, the predetermined value can be any number in the range 0-16 in relation to the scale 0-4 of the score used, but is not limited thereto. The predetermined number may range from 0.001 to 1 depending on the scale of the score applied in the relevant embodiment. Also, in order to obtain further predictions or observations, or to obtain more definitive predictions or observations, the derived ratios can be analyzed, or the derived ratios can be used for hepatocellular carcinoma or cholangiocarcinoma. It can be further mathematically associated with the expression level of other genes or proteins associated with.

As mentioned above, according to some embodiments, the first predicted survival outcome corresponds to an overall survival rate of greater than 40% to 60% over a given period of survival analysis and a second predicted survival outcome. Corresponds to an overall survival rate of 40% to 60% or less during the predetermined period of the survival analysis. In addition, the survival rates generated can be considered as indicators that provide empirical information about the likelihood of a test subject's response to common treatments for hepatocellular carcinoma or cholangiocarcinoma. If a subject is associated with a second predicted survival outcome or a relatively low survival rate, or is assigned a second predicted survival outcome or a relatively low survival rate, the subject will receive alternative treatment instead of conventional treatment. You may have to take it. The inventors of the present disclosure also find that for subjects with or of Asian descent, the disclosed methods provide more reliable or more robust results. Noticed.

In some embodiments, the predicted survival outcome indicates the survival rate of the subject over a predetermined period of time (preferably 6-48 months, more preferably 12-36 months), or only that survival rate.

According to another aspect of the present disclosure, a device for generating a predicted survival outcome for a subject suffering from liver cancer is disclosed. In particular, the disclosed devices include a first module capable of identifying or detecting transcriptome expression and / or protein expression of PLK1 and ECT2 in the subject's tissue sample, and specific on the prepared sample. It comprises a second module configured to derive the ratio of expression of PLK1 to expression of ECT2 in the region and to generate the predicted survival outcome of the subject based on the derived ratio. When the derived ratio is less than the predetermined value, the predicted survival outcome of the subject is set as the first predicted survival outcome, and when the derived ratio is greater than or equal to the predetermined value, the above. The subject's predicted survival outcome is configured to be the second predicted survival outcome. Preferably, the first predicted survival outcome corresponds to an overall survival rate greater than 40% to 60% over a given time period of the survival analysis, and the second predicted survival outcome corresponds to the given predetermined survival analysis. Corresponds to an overall survival rate of 40% to 60% or less over a period of time.

The first module described in the present disclosure is genetic, proteomics, immunochemistry and / or tissue to selectively enable expression of PLK1 and / or expression of ECT2 by human or machine reading. It can be any known platform that can operate using the principles of chemistry. Thus, in some embodiments, the disclosed device is formed to continuously or periodically emit a signal separable by the first module above with the expressed transcriptome or protein of PLK1 and ECT2. Further, a test module may be provided capable of performing at least one nucleic acid-based assay or proteomics-based assay for binding to a signal transduction moiety. In particular, the first module is configured to attach or bind a signaling moiety capable of emitting a detectable signal to the transcriptome and / or protein expressed for PLK1 and / or ECT2. , Emphasize the presence of the expressed transcriptome or protein associated with the signaling moiety. The first module of the present disclosure is a plurality of work stations arranged to carry out a continuous process for the preparation and processing of tissue samples in order to finally lead to the identification of the expression of PLK1 and / or the expression of ECT2. It may be composed of. These plurality of work stations may be interconnected or may operate individually, depending on the choice of disclosed embodiments. Preferably, in a further embodiment, the first module was designed to generate a digital image of the illuminated sample with the illumination device of the first module to emphasize the expression of PLK1 and / or ECT2. With an imaging device, the first module uploads captured images to the second module for analysis and derivation of predicted survival outcomes. Most preferably, the first module stores in a digital image generated in a connected database, which allows the stored image to be recalled and further analyzed. For example, the first module can be a gene expression microarray system or tissue microarray system for the detection and quantification of expression of PLK1 and / or ECT2 in tissue samples suffering from hepatocellular carcinoma or cholangiocarcinoma.

According to some embodiments, the second module derives the ratio of PLK1 expression to ECT2 expression, or the ratio of ECT2 expression to PLK1 expression, in a particular region on the prepared sample. It is designed to generate a predicted survival outcome for the subject, corresponding to the derived ratio. The second module can be a computing means in which a set of commands for analyzing captured images is installed in the form of a computing program. The characteristics of the captured image may be processed depending on the analysis performed. The second module preferably analyzed the treated tissue sample and more preferably the image of the treated tissue sample and showed various parameters or characteristics of the image (eg, the percentage of tumor cells, and shown). Provides a score for (indicative signal or staining intensity). The second module may further process the scores generated for various properties of the image to obtain more useful test results for subjects with hepatocellular carcinoma or bile duct cancer (eg, scores to each other). Multiply or enter the score into one or more algorithms). In a further embodiment, the second module determines the expression level of PLK1 or ECT2 of the calculated score obtained from different scores associated with each of the properties of the imaged image of the treated tissue sample. Define as a number. For example, the second module quantifies the expression level by multiplying the score of the percentage of tumor cells expressing the detected PLK1 or ECT2 by the score of the signal or staining intensity of the expressed PLK1 or ECT2. , This yields a calculated score that substantially represents the expression level of each target tumor marker. In such an embodiment, the disclosed device considers that the expression level of the tumor marker of interest corresponds to the calculated score. The scale of the score for each image characteristic can range from 1 to 15, more preferably 1 to 10, and most preferably 1 to 4. In some embodiments, in order to better represent the expression of PLK1 and / or ECT2 in the captured image, the second module generates scores for each of the multiple regions of the treated sample. And calculate the average of these individual scores.

In addition, the second module calculates or derives the ratio of PLK1 expression to ECT2 expression, or the ratio of ECT2 expression to PLK1 expression in a particular region of the prepared sample. The region preferably has dimensions of 0.1 to 1.0 mm in length, 0.1 to 1.0 mm in length, and 1 to 4 um in thickness. To simplify the process of deriving a subject's predicted survival outcome, the disclosed second module sets the subject's predicted survival outcome to the first predicted survival outcome if the derived ratio is less than a predetermined value. Is configured to be the first predicted survival outcome, and if the derived ratio is greater than or equal to a predetermined value, the subject's predicted survival outcome is defined as the second predicted survival outcome. The predetermined value can be any value in the range of 0 to 16, depending on the scale of the score adopted, but is not limited thereto. As mentioned above, the first predicted survival outcome corresponds to an overall survival rate greater than 40% to 60% over a given time period of the survival analysis, and the second predicted survival outcome corresponds to the given predetermined survival analysis. Corresponds to an overall survival rate of 40% to 60% or less over a period of time. Preferably, the predetermined period is preferably 6 to 48 months, more preferably 12 to 36 months.

In some embodiments, the nucleic acid-based assay is one of a microarray-based technique for DNA and / or RNA polynucleotide detection, as well as a polymerase chain reaction-based technique (including digital PCR). Or a combination.

In another embodiment, the proteomics-based assay is tissue microarray tissue microarray and / or immunohistochemical staining of human tissue.

The following examples are intended to further illustrate the invention and are not intended to limit the invention to the specific embodiments described in the examples.

Example 1
Cohort and Clinical Samples A set of 398 surgical pairs of tumor and non-tumor samples from 199 consecutive patients in the TIGER-LC cohort (130 CCA patients and 69 HCC patients). Used in this study. Tumor diagnosis was confirmed by pathological evaluation. Clinical, demographic, socio-economic and morbidity data were extracted from comprehensive questionnaires and chart records. A list of clinical variables evaluated in this study is given in Supplementary Table S1. The characteristics of 156 Asian HCC patients and 163 Caucasian HCC patients (TCGA Research Network: http: // cancerenome.nih.gov /) obtained from TCGA were also used in this study. Characteristics of 104 Caucasian CCA patients and 182 Japanese patients in an independent cohort have been recently reported ^14,25 . An HCC cohort of 247 Chinese patients from LCI was previously reported ³⁵ . Informed consent was obtained from all patients included in this study and was approved by the institutional review board at each institution.

RNA Isolation and Transcript Mix Total RNA was extracted from frozen tissue using TRIzol (Invitrogen) according to the manufacturer's protocol. Only RNA samples confirmed to have good RNA quality using the Agilent 2100 Bioanalyzer (Agilent Technologies) were included in the array test. Transcripts in paired tumor tissue and non-tumor tissue samples were measured using Affymetrix Human Transcriptome Array 2.0. Raw gene expression data were normalized using the Robust Multi-array Average (RMA) method and the sketch quantile normalization method. Mean gene expression was calculated for genes with multiple probe sets. The microarray platform and data were submitted to NCBI's Gene Expression Omnibus (GEO) public database in accordance with the MIAME guidelines (GEO Series GSE76297). Gene expression data for three independent cohorts are available from the GEO Series (GSE14520 and GSE26566), TCGA, and the European Genome-Genome Archive (EGA) database (EGA0000100950). All expression data were log ₂ converted. 2615 common to both the HCC cohort and the CCA cohort of TIGER-LC using consensus clustering (cCruster; hierarchical clustering; Pearson distance; complete linkage; 1000 resampling iteration). The most variable gene set (gene filter: 1.5-fold change from median gene, Log intensity variation p-value> 0.01) was used to define subtypes in HCC or CCA. CClastering of the validation cohort was performed using genes common to the individual cohorts and the 2615 most variable genes described above. Hierarchical clustering (Partek Generic Suite 6.6; Pearson distance; complete linkage) was used to determine and visualize the relationships between the various subgroups defined by cCruster. Using the unsupervised subclass mapping method (SubMap), the default settings found in the SubMap module were used to identify common subgroups between independent cohorts (http://genepattern.broadinstation. /) ¹¹ .

Previous studies have shown that HCC and CCA have high heterogeneity between tumors, but much less heterogeneity within tumors, with respect to the heterogeneity determined by transcriptome profiling. ^5-7 . To define molecularly uniform tumor subgroups in Thai CCA and HCC patients, Affymetrix Human Transcriptome Array 2.0 was performed on 398 surgical samples obtained from 199 patients. Of these, 153 samples passed quality control tests and were used for transcriptome analysis. 8 As mentioned above, consensus clustering (cCruster) was used to define the subtypes of HCC and CCA. In cClaster, four major subgroups of CCA and three major subgroups of HCC were revealed based on the consensus distribution and the corresponding consensus matrix (Fig. 1). Relationships between the various subgroups defined by cCruster can be visualized by unsupervised hierarchical clustering (FIGS. 2A-B). This analysis revealed a unique subgroup within HCC or CCA cases with different gene expression patterns.

Several recent studies have described transcriptome similarities between CCA and HCC ^9,10 . The present invention uses a subclass mapping (Subclass Mapping, SM), and to clarify the relationship between the CCA and HCC molecular subgroups ^11. This analysis revealed that the CCA-C1 and HCC-C1 subtypes have similar gene expression matrices, while the CCA-C2 and HCC-C2 / C3 subtypes are similar (). FIG. 2C). However, this relationship was not observed in non-tumor tissues, suggesting that subtype-related genes are tumor-specific. Analysis of gene expression matching in the C1 and C2 subtypes revealed that the upregulated genes in gene cluster 1 (GC1) are similar between CCA-C1 and HCC-C1 (Figure). 2A-B and FIG. 3). Similar results were observed in the GC2 gene cluster. Interestingly, the cases defined as CCA-C1 and HCC-C1 had poor survival, while the cases of CCA-C2 and HCC-C2 had higher survival as shown in FIG. The subtype-specific gene signature Gene Set Enrichment Analysis (GSEA) is enriched for both the CCA-C1 and HCC-C1 subtypes for the mitotic checkpoint signaling pathway, as shown in FIG. It was revealed that there was. This suggests that this subtype has high chromosomal instability. In contrast, the CCA-C2 and HCC-C2 subtypes are enriched for cell-mediated immune-related pathways, suggesting that the inflammatory response is associated with the C2 subtype. These results indicate that despite clear histological differences between CCA and HCC, there are common subtypes with similar gene expression matrices and tumor ecology.

Several gene signatures have been associated with HCC / CCA prognostic subtypes, cancer stem cell characteristics and tumor metastases 5, ^{6, 12-15} . Therefore, the authors, the association between HCC subtype known signature of Thai, was examined using the closest template prediction algorithm ^16. Both the CCA-C1 and HCC-C1 subtypes are enriched for S1-2-related genetics (based on Hoshida's study) ^6,11,16 , for stem cell genes. It was found that ^12,34 (based on Lee's study ^{) and 13} (based on Yamashita's study) were enriched for the EpCAM gene. Similarly, the CCA-C1 subtype is enriched for the S1-2 gene, stem cell gene and EpCAM gene (FIG. 6). In addition, both CCA-C2 and HCC-C2 are enriched for cases negative for the aforementioned signature, while the C3 / 4 subtype includes mixed cases. In contrast, Class 2 Signature ¹⁴ and Proliferation Signature ¹⁵ did not adequately separate common subtypes.

A major feature of liver cancer is its association with various pathogenic factors, as well as its high heterogeneity in clinical manifestations and the underlying tumor ecology. As a result, many patients with liver cancer are refractory to treatment, with unfavorable outcomes. One of the essential requirements to improve outcomes in patients with liver cancer is the precise definition of uniform molecular subtypes, each showing a unique tumor ecology and potentially draggable driver genes. To provide a diagnostic toolkit that can be used to make rational treatment choices based on molecular subtypes. Therefore, developing a well-annotated biobank for cancer patients, such as the efforts of TCGA and ICGC, is an important resource for further developing precision medicine goals. Currently, TCGA and ICGC mainly include HCC samples from North America, Europe and Japan. Given that HCCs and CCAs are much more prevalent in the Asian population, this disclosure establishes a TIGER-LC consortium for conducting case-control studies in liver cancer, especially in Thailand, where CCA is endemic.

The development of liver cancer is a complex process that results from decades of accumulation of genetic and epigenetic changes in various cancer drivers. Tumor evolution is expected to vary significantly between different tumor types, as each tumor cell must adapt to microenvironmental stress from different pathogenic factors. As a result, liver cancer tends to show particularly genetic heterogeneity. While the whole-genome / exome sequencing approach is potent in identifying potential cancer drivers, mutations in candidate HCC or CCA-related drivers are highly heterogeneous. This is because each tumor has a large number of low frequency mutated genes in various combinations ^{25, 31, 32} . Furthermore, most of these genetic abnormalities characterized by all stations Zoom sequencing, has been thought to be either passenger mutation or histological mutations have no functional effect on the tumor to be diagnosed ³³ .. For genes with relatively high mutation frequencies characterized as candidate drivers, a vast amount of tumor heterogeneity between different tumor lesions with different combinations of these candidate drivers is ^{evident31, 32} . It is conceivable that a combination of different cancer drivers may emerge as a new convergent adaptive pathway that is unique to different subtypes and has been tethered for cancer cell survival. It should be noted that cancer drivers can be obtained not only by somatic mutations, but also by epigenetic mechanisms (eg, DNA methylation of oncogenes or tumor suppressor genes). This may explain that it is very difficult to find stable tumor subtypes with unique tumor ecology by whole exome sequencing that captures only some of the tumor properties. In contrast, transcriptome-based approaches have been difficult to define key driver genes in various tumor subtypes, but transcriptome profiling is a stable HCC that can reflect the unique tumor ecology. Succeeded in defining subtypes ^6,13,34-36 . An integrated omics approach that correlates genomics, epigenomics, transcriptomics, proteomics, and metabolomics data can be key to addressing tumor heterogeneity and cancer drivers.

Example 2
DNA Isolation and Somatic Copy Number Alterations (SCNA)
Total DNA was extracted from frozen tissue using a phenol / chloroform extraction protocol. Samples confirmed to have a sufficient amount of double-stranded DNA using the Quant-iT PicoGreen dsDNA assay kit (Thermo Fisher Scientific) were included for array studies. Affymetrix Genome-Wide Human SNP Array 6.0 was used to determine Somatic Copy Number Alterations (SCNA) between paired tumor tissue and non-tumor tissue samples. Raw SCNA data is available from GEO Series GSE76213. The SCNA of CCA and HCC was analyzed using the Partek Genomics Suite 7.5 with reference to the paired non-tumor tissue of each patient. Segmented regions in CCA and HCC were found using the genomic segmentation algorithm in Partek. Pearson correlation values were calculated with copy count segmentation regions and transcriptome data from CCA or HCC tissues to identify genes that are consistently regulated by SCNA. Genes located in the segmented region with a positive correlation value and p-value ≤ 0.005 were considered to define matching genes specific for the C1 and C2 subtypes. To define the copy number matching genes for the C1 and C2 subtypes in CCA and HCC, we used: (i) p-values from Student's t-test, and (ii) C1 and C2. Multiple changes in _{rod 2} conversion expression among the most variable genes used for class prediction between subtypes of CCA or HCC. To identify C1 subtype-specific copy-number matching genes in CCA and HCC, the inventors of the present disclosure have p-valued by Student's t-test when comparing C1 and C2 subtypes. Selected genes with a value of ≤0.005.

Next, the inventors of the present disclosure used Affymetrix Genome-Wide Human SNP Array 6.0 to perform Simatic Copy Number Alterations (SCNA) between Thai tumor samples and paired non-tumor tissues as controls. )It was determined. Consistent with previously published study ¹⁷ , a typical SCNA profile with recurrent gain at 1q, 6p, 8q and recurrent loss at 4q, 8p, 13q, 16, 17p is shown in FIG. 7A. As shown in, it was apparent in the HCC sample. It was found that the SCNA profiles differed considerably between HCC and CCA (Fig. 7B). However, when we analyzed the SCNA profile based on the C1 and C2 subtypes, they were significantly higher in both CCA-C1 and HCC-C1 compared to CCA-C2 and HCC-C2. Recurrent gains and losses were seen (FIGS. 8A-B). This result is consistent with transcriptome and GSEA analysis, suggesting that CCA-C1 and HCC-C1 contain mitotic checkpoint deficiencies, which may result in a higher degree of aneuploidy. ..

To determine potential subtype-related driver genes, the inventors of the present disclosure, based on ^{concepts 17,18} that driver genes should have highly consistent changes between SCNA and gene expression. First made a Pearson correlation between SCNA and the transcriptome. This analysis found that there was a significant positive correlation between a subset of genes (Suppl Fig S5A-B), ie 239 genes for CCA-C1 and 89 genes for HCC-C1 tumor samples. Revealed. Among them, 51 genes overlapped between CCA-C1 and HCC-C1, suggesting the existence of a common subtype-specific driver in HCC and CCA. Consistently, as shown in FIG. 9, the C1 subtype is associated with higher copy count and increased expression than the C2 subtype. Consistent with the above results, network analysis of these 51 genes revealed enrichment of the mitotic checkpoint signaling pathway associated with PLK1 signaling. Furthermore, ECT2 was the highest ranking of genes with different expression levels between the C1 and C2 subtypes. Similar results were observed using the TCGA HCC dataset shown in FIG.

Example 3
Immunohistochemical Tissue Microarray (TMA) was constructed using a 1.0 mm core from formalin-fixed paraffin-embedded tissue examined by SH 44 as separate TMA for cholangiocarcinoma and hepatocellular carcinoma. As a reference, a pair of normal tissue TMAs from the patient was constructed. All TMAs contained internal control tissue. Immunohistochemistry was performed on a 5 μm TMA fragment. First, slides were deparaffinized in xylene and rehydrated in alcohol with a concentration gradient. Antigen recovery was performed in a pressure cooker for 20 minutes using pH 6 citrate buffer. Anti-PLK1 (mouse monoclonal, clone CN05-844, Millipore) diluted 1: 1000 with 2% skim milk solution was contacted at room temperature for 2 hours at 1: 1000 dilution. Anti-ECT2 (mouse monoclonal, "E-1" cat SC-514769, Santa Cruz Biotechnology) diluted 1: 500 with 2% skim milk solution was contacted. Antigen-antibody complexes were detected by Envision + (Dako) secondary and DAB. Slides were counterstained with hematoxylin, dehydrated, clarified and covered with a cover glass. Positive and negative controls were performed. A pair of normal liver TMAs were stained at the same time as the tumor TMAs. Interpretation of immunohistochemical staining was performed at a magnification of 200 times. Tumors were scored for the percentage of stained tumor cells (0-4 in quartile) and the intensity of staining (0-4). Multiplied by two values (range 0-16). TMA in normal tissue showed no staining of tumor markers. High or low levels of PLK1 or ECT2 were determined using median scores. The ECT2 / PLK1 ratio was calculated from the raw data and the cutpoints were determined to be ECT2 / PLK1 ratio ≤ 1 at low levels and> 1 at high levels, as in the previous 24, and plotted as a Kaplan-Meier plot. The Cox-Mantel log-rank test was used for statistical analysis.

Pathological analysis was performed using Gene Set Engineering analysis (GSEA) version 16 and Industry Pathway Analysis (IPA) version 24718999. To compare patient survival, Kaplan-Meier survival analysis was performed using GraphPad Prism 6 and statistical p-values were obtained by the Cox-Mantel log-rank test. All p values are two-sided, and statistical significance is defined as p <0.05 unless otherwise specified. The previously reported relationship between signatures and the Thai CCA and HCC cohorts is based on the false discovery rate (FDR) cutoff of 0.05 predictive reliability, which is the closest template prediction algorithm implemented in GenePattern. Determined using. Relative fractions of immune / inflammatory cell subsets were estimated from the tissue expression profile of CCA or HCC using CIBERSORT 30. Gene expression data were transformed by quantile normalization of the log2 scale expression matrix and the relative fractions of leukocytes in the C1 and C2 subtypes of CCA and HCC were analyzed using the integrated LM22 signature matrix (LM22). It was quantified according to the website (civersort.standford.edu/). Each of the 22 relative leukocyte fractions was compared between the C1 and C2 subtypes using the Welch Two Sample t-test. Pearson correlation analysis was used to determine the correlation between leukocytes and BMI in the C1 and C2 subtypes.

The data obtained suggested that ECT2 and PLK1 are clinically important functional biomarkers effective in detecting subtypes of HCC and CCA. This is because both genes were previously associated with tumor progression ^19-21 . Therefore, the inventors of the present disclosure evaluated ECT2 and PLK1 by immunohistochemistry (IHC) of tissue microarray (TMA) of 199 Thai patients. TMA of CCA and HCC was constructed. PLK1 is detected in the cytoplasm but not in normal hepatocytes (Fig. 11). ECT2 is expressed in the nucleus but is absent in normal hepatocytes. The expression detected by IHC correlated with the mRNA data shown in FIG. 12A. Further analysis (FIG. 12B) demonstrated a correlation between PLK1 expression and ECT2 expression at RNA and protein expression levels. Individual survival analysis of PLK1 and ECT2 showed that patients with high TMA scores for PLK1 or ECT2 were more likely to have poor outcomes (FIG. 13). Substantial correlation in expression between PLK1 and ECT2, and PLK1 Considering that ²² to phosphorylate ECT2 in vitro has been demonstrated, the present inventors have found that these in order to predict the outcome The combination of the two biomarkers was evaluated. ^{Based on previous examples 23,24} of the ratiometric combination of relevant biomarkers, the present disclosure generated the PLK1 / ECT2 ratio from the TMA score and performed the survival analysis shown in FIGS. 14A-B, which is the PLK1. We have demonstrated that the / ECT2 ratio is robust to identify outcomes (CCA, p = 0.01, HCC, p = 0.012).

Experiments performed have revealed that the C1 subtype is associated with mitotic checkpoint defects and that PLK1 and ECT2 are two important genes clinically relevant for C1. Expression of both PLK1 and ECT2 was found to be highly expressed in the C1 subtype and robust in the definition of tumor subtype using IHC. This is clinically significant as IHC is the preferred method for pathological diagnosis. The ECT2 gene is also preferentially amplified or mutated in the C1 subtype, which is consistent with the hypothesis that it is the driver gene for the C1 subtype. Consistently, both PLK1 and ECT2 are functionally associated with cancer. PLK1 is one of the mitotic serine / threonine protein kinases required for the initiation of mitosis and the regulation of mitotic spindle assembly. Overreactive PLK1 signaling has been found in human tumors (including HCC) and because of its key functional node ¹⁹ in the oncogeneic network, it has potential for the treatment of cancer. It has been proposed that it can be a target. Many PLK1 inhibitors are currently being tested in solid tumors ^37. ECT2 are classical oncogene was first identified in 1991 on the basis of the ^20, NIH3T3 cells belong to the GEF and GAP of Ras superfamily its ability ²¹ to transform. Since ECT2 is involved in promoting early recurrence of HCC, the role of ECT2 as a manageable target for the treatment of cancers including HCC has been proposed ^20,38 . The results obtained suggest that expression of PLK1 and ECT2 may serve as a biomarker and molecular target for patient stratification for the C1 subtypes of HCC and CCA in Asians.

Example 4
Common Race / Ethnicity Tumor Subtypes To determine if the common molecular subtypes of CCA and HCC observed in Thai samples are universal, the inventors of the present disclosure We reviewed several independent cohorts with transcriptome data available for patients residing in Asia, Europe, and North America. These cohorts include 247 HCC patients from China, 378 HCC patients from the United States (TCGA data), 104 CCA patients from Europe, and 128 CCA patients from Japan. Similarities between the various subtypes identified by the transcriptome were determined using SM. The authors observed C1 or C2 prognostic molecular subtypes in Chinese HCC patients, Asian American (AsA) HCC patients, and Japanese CCA patients, but European American (EA) HCC. We found that it was not observed in patients or CCA patients in Europe (Fig. 15). Similar results showing an association between 51 common subtype-related genes and race / ethnicity-related prognosis were also observed in HCC patients in the United States. Taken together, these results suggest that they are also associated with a common prognostic molecular subtype race / ethnicity of CCA and HCC.

Since gene mutation data were available for 128 Japanese CCA tumors, the disclosure presents the common subtypes identified in this study and the top 32 mutant genes described in ^{Nakamura et al. 25.} I tried to determine if there was any relationship between them. The present disclosure has found that mutation data do not clearly identify common subtypes and mutation frequencies range from 1-43% for each subtype, but in TP53, KRAS, MYC, and GNAS. Mutations (> 10% mutations) show enrichment in the poor prognosis C1 subtype. Also, mutations in BAP1 and IDH1 were more frequent in the X subtype, a subgroup that did not match the C1 / C2 subtype, but were associated with good prognosis. Notably, 15% of the X subtypes have the IDH1 mutation, but neither the C1 subtype nor the C2 subtype has the IDH1 mutation. This suggests that this unique CCA subtype, independent of C1 and C2, has a different genetic pattern and a better prognosis. Interestingly, a statistically significant number of genes (3 out of 51 driver genes, namely NRAS, PRKCI and ECT2) based on SCNA and transcriptome (Fig. 9), suddenly appear in these 32 genes. It was found to overlap with the mutant gene (p = 0.0004; hypergeometric test). Notably, 6% of the Japanese CCA-C1 subtypes also showed mutations in ECT2, consistent with our finding that ECT2 is a functional driver of the common C1 subtype.

Example 5
Obesity, Metabolomics, and Tumor Subtypes To determine if any of the etiologic / demographic / clinical features are associated with the identified tumor subtype, the present disclosure is available clinically. C1 and C2 subtypes were compared based on variables, including age, gender, tobacco and alcohol consumption, body mass index (BMI), and tumor characteristics. Only the age, BMI status and tumor size appeared to differ between the C1 and C2 subtypes of Thai HCC. Alcohol consumption, HBV and HCV status, cirrhosis (Child-Pugh score), alkaline phosphatase (ALP) levels, CA19-9 levels, α-fetoprotein (AFP) levels, and tumor staging are CCA and It should be noted that it was significantly different from HCC. Interestingly, these pathogenic factors are not associated with common molecular subtypes, but are associated with histological subtypes.

Small biochemical species in paired tumor tissue and non-tumor tissue samples were measured using Metabolon's Discover HD4 Platform. Both positive and negative mode (LC + / LC-) liquid chromatography / mass spectrometry and gas chromatography / mass spectrometry (GC / MS) were used. A total of 718 metabolites were measured. Missing values were imputed using the minimum value of each metabolite. The 178 most variable metabolites common to both the HCC and CCA cohorts were selected in each cohort (filter: 1.5-fold change from median metabolites, log intensity variation p-value> 0.01. ). Pearson correlations between selected metabolites and the most variable genes were calculated separately for each cohort. Only significantly correlated metabolites and genes (P <0.05) within the same sample were included in the further analysis. In addition, metabolites that were significantly associated with at least 20% of the genes were selected.

In addition, BMI profiles were determined in Thai, AsA and EA patients with available BMI data. The disclosure found that EA patients, whether living in Asia or the United States, tend to have a higher BMI than Asian patients. The C2 subtype tends to have a higher BMI than the C1 subtype, the difference being statistically significant (FIG. 16). To determine if differences in BMI are associated with changes in tumor metabolism, the authors found that 199 HCC and CCA tumor samples from Thai patients and paired non-tumor tissues. Untargeted metabolome profiling was performed. Incorporating the expression profile of metabolites and gene reduces the false positive, which is a powerful technique to improve the possibility of defining functional metabolites ^26. Therefore, the present disclosure has performed a global Pearson correlation analysis between the metabolite expression profile and the gene expression profile. The results obtained are shown in FIG. From a total of 178 most variable metabolites, a total of 77 metabolites for CCA and a total of 81 metabolites for HCC showed high correlations with gene expression (-0.5 <R <0. 5; P <0.05, associated with at least 20% of genes). A statistically significant number of duplicate metabolites (n = 46) were found between CCA and HCC (hypergeometric p = 0.0007). In addition, metabolites that showed high concordance with gene expression identified C1 and C2 subtypes, as illustrated in FIGS. 18A-B. The top networks of the most important metabolites from HCC and CCA were significantly similar (FIGS. 19A-B). The present disclosure has found that bile acid-related metabolites (eg, taurochenodeoxycholic acid and tauroursodeoxychoate (TUDCA)) are significantly more abundant in both HCC-C2 and CCA-C2 than in the C1 subtype (FIG. 20A). ~ B).

The C2 subtype is associated with high BMI. BMI is known to be associated with metabolic diseases and cell inflammation ^27,28 . Moreover, C2 subtype associated bile acid metabolites are known to be associated with inflammation and immune ^29. Therefore, the inventors of the present disclosure ^{used CIBERSORT 30} to examine infiltrative immune cells in Thai HCC and CCA. The activity of leukocyte infiltrates was found to be much higher in the C2 subtype than in the C1 subtype. The results are shown in FIGS. 21A-B. Notably, elevated CD4 + memory T cells and γδ T cells, as well as decreased Treg cells, are associated with C2 subtypes (FIGS. 22A-B). Taken together, this result includes a mitotic checkpoint deficiency with altered C1 subtypes PLK1 and ECT2, while the C2 subtype has elevated BMI, immune cell abnormalities, and abnormal bile acid metabolism. Show that.

The present disclosure has found that both CCA and HCC consist of several common molecular subtypes that are shared only among Asian patients, regardless of their histological differences and differences in associated pathogenesis. .. Interestingly, common CCA and HCC subtypes that share a unique gene expression matrix have similar prognostic results. This suggests the existence of a common molecular type that goes beyond traditional histological tumor subtypes. Systematic integration of the cancer transcriptome, SCNA, and metabolome has revealed key carcinogenic drivers associated with aggressive molecular subtypes. Specifically, the C1 subtype contains a mitotic checkpoint deficiency, while the C2 subtype is associated with inflammation, obesity and bile acid biosynthesis. These results suggest that treatment stratification should not be based on histological type, but rather on molecular type.

Metabolic liver disease and obesity are associated with liver inflammation and cancer ^27,28,39,40 . However, the molecular mechanisms underlying these relationships are unknown. The authors found that BMI has a high common clinical feature associated with the C2 subtype. This suggests a possible association with liver-related metabolic disorders. Since BMI is associated with inflammatory responses and metabolic disorders, the authors also examined tumor-related leukocytes in molecularly defined subtypes. It was found that lymphoid cells such as CD4 + memory T cells and γδ T cells were significantly elevated in the C2 subtype, whereas myeloid cells were not significantly elevated in the C2 subtype. These results are consistent with enriched C2 subtype gene expression data for cell-mediated immune-related pathways, reaffirming the idea that the inflammatory response is associated with C2 subtypes. It is also interesting that in both HCC and CCA, some bile acid metabolites (such as TUDCA, taurocholic acid and glycokenodeoxycholic acid) are consistently much more abundant in the C2 subtype than in the C1 subtype. Bile acids have recently emerged as versatile signaling molecules for regulating cholesterol metabolism, energy homeostasis, and glucose homeostasis, for the treatment of common metabolic and liver disorders. The basis for developing new drug targets ⁴¹ . It is believed that some of these agents may be applicable to treat the C2 subtype of liver cancer. The results obtained also show that deoxycholic acid, an obesity-inducing intestinal microbial metabolite, promotes liver carcinogenesis through the senescence secretome, and that diet alters the human intestinal microbiome, resulting in obesity. ^Consistent with recent studies 29,42 showing that it promotes such diet-related disorders. In addition, dietary fat-induced taurocholic acid promotes pasoviont growth and colitis in IL1-deficient mice, which is that in genetically susceptible hosts, diets containing certain saturated fats are immune-mediated. clear of inducing disease, a reasonable mechanistic basis ^43.

The results obtained indicate that common C1-C2 subtypes are predominantly associated with Asian patients, but the reason for the association with race / ethnicity is unknown. Western diets can induce different dysbiosis between Asians and Caucasians, and racial / ethnic differences in the intestinal microbiome, in cooperation with bile acid metabolism, are common to Asians. It is possible to induce the unique carcinogenic process observed in the C2 subtype of. An increase in invasive T cells in the C2 subtype also suggests that these tumors may be sensitive to immune checkpoint inhibitors. In summary, the integrated omics approach adopted in this disclosure defines common molecular subtypes of CCA and HCC across several Asian populations, and potential driver genes and metabolic processes associated with specific subtypes. Was identified.

It should be understood that the present invention may be implemented in other specific embodiments and is not limited to the embodiments described above. However, changes to the disclosed concept that can be easily conceived by those skilled in the art are intended to be included in the appended claims.

See 1. Theise, N.M. D. Liver cancer. (International Agency for Research on Cancer, 2014).

2. Wang, X. W. & Thrergeirsson, S.A. S. The Biological and clinical challenge of liver cancer. Hepatic Oncology 1, 5, doi: 10.217 / HEP. 14.18 (2014).

3. 3. El-Serg, H. et al. B. Hepatocellular carcinoma. The New England journal of medicine 365, 1118-1127, doi: 10.1056 / NEJMra1001683 (2011).

4. Sripa, B. et al. Liver fuke induces liveriocarcinoma. PLoS. Med. 4, e201 (2007).

5. Ye, Q. H. et al. Reading hepatitis B virus-possitive hepatitis hepatocellular carcinomas using gene expression profiling and supervised machine learning. Nat. Med. 9, 416-423 (2003).

6. Hoshida, Y.M. et al. Integrative transcriptome analysis revivals common molecular subclasses of human hepatocellular carcinoma. Cancer research 69, 7385-7392 (2009).

7. Rosessler, S.M. et al. Integrative genomic and transcriptomics characterization of matched liver and methastatic liver and colorectal cancer. International journal of biological sciences 11, 88-98, doi: 10.7150 / ijbs. 10583 (2015).

8. Monti, S.M. et al. Molecular profiling of diffuse large B-cell lymphoma identifies robot subtypes incurding one host inflammatory respawn. Blood 105, 1851-1861, doi: 10.1182 / blood-2004-07-2947 (2005).

9. Woo, H. G. et al. Identity of a cholangiocarcinoma-like gene expression trait in hepatocellular carcinoma. Cancer research 70, 303-3041 (2010).

10. Oishi, N.M. et al. Transcriptomics profiling revolutions hepatic stem-like gene genes and interplay of miR-200c and epithelial-mesenchimal training. Hepatology 56, 1792-1803 (2012).

11. Hoshida, Y.M. , Brunet, J. et al. P. , Tamayo, P.M. , Golub, T.I. R. & Mesirov, J.M. P. Subclass mapping: identifying community subtypes in independent disease data sets. PLOS one 2, e1195, doi: 10.131 / journal. pone. 0001195 (2007).

12. Lee, J.M. S. et al. Classification and prediction of revival in hepatocellular carcinoma by gene expression profiling. Hepatology 40, 667-676 (2004).

13. Yamashita, T.M. et al. EpCAM and alpha-fetoprotein expression defenses novel prognotic subtypes of hepatocellular carcinoma. Cancer Res. 68, 1451-1461 (2008).

14. Andersen, J. Mol. B. et al. Genomic and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors. Gastroenterology 142, 1021-1031 e1015, doi: 10.1053 / j. gastro. 2011.12.005 (2012).

15. Sia, D.I. et al. Integrative molecular analysis of intrahepatic cholangiocarcinoma revivals 2 classes that have different outcomes. Gastroenterology 144, 829-840, doi: 10.1053 / j. gastro. 2013.01.001 (2013).

16. Hoshida, Y.M. Nealest temple prediction: a single-sample-based flexible class prophecy with confidence assistance. PloS one 5, e15543, doi: 10.131 / journal. pone. 0015543 (2010).

17. Rosessler, S.M. et al. Integrative genomics of genes on 8p associated with hepatocellular carcinoma progression and patient surveillance. Gastroenterology 142, 957-966 (2012).

18. Woo, H. G. et al. Idealization of potential driver genes in human live carcinoma by genes screening. Cancer research 69, 4059-4066 (2009).

19. Strebhard, K. et al. Multifacated polo-like kinases: drug targets and antitargets for cancer therapy. Nat Rev Drag Discov 9, 643-660, doi: 10.1038 / nrd3184 (2010).

20. Vigil, D. , Cherfils, J. et al. , Rossmann, K.K. L. & Der, C.I. J. Ras superfamily GEFs and GAPs: validated and traumatic targets for cancer therapy? Nature reviews. Cancer 10, 842-857, doi: 10.1038 / nrc2960 (2010).

21. Cook, D. R. , Rossmann, K.K. L. & Der, C.I. J. Rho guanine nucleice exchange factorors: regulators of Rho GTPase activity in development and disease. Oncogene 33, 4021-4035, doi: 10.1038 / onc. 2013.362 (2014).

22. Niiya, F.M. , Tatsumoto, T.I. , Lee, K.K. S. & Miki, T.M. Phosphorylation of the cytokinesis regulator ECT2 at G2 / M phase stimulates association of the mitotic kinase Plk1 and acquisition of GTP Oncogene 25, 827-837, doi: 10.1038 / sj. onc. 1209124 (2006).

23. Chung, J.M. Y. et al. The expression of phospho-AKT, phospho-mTOR, and PTEN in extrahepatic cholangiocarcinoma. Clinical cancel search: an official journal of the American Association for Cancer Research 15, 660-667, doi: 10.1158 / 1078-0432. CCR-08-1084 (2009).

24. Kitano, H.M. et al. Profiling of phospho-AKT, phospho-mTOR, phospho-MAPK and EGFR in non-smal cell lung cancer. J Histochem Cytochem 62, 335-346, doi: 10.1369 / 0022155414523022 (2014).

25. Nakamura, H. et al. et al. Genomic spectrum of battery trac cancer. Nat Genet 47, 1003-1010, doi: 10.1038 / ng. 3375 (2015).

26. Budhu, A. et al. Integrated metabolite and gene expression profiling lipids lipid biomarkers associated with patients expression of hepatocellular carcinoma. Gastroenterology 144, 1066-1075 (2013).

27. Calle, E.I. E. , Rodriguez, C.I. , Walker-Turmond, K.K. & Thun, M.M. J. Overweight, obesity, and mortality from cancer in a productive studied studio of U.S.A. S. adults. N. Engl. J Med. 348, 1625-1638 (2003).

28. Park, E.I. J. et al. Dietary and genetics activity liver inflamemation and tumor necrosis by enhancing IL-6 and TNF expression. Cell 140, 197-208 (2010).

29. Yoshimoto, S.A. et al. Obesity-induced gut microbial metabolite features liver cancer liver sensence secretome. Nature 499, 97-101, doi: 10.1038 / nature12347 (2013).

30. Newman, A. M. et al. Robot enumeration of cell subsets from tissue expression profiling. Nature methods 12, 453-457, doi: 10.1038 / nmeth. 3337 (2015).

31. Guicard, C.I. et al. Integrated analysis of somatic mutations and focal copy-number channels identificies key genes and hepatocellular carcinoma. Nat. Genet. 44, 694-698 (2012).

32. Totoki, Y. et al. Trans-ancestry mutational landscape of hepatocellular carcinoma genome. Nat Genet 46, 1267-1273, doi: 10.1038 / ng. 3126 (2014).

33. Helday, T.M. , Eshdad, S.A. & Nick-Zainal, S.A. Mechanisms underlying mutations in human cancers. Nat Rev Genet 15, 585-598, doi: Doi 10.1038 / Nrg3729 (2014).

34. Lee, J.M. S. et al. A novel progenitor cell of human hepatocellular carcinoma derived from hepatic progenitor cells. Nat. Med. 12, 410-416 (2006).

35. Rosessler, S.M. et al. A unique metastasis gene signature envelopes prophecy of tumor relaxation in early-stage hepatocellular carcinoma patients. Cancer research 70, 10202-10212 (2010).

36. Nault, J. et al. C. et al. A hepatocellular carcinoma 5-gene score associated with survival of patients after liver reduction. Gastroenterology 145, 176-187, doi: 10.1053 / j. gastro. 2013.03.051 (2013).

37. Yim, H. Current clinical trials with poly-like kinase 1 inhibitor in solid tumors. Antique Drugs 24, 999-1006, doi: 10.1097 / CAD. 0000000000000007 (2013).

38. Chen, J.M. et al. ECT2 Regulates the Rho / ERK signaling axes to promote early recognition in human hepatocellular carcinoma. Journal of hepatology 62, 1287-1295, doi: 10.016 / j. jhep. 2015.01.014 (2015).

39. Cohen, J.M. C. , Horton, J. et al. D. & Hobbs, H. H. Human fatty liver disease: old questions and new insights. Science (New York, NY) 332, 1519-1523, doi: 10.1126 / science. 1204265 (2011).

40. Welzel, T.M. M. et al. Metabolic syndrome increases the risk of liver cancer in the United States: a study in the SEER-Medicare database. Hepatology 54, 463-471, doi: 10.10012 / hep. 24397 (2011).

41. Thomas, C.I. , Pelicciari, R.M. , Pruzanki, M. et al. , Auwerx, J. et al. & Schoonjans, K.K. Targeting bile-acid signaling for metabolic diseases. Nat Rev Drag Discov 7, 678-693, doi: 10.1038 / nrd2619 (2008).

42. David, L. et al. A. et al. Diet rapidly and reproducible alters the human gut microbiome. Nature 505, 559-563, doi: 10.1038 / nature12820 (2014).

43. Devkota, S.M. et al. Dietary-fat-induced taurocholic acid technologies pasovation expansion and colitis in Il10 − / − meeting. Nature 487, 104-108, doi: 10.1038 / nature11225 (2012).

44. Kononen, J.M. et al. Tissue microarrays for high-throughput preview molecular profiling of tumor specimens. Nature Medicine 4, 844-847 (1998).

45. Chen, J.M. et al. ECT2 REGULATES THE Rho / ERK signaling axes to promotion in human hepatocellular carcinoma. J Hepatol. 62 (6): 1287-95 (2015).

46. Sun, W. et al. High expression of Polo-Like Kinase 1 is assisted with early development of hepatocellular carcinoma. International Journal of Genetics, volume 2014, Article ID 312130.

47. Takai, N.M. et al. Polo-like Kinases (PLKs) and cancer. Oncogene 24, 287-291 (2005).

Claims (12)

A method for generating a predicted survival outcome for subjects with liver cancer,
A step of preparing a tissue sample of one of hepatocellular carcinoma or cholangiocarcinoma collected from the above subjects, and
At least one nucleic acid-based assay and / or to quantify transcriptome expression and / or protein expression of polo-like kinase 1 (PLK1) and epithelial cell transformation 2 (ECT2) in the prepared sample. The process of performing a proteomics-based assay and
A step of generating the predicted survival outcome of the subject based on the individual expression levels of PLK1 or ECT2 in a particular region on the prepared sample or using the relative expression of PLK1 and ECT2. ;
(A) When the median expression level was used as the cutoff, it was in the group with low expression levels of PLK1 or ECT2 in the subject and / or
(B) The relative expression of PLK1 and ECT2 of the subject is less than a predetermined value within the range of 0.001 to 1.
In this case, the predicted survival outcome of the subject is defined as the first predicted survival outcome.
(A) When the median expression level was used as the cutoff, it was in the group with high expression levels of PLK1 or ECT2 in the subject and / or
(B) The relative expression of PLK1 and ECT2 of the subject is equal to or higher than a predetermined value within the range of 0.001 to 1.
In the case, the step of setting the predicted survival outcome of the subject as the second predicted survival outcome is included.
The first predicted survival outcome corresponds to an overall survival rate of greater than 40% to 60% in the predetermined period of survival analysis, and the second predicted survival outcome corresponds to the predetermined period of survival analysis. A method corresponding to an overall survival rate of 40% to 60% or less. The method according to claim 1, wherein the predetermined period is 12 to 36 months. The method of claim 1, wherein the nucleic acid-based assay is any one or combination of a microarray-based technique and a polymerase chain reaction-based technique. The method of claim 1, wherein the proteomics-based assay is tissue microarray and / or immunohistochemical staining. The method according to claim 1, wherein the subject is of Asian descent. A device for generating predicted survival outcomes for subjects with liver cancer.
A first module capable of detecting transcriptome expression and / or protein expression of polo-like kinase 1 (PLK1) and epithelial cell transformation 2 (ECT2) in the tissue sample of the subject,
A second configured to derive the ratio of PLK1 expression to ECT2 expression in a particular region on the prepared sample and to generate the predicted survival outcome of the subject based on the derived ratio. With modules,
The second module above is
(A) When the median expression level was used as the cutoff, it was in the group with low expression levels of PLK1 or ECT2 in the subject and / or
(B) The relative expression of PLK1 and ECT2 of the subject is less than a predetermined value within the range of 0.001 to 1.
In this case, the predicted survival outcome of the subject is defined as the first predicted survival outcome, or
(A) When the median expression level was used as the cutoff, it was in the group with high expression levels of PLK1 or ECT2 in the subject and / or
(B) The relative expression of PLK1 and ECT2 of the subject is equal to or higher than a predetermined value within the range of 0.001 to 1.
In this case, the predicted survival outcome of the subject is configured to be the second predicted survival outcome.
The first predicted survival outcome corresponds to an overall survival rate of greater than 40% to 60% in the predetermined period of survival analysis, and the second predicted survival outcome corresponds to the predetermined period of survival analysis. A device that supports an overall survival rate of 40% to 60% or less. At least one for binding the expressed transcriptome or protein of PLK1 and ECT2 to a signaling moiety formed to emit a signal separable continuously or periodically by the first module. The apparatus of claim 6, further comprising a test module capable of performing a nucleic acid-based assay or a proteomics-based assay. The device according to claim 6, wherein the predetermined period is 12 to 36 months. The apparatus according to claim 7, wherein the nucleic acid-based assay is any one or combination of a microarray-based technique and a polymerase chain reaction-based technique. The apparatus of claim 7, wherein the proteomics-based assay is tissue microarray and / or immunohistochemical staining. The device according to claim 6, wherein the subject is of Asian descent. The apparatus according to claim 6, wherein the tissue sample of the subject is a carcinoma, which is one of hepatocellular carcinoma or cholangiocarcinoma.

Images