CN109081866B

CN109081866B - T cell subpopulations in cancer and genes characteristic thereof

Info

Publication number: CN109081866B
Application number: CN201710443584.3A
Authority: CN
Inventors: 张泽民; 郑春红; 郑良涛; 张园园; 郭心怡; 胡学达
Original assignee: Peking University
Current assignee: Peking University
Priority date: 2017-06-13
Filing date: 2017-06-13
Publication date: 2021-07-16
Anticipated expiration: 2037-06-13
Also published as: CN109081866A

Abstract

The present invention utilizes a single cell transcriptome analysis technique to isolate and characterize T cell subsets that reflect the immune status of body tumors, i.e., the depleted CD8 expressing genes WARS and ACP5, by analyzing the single cell gene expression profiles of infiltrating T cells in cancer tissues⁺T cells, and regulatory T cells expressing the genes STAM and BATF. And further research confirms that the new characteristic genes WARS and ACP5 expressed by the cell subsets and the relation between STAM and BATF and tumor prognosis can be used for diagnosing and monitoring tumor prognosis and can be used as a new tumor immunotherapy target.

Description

T cell subpopulations in cancer and genes characteristic thereof

Technical Field

The present invention relates to the field of biotechnology, and in particular to the identification, characterization and enrichment of cells or cell subsets in general, and T cell subsets derived from tumor tissue in particular. The invention also relates to a series of markers, methods of isolation and resulting formulations for isolating T cell subsets in tumors, and the use of isolated T cell subsets for drug research and development and for diagnostic, therapeutic and other clinical and non-clinical applications.

Background

It is generally accepted that the adaptive immune response in which T lymphocytes participate is one of the major pathways in humans against tumor cells.

Immunotherapy has become an indispensable link in clinical treatment of tumors at present. The drugs and regimens for immunotherapy involve various stages of the body's immune system recognizing and attacking cancer cells. Existing drugs for tumor immunization include several types: cancer cell-targeting antibodies, adoptive cell therapy, oncolytic viruses, dendritic cell-related therapy, tumor vaccines at DNA and protein levels, immune activating cytokines, and other immunomodulatory compounds. These drugs are involved in the body's immune system recognizing and killing various stages of the tumor. Among them, antibody drugs against T cell checkpoint inhibitory proteins and T cell adoptive therapies specific to tumor antigens have made a breakthrough in recent years and have attracted much attention.

Checkpoint inhibitors against T cells are an antibody-based immunotherapy that monitors and kills tumor cells by restoring Cytotoxic T Lymphocyte (CTL) activity. Effector T cells in the tumor microenvironment are inhibited from killing tumor cells due to the presence of multiple immune escape mechanisms in the tumor. Antibodies that either potentiate or block the relevant pathways are designed against known immune escape mechanisms, thereby enhancing the action of the immune system. For example, the drug Iplilimumab targeting CTLA-4 acts to abrogate the CTL inhibition by regulatory T cells (Tregs) by blocking CTLA-4. PD-1 is a receptor of PD-L1/L2 expressed on the surface of T cells, and the latter is expressed on the surface of tumor cells, and the combination of the two can activate the apoptosis pathway of T cells. The antibody medicament Nivolumab inhibits the combination of the Nivolumab and PD-L1/L2 by blocking PD-1 on the surface of the T cell, thereby inhibiting the inactivation or apoptosis of the T cell and ensuring that CTL has immune activity all the time. In turn, MDX-1151, which targets PD-L1, is also effective in increasing the survival rates of lung, melanoma and renal cancers. However, similar to targeted therapies and the like, the target of the T cell checkpoint inhibitor still has room for further expansion; at the same time, how to find more effective patients also requires a biological systematic understanding of the T cell response process in cancer tissues.

On the other hand, the early diagnosis, early treatment and prognosis judgment of the tumor have important significance for improving the treatment effect of the tumor and the survival rate of patients. The reliable tumor marker is determined by utilizing the molecular biology technology, and the method has important significance for predicting the clinical outcome of the tumor, targeting treatment and evaluating the curative effect. The tumor prognosis markers discovered at present mainly comprise cancer suppressor genes, oncogenes, cell cycle regulatory factors, apoptosis regulatory factors, invasion and metastasis related markers, telomerase and the like. There are few molecular markers that can be used to monitor and assess the immune status of tumors.

Understanding the interaction of the immune system with cancerous tissues is key to the development and improvement of immunotherapy and also provides a means to assess and monitor the immune status of tumors in the body. The classification, differentiation and the pathway of inhibition of T cells (TILs) infiltrating into cancer tissues has not yet been fully understood. It is currently known that depleted T cells, disabled T cells or aged T cell populations exist in TILs.

Considering the complexity of human immune cells, especially T cell types, the classification, differentiation process and functional genes of the immune cells related to tumors need to be recognized from higher precision, so as to further provide more potential immune targets for the development of tumor immunotherapy drugs and molecular markers for tumor prognosis.

In recent years, the rapidly developed single cell sequencing technology provides a powerful support for the gene expression difference research of the population cells and the single cells. Single-Cell transcriptome deep sequencing methods such as STRT-Seq (single-Cell labeled reverse transcription sequencing), Smart-Seq and Smart-Seq2, Cell-Seq (Cell expression by linear amplification and sequencing), and PMA-Seq (Phi29-mRNA amplification and sequencing), and various bioinformatics analysis methods matched with the same have been developed.

Disclosure of Invention

The present inventors have completed the present invention based on the fact that a single-cell transcriptome analysis technique is used to analyze the single-cell gene expression profile of infiltrating T cells in cancer tissues, to isolate and characterize T cell subsets that reflect the immune status of tumors in the body, and further study and determine novel characteristic genes expressed by the cell subsets and the relationship between the characteristic genes and the prognosis of tumors.

Thus, the present invention is directed broadly to a series of markers, methods, compounds, compositions and articles of manufacture that can be used to identify or characterize, and optionally to isolate, partition, separate or enrich, T cell subsets that are relevant to tumor immunity. Said T cells comprise CD8⁺T cells and CD4⁺T cells, particularly depleting CD8⁺T cell and regulatory CD4⁺T cells.

More specifically, the inventors of the present application have discovered a series of markers that can be used independently or collectively to accurately identify, sort, enrich for and/or characterize T cell subsets from tumors. Using selected biochemical techniques, T cell subsets reflecting the immune status of a tumor can be enriched, isolated or purified by association of the markers of the invention with T cells in tumor tissue.

The markers disclosed in the invention can identify or identify T cell subsets from tumors, constitute a general characterization of immune cells of the tumors, and can be used for the clarification of therapeutic targets and the screening of drug compounds. Moreover, it can be further used in clinical and non-clinical applications for diagnosis, prognosis, classification, monitoring and management of tumor patients, and to provide related kits or other articles of manufacture.

A first aspect of the present invention provides:

CD8 for diagnosis or monitoring⁺A biomarker panel for T cells comprising the genes WARS and ACP5, or proteins or protein fragments expressed by the genes.

Preferably, the biomarker panel further comprises at least one of the genes HAVCR2, PDCD1 and LAG3, or a protein or protein fragment expressed by said at least one gene.

For example, the biomarker panel comprises the genes WARS, ACP5, and HAVCR2, or the proteins or protein fragments expressed by the genes; alternatively, the biomarker panel comprises the genes WARS, ACP5, and PDCD1, or a protein or protein fragment expressed by the genes; alternatively, the biomarker panel comprises the genes WARS, ACP5, and LAG3, or the proteins or protein fragments expressed by the genes; alternatively, the biomarker panel comprises the genes WARS, ACP5, HAVCR2, and PDCD1, or proteins or protein fragments expressed by the genes; alternatively, the biomarker panel comprises the genes WARS, ACP5, HAVCR2, and LAG3, or the proteins or protein fragments expressed by the genes; alternatively, the biomarker panel comprises the genes WARS, ACP5, PDCD1, and LAG3, or the proteins or protein fragments expressed by the genes; alternatively, the biomarker panel comprises the genes WARS, ACP5, HAVCR2, PDCD1, and LAG3, or a protein or protein fragment expressed by the genes.

Most preferably, the biomarker panel comprises the genes WARS, ACP5, HAVCR2, PDCD1 and LAG3, or the proteins or protein fragments expressed by the genes.

The diagnosis or monitoring may be diagnosis or monitoring of CD8⁺Depletion state of T cells. The diagnosis or monitoring may be a diagnosis or monitoring of prognosis of the tumor.

And CD8 in normal tissues or peripheral blood of patients⁺An increase in expression or expression of at least one of the five genes as compared to T cells indicates CD8⁺The T cells are in a depleted state, which further indicates that the prognosis of the patient is poor.

CD4 for diagnosis or monitoring⁺A biomarker panel for T cells comprising the genes STAM and BATF, or a protein or protein fragment expressed by said genes.

Preferably, the biomarker panel further comprises the gene TNFRSF4 and/or the gene CTLA4, or a protein or protein fragment expressed by said gene.

For example, the biomarker panel comprises the genes STAM, BATF and TNFRSF4, or a protein or protein fragment expressed by the genes; alternatively, the biomarker panel comprises the genes STAM, BATF and CTLA4, or a protein or protein fragment expressed by the genes; alternatively, the biomarker panel comprises the genes stamp, bat, TNFRSF4 and CTLA4, or proteins or protein fragments expressed by said genes.

Most preferably, the biomarker panel comprises the genes stamp, bat, TNFRSF4 and CTLA4, or proteins or protein fragments expressed by said genes.

The diagnosis or monitoring may be diagnosis or monitoring of CD4⁺Whether the T cell is an active regulatory T cell. The diagnosis or monitoring may be a diagnosis or monitoring of prognosis of the tumor.

And CD4 in normal tissues or peripheral blood of patients⁺Increased expression or expression of said gene in T cells, indicative of said CD4⁺T cells are active regulatory T cells, further, indicating poor prognosis for the patient.

In one embodiment of the invention, the tumor comprises liver cancer, in particular hepatocellular carcinoma.

A second aspect of the present invention provides:

a kit for diagnosing or monitoring the prognosis of a tumour, said kit comprising CD8⁺A binding agent that binds to the T cell gene WARS or a protein or protein fragment expressed therefrom, and a polypeptide capable of binding to CD8⁺A binding agent to which the T cell gene ACP5, or an expressed protein or protein fragment thereof, binds.

Preferably, the kit further comprises a CD 8-binding agent⁺A binding agent which binds to the T cell gene HAVCR2 or a protein or protein fragment expressed therefrom, and/or which binds to CD8⁺A binding agent which is combined with a gene PDCD1 of a T cell or an expressed protein or protein fragment thereof, and/or can be combined with CD8⁺A binding agent to which the gene LAG3 of a T cell, or an expressed protein or protein fragment thereof, binds.

Most preferably, the kit comprises a CD 8-binding agent⁺Binding agent capable of binding to CD8 and binding to WARS gene of T cell or expressed protein or protein fragment thereof⁺Binding agent for binding to ACP5 gene of T cell or protein fragment thereof capable of binding to CD8⁺Binding agent for binding T cell gene HAVCR2 or its expressed protein or protein fragment, and CD8⁺A binding agent which is combined with a gene PDCD1 of a T cell or an expressed protein or a protein fragment thereof, and can be combined with CD8⁺T cell gene LAG3 or protein fragment expressed by sameThe binder of (1).

A kit for diagnosing or monitoring the prognosis of a tumour, said kit comprising CD4⁺A binding agent that binds to the gene STAM of T cells or a protein or protein fragment expressed therefrom, and a polypeptide capable of binding to CD4⁺A binding agent to which the gene BATF of T cells or a protein or protein fragment expressed therefrom binds.

Preferably, the kit further comprises a CD 4-binding agent⁺A binding agent which binds to the T cell gene TNFRSF4 or a protein or protein fragment expressed therefrom, and/or which binds to CD4⁺A binding agent to which the gene CTLA4 of T cells or expressed proteins or protein fragments thereof bind.

The binding agent includes nucleic acid, ligand, enzyme, substrate, antibody, etc.

In one embodiment of the invention, the kit comprises a container or containers comprising one or more of the binding agents. Further preferably, the kit further comprises instructional materials, such as instructions, for using the kit.

In a preferred embodiment of the invention, the tumor comprises liver cancer, in particular hepatocellular carcinoma.

A third aspect of the present invention provides:

use of an inhibitor for inhibiting the expression of a target gene or a protein inhibiting the expression of a target gene selected from the group consisting of: WARS, ACP5, STAM and BATF.

The expression or high expression of WARS or ACP5 can cause the corresponding CD8⁺T cells are depleted and expression or high expression of STAM or BATF leads to CD4⁺Conversion of T cells into regulatory T cells, inhibition of expression of these genes or inhibition of activity of the expressed proteins, would be beneficial to restoring CD8 in a depleted state⁺Activity of T cells, or inhibition of CD8⁺Inactivation or apoptosis of T cells, or release of cytotoxic T cell inhibition by regulatory T cells, to effectThe effect of treating the tumor is achieved.

A fourth aspect of the present invention provides:

a method of screening for a drug, the method comprising the steps of:

1) mixing a test chemical with said gene or protein expressed thereby, or mixing a test chemical with CD8 expressing said gene⁺T cells or CD4⁺Mixing T cells;

2) detecting a change in the activity of the expressed protein, or whether the test chemical binds to the gene or protein expressed therefrom, or a change in the activity of the cell, or a change in the amount of expression of the gene by the cell;

the gene is selected from the group consisting of: WARS, ACP5, STAM and BATF.

In a preferred embodiment of the invention, the medicament is for the treatment of tumors. In one embodiment of the invention, the tumor comprises liver cancer, in particular hepatocellular carcinoma.

A fifth aspect of the present invention provides:

CD8⁺Subpopulations of T cells expressing genes WARS and ACP 5.

The CD8⁺The subpopulation of T cells may further express at least one of the genes HAVCR2, PDCD1 and LAG 3.

For example, the CD8⁺T cell subsets express the genes WARS, ACP5 and HAVC 2; alternatively, the CD8⁺The T cell subsets express genes WARS, ACP5, and PDCD 1; alternatively, the CD8⁺Subpopulations of T cells express genes WARS, ACP5, and LAG 3; alternatively, the CD8⁺The T cell subsets express genes WARS, ACP5, HAVCR2 and PDCD 1; alternatively, the CD8⁺T cell subpopulations express genes WARS, ACP5, HAVCR2, and LAG 3; alternatively, the CD8⁺T cell subsets express genes WARS, ACP5, PDCD1, and LAG 3; alternatively, the CD8+ T cell subpopulation expresses the genes WARS, ACP5, HAVCR2, PDCD1, and LAG 3.

The CD8⁺The T cell subpopulation is depleted T cells. The inventor of the invention firstly discovers that the genes WARS and ACP5 are expressed by C through comparative researchD8⁺The T cells are in a depleted state. The CD8⁺The larger the proportion of the T cell subgroup in the T cells infiltrated into the tumor is, the weaker the capability of the T cells infiltrated into the tumor of the patient to kill the tumor cells is, and the worse the prognosis of the patient is; the CD8⁺The more the T cell subsets express the gene types of the five genes WARS, ACP5, HAVCR2, PDCD1 and LAG3, the more accurate the diagnosis of the tumor immune state of the patient is and the more accurate the prognosis judgment of the patient is.

CD4⁺T cell subpopulations expressing genes STAM and BATF.

The CD4⁺The T cell subsets may further express the genes TNFRSF4, and/or CTLA 4.

For example, the CD4⁺The T cell subset expresses genes STAM, BATF and TNFRSF 4; alternatively, the CD4⁺The T cell subset expresses genes STAM, BATF and CTLA 4; alternatively, the CD4⁺The T cell subsets express the genes STAM, BATF, TNFRSF4 and CTLA 4.

The CD4⁺The T cell subpopulation is regulatory T cells. The inventor of the invention firstly discovers that the CD4 expressing genes STAM and BATF through comparison research⁺T cells are regulatory T cells and exert inhibitory regulatory effects on cytotoxic T cells. The CD4⁺The larger the proportion of the T cell subgroup in the T cells infiltrated into the tumor is, the weaker the capability of the T cells infiltrated into the tumor of the patient to kill the tumor cells is, and the worse the prognosis of the patient is; the CD4⁺The more the T cell subsets express gene types of the four genes of STAM, BATF, TNFRSF4 and CTLA4, the more accurate the diagnosis of the tumor immune state of the patient is, and the more accurate the prognosis of the patient is.

In a preferred embodiment of the invention, both T cell subsets are derived from a tumor. In one embodiment of the invention, the tumor is liver cancer, in particular hepatocellular carcinoma.

A sixth aspect of the present invention provides:

a method of enriching a subpopulation of T cells according to the fifth aspect.

The method comprises the following steps: contacting the immune cell population infiltrating the tumor with a binding agent that binds to at least one of the above genes or a protein or protein fragment expressed therefrom; and sorting the immune cells bound to the binding agent to provide an enriched T cell subpopulation.

As an embodiment of the present invention, a CD8 is enriched⁺A method of T cell subpopulation, the method comprising the steps of: contacting a population of immune cells infiltrating the tumor with a binding agent that binds to the gene of interest or a protein or protein fragment expressed therefrom; and sorting the immune cells bound to the binding agent to provide enriched CD8⁺T cell subsets.

The target genes include WARS and ACP 5. Preferably, the target gene further comprises at least one of HAVCR2, PDCD1 and LAG 3.

As another embodiment of the invention, a CD4 is enriched⁺A method of T cell subpopulation, the method comprising the steps of: contacting a population of immune cells infiltrating the tumor with a binding agent that binds to the gene of interest or a protein or protein fragment expressed therefrom; and sorting the immune cells bound to the binding agent to provide enriched CD4⁺T cell subsets.

The target genes include STAM and BATF. Preferably, the target gene further comprises TNFRSF4, and/or CTLA 4.

In a preferred embodiment of the invention, the sorting step comprises fluorescence activated cell sorting, magnetic assisted cell sorting, substrate assisted cell sorting, laser mediated cleavage, fluorimetry, flow cytometry or microscopy.

In a preferred embodiment of the invention, the tumor is liver cancer, in particular hepatocellular carcinoma.

A seventh aspect of the present invention provides:

a method of diagnosis or assessment, the method comprising the steps of:

1) immunization of tumors with infiltration obtained from a subjectCell sample classified as CD8⁺T cells and CD4⁺A T cell; and the combination of (a) and (b),

2) the separated CD8⁺Contacting the T cells with at least one binding agent that binds to a target gene, or protein fragment expressed therefrom, said target gene comprising WARS and ACP 5;

and/or the presence of a gas in the gas,

3) the separated CD4⁺Contacting the T cells with at least one binding agent that binds to a target gene or protein fragment thereof expressed, said target gene comprising STAM and BATF.

Preferably, in the step 2), the target gene further comprises at least one of HAVCR2, PDCD1 and LAG 3.

Preferably, in step 3), the target gene further comprises TNFRSF4, and/or CTLA 4.

In one embodiment of the invention, the method is performed prior to the patient receiving treatment. In other embodiments, the method can be performed after the subject receives treatment, e.g., after chemotherapy, after radiation therapy, after surgical treatment, etc.

In one embodiment of the invention, the method is used to determine the prognosis of a subject.

An eighth aspect of the present invention provides:

a method for single cell transcriptome analysis of T cells, the method comprising the steps of: (1) obtaining individual T cells; (2) constructing a cDNA library of each T cell and sequencing to obtain the expression quantity of each gene of each cell; (3) carrying out unsupervised clustering on the T cells according to the expression quantity of each gene of each cell; (4) each type of T cell was analyzed for characteristic expressed genes.

According to the present invention, individual T cells can be obtained using various methods known in the art, for example, for individual T cells in blood, density gradient centrifugation; for individual T cells in the tissue, milling may be used.

According to the present invention, various methods known in the art for constructing cDNA libraries of transcriptomes of single cells can be used to construct cDNA libraries of each T cell and sequence the cDNA libraries to obtain the expression level of each gene of each cell, for example: tom enrichment 2009 created methods (Tang, F.et al. RNA-Seq analysis to capture the transformed same will be of single Cell Nat. Protococ.5, 516-535 (2010)), STRT-Seq (single-Cell tagged reverse transcription sequencing), Smart-Seq and Smart-Seq2, Cell-Seq (Cell expression by linear amplification and sequencing), and PMA-Seq (Phi29-mRNA amplification and sequencing).

In a preferred embodiment of the present invention, a cDNA library of each T cell is constructed using Smart-Seq2 and sequenced to obtain the expression level of each gene of each cell.

The inventor of the invention relatively researches a method established in 2009 for soup remuneration and Smart-Seq2, and finds that the Smart-Seq2 method can detect more genes under the condition of ensuring the sequencing quality, wherein the genes comprise a marker CD3 gene shared by T lymphocytes; and the Smart-seq2 method is more beneficial to amplifying complete cDNA and is more suitable for T cell single cell transcriptome amplification.

Through experimental research, the inventor of the invention further improves the operating conditions in Smart-seq2, and improves the reverse transcription yield of mRNA and the purification efficiency of products after PCR amplification.

In the specific embodiment of the invention, when the Smart-seq2 method is adopted for reverse transcription, the following reverse transcription conditions are adopted, so that the yield of reverse transcription cDNA and the proportion of the whole length of the cDNA are improved:

compared with the common reverse transcription condition of 30 minutes at 50 ℃, the improved reverse transcription condition can improve the cDNA yield by 16-23 percent and the average length of the whole cDNA length by about 20 percent.

In the specific implementation mode of the invention, the method for purifying the PCR amplification product by adopting the Smart-seq2 method is as follows, greatly improves the purity of the PCR product, and is beneficial to the improvement of the subsequent sequencing and library construction quality: and (3) purifying by using magnetic beads, wherein the volume of the added magnetic beads is the same as that of the PCR amplification product during the first purification, and the volume of the added magnetic beads is 2 times of that of the PCR amplification product during the second purification.

According to the present invention, when the analysis of step (3) is performed, the biological information data obtained in step (2) is compared and quality-controlled, removing the low-quality part.

According to the invention, the method for controlling the data quality of sequencing reads (reads) of cDNA comprises the following steps: sequencing reads that met the following conditions were retained: the unknown base accounts for no more than 10% of the total sequence of the given read, the base with the Phred mass value lower than 5 does not exceed 50%, and the sequence does not contain a linker.

According to the invention, the cell quality control method is to remove cells with low data quantity and data quality and keep the cells meeting the following conditions: the TPM of CD3D is larger than 3; ② when separating CD4⁺For T cells, the TPM of CD4 needs to be greater than 3, while the TPM of CD8 is less than 30; ③ separating CD8⁺For T cells, the TPM of CD8 needs to be greater than 3, while the TPM of CD4 is less than 30; (iv) the ratio of reads on the mitochondrial gene to all reads is not higher than 10%. Wherein, the definition of TPM value is:

wherein C is_ijExpressed as the number of reads of gene i in cell j.

According to the present invention, the quality control method of the gene expression level of a single cell for analysis comprises: the average number of reads detected for a gene in all cells was greater than 1 before use in subsequent analyses.

According to the present invention, log2 transformation was performed on the expression levels of all genes, and the mean value after log2 transformation for each gene expression in each patient was set to 0, thereby facilitating comparison of the expression difference of a given gene in different patients.

According to the present invention, the expressed genes may be subjected to alignment analysis in step (3) using various software known in the art, including but not limited to GSNAP software; the assignment of each read to a gene can be analyzed using various software known in the art, including but not limited to the R language package "findOverlaps".

According to the invention, the unsupervised cluster analysis of step (3) is performed with the software SC 3.

The working principle and steps of the software SC3 are as follows: screening n genes with the largest variance for spectral clustering: calculating Spearman correlation coefficient as the measure of the cell-to-cell distance; transforming the cells into a d-dimensional feature map space according to the distance between the cells; and performing k-means clustering in the d-dimensional feature map space. And obtaining an operation result every time one d is taken, and averaging the operation results of all the values of d to obtain a consistency matrix. The elements of the consistency matrix indicate how many proportions of the cells of the corresponding row and the cells of the corresponding column are grouped together in a class in the result of the operation. And then, obtaining k classes by utilizing hierarchical clustering on the consistency matrix. And selecting an appropriate clustering result according to the consistency matrix, the Silhouette statistic, the Dunn statistic and the like.

In a particular embodiment of the invention, the dimension number d ranges from 4% to 7% of the total number of cells.

In the present embodiment, different SC3 operating results were obtained with attempts to have k from 2 to 6 and n at 1500, 2000 and 3000.

In a preferred embodiment of the present invention, after the clustering result is obtained, the same clustering analysis can be performed on each class to obtain a finer clustering result.

According to the present invention, the analysis of step (4) can be performed using various types of software known in the art for analyzing characteristic expressed genes.

In one embodiment of the invention, the R language package (aov) is used to identify the characteristic expressed genes in each group, and the criteria for differential gene expression are: f test for BH correction, P < 0.05; the honeyst significance difference test (difference between any two groups), P < 0.01.

In the inventionIn one embodiment, to identify CD8⁺The genes characteristic of the T cells depleted in T cells were differentially analyzed for depleted and non-depleted T cells using the software limma. Setting the standard as expression quantity absolute value difference greater than 4 times, BH corrected P<0.01。

In one embodiment of the invention, to identify active regulatory CD4⁺Genes were expressed characteristic of T cells and differential expression analysis was performed using limma software. The criteria were set for expression levels between active regulatory T cells and inactive regulatory T cells, as well as regulatory T cells and other CD4⁺Absolute differences between T cells need to be greater than 4-fold, BH corrected P<0.01。

The invention has the beneficial effects that:

the invention utilizes the single cell transcriptome analysis technology, and discovers a new T cell gene capable of reflecting the tumor immune state of an organism by analyzing the single cell gene expression profile of the infiltrated T cells in the cancer tissues.

The invention discovers that the T cells express genes WARS and ACP5, which means that the T cells are in a depletion state, genes WARS and ACP5 are genes which are firstly discovered by the invention and are related to the depletion state of the T cells, and CD8 expressing the WARS and ACP5 genes⁺There is a correlation between T cells and tumor prognosis.

The present inventors have found that CD4⁺The T cells express genes STAM and BATF, which means that the regulatory T cells are converted into a state of activity in cancer tissues, the genes STAM and BATF are genes related to the activity state of the regulatory T cells and express CD4 of STAM and BATF⁺There is a correlation between T cells and tumor prognosis.

On the basis of the CD8⁺WARS and ACP5 genes of T cells, CD4⁺The STAM and BATF genes of the T cell can become functional molecules of the T cell, inhibit the expression of the genes or the activity of expression proteins of the genes, and can be used for the immunotherapy of tumors.

In addition, CD8⁺WARS and ACP5 genes of T cells, CD4⁺STAM and BATF genes of T cells, and can also be used as diagnostic marker set for tumor prognosis。

In addition, in the research process, the invention develops and debugs the single-cell transcriptome analysis technology of the T cell, improves the reverse transcription yield of the single-cell mRNA, and improves the accuracy and the effectiveness of the analysis method.

The terms in the present invention describe:

HAVCR2, also known as Tim3, encodes proteins belonging to the immunoglobulin superfamily and the Tim family. The gene is known to be expressed on Th1 cells, and can regulate macrophage activity, and inhibit Th1 cell-mediated autoimmune reaction, thereby promoting immune tolerance.

PDCD1, also known as PD1, encodes a protein belonging to the immunoglobulin superfamily that is a transmembrane protein. This gene is expressed on precursor B cells and plays a role in B cell differentiation. Deletion of this gene in mice will lead to dilated cardiomyopathy and congestive heart failure, and PDCD1 also plays an important role in suppressing autoimmune responses in T cells. There are tumor immune antibody drugs against PDCD 1.

WARS, an aminoacyl tRNA synthetase, its encoded protein catalyzes the aminoacylation of tRNA. WARS is expressed under interferon-mediated conditions, responsible for adding tryptophan residues to tRNA, and is an evolutionarily well-conserved gene.

ACP5, a protein encoded by the gene that is a glycoprotein containing metal elements responsible for the catalytic reaction of converting monophosphoryl lipids to alcohol and phosphate. Impairment of the ACP5 gene is associated with spondylonchondysplasia-related immunodeficiency.

LAG3, the encoded protein belonging to the immunoglobulin superfamily, is known from CD4⁺Is specifically expressed on T and is a co-suppression molecule of T cells.

TNFRSF4, also known as OX40, encodes a protein belonging to the Tumor Necrosis Factor (TNF) receptor superfamily, which is a costimulatory molecule for T cell activation. NF-kappaB (a transcription factor) can be activated by interaction with TRAF2 and TRAF 5.

STAMs, the encoded proteins of which belong to the molecular family of signal transduction linkers, play a role in the action of intercellular cytokines and growth factors, and possibly in the T cell differentiation process.

BATF, a transcription factor belonging to the AP-1 family. The expression of the gene in T cells plays an important role in T cell differentiation, and comprises Th17 and Tfh (lamellar T helper) cells. The leucine zinc finger protein structure of this protein may aid dimerization of JUN proteins.

CTLA4, the encoded protein belonging to the immunoglobulin superfamily, mediates signal molecules that inhibit T cell activity and includes V, transmembrane and intracellular domains. Mutations in this gene are known to be associated with a variety of diseases, including autoimmune diseases such as insulin-dependent diabetes mellitus, Graves disease, celiac disease, systemic lupus erythematosus, thyroid-associated orbital disease, and the like. There are tumor immune antibody drugs against CTLA 4.

The basic information of the above genes is as follows (the gene names are based on HGNC, and the NG, NM and NP numbers are based on RefSeq database):

in the present invention, the terms "selecting", "sorting", "dividing" or "isolating" a selected cell, cell population or cell subpopulation may be used interchangeably and, unless the context indicates otherwise, refer to removing a selected cell or defined subset of cells from a tissue sample.

The term "enriched" can be broadly construed as a treated cell population that contains a higher percentage of a selected cell type than in an untreated, otherwise equivalent cell population or sample. In some preferred embodiments, enriching a cell population refers to increasing the percentage of one cell type in the cell population by about 50% or greater than 50% as compared to the starting cell population. In other preferred embodiments, the enriched cell population of the invention will comprise at least 50%, 60%, 70%, 80%, 85%, 90%, 95%, 98% or 99% of the selected cell type.

The term "substantially pure" with respect to a particular cell population refers to a cell population that is at least about 75%, preferably at least about 85%, more preferably at least about 90%, and most preferably at least about 95% pure relative to the cells that make up the total cell population.

The terms "marker", "marker" or "cellular marker" are synonymous and refer to any trait or characteristic in the form of a chemical or biological entity. The label may be morphological, functional or biochemical in nature. In preferred embodiments, the label is differentially or preferentially expressed by a particular cell type (e.g., depleting CD 8)⁺T cells), or cytokines or surface antigens or membrane proteins or cytoplasmic proteins expressed by the cells under certain conditions (e.g., at a particular point in the cell cycle or at a particular extracellular matrix). In the present invention more specifically those markers which indicate a cell or a cell subset by virtue of their presence (positive) or absence (negative).

Similarly, in the context of a tissue, cell, or cell population, the term "marker phenotype" means any marker or combination of markers that can be used to characterize, identify, quantify, separate, isolate, purify, or enrich for a particular cell or cell population. In a particularly preferred embodiment, the marker phenotype is a cell surface phenotype that can be determined by detecting or identifying the expression of a combination of cell surface markers.

The terms "binding agent", "binding molecule" and "binding entity" are synonymous and may be used interchangeably. In the context of the present invention, the binding agent binds to, recognizes, interacts with, or otherwise associates with a selectable marker on said subpopulation of T cells. Exemplary binding agents may include, but are not limited to, antibodies or fragments thereof, antigens, aptamers, nucleic acids (e.g., DNA and RNA), proteins (e.g., receptors, enzymes, enzyme inhibitors, enzyme substrates, ligands), peptides, lectins, fatty acids or lipids, and polysaccharides. For example, in some embodiments of the invention, the binding agent comprises an antibody or fragment thereof, a nucleic acid (e.g., DNA and RNA).

The term "antibody" is used in the broadest sense and specifically covers synthetic antibodies, monoclonal antibodies, polyclonal antibodies, recombinant antibodies, intrabodies, multispecific antibodies, bispecific antibodies, monovalent antibodies, multivalent antibodies, human antibodies, humanized antibodies, chimeric antibodies, primatized antibodies, Fab fragments, F (ab') fragments, single chain fvfc (scfvffc), single chain fv (scfv), anti-idiotypic (anti-Id) antibodies, and any other immunologically active antibody fragments, so long as they exhibit the desired biological activity (i.e., label-related or binding). In a broader sense, the antibodies of the invention include immunoglobulin molecules and immunologically active fragments of immunoglobulin molecules (i.e., molecules that contain an antigen binding site), where these fragments may or may not be fused to another immunoglobulin domain, including but not limited to an Fc region or fragment thereof. Furthermore, as outlined herein in more detail, the term antibody and antibodies specifically includes Fc variants or fragments thereof, including full length antibodies and variant Fc-fusions comprising an Fc region, optionally comprising at least one amino acid residue modification and fused to an immunologically active fragment of an immunoglobulin.

By "fragment" of a molecule is meant any contiguous polypeptide or subset of nucleotides of the molecule. For example, a fragment of a transmembrane protein may comprise a construct that comprises only the extracellular domain or some portion thereof. For purposes of the present invention, a marker fragment or derivative may include any immunoreactive or immunologically active portion of a selectable marker.

"analog" of a molecule (e.g., a label) means a molecule that is functionally similar to the entire molecule or to a fragment thereof. As used herein, a molecule is referred to as a "chemical derivative" of another molecule when the molecule contains additional chemical moieties that are not normally part of the molecule. Such moieties may improve the solubility, absorption, biological half-life, etc. of the molecule. These moieties may alternatively reduce the toxicity of the molecule, eliminate or attenuate any adverse side effects of the molecule, and the like.

The terms "subject" or "patient" are used interchangeably and include, but are not limited to, humans, non-human animals, e.g., non-human primates such as chimpanzees and other apes and monkey species; farm animals such as cattle, sheep, pigs, goats, and horses; domestic subjects such as dogs and cats; laboratory animals, including rodents such as mice, rats and guinea pigs, and the like. The term does not indicate a specific age or gender.

The terms "malignancy," "tumor," and "cancer" are used interchangeably to refer to a disease or disorder characterized by uncontrolled, hyperproliferative or abnormal growth or metastasis of cells.

The term "diagnostic agent" refers to any molecule, compound, and/or substance used for the purpose of diagnosing a disease or disorder. In a preferred embodiment, the diagnostic agent should comprise a binding agent that binds to the reporter molecule. Other non-limiting examples of diagnostic agents include antibodies, antibody fragments, or other proteins, including those that bind to a detection reagent. The term "detection reagent" or "reporter" refers to any molecule, compound, and/or substance detectable by any methodology available to those of skill in the art, non-limiting examples of which include dyes, fluorescent labels, gases, metals, or radioisotopes.

The term "test chemical" refers to a chemical whose activity is being measured and includes, but is not limited to, small molecule compounds, nucleic acids (DNA or RNA), proteins or polypeptides (e.g., ligands, antibodies, fusion proteins, etc.), polysaccharides, and the like.

The term "inhibit" refers to a decrease in the activity of a protein or cell as compared to the absence of an inhibitor. In some embodiments, the term "inhibit" refers to a decrease in activity of at least about 25%, at least about 50%, at least about 60%, at least about 70%, at least about 80%, at least about 90%, or at least about 95%. In other embodiments, inhibition refers to a decrease in activity of about 25% to about 50%, about 50% to about 75%, or about 75% to 100%. In some embodiments, inhibition refers to a decrease in activity of about 95% to 100%, e.g., a decrease in activity of 95%, 96%, 97%, 98%, 99%, or 100%. Such a reduction can be measured using a variety of techniques known to those skilled in the art.

The terms "expression" and "gene expression" are synonymous and mean that a cell converts genetic information stored in a DNA sequence during its life through transcription and translation into a biologically active protein molecule.

The terms "increased expression" and "high expression" are synonymous and refer to an increased copy number of a gene transcript, and/or increased translation, as compared to normal levels.

The term "prognosis" refers to the prediction of the likely course and outcome of a disease, and includes both the determination of the particular outcome of a disease (e.g., recovery, the appearance or disappearance of other abnormalities such as certain symptoms, signs, and complications, and death), and the provision of a time cue, such as the prediction of the likelihood of a certain outcome occurring over a certain period of time.

Drawings

FIG. 1 is a graph of flow cytometric analysis of tumor tissue, normal tissue and peripheral blood T cells. First row: t cells in peripheral blood; a second row: t cells in normal tissue; third row: t cells in tumor tissue.

FIG. 2-1, FIG. 2-2 are examples of qualified single-cell cDNA. 96 cells were simultaneously examined using a LifeTech Real-Time PCR instrument 7500. CT values of subject RT-PCR were below 26 (shown in black box in FIG. 2-1); the peak of the subject dissolution curve was between 85 ℃ and 90 ℃ (shown in fig. 2-2 black box).

FIG. 3 is an example of a qualified single-cell cDNA. Detection results of Fragment Analysis based on capillary electrophoresis. The peak around 1700 is the fragment size of the full length transcriptome and the peak around 1100 is the ERCC as the internal reference.

FIG. 4 is an example of a qualified cDNA library. Results of Fragment Analysis based on capillary electrophoresis.

Dunn value and Silhouette value for different clustering software of FIGS. 5-1 and 5-2. Wherein, FIG. 5-1 shows Dunn values of different clustering methods when different k values (number of classes of clustering) are taken; fig. 5-2 shows the values of silouette for different clustering methods when different values of k (number of classes clustered) are taken. The points in the circle are the results obtained for the optimal k value.

FIG. 6CD8⁺Heat map of T cell gene expression. According to the gene expression, 5 kinds of CD8 are formed⁺T cell subsets. Wherein cluster4 is a depleted T cell.

FIG. 7CD4⁺Heat map of T cell gene expression. According toThe gene expression condition is aggregated into 6 types of CD4⁺T cell subsets. Wherein cluster3 is an active regulatory T cell.

FIGS. 8-1 to 8-9 are boxplots showing the expression levels of ACP5, BATF, CTLA4, HAVCR2, LAG3, PDCD1, STAM, TNFRSF4, and WARS in different cell subsets, respectively. Each point in the figure is a cell, the abscissa is the class of 9T cell subsets obtained by clustering, and the ordinate is the expression level TPM (log2 transformation) of the corresponding gene.

FIG. 9 regulatory CD4⁺The expression of T cell characteristic genes and the survival correlation Kaplan-Meier curve of a liver cancer patient, wherein the characteristic genes are TNFRSF4, STAM, BATF and CTLA 4.

FIG. 10 exhaustive CD8⁺The expression of T cell characteristic genes is related to the survival of the liver cancer patient by a Kaplan-Meier curve, and the characteristic genes are HAVCR2, PDCD1, WARS, ACP5 and LAG 3.

FIG. 11 is a statistical map of the amplification bias of the method of Tang 2009.

FIG. 12Smart-seq2 shows a statistical bias for amplification.

Detailed Description

The present invention is further described below with reference to examples.

The following examples are illustrative of methods for analyzing single-cell transcriptome information of T cells performed in liver cancer patients.

It should be noted that the examples are not intended to limit the scope of the present invention, and those skilled in the art will appreciate that any modifications and variations based on the present invention are within the scope of the present invention.

The chemical reagents used in the following examples are conventional and are commercially available.

The analytical software used and its source were as follows:

GSNAP(http://research-pub.gene.com/gmap/)；

statistical software R (https:// www.r-project. org /);

HTSeqGenie,DESeq2,SC3,monocle,ComplexHeatmap,ggplot2,Rtsne and survival(https://www.bioconductor.org/)

HCC sequence database: EGA (https:// www.ebi.ac.uk/EGA/home) access number EGAS00001002072

TCGA data: cBioportal (http:// www.cbioportal.org /) and (https:// gdc. cancer. gov /).

Example 1 Single cell transcriptome data acquisition of T cells

1. Clinical specimen collection

Surgical tissues and peripheral blood including cancer tissues (diameter 3-5 mm) and paracancer normal tissues of patients were collected from 2014 to 2015 6 at the Beijing university people hospital and the Beijing century bed hospital, and peripheral blood (3ml) was collected. The patients were hepatocellular carcinoma and did not undergo adjuvant radiotherapy or chemotherapy before surgery, for 5 cases. The study was in compliance with the medical ethical standards announced by helsinki and was approved by the medical ethical committee of the university of beijing.

Blood samples were collected in EDTA anticoagulation tubes before surgery and temporarily stored on ice; collecting cancer tissue and paracancerous normal tissue samples during surgery, wherein the cancer tissue is depleted of necrotic tissue; the paracancerous tissue is normal tissue at least 5cm away from the cancerous tissue. The cancer tissue and the para-carcinoma tissue were placed on ice and RNAlater (Qiagen) solution within 30 minutes ex vivo and the single cell isolation procedure was completed within the day.

2. Single cell suspension preparation

Peripheral blood: peripheral blood mononuclear cells were separated by density gradient centrifugation. The specific operation is to slowly add 3ml of whole blood to 3ml

1077 on isolate (Sigma, cat. No.1077), 400g were centrifuged at room temperature for 30 minutes, white layer monocytes were carefully aspirated, washed with 10ml PBS, centrifuged at 4 ℃ for 15 minutes and the above washing procedure was repeated once. Finally, cells were dissolved in 0.5ml PBS and 1% bovine serum (FBS) was added.

Cancer tissue and paracancerous normal tissue: grinding to obtain single cells of cancer tissue and paracancer normal tissue. Firstly, the tissue in vitro of the operation is cut into 1mm³The size pieces were soaked in RPMI-1640 medium and 10% calf serum was added. Rapid grinding of tissue using copper meshTissue debris was removed by 40 μm sieving and single cell suspensions were collected by centrifugation at 400g for 10 min. The erythrocytes mixed in the tissue were further removed using an erythrocyte lysate. The cells were also washed twice with 10ml PBS and finally lysed in 0.5ml PBS and 1% calf serum was added.

3. Single cell isolation of T cells of interest

The isolated cells of interest include cytotoxic T cells (CD3 positive, CD8 positive), helper T cells (CD3 positive, CD4 positive, CD25 negative) and regulatory T cells (CD3 positive, CD4 positive, CD25 positive).

The three cells were fluorescently labeled with different antibodies from eBioscience, 10 each⁶Each cell was treated with 5 μ l antibody:

rabbit anti-CD 3 antibody (FACS, Cat #48-0037-41)

Rabbit anti-CD 4 antibody (FACS, Cat #11-0048-41)

Mouse anti-CD 8 antibody (FACS, Cat #17-0086-41)

Mouse anti-CD 25 antibody (FACS, Cat #12-0259-42)

7AAD (FACS, Cat #00-6993-50), 7AAD was used to mark dead cells.

The reaction solution was previously added to each well of a 96-well plate:

the primer sequence is as follows:

AAGCAGTGGTATCAACGCAGAGTACTTTTTTTTTTTTTTTTT

TTTTTTTTTTTTTVN

the isolation of the T cells of interest is shown in FIG. 1. Cytotoxic T cells, helper T cells and regulatory T cells were selected based on molecular markers on the cell surface, and individual cells were individually collected into each well of a corresponding 96-well plate using a flow cytometer.

4. mRNA reverse transcription and cDNA amplification

The procedure followed for reverse transcription of single cells isolated in 96-well plates to obtain cDNA was as follows according to the SMART-seq2 method (Picelli, S.et al. full-length RNA-seq free cells using Smart-seq2.nat. Protoc.9, 171-181 (2014)):

1) single cell lysis: the single cells in the above solution were vortexed for at least 10 seconds. Incubate for 3 minutes at 72 ℃ on a PCR instrument.

2) Internal reference RNA (ERCC RNA Spike-In Mix, Invitrogen, cat. No.4456740) was added. It was diluted 350 times in advance, and 1. mu.l was added. The reference RNA is useful for quantitative calculation of the gene expression level.

3) Reverse transcription: the reaction system is as follows:

the sequence of the TSO primer is: AAGCAGTGGTATCAACGCAGAGAGTACATrGrG + G

The reaction conditions are as follows:

4) and (3) PCR amplification: the reaction system is as follows:

KAPA HiFi HotStart ReadyMix(2x) 12.5μl

IS PCR primer (10. mu.M) 0.25. mu.l

Nuclease-free ultrapure water 2.25. mu.l.

The IS PCR primer sequence IS: AAGCAGTGGTATCAACGCAGAGT

The reaction conditions are as follows:

the amplified PCR product was purified using Agencour AMPure XP magnetic beads (Beckman) as follows:

(1) adding 25 mul of magnetic beads into 25 mul of the reaction solution in the previous step, and uniformly mixing by blowing;

(2) standing at room temperature for 5 minutes;

(3) placing the test tube or plate containing the solution on a magnetic frame for 5 minutes;

(4) removing the liquid;

(5) washing the magnetic beads with 100. mu.l of 80% ethanol, standing for 30 seconds, removing the magnetic beads, and repeating the process once;

(6) taking down the magnetic frame, adding 20 mul EB solution, blowing and sucking and mixing evenly;

(7) after standing for 2 minutes, the mixture was placed on a magnetic stand, and after standing for 2 minutes, the liquid was aspirated.

It has been found that primers remaining in solution during the above process reduce the efficiency of library construction, allowing the library to contain components of the acellular cDNA. For this purpose, an additional purification operation was required, and the purification process was identical to that described above except that the amount of the magnetic beads was changed to 50. mu.l.

And performing quality detection, namely detecting the gene CD3 specifically expressed by the T cells through RT-PCR to judge the effectiveness of amplification.

The reaction system is as follows:

the primer sequence for CD3 was: TCATTGCCACTCTGCTCC (forward direction) and

GTTCACTTGTTCCGAGCC (reverse).

The reaction conditions are as follows:

there are two criteria for determining whether cDNA is available: the CT value of the RT-PCR of the object is lower than 26; the second is that the peak of the dissolution curve of the subject is between 85 ℃ and 90 ℃. Examples of qualified cDNAs obtained in this example are shown in FIGS. 2-1 and 2-2.

Another quality control means is Fragment Analysis, which detects the size and concentration of fragments of sample DNA based on capillary electrophoresis. FIG. 3 shows an example of a qualified cDNA obtained in this example.

5. Sequencing library construction

cDNA library constructsThe TruePrepTM DNA Library Prep Kit V2for was used

Kit (vazyme, cat. No. td501/502/503); matching the double-end Index to TruePrepTM Index Kit V2for

(vazyme, cat No. td202). The library was started with 1. mu.g of cDNA according to the kit instructions. Selecting the size of the Fragment by using the magnetic bead to obtain a cDNA library with the size of the target Fragment of 400 bp-600 bp, finally performing quality control by Fragment Analysis, and determining that the library is qualified in construction, wherein the corresponding Analysis result is shown in figure 4.

The Illumina Hiseq4000 was used for sequencing in a paired end 150bp mode, and typically 1 million reads were required for data size of one cell.

Example 2 analysis of biological information

1. Data comparison and quality control

For reads obtained from the sequencer (reads), the low quality fractions are first removed, with the following retention criteria: firstly, the unknown base accounts for not more than 10 percent of the total sequence of the given read, secondly, the base with the mass value of less than 5 does not exceed 50 percent, and thirdly, the unknown base cannot contain a linker sequence. Alignment was done using GSNAP software. Txt "from UCSC, using R language package" findOverlaps "to count the assignment of reads on genes, using TPM value to calibrate the expression level of each gene in each cell, using the formula:

wherein C is_ijExpressed as the number of reads of gene i in cell j.

T cells with low data volume and data quality need to be filtered out. Cells meeting the following criteria were retained: the TPM of CD3D is larger than 3; ② when separating CD4⁺For T cells, the TPM of CD4 needs to be greater than 3, while the TPM of CD8 is less than 30; ③ separating CD8⁺T is thinWhen the cell is detected, the TPM of the CD8 needs to be more than 3, and the TPM of the CD4 needs to be less than 30; (iv) the ratio of reads on the mitochondrial gene to all reads is not higher than 10%.

In addition, some reference standards were also set on library capacity (library size) and gene expression quantity. The average number of reads detected for a gene in all cells was greater than 1 before use in subsequent analyses. The log2 transformation was performed for all genes and the log2 mean value for each gene expressed in each patient was set to 0, thereby facilitating comparison of differences in expression of a given gene among different patients, using the R language package computers genes. Generally, the number of expressed genes detected per cell is about 3000 or so.

2. Single cell expression profiling identification

And (4) judging the type of the T cell according to the matrix of the expression quantity of each gene of each cell obtained in the last step.

First an iterative unsupervised clustering was performed with the software SC3 (Kiselev VY, Kirschner K, Schaub MT, Andrews T, Yiu A, Chandra T1, Natarajan KN, Reik W, Barahona M8, Green AR, Hemberg M.SC3: consensus clustering of single-cell RNA-seq data. nat methods.2017 May; 14(5):483-486.doi:10.1038/nmeth.4236.Epub 2017Mar 27.). Specifically, the n genes with the largest variance are screened for spectral clustering: calculating Spearman correlation coefficient as the measure of the cell-to-cell distance; the distance between cells can be represented by a graph, and the cells can be transformed into a d-dimensional feature map space through a Laplace matrix and a feature vector of the graph; and performing k-means clustering in the d-dimensional feature map space. The value of the dimension number d is from 4% to 7% of the total number of cells, an operation result is obtained by taking one d every time, and a consistency matrix can be obtained by averaging the operation results of all the values of d. The elements of the consistency matrix indicate how many proportions of the cells of the corresponding row and the cells of the corresponding column are grouped together in a class in the result of the operation. And then k classes are obtained on the consistency matrix by utilizing hierarchical clustering. k is a parameter of the SC3 software. Trying k from 2 to 6, n is 1500, 2000 or 3000, from these different SC3 run results, one of the best is selected as a clustering result based on the consistency matrix, the silouette statistics, the Dunn statistics, etc. After the clustering result is obtained, the above process is carried out on each class to obtain finer clustering until a meaningful clustering result is found.

Meanwhile, the same data are contrastively analyzed by adopting a clustering method based on Hcluster and K-means, and the preprocessing of whether principal component analysis is carried out is respectively included. The comparative method was to calculate the values for Dunn and Silhouette. The formula for Dunn is

I.e. the most distant point in the same class is compared with the closest point in a different class. And the calculation formula of Silhouuette is S_i＝(b_i-a_i)/max(a_i，b_i) I.e. comparing the distance of any two points in a group with the farthest cluster. The larger these two values represent a clearer boundary between cell classes, while the more dense the cells within the class.

Through comparison, different clustering methods are found, namely: SC3, Hcluster and k-means, obtained Dunn and Silhouette values were different, and SC3 results were superior to the Hcluster and k-means methods, see FIG. 5-1 and FIG. 5-2 for comparison results.

When the SC3 analysis is used, the result is relatively optimal when k is 4, and therefore, the clustering result is used to identify the subsequent signature gene.

FIGS. 6 and 7 show the use of SC3 software for CD8⁺T cells and CD4⁺And (4) clustering the T cells to obtain matrix images of various subgroups.

The R language package (aov) is used to identify characteristic expressed genes in each group, from which the functional class of cells is inferred, including cytotoxic T cells, regulatory T cells, etc. To identify CD8⁺Characterization of T-cell depleting in T-cells we performed differential analysis of T-cells with the software limma for depleting and non-depleting cells. Setting the standard as expression quantity absolute value difference greater than 4 times, BH corrected P<0.01. Finally, 102 characteristic expression genes of the exhausted T cells in the liver cancer are obtained (Table 1, E in the table is a scientific counting method).

TABLE 1 genes characteristic of infiltrating depleted T cells in liver cancer tissue

Likewise, to identify active regulatory CD4⁺Genes were expressed characteristic of T cells and differential expression analysis was performed using limma software. The criteria were set for expression levels between active regulatory T cells and inactive regulatory T cells, as well as regulatory T cells and other CD4⁺Absolute differences between T cells need to be greater than 4-fold, BH corrected P<0.01. Finally, 94 genes expressing characteristics of regulatory T cells in liver cancer were obtained (Table 2, E in the table is scientific counting method).

TABLE 2 genes characteristic of infiltrating regulatory T cells in liver cancer tissue

Among the genes expressed by the signature are known immunotherapeutic target genes, such as PDCD1 and CTLA4, as well as previously unknown genes associated with cellular status and function. Through research, BATF and STAM are found to be novel genes expressed by characteristics of active regulatory T cells, and WARS and ACP5 are novel genes expressed by characteristics of exhausted T cells.

3. Utility in prognosis determination

The identified characteristic genes of the T cell functional group are analyzed for the use value in the disease prognosis of patients. The dataset used was data for tcga (the Cancer Genome atlas) liver Cancer (LIHC, hepatocellular carcinoma) as this study collected follow-up information collated with patients. Data on the gene expression levels of patient cancer tissues were downloaded from UCSC Xena (http:// Xena. UCSC. edu /), and patient follow-up (survival) information was downloaded from GDC Data Portal (https:// GDC-Portal. nci. gov /).

In order to exclude interference of the amount of expression of a certain gene in the cells of non-tumor infiltrating T cells of the patient on the analysis, the depleted CD8 identified by the above work was subjected to⁺T cell and activity modulating CD4⁺T cell characteristic gene, gene expression information obtained from TCGA database, gene expression amount after z-score conversion was averaged in each sample, and then divided by the expression value of CD3 in each sample to reflect the degree of T cell infiltration in cancer tissue.

To maximize the differentiation of patient prognosis, the characteristic genes were analyzed individually and in random combinations. For individual genes, patients will be divided into high expression and low expression groups, grouped at the median. For the combined genes, the gene expression values in one group are averaged, and then divided into a high expression value group and a low expression value group, and the groups are divided into a median value.

The candidate gene range was the exhausted CD8 obtained by the analysis of step 2 above⁺Gene specifically highly expressed in T cells and activity-regulated CD4⁺A gene specifically expressed in T cells. Patient survival differences were expressed using the Kaplan-Meier curve (FIGS. 9 and 10), and log-rank was used to test whether the difference was significant (P < 0.01).

TABLE 3 statistics of the prognosis of patient survival differences for characteristic expression genes of T-depleted cells

Gene combinations	P value	Ratio of risks
			HAVCR2_PDCD1_WARS_ACP5_LAG3	0.0022	1.7207
HAVCR2_CTLA4	0.0039	1.6622
			HAVCR2_PDCD1_SIRPG_WARS_ACP5_LAG3	0.0070	1.6092
HAVCR2_PDCD1_WARS_ACP5_DUSP4_LAYN	0.0082	1.5948
			SIRPG_WARS_ACP5_LAG3	0.0087	1.5923
PDCD1_WARS_ACP5_LAG3	0.0093	1.5858
			TIGIT_WARS_ACP5_LAG3	0.0096	1.5798

As can be seen from the above table, CD8⁺The liver cancer patient with high WARS and ACP5 expression of T cells has short survival time and important significance in prognosis diagnosis of liver cancer patient, and the CD8 containing WARS and ACP5⁺The gene combination expressed by the T cells can be used for diagnosing the prognosis of the liver cancer patient.

TABLE 4 statistics of genes expressed characteristic of activity-regulated T cells for prognosis of patient survival differences

Gene combinations	P value	Ratio of risks
			TNFRSF4_CTLA4_STAM_BATF	0.0003	1.9009
TNFRSF4_STAM_BATF	0.0004	1.8816
			CTLA4_STAM_BATF	0.0026	1.6954
STAM_BATF	0.0054	1.6299

As can be seen from the above table, the liver cancer patients with high expression of STAM and BATF by the activity regulatory T cells have short survival time and have important significance in prognosis diagnosis of the liver cancer patients, and the genome assembly containing the expression of the activity regulatory T cells of STAM and BATF can be used for diagnosis of prognosis of the liver cancer patients.

Wherein, as can be seen from the above list: exhausted CD8⁺The expression of the following 5 genes of the T cell is most obviously related to the prognosis of the liver cancer patient: HAVCR2, PDCD1, WARS, ACP5 and LAG 3. Activity-modulating CD4⁺The expression of the following 4 genes of the T cell is most obviously related to the prognosis of the liver cancer patient: TNFRSF4, tam, BATF, CTLA 4.

Comparative example:

1. comparison of cDNA library construction methods

The Smart-Seq2 method and the soup remuneration creation method (for convenience of description, hereinafter referred to as "Tang 2009", Tang, F.et al. RNA-Seq analysis to capture the transcripto-me landscaping of a single cell. Nat. Protoc.5, 516-535 (2010))

In terms of efficiency of cDNA amplification:

after completion of the cDNA amplification, the amplification efficiency was examined by detecting the expression of housekeeping gene β -actin, GAPDH or CD3 gene by RT-PCR using ultrapure water without nucleic acid as a negative control. There are two criteria for determining whether a gene is expressed: firstly, the CT value of the cell sample RT-PCR is obviously smaller than that of a negative control; secondly, the peak of the lysis curve for the cell sample is between 85 ℃ and 90 ℃ (negative control is about 78 ℃).

According to the experimental results, after the amplification by Smart-seq2 method, beta-actin, GAPDH and CD3 can be detected in most cells. After amplification by Tang2009, expression of β -actin and GAPDH was detected in most cells, and CD3 was detected only in very few cells.

In terms of cDNA library quality:

the construction of the library is not influenced by an amplification method, the amount of cDNA required by the construction of the library can be provided by the amplification of the Tang2009 method and the Smart-seq2 method, and the sizes of the fragments of the constructed library meet the requirements of an Illumina sequencer.

However, if the ratio of the number of the cDNA library finally constructed and the number of the initial single cells for library construction is compared, the success rate of T cell amplification by the Smart-seq2 method is generally higher than that by Tang2009 method in comparison with samples from three liver cancer patients, as shown in the following table.

Construction of cDNA library Power Table

Patient numbering

Library construction method

PTC

PTH

PTR

TTC

TTH

TTR

NTC

NTH

NTR

20141202

Tang2009

46％

24％

26％

62％

52％

78％

/

20150205

Smart-seq2

80％

67％

80％

60％

67％

/

20150508

Smart-seq2

92％

74％

92％

95％

92％

88％

70％

86％

/

Note: "/" indicates no samples of this type. PTC is cytotoxic T cells in peripheral blood, PTH is helper T cells in peripheral blood, and PTR is regulatory T cells in peripheral blood. TTC is a cytotoxic T cell in cancer tissue, TTH is a helper T cell in cancer tissue, and TTR is a regulatory T cell in cancer tissue. NTC is cytotoxic T cells in normal liver tissue, NTH is helper T cells in normal liver tissue, and NTR is regulatory T cells in normal liver tissue. The abbreviations for each of the tables below are the same.

In terms of sequencing quality:

taking samples of three liver cancer patients as an example, the specific parameters are shown in the following table.

Tang2009 method (patient 20141202) sample average sequencing quality Table

Cell type	Total read	High quality read rate	Comparison rate	Expression of gene factors
					PTC	1365867	90.02％	95.36％	1329
PTH	1912263	54.62％	98.36％	1189
					PTR	1380273	94.55％	98.85％	2376
TTC	1418412	88.63％	98.13％	1987
					TTH	2155667	93.54％	94.95％	1274
TTR	1446343	93.86％	97.52％	1698

Smart-seq2 method (patient 20150205) sample average sequencing quality Table

Cell type	Total read	High quality read rate	Comparison rate	Expression of gene factors
					PTC	1561419	97.88％	62.47％	3042
PTH	1610403	98.33％	79.72％	2553
					PTR	1482323	98.42％	60.33％	3103
TTC	1442573	99.51％	51.83％	3071
					TTH	1890069	77.07％	66.7％	2478
TTR	1558685	98.49％	67.87％	3478

Smart-seq2 method (patient 20150508) sample average sequencing quality Table

Cell type	Total read	Comparison rate	High quality read rate	Expression of gene factors
					NTC	1819830	97.10％	99.25％	2187
NTH	1431656	93.38％	99.44％	2267
					PTC	1476204.5	95.40％	99.38％	2745
PTH	1530590	86.35％	99.39％	2026
					PTR	1513190	96.09％	99.43％	2295.5
TTC	1602175	95.83％	99.38％	2823
					TTH	1562359.5	94.52％	99.38％	2795.5
TTR	1514486	96.33％	99.43％	2907

As the data of the single cell RNA-seq, the high-quality reading rate and the comparison rate of most cells are higher, which indicates that the sequencing quality is good, and the obtained data is suitable for analyzing the expression quantity and the expression sequence. From the aspect of the detected gene number, the gene number detected by the Smart-seq2 method is obviously higher than that detected by the Tang2009 method.

In the amplification bias, there are different tendencies in amplifying cDNA by the Tang2009 method and the Smart-seq2 method. As can be readily seen from the analysis of the sequencing results, the Tang2009 approach is more prone to amplification of the 3' end of the cDNA, which is a bias against the assembly of the entire TCR structure, whereas the Smart-seq2 approach is more homogeneous in amplifying the cDNA, facilitating the assembly of the entire TCR sequence (FIGS. 11 and 12).

Claims

1. CD8⁺Use of biomarker panels for T cells in the preparation of a diagnostic or monitoring CD8⁺Use of a T cell depletion state or a kit for diagnosing or monitoring the prognosis of liver cancer, the biomarker panel comprising the genes WARS and ACP5, or proteins or protein fragments expressed by said genes.

2. The use of claim 1, wherein the panel of biomarkers further comprises at least one of the genes HAVCR2, PDCD1 and LAG3, or a protein or protein fragment expressed by said at least one gene.

3. The use of claim 2, wherein the biomarker panel is the genes WARS, ACP5, HAVCR2, PDCD1 and LAG3, or the proteins or protein fragments expressed by said genes.

4. The use of any one of claims 1-3, wherein the liver cancer is hepatocellular carcinoma.

5. Can be combined with CD8⁺A binding agent that binds to the T cell gene WARS or a protein or protein fragment expressed therefrom, and a polypeptide capable of binding to CD8⁺Binding agent for binding ACP5 gene of T cell or protein fragment thereof for preparing gene for diagnosing or monitoring CD8⁺The exhaustion state of T cells, or the use of a kit for diagnosing or monitoring the prognosis of liver cancer.

6. The use of claim 5, wherein said kit further comprises a CD8 conjugate⁺A binding agent which binds to the T cell gene HAVCR2 or a protein or protein fragment expressed therefrom, and/or which binds to CD8⁺T cell gene PDCD1 or its expressed protein orBinding agents which bind to protein fragments, and/or which bind to CD8⁺A binding agent to which the gene LAG3 of a T cell, or an expressed protein or protein fragment thereof, binds.

7. The use of claim 5, wherein the kit comprises a CD8 conjugate⁺Binding agent capable of binding to CD8 and binding to WARS gene of T cell or expressed protein or protein fragment thereof⁺Binding agent for binding to ACP5 gene of T cell or protein fragment thereof capable of binding to CD8⁺Binding agent for binding T cell gene HAVCR2 or its expressed protein or protein fragment, and CD8⁺A binding agent which is combined with a gene PDCD1 of a T cell or an expressed protein or a protein fragment thereof, and can be combined with CD8⁺A binding agent to which the gene LAG3 of a T cell, or an expressed protein or protein fragment thereof, binds.

8. The use of any one of claims 5-7, wherein the liver cancer is hepatocellular carcinoma.

9. The use of any one of claims 5 to 7, wherein the binding agent comprises a nucleic acid, a ligand, an enzyme, a substrate, and/or an antibody.